mirror of
https://github.com/facebook/rocksdb.git
synced 2026-11-06 11:02:36 -05:00
Compare commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
18a3d839d7 | ||
|
|
1858ff2a30 | ||
|
|
f803cc0e3c | ||
|
|
b871cc11e2 | ||
|
|
c377f3a35a | ||
|
|
907c391fa5 | ||
|
|
6df0b75137 | ||
|
|
736d2e71df | ||
|
|
8f0c83ae72 |
No files matched your search
@@ -32,7 +32,6 @@ build/
|
||||
ldb
|
||||
manifest_dump
|
||||
sst_dump
|
||||
column_aware_encoding_exp
|
||||
util/build_version.cc
|
||||
build_tools/VALGRIND_LOGS/
|
||||
coverage/COVERAGE_REPORT
|
||||
@@ -43,7 +42,6 @@ unity.a
|
||||
tags
|
||||
rocksdb_dump
|
||||
rocksdb_undump
|
||||
db_test2
|
||||
|
||||
java/out
|
||||
java/target
|
||||
@@ -65,7 +63,3 @@ java/javadoc
|
||||
scan_build_report/
|
||||
t
|
||||
LOG
|
||||
|
||||
db_logs/
|
||||
tp2/
|
||||
fbcode/
|
||||
+26
-24
@@ -1,39 +1,41 @@
|
||||
sudo: false
|
||||
language: cpp
|
||||
os:
|
||||
- linux
|
||||
- osx
|
||||
compiler:
|
||||
- clang
|
||||
|
||||
addons:
|
||||
apt:
|
||||
sources: ['ubuntu-toolchain-r-test', 'llvm-toolchain-precise-3.6']
|
||||
packages: ['clang-3.6' , 'zlib1g-dev', 'libbz2-dev', 'libsnappy-dev', 'curl', 'gflags']
|
||||
env:
|
||||
# Run all tests before db_block_cache_test (db_test, db_test2)
|
||||
- JOB_NAME=unittests ROCKSDBTESTS_END=db_block_cache_test
|
||||
# Run all tests starting from db_block_cache_test (db_block_cache_test, db_iter_test, ...)
|
||||
- JOB_NAME=unittests ROCKSDBTESTS_START=db_block_cache_test
|
||||
# Run java tests
|
||||
- JOB_NAME=java_test
|
||||
# Build ROCKSDB_LITE
|
||||
- JOB_NAME=lite_build
|
||||
matrix:
|
||||
include:
|
||||
- os: linux
|
||||
compiler: clang
|
||||
env: COMPILER=clang++-3.6
|
||||
addons:
|
||||
apt:
|
||||
sources: ['ubuntu-toolchain-r-test', 'llvm-toolchain-precise-3.6']
|
||||
packages: ['clang-3.6', 'clang-format-3.6', 'zlib1g-dev', 'libbz2-dev', 'libsnappy-dev', 'curl']
|
||||
- os: osx
|
||||
compiler: clang
|
||||
|
||||
install:
|
||||
# Build gflags
|
||||
# TODO(noetzli): Remove when gflags available through Travis
|
||||
- pushd /tmp/ && curl -L https://github.com/gflags/gflags/archive/v2.1.2.tar.gz -o gflags.tar.gz && tar xfz gflags.tar.gz && cd gflags-2.1.2 && cmake . && make && popd
|
||||
# Download clang-format-diff.py to check source code formatting
|
||||
- pushd /tmp/ && curl -L http://llvm.org/svn/llvm-project/cfe/trunk/tools/clang-format/clang-format-diff.py -o clang-format-diff.py && chmod +x clang-format-diff.py && popd
|
||||
|
||||
before_script:
|
||||
- if [[ "${TRAVIS_OS_NAME}" == 'linux' ]]; then CXX=clang++-3.6; fi
|
||||
# Limit the maximum number of open file descriptors to 8192
|
||||
- ulimit -n 8192 || true
|
||||
# Add gflags to include/library paths
|
||||
# TODO(noetzli): Remove when gflags available through Travis
|
||||
- export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/tmp/gflags-2.1.2/lib"
|
||||
- export LIBRARY_PATH="$LIBRARY_PATH:/tmp/gflags-2.1.2/lib"
|
||||
- export CPLUS_INCLUDE_PATH="$CPLUS_INCLUDE_PATH:/tmp/gflags-2.1.2/include"
|
||||
- if [ -n "${COMPILER}" ]; then CXX=${COMPILER}; fi
|
||||
- if [[ "${TRAVIS_OS_NAME}" == 'osx' ]]; then brew install gflags snappy; fi
|
||||
- ulimit -n 2000 || true
|
||||
|
||||
# Lousy hack to disable use and testing of fallocate, which doesn't behave quite
|
||||
# as EnvPosixTest::AllocateTest expects within the Travis OpenVZ environment.
|
||||
script:
|
||||
- if [[ "${JOB_NAME}" == 'unittests' ]]; then OPT=-DTRAVIS V=1 make -j4 check_some; fi
|
||||
- if [[ "${JOB_NAME}" == 'java_test' ]]; then OPT=-DTRAVIS V=1 make clean jclean rocksdbjava jtest; fi
|
||||
- if [[ "${JOB_NAME}" == 'lite_build' ]]; then OPT="-DTRAVIS -DROCKSDB_LITE" V=1 make -j4 static_lib; fi
|
||||
- if [[ "${TRAVIS_OS_NAME}" == 'linux' ]]; then OPT=-DTRAVIS CLANG_FORMAT_DIFF=/tmp/clang-format-diff.py make format || true; fi
|
||||
- OPT=-DTRAVIS V=1 make -j4 check && OPT=-DTRAVIS V=1 make clean jclean rocksdbjava jtest
|
||||
|
||||
notifications:
|
||||
email:
|
||||
- leveldb@fb.com
|
||||
|
||||
+49
-174
@@ -13,38 +13,26 @@
|
||||
# cd build
|
||||
# 3. Run cmake to generate project files for Windows, add more options to enable required third-party libraries.
|
||||
# See thirdparty.inc for more information.
|
||||
# sample command: cmake -G "Visual Studio 12 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 -DJNI=1 ..
|
||||
# OR for VS Studio 15 cmake -G "Visual Studio 14 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 -DJNI=1 ..
|
||||
# 4. Then build the project in debug mode (you may want to add /m[:<N>] flag to run msbuild in <N> parallel threads
|
||||
# or simply /m ot use all avail cores)
|
||||
# msbuild rocksdb.sln
|
||||
#
|
||||
# rocksdb.sln build features exclusions of test only code in Release. If you build ALL_BUILD then everything
|
||||
# will be attempted but test only code does not build in Release mode.
|
||||
#
|
||||
# sample command: cmake -G "Visual Studio 12 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 ..
|
||||
# 4. Then build the project in debug mode (you may want to add /m:<N> flag to run msbuild in <N> parallel threads)
|
||||
# msbuild ALL_BUILD.vcxproj
|
||||
# 5. And release mode (/m[:<N>] is also supported)
|
||||
# msbuild rocksdb.sln /p:Configuration=Release
|
||||
# msbuild ALL_BUILD.vcxproj /p:Configuration=Release
|
||||
#
|
||||
|
||||
cmake_minimum_required(VERSION 2.6)
|
||||
project(rocksdb)
|
||||
|
||||
include(${CMAKE_CURRENT_SOURCE_DIR}/thirdparty.inc)
|
||||
execute_process(COMMAND powershell -Command "Get-Date -format MM_dd_yyyy" OUTPUT_VARIABLE DATE)
|
||||
execute_process(COMMAND powershell -Command "Get-Date -format HH:mm:ss" OUTPUT_VARIABLE TIME)
|
||||
string(REGEX REPLACE "(..)_(..)_..(..).*" "\\1/\\2/\\3" DATE "${DATE}")
|
||||
string(REGEX REPLACE "(..):(.....).*" " \\1:\\2" TIME "${TIME}")
|
||||
|
||||
execute_process(COMMAND $ENV{COMSPEC} " /C date /T" OUTPUT_VARIABLE DATE)
|
||||
execute_process(COMMAND $ENV{COMSPEC} " /C time /T" OUTPUT_VARIABLE TIME)
|
||||
string(REGEX REPLACE "(..)/(..)/..(..).*" "\\1/\\2/\\3" DATE ${DATE})
|
||||
string(REGEX REPLACE "(..):(.....).*" " \\1:\\2" TIME ${TIME})
|
||||
string(CONCAT GIT_DATE_TIME ${DATE} ${TIME})
|
||||
|
||||
find_package(Git)
|
||||
|
||||
if (GIT_FOUND AND EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
|
||||
execute_process(COMMAND $ENV{COMSPEC} /C ${GIT_EXECUTABLE} -C ${CMAKE_CURRENT_SOURCE_DIR} rev-parse HEAD OUTPUT_VARIABLE GIT_SHA)
|
||||
else()
|
||||
set(GIT_SHA 0)
|
||||
endif()
|
||||
|
||||
string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA "${GIT_SHA}")
|
||||
execute_process(COMMAND $ENV{COMSPEC} " /C git rev-parse HEAD 2>nil" OUTPUT_VARIABLE GIT_SHA)
|
||||
string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA ${GIT_SHA})
|
||||
|
||||
set(BUILD_VERSION_CC ${CMAKE_CURRENT_SOURCE_DIR}/util/build_version.cc)
|
||||
|
||||
@@ -58,8 +46,7 @@ add_custom_command(OUTPUT ${BUILD_VERSION_CC}
|
||||
add_custom_target(GenerateBuildVersion DEPENDS ${BUILD_VERSION_CC})
|
||||
|
||||
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
|
||||
|
||||
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /WX /wd4127 /wd4800 /wd4996 /wd4351")
|
||||
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /WX /wd4018 /wd4100 /wd4101 /wd4127 /wd4189 /wd4200 /wd4244 /wd4267 /wd4296 /wd4305 /wd4307 /wd4309 /wd4512 /wd4701 /wd4702 /wd4800 /wd4804 /wd4996")
|
||||
|
||||
# Used to run CI build and tests so we can run faster
|
||||
set(OPTIMIZE_DEBUG_DEFAULT 0) # Debug build is unoptimized by default use -DOPTDBG=1 to optimize
|
||||
@@ -82,7 +69,7 @@ set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oxt /Zp8 /Gm- /Gy /MD")
|
||||
set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} /DEBUG")
|
||||
set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /DEBUG")
|
||||
|
||||
add_definitions(-DWIN32 -DOS_WIN -D_MBCS -DWIN64 -DNOMINMAX)
|
||||
add_definitions(-DWIN32 -DOS_WIN -D_MBCS -DWIN64)
|
||||
|
||||
include_directories(${PROJECT_SOURCE_DIR})
|
||||
include_directories(${PROJECT_SOURCE_DIR}/include)
|
||||
@@ -90,15 +77,13 @@ include_directories(${PROJECT_SOURCE_DIR}/third-party/gtest-1.7.0/fused-src)
|
||||
|
||||
set(ROCKSDB_LIBS rocksdblib${ARTIFACT_SUFFIX})
|
||||
set(THIRDPARTY_LIBS ${THIRDPARTY_LIBS} gtest)
|
||||
set(SYSTEM_LIBS ${SYSTEM_LIBS} Shlwapi.lib Rpcrt4.lib)
|
||||
set(SYSTEM_LIBS Shlwapi.lib Rpcrt4.lib)
|
||||
|
||||
set(LIBS ${ROCKSDB_LIBS} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
|
||||
|
||||
add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
|
||||
|
||||
# Main library source code
|
||||
set(SOURCES
|
||||
db/auto_roll_logger.cc
|
||||
db/builder.cc
|
||||
db/c.cc
|
||||
db/column_family.cc
|
||||
@@ -113,9 +98,7 @@ set(SOURCES
|
||||
db/db_impl.cc
|
||||
db/db_impl_debug.cc
|
||||
db/db_impl_experimental.cc
|
||||
db/db_impl_add_file.cc
|
||||
db/db_impl_readonly.cc
|
||||
db/db_info_dumper.cc
|
||||
db/db_iter.cc
|
||||
db/event_helpers.cc
|
||||
db/experimental.cc
|
||||
@@ -134,6 +117,7 @@ set(SOURCES
|
||||
db/merge_helper.cc
|
||||
db/merge_operator.cc
|
||||
db/repair.cc
|
||||
db/slice.cc
|
||||
db/snapshot_impl.cc
|
||||
db/table_cache.cc
|
||||
db/table_properties_collector.cc
|
||||
@@ -146,19 +130,10 @@ set(SOURCES
|
||||
db/write_batch_base.cc
|
||||
db/write_controller.cc
|
||||
db/write_thread.cc
|
||||
db/xfunc_test_points.cc
|
||||
memtable/hash_cuckoo_rep.cc
|
||||
memtable/hash_linklist_rep.cc
|
||||
memtable/hash_skiplist_rep.cc
|
||||
memtable/skiplistrep.cc
|
||||
memtable/vectorrep.cc
|
||||
port/stack_trace.cc
|
||||
port/win/io_win.cc
|
||||
port/win/env_win.cc
|
||||
port/win/env_default.cc
|
||||
port/win/port_win.cc
|
||||
port/win/win_logger.cc
|
||||
port/win/xpress_win.cc
|
||||
table/adaptive_table_factory.cc
|
||||
table/block.cc
|
||||
table/block_based_filter_block.cc
|
||||
@@ -166,6 +141,7 @@ set(SOURCES
|
||||
table/block_based_table_factory.cc
|
||||
table/block_based_table_reader.cc
|
||||
table/block_builder.cc
|
||||
table/block_hash_index.cc
|
||||
table/block_prefix_index.cc
|
||||
table/bloom_block.cc
|
||||
table/cuckoo_table_builder.cc
|
||||
@@ -179,59 +155,59 @@ set(SOURCES
|
||||
table/merger.cc
|
||||
table/sst_file_writer.cc
|
||||
table/meta_blocks.cc
|
||||
table/mock_table.cc
|
||||
table/plain_table_builder.cc
|
||||
table/plain_table_factory.cc
|
||||
table/plain_table_index.cc
|
||||
table/plain_table_key_coding.cc
|
||||
table/plain_table_reader.cc
|
||||
table/persistent_cache_helper.cc
|
||||
table/table_properties.cc
|
||||
table/two_level_iterator.cc
|
||||
tools/sst_dump_tool.cc
|
||||
tools/db_bench_tool.cc
|
||||
tools/dump/db_dump_tool.cc
|
||||
util/arena.cc
|
||||
util/auto_roll_logger.cc
|
||||
util/bloom.cc
|
||||
util/build_version.cc
|
||||
util/cf_options.cc
|
||||
util/clock_cache.cc
|
||||
util/cache.cc
|
||||
util/coding.cc
|
||||
util/compaction_job_stats_impl.cc
|
||||
util/comparator.cc
|
||||
util/concurrent_arena.cc
|
||||
util/crc32c.cc
|
||||
util/delete_scheduler.cc
|
||||
util/db_info_dumper.cc
|
||||
util/delete_scheduler_impl.cc
|
||||
util/db_test_util.cc
|
||||
util/dynamic_bloom.cc
|
||||
util/env.cc
|
||||
util/env_chroot.cc
|
||||
util/env_hdfs.cc
|
||||
util/event_logger.cc
|
||||
util/file_util.cc
|
||||
util/file_reader_writer.cc
|
||||
util/sst_file_manager_impl.cc
|
||||
util/filter_policy.cc
|
||||
util/hash.cc
|
||||
util/hash_cuckoo_rep.cc
|
||||
util/hash_linklist_rep.cc
|
||||
util/hash_skiplist_rep.cc
|
||||
util/histogram.cc
|
||||
util/histogram_windowing.cc
|
||||
util/instrumented_mutex.cc
|
||||
util/iostats_context.cc
|
||||
util/lru_cache.cc
|
||||
tools/ldb_cmd.cc
|
||||
tools/ldb_tool.cc
|
||||
util/ldb_cmd.cc
|
||||
util/ldb_tool.cc
|
||||
util/logging.cc
|
||||
util/log_buffer.cc
|
||||
util/memenv.cc
|
||||
util/mock_env.cc
|
||||
util/murmurhash.cc
|
||||
util/mutable_cf_options.cc
|
||||
util/options.cc
|
||||
util/options_builder.cc
|
||||
util/options_helper.cc
|
||||
util/options_parser.cc
|
||||
util/options_sanity_check.cc
|
||||
util/perf_context.cc
|
||||
util/perf_level.cc
|
||||
util/random.cc
|
||||
util/rate_limiter.cc
|
||||
util/sharded_cache.cc
|
||||
util/skiplistrep.cc
|
||||
util/slice.cc
|
||||
util/sst_dump_tool.cc
|
||||
util/statistics.cc
|
||||
util/status.cc
|
||||
util/status_message.cc
|
||||
@@ -240,42 +216,27 @@ set(SOURCES
|
||||
util/testharness.cc
|
||||
util/testutil.cc
|
||||
util/thread_local.cc
|
||||
util/threadpool_imp.cc
|
||||
util/thread_status_impl.cc
|
||||
util/thread_status_updater.cc
|
||||
util/thread_status_updater_debug.cc
|
||||
util/thread_status_util.cc
|
||||
util/thread_status_util_debug.cc
|
||||
util/transaction_test_util.cc
|
||||
util/vectorrep.cc
|
||||
util/xfunc.cc
|
||||
util/xxhash.cc
|
||||
utilities/backupable/backupable_db.cc
|
||||
utilities/blob_db/blob_db.cc
|
||||
utilities/checkpoint/checkpoint.cc
|
||||
utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc
|
||||
utilities/date_tiered/date_tiered_db_impl.cc
|
||||
utilities/document/document_db.cc
|
||||
utilities/document/json_document.cc
|
||||
utilities/document/json_document_builder.cc
|
||||
utilities/env_mirror.cc
|
||||
utilities/env_registry.cc
|
||||
utilities/flashcache/flashcache.cc
|
||||
utilities/geodb/geodb_impl.cc
|
||||
utilities/leveldb_options/leveldb_options.cc
|
||||
utilities/memory/memory_util.cc
|
||||
utilities/merge_operators/string_append/stringappend.cc
|
||||
utilities/merge_operators/string_append/stringappend2.cc
|
||||
utilities/merge_operators/put.cc
|
||||
utilities/merge_operators/max.cc
|
||||
utilities/merge_operators/uint64add.cc
|
||||
utilities/option_change_migration/option_change_migration.cc
|
||||
utilities/options/options_util.cc
|
||||
utilities/persistent_cache/block_cache_tier.cc
|
||||
utilities/persistent_cache/block_cache_tier_file.cc
|
||||
utilities/persistent_cache/block_cache_tier_metadata.cc
|
||||
utilities/persistent_cache/persistent_cache_tier.cc
|
||||
utilities/persistent_cache/volatile_tier_impl.cc
|
||||
utilities/redis/redis_lists.cc
|
||||
utilities/simulator_cache/sim_cache.cc
|
||||
utilities/spatialdb/spatial_db.cc
|
||||
utilities/table_properties_collectors/compact_on_deletion_collector.cc
|
||||
utilities/transactions/optimistic_transaction_impl.cc
|
||||
@@ -289,19 +250,6 @@ set(SOURCES
|
||||
utilities/ttl/db_ttl_impl.cc
|
||||
utilities/write_batch_with_index/write_batch_with_index.cc
|
||||
utilities/write_batch_with_index/write_batch_with_index_internal.cc
|
||||
utilities/col_buf_encoder.cc
|
||||
utilities/col_buf_decoder.cc
|
||||
utilities/column_aware_encoding_util.cc
|
||||
)
|
||||
|
||||
# For test util library that is build only in DEBUG mode
|
||||
# and linked to tests. Add test only code that is not #ifdefed for Release here.
|
||||
set(TESTUTIL_SOURCE
|
||||
db/db_test_util.cc
|
||||
table/mock_table.cc
|
||||
util/mock_env.cc
|
||||
util/fault_injection_test_env.cc
|
||||
util/thread_status_updater_debug.cc
|
||||
)
|
||||
|
||||
add_library(rocksdblib${ARTIFACT_SUFFIX} ${SOURCES})
|
||||
@@ -313,84 +261,53 @@ set_target_properties(rocksdb${ARTIFACT_SUFFIX} PROPERTIES COMPILE_FLAGS "-DROCK
|
||||
add_dependencies(rocksdb${ARTIFACT_SUFFIX} GenerateBuildVersion)
|
||||
target_link_libraries(rocksdb${ARTIFACT_SUFFIX} ${LIBS})
|
||||
|
||||
if (DEFINED JNI)
|
||||
if (${JNI} EQUAL 1)
|
||||
message(STATUS "JNI library is enabled")
|
||||
add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/java)
|
||||
else()
|
||||
message(STATUS "JNI library is disabled")
|
||||
endif()
|
||||
else()
|
||||
message(STATUS "JNI library is disabled")
|
||||
endif()
|
||||
|
||||
set(APPS
|
||||
tools/db_bench.cc
|
||||
db/db_bench.cc
|
||||
db/memtablerep_bench.cc
|
||||
table/table_reader_bench.cc
|
||||
tools/db_stress.cc
|
||||
tools/write_stress.cc
|
||||
tools/db_repl_stress.cc
|
||||
tools/ldb.cc
|
||||
tools/sst_dump.cc
|
||||
tools/dump/rocksdb_dump.cc
|
||||
tools/dump/rocksdb_undump.cc
|
||||
util/cache_bench.cc
|
||||
utilities/persistent_cache/hash_table_bench.cc
|
||||
utilities/column_aware_encoding_exp.cc
|
||||
)
|
||||
|
||||
set(C_TESTS db/c_test.c)
|
||||
|
||||
set(TESTS
|
||||
db/auto_roll_logger_test.cc
|
||||
db/column_family_test.cc
|
||||
db/compact_files_test.cc
|
||||
db/compaction_iterator_test.cc
|
||||
db/compaction_job_test.cc
|
||||
db/compaction_job_stats_test.cc
|
||||
db/compaction_picker_test.cc
|
||||
db/comparator_db_test.cc
|
||||
db/corruption_test.cc
|
||||
db/cuckoo_table_db_test.cc
|
||||
db/db_iter_test.cc
|
||||
db/db_test.cc
|
||||
db/db_compaction_filter_test.cc
|
||||
db/db_compaction_test.cc
|
||||
db/db_dynamic_level_test.cc
|
||||
db/db_inplace_update_test.cc
|
||||
db/db_iter_test.cc
|
||||
db/db_log_iter_test.cc
|
||||
db/db_options_test.cc
|
||||
db/db_properties_test.cc
|
||||
db/db_table_properties_test.cc
|
||||
db/db_tailing_iter_test.cc
|
||||
db/db_test.cc
|
||||
db/db_test2.cc
|
||||
db/db_block_cache_test.cc
|
||||
db/db_bloom_filter_test.cc
|
||||
db/db_flush_test.cc
|
||||
db/db_iterator_test.cc
|
||||
db/db_sst_test.cc
|
||||
db/external_sst_file_test.cc
|
||||
db/db_universal_compaction_test.cc
|
||||
db/db_wal_test.cc
|
||||
db/db_tailing_iter_test.cc
|
||||
db/dbformat_test.cc
|
||||
db/deletefile_test.cc
|
||||
db/fault_injection_test.cc
|
||||
db/file_indexer_test.cc
|
||||
db/filename_test.cc
|
||||
db/flush_job_test.cc
|
||||
db/inlineskiplist_test.cc
|
||||
db/listener_test.cc
|
||||
db/log_test.cc
|
||||
db/manual_compaction_test.cc
|
||||
db/memtable_list_test.cc
|
||||
db/merge_test.cc
|
||||
db/merge_helper_test.cc
|
||||
db/options_file_test.cc
|
||||
db/perf_context_test.cc
|
||||
db/plain_table_db_test.cc
|
||||
db/prefix_test.cc
|
||||
db/repair_test.cc
|
||||
db/skiplist_test.cc
|
||||
db/table_properties_collector_test.cc
|
||||
db/version_builder_test.cc
|
||||
@@ -400,8 +317,8 @@ set(TESTS
|
||||
db/write_batch_test.cc
|
||||
db/write_callback_test.cc
|
||||
db/write_controller_test.cc
|
||||
db/db_io_failure_test.cc
|
||||
table/block_based_filter_block_test.cc
|
||||
table/block_hash_index_test.cc
|
||||
table/block_test.cc
|
||||
table/cuckoo_table_builder_test.cc
|
||||
table/cuckoo_table_reader_test.cc
|
||||
@@ -409,48 +326,37 @@ set(TESTS
|
||||
table/merger_test.cc
|
||||
table/table_test.cc
|
||||
tools/db_sanity_test.cc
|
||||
tools/ldb_cmd_test.cc
|
||||
tools/reduce_levels_test.cc
|
||||
tools/sst_dump_test.cc
|
||||
util/arena_test.cc
|
||||
util/autovector_test.cc
|
||||
util/auto_roll_logger_test.cc
|
||||
util/bloom_test.cc
|
||||
util/cache_test.cc
|
||||
util/coding_test.cc
|
||||
util/crc32c_test.cc
|
||||
util/delete_scheduler_test.cc
|
||||
util/dynamic_bloom_test.cc
|
||||
util/env_basic_test.cc
|
||||
util/env_test.cc
|
||||
util/event_logger_test.cc
|
||||
util/filelock_test.cc
|
||||
util/file_reader_writer_test.cc
|
||||
util/heap_test.cc
|
||||
util/histogram_test.cc
|
||||
util/iostats_context_test.cc
|
||||
util/lru_cache_test.cc
|
||||
util/ldb_cmd_test.cc
|
||||
util/manual_compaction_test.cc
|
||||
util/memenv_test.cc
|
||||
util/mock_env_test.cc
|
||||
util/options_settable_test.cc
|
||||
util/options_test.cc
|
||||
util/rate_limiter_test.cc
|
||||
util/slice_transform_test.cc
|
||||
util/statistics_test.cc
|
||||
util/sst_dump_test.cc
|
||||
util/thread_list_test.cc
|
||||
util/thread_local_test.cc
|
||||
utilities/backupable/backupable_db_test.cc
|
||||
utilities/blob_db/blob_db_test.cc
|
||||
utilities/checkpoint/checkpoint_test.cc
|
||||
utilities/date_tiered/date_tiered_test.cc
|
||||
utilities/document/document_db_test.cc
|
||||
utilities/document/json_document_test.cc
|
||||
utilities/env_registry_test.cc
|
||||
utilities/geodb/geodb_test.cc
|
||||
utilities/memory/memory_test.cc
|
||||
utilities/merge_operators/string_append/stringappend_test.cc
|
||||
utilities/option_change_migration/option_change_migration_test.cc
|
||||
utilities/options/options_util_test.cc
|
||||
utilities/persistent_cache/hash_table_test.cc
|
||||
utilities/persistent_cache/persistent_cache_test.cc
|
||||
utilities/redis/redis_lists_test.cc
|
||||
utilities/spatialdb/spatial_db_test.cc
|
||||
utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
|
||||
@@ -458,10 +364,9 @@ set(TESTS
|
||||
utilities/transactions/transaction_test.cc
|
||||
utilities/ttl/ttl_test.cc
|
||||
utilities/write_batch_with_index/write_batch_with_index_test.cc
|
||||
utilities/column_aware_encoding_test.cc
|
||||
)
|
||||
|
||||
set(EXES ${APPS})
|
||||
set(EXES ${APPS} ${TESTS})
|
||||
|
||||
foreach(sourcefile ${EXES})
|
||||
string(REPLACE ".cc" "" exename ${sourcefile})
|
||||
@@ -470,42 +375,12 @@ foreach(sourcefile ${EXES})
|
||||
target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS})
|
||||
endforeach(sourcefile ${EXES})
|
||||
|
||||
# test utilities are only build in debug
|
||||
set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
|
||||
add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
|
||||
set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
|
||||
set_target_properties(${TESTUTILLIB}
|
||||
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
|
||||
)
|
||||
|
||||
# Tests are excluded from Release builds
|
||||
set(TEST_EXES ${TESTS})
|
||||
|
||||
foreach(sourcefile ${TEST_EXES})
|
||||
string(REPLACE ".cc" "" exename ${sourcefile})
|
||||
string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
|
||||
add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
|
||||
set_target_properties(${exename}${ARTIFACT_SUFFIX}
|
||||
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
|
||||
)
|
||||
target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS} testutillib${ARTIFACT_SUFFIX})
|
||||
endforeach(sourcefile ${TEST_EXES})
|
||||
|
||||
# C executables must link to a shared object
|
||||
set(C_TEST_EXES ${C_TESTS})
|
||||
set(C_EXES ${C_TESTS})
|
||||
|
||||
foreach(sourcefile ${C_TEST_EXES})
|
||||
foreach(sourcefile ${C_EXES})
|
||||
string(REPLACE ".c" "" exename ${sourcefile})
|
||||
string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
|
||||
add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
|
||||
set_target_properties(${exename}${ARTIFACT_SUFFIX}
|
||||
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
|
||||
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
|
||||
)
|
||||
target_link_libraries(${exename}${ARTIFACT_SUFFIX} rocksdb${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX})
|
||||
endforeach(sourcefile ${C_TEST_EXES})
|
||||
target_link_libraries(${exename}${ARTIFACT_SUFFIX} rocksdb${ARTIFACT_SUFFIX})
|
||||
endforeach(sourcefile ${C_TESTS})
|
||||
@@ -1,14 +0,0 @@
|
||||
# RocksDB default options change log
|
||||
## 4.8.0 (5/2/2016)
|
||||
* options.max_open_files changes from 5000 to -1. It improves performance, but users need to set file descriptor limit to be large enough and watch memory usage for index and bloom filters.
|
||||
* options.base_background_compactions changes from max_background_compactions to 1. When users set higher max_background_compactions but the write throughput is not high, the writes are less spiky to disks.
|
||||
* options.wal_recovery_mode changes from kTolerateCorruptedTailRecords to kPointInTimeRecovery. Avoid some false positive when file system or hardware reorder the writes for file data and metadata.
|
||||
|
||||
## 4.7.0 (4/8/2016)
|
||||
* options.write_buffer_size changes from 4MB to 64MB.
|
||||
* options.target_file_size_base changes from 2MB to 64MB.
|
||||
* options.max_bytes_for_level_base changes from 10MB to 256MB.
|
||||
* options.soft_pending_compaction_bytes_limit changes from 0 (disabled) to 64GB.
|
||||
* options.hard_pending_compaction_bytes_limit changes from 0 (disabled) to 256GB.
|
||||
* table_cache_numshardbits changes from 4 to 6.
|
||||
* max_file_opening_threads changes from 1 to 16.
|
||||
+8
-119
@@ -1,115 +1,4 @@
|
||||
# Rocksdb Change Log
|
||||
## 4.12.0 (9/12/2016)
|
||||
### Public API Change
|
||||
* CancelAllBackgroundWork() flushes all memtables for databases containing writes that have bypassed the WAL (writes issued with WriteOptions::disableWAL=true) before shutting down background threads.
|
||||
* Merge options source_compaction_factor, max_grandparent_overlap_bytes and expanded_compaction_factor into max_compaction_bytes.
|
||||
* Remove ImmutableCFOptions.
|
||||
* Add a compression type ZSTD, which can work with ZSTD 0.8.0 or up. Still keep ZSTDNotFinal for compatibility reasons.
|
||||
|
||||
### New Features
|
||||
* Introduce NewClockCache, which is based on CLOCK algorithm with better concurrent performance in some cases. It can be used to replace the default LRU-based block cache and table cache. To use it, RocksDB need to be linked with TBB lib.
|
||||
* Change ticker/histogram statistics implementations to accumulate data in thread-local storage, which improves CPU performance by reducing cache coherency costs. Callers of CreateDBStatistics do not need to change anything to use this feature.
|
||||
* Block cache mid-point insertion, where index and filter block are inserted into LRU block cache with higher priority. The feature can be enabled by setting BlockBasedTableOptions::cache_index_and_filter_blocks_with_high_priority to true and high_pri_pool_ratio > 0 when creating NewLRUCache.
|
||||
|
||||
## 4.11.0 (8/1/2016)
|
||||
### Public API Change
|
||||
* options.memtable_prefix_bloom_huge_page_tlb_size => memtable_huge_page_size. When it is set, RocksDB will try to allocate memory from huge page for memtable too, rather than just memtable bloom filter.
|
||||
|
||||
### New Features
|
||||
* A tool to migrate DB after options change. See include/rocksdb/utilities/option_change_migration.h.
|
||||
* Add ReadOptions.background_purge_on_iterator_cleanup. If true, we avoid file deletion when destorying iterators.
|
||||
|
||||
## 4.10.0 (7/5/2016)
|
||||
### Public API Change
|
||||
* options.memtable_prefix_bloom_bits changes to options.memtable_prefix_bloom_bits_ratio and deprecate options.memtable_prefix_bloom_probes
|
||||
* enum type CompressionType and PerfLevel changes from char to unsigned char. Value of all PerfLevel shift by one.
|
||||
* Deprecate options.filter_deletes.
|
||||
|
||||
### New Features
|
||||
* Add avoid_flush_during_recovery option.
|
||||
* Add a read option background_purge_on_iterator_cleanup to avoid deleting files in foreground when destroying iterators. Instead, a job is scheduled in high priority queue and would be executed in a separate background thread.
|
||||
* RepairDB support for column families. RepairDB now associates data with non-default column families using information embedded in the SST/WAL files (4.7 or later). For data written by 4.6 or earlier, RepairDB associates it with the default column family.
|
||||
* Add options.write_buffer_manager which allows users to control total memtable sizes across multiple DB instances.
|
||||
|
||||
## 4.9.0 (6/9/2016)
|
||||
### Public API changes
|
||||
* Add bottommost_compression option, This option can be used to set a specific compression algorithm for the bottommost level (Last level containing files in the DB).
|
||||
* Introduce CompactionJobInfo::compression, This field state the compression algorithm used to generate the output files of the compaction.
|
||||
* Deprecate BlockBaseTableOptions.hash_index_allow_collision=false
|
||||
* Deprecate options builder (GetOptions()).
|
||||
|
||||
### New Features
|
||||
* Introduce NewSimCache() in rocksdb/utilities/sim_cache.h. This function creates a block cache that is able to give simulation results (mainly hit rate) of simulating block behavior with a configurable cache size.
|
||||
|
||||
## 4.8.0 (5/2/2016)
|
||||
### Public API Change
|
||||
* Allow preset compression dictionary for improved compression of block-based tables. This is supported for zlib, zstd, and lz4. The compression dictionary's size is configurable via CompressionOptions::max_dict_bytes.
|
||||
* Delete deprecated classes for creating backups (BackupableDB) and restoring from backups (RestoreBackupableDB). Now, BackupEngine should be used for creating backups, and BackupEngineReadOnly should be used for restorations. For more details, see https://github.com/facebook/rocksdb/wiki/How-to-backup-RocksDB%3F
|
||||
* Expose estimate of per-level compression ratio via DB property: "rocksdb.compression-ratio-at-levelN".
|
||||
* Added EventListener::OnTableFileCreationStarted. EventListener::OnTableFileCreated will be called on failure case. User can check creation status via TableFileCreationInfo::status.
|
||||
|
||||
### New Features
|
||||
* Add ReadOptions::readahead_size. If non-zero, NewIterator will create a new table reader which performs reads of the given size.
|
||||
|
||||
## 4.7.0 (4/8/2016)
|
||||
### Public API Change
|
||||
* rename options compaction_measure_io_stats to report_bg_io_stats and include flush too.
|
||||
* Change some default options. Now default options will optimize for server-workloads. Also enable slowdown and full stop triggers for pending compaction bytes. These changes may cause sub-optimal performance or significant increase of resource usage. To avoid these risks, users can open existing RocksDB with options extracted from RocksDB option files. See https://github.com/facebook/rocksdb/wiki/RocksDB-Options-File for how to use RocksDB option files. Or you can call Options.OldDefaults() to recover old defaults. DEFAULT_OPTIONS_HISTORY.md will track change history of default options.
|
||||
|
||||
## 4.6.0 (3/10/2016)
|
||||
### Public API Changes
|
||||
* Change default of BlockBasedTableOptions.format_version to 2. It means default DB created by 4.6 or up cannot be opened by RocksDB version 3.9 or earlier.
|
||||
* Added strict_capacity_limit option to NewLRUCache. If the flag is set to true, insert to cache will fail if no enough capacity can be free. Signature of Cache::Insert() is updated accordingly.
|
||||
* Tickers [NUMBER_DB_NEXT, NUMBER_DB_PREV, NUMBER_DB_NEXT_FOUND, NUMBER_DB_PREV_FOUND, ITER_BYTES_READ] are not updated immediately. The are updated when the Iterator is deleted.
|
||||
* Add monotonically increasing counter (DB property "rocksdb.current-super-version-number") that increments upon any change to the LSM tree.
|
||||
|
||||
### New Features
|
||||
* Add CompactionPri::kMinOverlappingRatio, a compaction picking mode friendly to write amplification.
|
||||
* Deprecate Iterator::IsKeyPinned() and replace it with Iterator::GetProperty() with prop_name="rocksdb.iterator.is.key.pinned"
|
||||
|
||||
## 4.5.0 (2/5/2016)
|
||||
### Public API Changes
|
||||
* Add a new perf context level between kEnableCount and kEnableTime. Level 2 now does not include timers for mutexes.
|
||||
* Statistics of mutex operation durations will not be measured by default. If you want to have them enabled, you need to set Statistics::stats_level_ to kAll.
|
||||
* DBOptions::delete_scheduler and NewDeleteScheduler() are removed, please use DBOptions::sst_file_manager and NewSstFileManager() instead
|
||||
|
||||
### New Features
|
||||
* ldb tool now supports operations to non-default column families.
|
||||
* Add kPersistedTier to ReadTier. This option allows Get and MultiGet to read only the persited data and skip mem-tables if writes were done with disableWAL = true.
|
||||
* Add DBOptions::sst_file_manager. Use NewSstFileManager() in include/rocksdb/sst_file_manager.h to create a SstFileManager that can be used to track the total size of SST files and control the SST files deletion rate.
|
||||
|
||||
## 4.4.0 (1/14/2016)
|
||||
### Public API Changes
|
||||
* Change names in CompactionPri and add a new one.
|
||||
* Deprecate options.soft_rate_limit and add options.soft_pending_compaction_bytes_limit.
|
||||
* If options.max_write_buffer_number > 3, writes will be slowed down when writing to the last write buffer to delay a full stop.
|
||||
* Introduce CompactionJobInfo::compaction_reason, this field include the reason to trigger the compaction.
|
||||
* After slow down is triggered, if estimated pending compaction bytes keep increasing, slowdown more.
|
||||
* Increase default options.delayed_write_rate to 2MB/s.
|
||||
* Added a new parameter --path to ldb tool. --path accepts the name of either MANIFEST, SST or a WAL file. Either --db or --path can be used when calling ldb.
|
||||
|
||||
## 4.3.0 (12/8/2015)
|
||||
### New Features
|
||||
* CompactionFilter has new member function called IgnoreSnapshots which allows CompactionFilter to be called even if there are snapshots later than the key.
|
||||
* RocksDB will now persist options under the same directory as the RocksDB database on successful DB::Open, CreateColumnFamily, DropColumnFamily, and SetOptions.
|
||||
* Introduce LoadLatestOptions() in rocksdb/utilities/options_util.h. This function can construct the latest DBOptions / ColumnFamilyOptions used by the specified RocksDB intance.
|
||||
* Introduce CheckOptionsCompatibility() in rocksdb/utilities/options_util.h. This function checks whether the input set of options is able to open the specified DB successfully.
|
||||
|
||||
### Public API Changes
|
||||
* When options.db_write_buffer_size triggers, only the column family with the largest column family size will be flushed, not all the column families.
|
||||
|
||||
## 4.2.0 (11/9/2015)
|
||||
### New Features
|
||||
* Introduce CreateLoggerFromOptions(), this function create a Logger for provided DBOptions.
|
||||
* Add GetAggregatedIntProperty(), which returns the sum of the GetIntProperty of all the column families.
|
||||
* Add MemoryUtil in rocksdb/utilities/memory.h. It currently offers a way to get the memory usage by type from a list rocksdb instances.
|
||||
|
||||
### Public API Changes
|
||||
* CompactionFilter::Context includes information of Column Family ID
|
||||
* The need-compaction hint given by TablePropertiesCollector::NeedCompact() will be persistent and recoverable after DB recovery. This introduces a breaking format change. If you use this experimental feature, including NewCompactOnDeletionCollectorFactory() in the new version, you may not be able to directly downgrade the DB back to version 4.0 or lower.
|
||||
* TablePropertiesCollectorFactory::CreateTablePropertiesCollector() now takes an option Context, containing the information of column family ID for the file being written.
|
||||
* Remove DefaultCompactionFilterFactory.
|
||||
|
||||
|
||||
## 4.1.0 (10/8/2015)
|
||||
### New Features
|
||||
@@ -175,8 +64,8 @@
|
||||
* options.hard_rate_limit is deprecated.
|
||||
* When options.soft_rate_limit or options.level0_slowdown_writes_trigger is triggered, the way to slow down writes is changed to: write rate to DB is limited to to options.delayed_write_rate.
|
||||
* DB::GetApproximateSizes() adds a parameter to allow the estimation to include data in mem table, with default to be not to include. It is now only supported in skip list mem table.
|
||||
* DB::CompactRange() now accept CompactRangeOptions instead of multiple parameters. CompactRangeOptions is defined in include/rocksdb/options.h.
|
||||
* CompactRange() will now skip bottommost level compaction for level based compaction if there is no compaction filter, bottommost_level_compaction is introduced in CompactRangeOptions to control when it's possible to skip bottommost level compaction. This mean that if you want the compaction to produce a single file you need to set bottommost_level_compaction to BottommostLevelCompaction::kForce.
|
||||
* DB::CompactRange() now accept CompactRangeOptions instead of multiple paramters. CompactRangeOptions is defined in include/rocksdb/options.h.
|
||||
* CompactRange() will now skip bottommost level compaction for level based compaction if there is no compaction filter, bottommost_level_compaction is introduced in CompactRangeOptions to control when it's possbile to skip bottommost level compaction. This mean that if you want the compaction to produce a single file you need to set bottommost_level_compaction to BottommostLevelCompaction::kForce.
|
||||
* Add Cache.GetPinnedUsage() to get the size of memory occupied by entries that are in use by the system.
|
||||
* DB:Open() will fail if the compression specified in Options is not linked with the binary. If you see this failure, recompile RocksDB with compression libraries present on your system. Also, previously our default compression was snappy. This behavior is now changed. Now, the default compression is snappy only if it's available on the system. If it isn't we change the default to kNoCompression.
|
||||
* We changed how we account for memory used in block cache. Previously, we only counted the sum of block sizes currently present in block cache. Now, we count the actual memory usage of the blocks. For example, a block of size 4.5KB will use 8KB memory with jemalloc. This might decrease your memory usage and possibly decrease performance. Increase block cache size if you see this happening after an upgrade.
|
||||
@@ -208,7 +97,7 @@
|
||||
Lower numbered levels will be placed earlier in the db_paths and higher
|
||||
numbered levels will be placed later in the db_paths vector.
|
||||
* Potentially big performance improvements if you're using RocksDB with lots of column families (100-1000)
|
||||
* Added BlockBasedTableOptions.format_version option, which allows user to specify which version of block based table he wants. As a general guideline, newer versions have more features, but might not be readable by older versions of RocksDB.
|
||||
* Added BlockBasedTableOptions.format_version option, which allows user to specify which version of block based table he wants. As a general guidline, newer versions have more features, but might not be readable by older versions of RocksDB.
|
||||
* Added new block based table format (version 2), which you can enable by setting BlockBasedTableOptions.format_version = 2. This format changes how we encode size information in compressed blocks and should help with memory allocations if you're using Zlib or BZip2 compressions.
|
||||
* MemEnv (env that stores data in memory) is now available in default library build. You can create it by calling NewMemEnv().
|
||||
* Add SliceTransform.SameResultWhenAppended() to help users determine it is safe to apply prefix bloom/hash.
|
||||
@@ -284,7 +173,7 @@
|
||||
|
||||
## 3.5.0 (9/3/2014)
|
||||
### New Features
|
||||
* Add include/utilities/write_batch_with_index.h, providing a utility class to query data out of WriteBatch when building it.
|
||||
* Add include/utilities/write_batch_with_index.h, providing a utilitiy class to query data out of WriteBatch when building it.
|
||||
* Move BlockBasedTable related options to BlockBasedTableOptions from Options. Change corresponding JNI interface. Options affected include:
|
||||
no_block_cache, block_cache, block_cache_compressed, block_size, block_size_deviation, block_restart_interval, filter_policy, whole_key_filtering. filter_policy is changed to shared_ptr from a raw pointer.
|
||||
* Remove deprecated options: disable_seek_compaction and db_stats_log_interval
|
||||
@@ -298,7 +187,7 @@
|
||||
* Support Multiple DB paths in universal style compactions
|
||||
* Add feature of storing plain table index and bloom filter in SST file.
|
||||
* CompactRange() will never output compacted files to level 0. This used to be the case when all the compaction input files were at level 0.
|
||||
* Added iterate_upper_bound to define the extent upto which the forward iterator will return entries. This will prevent iterating over delete markers and overwritten entries for edge cases where you want to break out the iterator anyways. This may improve performance in case there are a large number of delete markers or overwritten entries.
|
||||
* Added iterate_upper_bound to define the extent upto which the forward iterator will return entries. This will prevent iterating over delete markers and overwritten entries for edge cases where you want to break out the iterator anyways. This may improve perfomance in case there are a large number of delete markers or overwritten entries.
|
||||
|
||||
### Public API changes
|
||||
* DBOptions.db_paths now is a vector of a DBPath structure which indicates both of path and target size
|
||||
@@ -313,7 +202,7 @@
|
||||
### New Features
|
||||
* Added JSON API prototype.
|
||||
* HashLinklist reduces performance outlier caused by skewed bucket by switching data in the bucket from linked list to skip list. Add parameter threshold_use_skiplist in NewHashLinkListRepFactory().
|
||||
* RocksDB is now able to reclaim storage space more effectively during the compaction process. This is done by compensating the size of each deletion entry by the 2X average value size, which makes compaction to be triggered by deletion entries more easily.
|
||||
* RocksDB is now able to reclaim storage space more effectively during the compaction process. This is done by compensating the size of each deletion entry by the 2X average value size, which makes compaction to be triggerred by deletion entries more easily.
|
||||
* Add TimeOut API to write. Now WriteOptions have a variable called timeout_hint_us. With timeout_hint_us set to non-zero, any write associated with this timeout_hint_us may be aborted when it runs longer than the specified timeout_hint_us, and it is guaranteed that any write completes earlier than the specified time-out will not be aborted due to the time-out condition.
|
||||
* Add a rate_limiter option, which controls total throughput of flush and compaction. The throughput is specified in bytes/sec. Flush always has precedence over compaction when available bandwidth is constrained.
|
||||
|
||||
@@ -328,11 +217,11 @@
|
||||
2) It added some complexity to the important code-paths,
|
||||
3) None of our internal customers were really using it.
|
||||
Because of that, Options::disable_seek_compaction is now obsolete. It is still a parameter in Options, so it does not break the build, but it does not have any effect. We plan to completely remove it at some point, so we ask users to please remove this option from your code base.
|
||||
* Add two parameters to NewHashLinkListRepFactory() for logging on too many entries in a hash bucket when flushing.
|
||||
* Add two paramters to NewHashLinkListRepFactory() for logging on too many entries in a hash bucket when flushing.
|
||||
* Added new option BlockBasedTableOptions::hash_index_allow_collision. When enabled, prefix hash index for block-based table will not store prefix and allow hash collision, reducing memory consumption.
|
||||
|
||||
### New Features
|
||||
* PlainTable now supports a new key encoding: for keys of the same prefix, the prefix is only written once. It can be enabled through encoding_type parameter of NewPlainTableFactory()
|
||||
* PlainTable now supports a new key encoding: for keys of the same prefix, the prefix is only written once. It can be enabled through encoding_type paramter of NewPlainTableFactory()
|
||||
* Add AdaptiveTableFactory, which is used to convert from a DB of PlainTable to BlockBasedTabe, or vise versa. It can be created using NewAdaptiveTableFactory()
|
||||
|
||||
### Performance Improvements
|
||||
|
||||
+5
-11
@@ -1,27 +1,22 @@
|
||||
## Compilation
|
||||
|
||||
**Important**: If you plan to run RocksDB in production, don't compile using default
|
||||
`make` or `make all`. That will compile RocksDB in debug mode, which is much slower
|
||||
than release mode.
|
||||
|
||||
RocksDB's library should be able to compile without any dependency installed,
|
||||
although we recommend installing some compression libraries (see below).
|
||||
We do depend on newer gcc/clang with C++11 support.
|
||||
|
||||
There are few options when compiling RocksDB:
|
||||
|
||||
* [recommended] `make static_lib` will compile librocksdb.a, RocksDB static library. Compiles static library in release mode.
|
||||
* [recommended] `make static_lib` will compile librocksdb.a, RocksDB static library.
|
||||
|
||||
* `make shared_lib` will compile librocksdb.so, RocksDB shared library. Compiles shared library in release mode.
|
||||
* `make shared_lib` will compile librocksdb.so, RocksDB shared library.
|
||||
|
||||
* `make check` will compile and run all the unit tests. `make check` will compile RocksDB in debug mode.
|
||||
* `make check` will compile and run all the unit tests
|
||||
|
||||
* `make all` will compile our static library, and all our tools and unit tests. Our tools
|
||||
depend on gflags. You will need to have gflags installed to run `make all`. This will compile RocksDB in debug mode. Don't
|
||||
use binaries compiled by `make all` in production.
|
||||
depend on gflags. You will need to have gflags installed to run `make all`.
|
||||
|
||||
* By default the binary we produce is optimized for the platform you're compiling on
|
||||
(-march=native or the equivalent). If you want to build a portable binary, add 'PORTABLE=1' before
|
||||
(-march=native). If you want to build a portable binary, add 'PORTABLE=1' before
|
||||
your make commands, like this: `PORTABLE=1 make static_lib`
|
||||
|
||||
## Dependencies
|
||||
@@ -87,5 +82,4 @@ your make commands, like this: `PORTABLE=1 make static_lib`
|
||||
* Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define two important pre-processing macros: `ROCKSDB_LITE` and `IOS_CROSS_COMPILE`.
|
||||
|
||||
* **Windows**:
|
||||
* For building with MS Visual Studio 13 you will need Update 4 installed.
|
||||
* Read and follow the instructions at CMakeLists.txt
|
||||
@@ -1,12 +0,0 @@
|
||||
This is the list of all known third-party language bindings for RocksDB. If something is missing, please open a pull request to add it.
|
||||
|
||||
* Java - https://github.com/facebook/rocksdb/tree/master/java
|
||||
* Python - http://pyrocksdb.readthedocs.org/en/latest/
|
||||
* Perl - https://metacpan.org/pod/RocksDB
|
||||
* Node.js - https://npmjs.org/package/rocksdb
|
||||
* Go - https://github.com/tecbot/gorocksdb
|
||||
* Ruby - http://rubygems.org/gems/rocksdb-ruby
|
||||
* Haskell - https://hackage.haskell.org/package/rocksdb-haskell
|
||||
* PHP - https://github.com/Photonios/rocksdb-php
|
||||
* C# - https://github.com/warrenfalk/rocksdb-sharp
|
||||
* Rust - https://github.com/spacejam/rust-rocksdb
|
||||
@@ -2,7 +2,7 @@ BSD License
|
||||
|
||||
For rocksdb software
|
||||
|
||||
Copyright (c) 2011-present, Facebook, Inc.
|
||||
Copyright (c) 2014, Facebook, Inc.
|
||||
All rights reserved.
|
||||
---------------------------------------------------------------------
|
||||
|
||||
|
||||
@@ -1,8 +1,6 @@
|
||||
## RocksDB: A Persistent Key-Value Store for Flash and RAM Storage
|
||||
|
||||
[](https://travis-ci.org/facebook/rocksdb)
|
||||
[](https://ci.appveyor.com/project/Facebook/rocksdb/branch/master)
|
||||
|
||||
|
||||
RocksDB is developed and maintained by Facebook Database Engineering Team.
|
||||
It is built on earlier work on LevelDB by Sanjay Ghemawat (sanjay@google.com)
|
||||
|
||||
@@ -1,21 +1,15 @@
|
||||
This document lists users of RocksDB and their use cases. If you are using RocksDB, please open a pull request and add yourself to the list.
|
||||
|
||||
## Facebook
|
||||
At Facebook, we use RocksDB as storage engines in multiple data management services and a backend for many different stateful services, including:
|
||||
At Facebook, we use RocksDB as a backend for many different stateful services. We're also experimenting with running RocksDB as a storage engine for two databases:
|
||||
|
||||
1. MyRocks -- https://github.com/MySQLOnRocksDB/mysql-5.6
|
||||
2. MongoRocks -- https://github.com/mongodb-partners/mongo-rocks
|
||||
3. ZippyDB -- Facebook's distributed key-value store with Paxos-style replication, built on top of RocksDB.[*] https://www.youtube.com/watch?v=DfiN7pG0D0khtt
|
||||
4. Laser -- Laser is a high query throughput, low (millisecond) latency, key-value storage service built on top of RocksDB.[*]
|
||||
4. Dragan -- a distributed graph query engine. https://code.facebook.com/posts/1737605303120405/dragon-a-distributed-graph-query-engine/
|
||||
5. Stylus -- a low-level stream processing framework writtenin C++.[*]
|
||||
|
||||
[*] https://research.facebook.com/publications/realtime-data-processing-at-facebook/
|
||||
|
||||
## LinkedIn
|
||||
Two different use cases at Linkedin are using RocksDB as a storage engine:
|
||||
|
||||
1. LinkedIn's follow feed for storing user's activities. Check out the blog post: https://engineering.linkedin.com/blog/2016/03/followfeed--linkedin-s-feed-made-faster-and-smarter
|
||||
1. LinkedIn's follow feed for storing user's activities
|
||||
2. Apache Samza, open source framework for stream processing
|
||||
|
||||
Learn more about those use cases in a Tech Talk by Ankit Gupta and Naveen Somasundaram: http://www.youtube.com/watch?v=plqVp_OnSzg
|
||||
@@ -43,31 +37,3 @@ Check out our RocksDB Protobuf merge operator at: https://github.com/vladb38/roc
|
||||
|
||||
## Santanader UK/Cloudera Profession Services
|
||||
Check out their blog post: http://blog.cloudera.com/blog/2015/08/inside-santanders-near-real-time-data-ingest-architecture/
|
||||
|
||||
## Airbnb
|
||||
Airbnb is using RocksDB as a storage engine for their personalized search service. You can learn more about it here: https://www.youtube.com/watch?v=ASQ6XMtogMs
|
||||
|
||||
## Pinterest
|
||||
Pinterest's Object Retrieval System uses RocksDB for storage: https://www.youtube.com/watch?v=MtFEVEs_2Vo
|
||||
|
||||
## Smyte
|
||||
[Smyte](https://www.smyte.com/) uses RocksDB as the storage layer for their core key-value storage, high-performance counters and time-windowed HyperLogLog services.
|
||||
|
||||
## Rakuten Marketing
|
||||
[Rakuten Marketing](https://marketing.rakuten.com/) uses RocksDB as the disk cache layer for the real-time bidding service in their Performance DSP.
|
||||
|
||||
## VWO, Wingify
|
||||
[VWO's](https://vwo.com/) Smart Code checker and URL helper uses RocksDB to store all the URLs where VWO's Smart Code is installed.
|
||||
|
||||
## quasardb
|
||||
[quasardb](https://www.quasardb.net) is a high-performance, distributed, transactional key-value database that integrates well with in-memory analytics engines such as Apache Spark.
|
||||
quasardb uses a heavily tuned RocksDB as its persistence layer.
|
||||
|
||||
## Netflix
|
||||
[Netflix](http://techblog.netflix.com/2016/05/application-data-caching-using-ssds.html) uses RocksDB on spinning disks to cache application data.
|
||||
|
||||
## TiKV
|
||||
[TiKV](https://github.com/pingcap/tikv) is a GEO-replicated, high-performance, distributed, transactional key-value database. TiKV is powered by Rust and Raft. TiKV uses RocksDB as its persistence layer.
|
||||
|
||||
## Apache Flink
|
||||
[Apache Flink](https://flink.apache.org/news/2016/03/08/release-1.0.0.html) uses RocksDB to store state locally on a machine.
|
||||
+3
-9
@@ -1,17 +1,11 @@
|
||||
version: 1.0.{build}
|
||||
image: Visual Studio 2015
|
||||
before_build:
|
||||
- md %APPVEYOR_BUILD_FOLDER%\build
|
||||
- cd %APPVEYOR_BUILD_FOLDER%\build
|
||||
- cmake -G "Visual Studio 14 2015 Win64" -DOPTDBG=1 -DXPRESS=1 ..
|
||||
- cmake -G "Visual Studio 12 Win64" ..
|
||||
- cd ..
|
||||
build:
|
||||
project: build\rocksdb.sln
|
||||
project: build\ALL_BUILD.vcxproj
|
||||
parallel: true
|
||||
verbosity: minimal
|
||||
test:
|
||||
test_script:
|
||||
- ps: build_tools\run_ci_db_test.ps1 -EnableRerun -Run db_test2 -Concurrency 8
|
||||
- ps: build_tools\run_ci_db_test.ps1 -EnableRerun -Run db_test -Exclude DBTest.GroupCommitTest -Concurrency 10
|
||||
- ps: build_tools\run_ci_db_test.ps1 -Run env_test -Concurrency 1
|
||||
|
||||
test: off
|
||||
@@ -0,0 +1,22 @@
|
||||
version: 1.0.{build}
|
||||
before_build:
|
||||
- md %APPVEYOR_BUILD_FOLDER%\build
|
||||
- cd %APPVEYOR_BUILD_FOLDER%\build
|
||||
- cmake -G "Visual Studio 12 Win64" -DOPTDBG=1 ..
|
||||
- cd ..
|
||||
build:
|
||||
project: build\ALL_BUILD.vcxproj
|
||||
parallel: true
|
||||
verbosity: minimal
|
||||
test:
|
||||
test_script:
|
||||
- ps: build_tools\run_ci_db_test.ps1
|
||||
notifications:
|
||||
- provider: Email
|
||||
to:
|
||||
- svmtrocksdb@microsoft.com
|
||||
subject: "Build {{status}}"
|
||||
message: "{{message}}, {{commitId}}, ..."
|
||||
on_build_success: false
|
||||
on_build_failure: true
|
||||
on_build_status_changed: true
|
||||
@@ -6,66 +6,33 @@
|
||||
* @phutil-library-version 2
|
||||
*/
|
||||
|
||||
if (class_exists('ArcanistWorkflow')) {
|
||||
phutil_register_library_map(array(
|
||||
'__library_version__' => 2,
|
||||
'class' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'cpp_linter/ArcanistCpplintLinter.php',
|
||||
'BaseDirectoryScopedFormatLinter' => 'cpp_linter/BaseDirectoryScopedFormatLinter.php',
|
||||
'FacebookArcanistConfiguration' => 'config/FacebookArcanistConfiguration.php',
|
||||
'FacebookFbcodeLintEngine' => 'lint_engine/FacebookFbcodeLintEngine.php',
|
||||
'FacebookFbcodeUnitTestEngine' => 'unit_engine/FacebookFbcodeUnitTestEngine.php',
|
||||
'FacebookHowtoevenLintEngine' => 'lint_engine/FacebookHowtoevenLintEngine.php',
|
||||
'FacebookHowtoevenLinter' => 'cpp_linter/FacebookHowtoevenLinter.php',
|
||||
'FbcodeClangFormatLinter' => 'cpp_linter/FbcodeClangFormatLinter.php',
|
||||
'FbcodeCppLinter' => 'cpp_linter/FbcodeCppLinter.php',
|
||||
),
|
||||
'function' =>
|
||||
array(
|
||||
),
|
||||
'xmap' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'ArcanistLinter',
|
||||
'BaseDirectoryScopedFormatLinter' => 'ArcanistLinter',
|
||||
'FacebookArcanistConfiguration' => 'ArcanistConfiguration',
|
||||
'FacebookFbcodeLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookFbcodeUnitTestEngine' => 'ArcanistBaseUnitTestEngine',
|
||||
'FacebookHowtoevenLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookHowtoevenLinter' => 'ArcanistLinter',
|
||||
'FbcodeClangFormatLinter' => 'BaseDirectoryScopedFormatLinter',
|
||||
'FbcodeCppLinter' => 'ArcanistLinter',
|
||||
),
|
||||
));
|
||||
} else {
|
||||
phutil_register_library_map(array(
|
||||
'__library_version__' => 2,
|
||||
'class' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'cpp_linter/ArcanistCpplintLinter.php',
|
||||
'BaseDirectoryScopedFormatLinter' => 'cpp_linter/BaseDirectoryScopedFormatLinter.php',
|
||||
'FacebookArcanistConfiguration' => 'config/FacebookOldArcanistConfiguration.php',
|
||||
'FacebookFbcodeLintEngine' => 'lint_engine/FacebookFbcodeLintEngine.php',
|
||||
'FacebookFbcodeUnitTestEngine' => 'unit_engine/FacebookOldFbcodeUnitTestEngine.php',
|
||||
'FacebookHowtoevenLintEngine' => 'lint_engine/FacebookHowtoevenLintEngine.php',
|
||||
'FacebookHowtoevenLinter' => 'cpp_linter/FacebookHowtoevenLinter.php',
|
||||
'FbcodeClangFormatLinter' => 'cpp_linter/FbcodeClangFormatLinter.php',
|
||||
'FbcodeCppLinter' => 'cpp_linter/FbcodeCppLinter.php',
|
||||
),
|
||||
'function' =>
|
||||
array(
|
||||
),
|
||||
'xmap' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'ArcanistLinter',
|
||||
'BaseDirectoryScopedFormatLinter' => 'ArcanistLinter',
|
||||
'FacebookArcanistConfiguration' => 'ArcanistConfiguration',
|
||||
'FacebookFbcodeLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookFbcodeUnitTestEngine' => 'ArcanistBaseUnitTestEngine',
|
||||
'FacebookHowtoevenLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookHowtoevenLinter' => 'ArcanistLinter',
|
||||
'FbcodeClangFormatLinter' => 'BaseDirectoryScopedFormatLinter',
|
||||
'FbcodeCppLinter' => 'ArcanistLinter',
|
||||
),
|
||||
));
|
||||
}
|
||||
phutil_register_library_map(array(
|
||||
'__library_version__' => 2,
|
||||
'class' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'cpp_linter/ArcanistCpplintLinter.php',
|
||||
'BaseDirectoryScopedFormatLinter' => 'cpp_linter/BaseDirectoryScopedFormatLinter.php',
|
||||
'FacebookArcanistConfiguration' => 'config/FacebookArcanistConfiguration.php',
|
||||
'FacebookFbcodeLintEngine' => 'lint_engine/FacebookFbcodeLintEngine.php',
|
||||
'FacebookFbcodeUnitTestEngine' => 'unit_engine/FacebookFbcodeUnitTestEngine.php',
|
||||
'FacebookHowtoevenLintEngine' => 'lint_engine/FacebookHowtoevenLintEngine.php',
|
||||
'FacebookHowtoevenLinter' => 'cpp_linter/FacebookHowtoevenLinter.php',
|
||||
'FbcodeClangFormatLinter' => 'cpp_linter/FbcodeClangFormatLinter.php',
|
||||
'FbcodeCppLinter' => 'cpp_linter/FbcodeCppLinter.php',
|
||||
),
|
||||
'function' =>
|
||||
array(
|
||||
),
|
||||
'xmap' =>
|
||||
array(
|
||||
'ArcanistCpplintLinter' => 'ArcanistLinter',
|
||||
'BaseDirectoryScopedFormatLinter' => 'ArcanistLinter',
|
||||
'FacebookArcanistConfiguration' => 'ArcanistConfiguration',
|
||||
'FacebookFbcodeLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookFbcodeUnitTestEngine' => 'ArcanistBaseUnitTestEngine',
|
||||
'FacebookHowtoevenLintEngine' => 'ArcanistLintEngine',
|
||||
'FacebookHowtoevenLinter' => 'ArcanistLinter',
|
||||
'FbcodeClangFormatLinter' => 'BaseDirectoryScopedFormatLinter',
|
||||
'FbcodeCppLinter' => 'ArcanistLinter',
|
||||
),
|
||||
));
|
||||
@@ -4,33 +4,32 @@
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
require('RocksDBCommonHelper.php');
|
||||
|
||||
define("DIFF_COMMAND", "diff");
|
||||
|
||||
class FacebookArcanistConfiguration extends ArcanistConfiguration {
|
||||
|
||||
public function didRunWorkflow($command,
|
||||
ArcanistWorkflow $workflow,
|
||||
ArcanistBaseWorkflow $workflow,
|
||||
$error_code) {
|
||||
// Default options don't terminate on failure, but that's what we want. In
|
||||
// the current case we use assertions intentionally as "terminate on failure
|
||||
// invariants".
|
||||
assert_options(ASSERT_BAIL, true);
|
||||
|
||||
assert($workflow);
|
||||
assert(strlen($command) > 0);
|
||||
|
||||
if ($command == DIFF_COMMAND && !$workflow->isRawDiffSource()) {
|
||||
$diffID = $workflow->getDiffId();
|
||||
|
||||
// When submitting a diff this code path gets executed multiple times in
|
||||
// a row. We only care about the case when ID for the diff is provided
|
||||
// because that's what we need to apply the diff and trigger the tests.
|
||||
if (strlen($diffID) > 0) {
|
||||
assert(is_numeric($diffID));
|
||||
startTestsInSandcastle(true /* $applyDiff */, $workflow, $diffID);
|
||||
}
|
||||
if ($command == 'diff' && !$workflow->isRawDiffSource()) {
|
||||
$this->maybePushToJenkins($workflow);
|
||||
}
|
||||
}
|
||||
|
||||
//////////////////////////////////////////////////////////////////////
|
||||
/* Send off builds to jenkins */
|
||||
function maybePushToJenkins($workflow) {
|
||||
$diffID = $workflow->getDiffID();
|
||||
if ($diffID === null) {
|
||||
return;
|
||||
}
|
||||
|
||||
$results = $workflow->getTestResults();
|
||||
if (!$results) {
|
||||
return;
|
||||
}
|
||||
|
||||
$url = "https://ci-builds.fb.com/view/rocksdb/job/rocksdb_diff_check/"
|
||||
."buildWithParameters?token=AUTH&DIFF_ID=$diffID";
|
||||
system("curl --noproxy '*' \"$url\" > /dev/null 2>&1");
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,36 +0,0 @@
|
||||
<?php
|
||||
// Copyright 2004-present Facebook. All Rights Reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
require('RocksDBCommonHelper.php');
|
||||
|
||||
define("DIFF_COMMAND", "diff");
|
||||
|
||||
class FacebookArcanistConfiguration extends ArcanistConfiguration {
|
||||
|
||||
public function didRunWorkflow($command,
|
||||
ArcanistBaseWorkflow $workflow,
|
||||
$error_code) {
|
||||
// Default options don't terminate on failure, but that's what we want. In
|
||||
// the current case we use assertions intentionally as "terminate on failure
|
||||
// invariants".
|
||||
assert_options(ASSERT_BAIL, true);
|
||||
|
||||
assert($workflow);
|
||||
assert(strlen($command) > 0);
|
||||
|
||||
if ($command == DIFF_COMMAND && !$workflow->isRawDiffSource()) {
|
||||
$diffID = $workflow->getDiffId();
|
||||
|
||||
// When submitting a diff this code path gets executed multiple times in
|
||||
// a row. We only care about the case when ID for the diff is provided
|
||||
// because that's what we need to apply the diff and trigger the tests.
|
||||
if (strlen($diffID) > 0) {
|
||||
assert(is_numeric($diffID));
|
||||
startTestsInSandcastle(true /* $applyDiff */, $workflow, $diffID);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1,337 +0,0 @@
|
||||
<?php
|
||||
// Copyright 2004-present Facebook. All Rights Reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
// Name of the environment variables which need to be set by the entity which
|
||||
// triggers continuous runs so that code at the end of the file gets executed
|
||||
// and Sandcastle run starts.
|
||||
define("ENV_POST_RECEIVE_HOOK", "POST_RECEIVE_HOOK");
|
||||
define("ENV_HTTPS_APP_VALUE", "HTTPS_APP_VALUE");
|
||||
define("ENV_HTTPS_TOKEN_VALUE", "HTTPS_TOKEN_VALUE");
|
||||
|
||||
define("PRIMARY_TOKEN_FILE", '/home/krad/.sandcastle');
|
||||
define("SECONDARY_TOKEN_FILE", '$HOME/.sandcastle');
|
||||
define("CONT_RUN_ALIAS", "leveldb");
|
||||
|
||||
//////////////////////////////////////////////////////////////////////
|
||||
/* Run tests in sandcastle */
|
||||
function postURL($diffID, $url) {
|
||||
assert(strlen($diffID) > 0);
|
||||
assert(is_numeric($diffID));
|
||||
assert(strlen($url) > 0);
|
||||
|
||||
$cmd = 'echo \'{"diff_id": "' . $diffID . '", '
|
||||
. '"name":"click here for sandcastle tests for D' . $diffID . '", '
|
||||
. '"link":"' . $url . '"}\' | '
|
||||
. 'http_proxy=fwdproxy.any.facebook.com:8080 '
|
||||
. 'https_proxy=fwdproxy.any.facebook.com:8080 arc call-conduit '
|
||||
. 'differential.updateunitresults';
|
||||
shell_exec($cmd);
|
||||
}
|
||||
|
||||
function buildUpdateTestStatusCmd($diffID, $test, $status) {
|
||||
assert(strlen($diffID) > 0);
|
||||
assert(is_numeric($diffID));
|
||||
assert(strlen($test) > 0);
|
||||
assert(strlen($status) > 0);
|
||||
|
||||
$cmd = 'echo \'{"diff_id": "' . $diffID . '", '
|
||||
. '"name":"' . $test . '", '
|
||||
. '"result":"' . $status . '"}\' | '
|
||||
. 'http_proxy=fwdproxy.any.facebook.com:8080 '
|
||||
. 'https_proxy=fwdproxy.any.facebook.com:8080 arc call-conduit '
|
||||
. 'differential.updateunitresults';
|
||||
return $cmd;
|
||||
}
|
||||
|
||||
function updateTestStatus($diffID, $test) {
|
||||
assert(strlen($diffID) > 0);
|
||||
assert(is_numeric($diffID));
|
||||
assert(strlen($test) > 0);
|
||||
|
||||
shell_exec(buildUpdateTestStatusCmd($diffID, $test, "waiting"));
|
||||
}
|
||||
|
||||
function getSteps($applyDiff, $diffID, $username, $test) {
|
||||
assert(strlen($username) > 0);
|
||||
assert(strlen($test) > 0);
|
||||
|
||||
if ($applyDiff) {
|
||||
assert(strlen($diffID) > 0);
|
||||
assert(is_numeric($diffID));
|
||||
|
||||
$arcrc_content = (PHP_OS == "Darwin" ?
|
||||
exec("cat ~/.arcrc | gzip -f | base64") :
|
||||
exec("cat ~/.arcrc | gzip -f | base64 -w0"));
|
||||
assert(strlen($arcrc_content) > 0);
|
||||
|
||||
// Sandcastle machines don't have arc setup. We copy the user certificate
|
||||
// and authenticate using that in Sandcastle.
|
||||
$setup = array(
|
||||
"name" => "Setup arcrc",
|
||||
"shell" => "echo " . $arcrc_content . " | base64 --decode"
|
||||
. " | gzip -d > ~/.arcrc",
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
// arc demands certain permission on its config.
|
||||
// also fix the sticky bit issue in sandcastle
|
||||
$fix_permission = array(
|
||||
"name" => "Fix environment",
|
||||
"shell" => "chmod 600 ~/.arcrc && chmod +t /dev/shm",
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
// Construct the steps in the order of execution.
|
||||
$steps[] = $setup;
|
||||
$steps[] = $fix_permission;
|
||||
}
|
||||
|
||||
// fbcode is a sub-repo. We cannot patch until we add it to ignore otherwise
|
||||
// Git thinks it is an uncommited change.
|
||||
$fix_git_ignore = array(
|
||||
"name" => "Fix git ignore",
|
||||
"shell" => "echo fbcode >> .git/info/exclude",
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
$steps[] = $fix_git_ignore;
|
||||
|
||||
// This will be the command used to execute particular type of tests.
|
||||
$cmd = "";
|
||||
|
||||
if ($applyDiff) {
|
||||
// Patch the code (keep your fingures crossed).
|
||||
$patch = array(
|
||||
"name" => "Patch " . $diffID,
|
||||
"shell" => "HTTPS_PROXY=fwdproxy:8080 arc --arcrc-file ~/.arcrc "
|
||||
. "patch --nocommit --diff " . $diffID,
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
$steps[] = $patch;
|
||||
|
||||
updateTestStatus($diffID, $test);
|
||||
$cmd = buildUpdateTestStatusCmd($diffID, $test, "running") . "; ";
|
||||
}
|
||||
|
||||
// Run the actual command.
|
||||
$cmd = $cmd . "J=$(nproc) ./build_tools/precommit_checker.py " . $test
|
||||
. "; exit_code=$?; ";
|
||||
|
||||
if ($applyDiff) {
|
||||
$cmd = $cmd . "([[ \$exit_code -eq 0 ]] &&"
|
||||
. buildUpdateTestStatusCmd($diffID, $test, "pass") . ")"
|
||||
. "||" . buildUpdateTestStatusCmd($diffID, $test, "fail")
|
||||
. "; ";
|
||||
}
|
||||
|
||||
// shell command to sort the tests based on exit code and print
|
||||
// the output of the log files.
|
||||
$cat_sorted_logs = "
|
||||
while read code log_file;
|
||||
do echo \"################ cat \$log_file [exit_code : \$code] ################\";
|
||||
cat \$log_file;
|
||||
done < <(tail -n +2 LOG | sort -k7,7n -k4,4gr | awk '{print \$7,\$NF}')";
|
||||
|
||||
// Shell command to cat all log files
|
||||
$cat_all_logs = "for f in `ls t/!(run-*)`; do echo \$f;cat \$f; done";
|
||||
|
||||
// If LOG file exist use it to cat log files sorted by exit code, otherwise
|
||||
// cat everything
|
||||
$logs_cmd = "if [ -f LOG ]; then {$cat_sorted_logs}; else {$cat_all_logs}; fi";
|
||||
|
||||
$cmd = $cmd . " cat /tmp/precommit-check.log"
|
||||
. "; shopt -s extglob; {$logs_cmd}"
|
||||
. "; shopt -u extglob; [[ \$exit_code -eq 0 ]]";
|
||||
assert(strlen($cmd) > 0);
|
||||
|
||||
$run_test = array(
|
||||
"name" => "Run " . $test,
|
||||
"shell" => $cmd,
|
||||
"user" => "root",
|
||||
"parser" => "python build_tools/error_filter.py " . $test,
|
||||
);
|
||||
|
||||
$steps[] = $run_test;
|
||||
|
||||
if ($applyDiff) {
|
||||
// Clean up the user arc config we are using.
|
||||
$cleanup = array(
|
||||
"name" => "Arc cleanup",
|
||||
"shell" => "rm -f ~/.arcrc",
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
$steps[] = $cleanup;
|
||||
}
|
||||
|
||||
assert(count($steps) > 0);
|
||||
return $steps;
|
||||
}
|
||||
|
||||
function getSandcastleConfig() {
|
||||
$sandcastle_config = array();
|
||||
|
||||
// This is a case when we're executed from a continuous run. Fetch the values
|
||||
// from the environment.
|
||||
if (getenv(ENV_POST_RECEIVE_HOOK)) {
|
||||
$sandcastle_config[0] = getenv(ENV_HTTPS_APP_VALUE);
|
||||
$sandcastle_config[1] = getenv(ENV_HTTPS_TOKEN_VALUE);
|
||||
} else {
|
||||
// This is a typical `[p]arc diff` case. Fetch the values from the specific
|
||||
// configuration files.
|
||||
assert(file_exists(PRIMARY_TOKEN_FILE) ||
|
||||
file_exists(SECONDARY_TOKEN_FILE));
|
||||
|
||||
// Try the primary location first, followed by a secondary.
|
||||
if (file_exists(PRIMARY_TOKEN_FILE)) {
|
||||
$cmd = 'cat ' . PRIMARY_TOKEN_FILE;
|
||||
} else {
|
||||
$cmd = 'cat ' . SECONDARY_TOKEN_FILE;
|
||||
}
|
||||
|
||||
assert(strlen($cmd) > 0);
|
||||
$sandcastle_config = explode(':', rtrim(shell_exec($cmd)));
|
||||
}
|
||||
|
||||
// In this case be very explicit about the implications.
|
||||
if (count($sandcastle_config) != 2) {
|
||||
echo "Sandcastle configuration files don't contain valid information " .
|
||||
"or the necessary environment variables aren't defined. Unable " .
|
||||
"to validate the code changes.";
|
||||
exit(1);
|
||||
}
|
||||
|
||||
assert(strlen($sandcastle_config[0]) > 0);
|
||||
assert(strlen($sandcastle_config[1]) > 0);
|
||||
assert(count($sandcastle_config) > 0);
|
||||
|
||||
return $sandcastle_config;
|
||||
}
|
||||
|
||||
// This function can be called either from `[p]arc diff` command or during
|
||||
// the Git post-receive hook.
|
||||
function startTestsInSandcastle($applyDiff, $workflow, $diffID) {
|
||||
// Default options don't terminate on failure, but that's what we want. In
|
||||
// the current case we use assertions intentionally as "terminate on failure
|
||||
// invariants".
|
||||
assert_options(ASSERT_BAIL, true);
|
||||
|
||||
// In case of a diff we'll send notificatios to the author. Else it'll go to
|
||||
// the entire team because failures indicate that build quality has regressed.
|
||||
$username = $applyDiff ? exec("whoami") : CONT_RUN_ALIAS;
|
||||
assert(strlen($username) > 0);
|
||||
|
||||
if ($applyDiff) {
|
||||
assert($workflow);
|
||||
assert(strlen($diffID) > 0);
|
||||
assert(is_numeric($diffID));
|
||||
}
|
||||
|
||||
// List of tests we want to run in Sandcastle.
|
||||
$tests = array("unit", "unit_non_shm", "unit_481", "clang_unit", "tsan",
|
||||
"asan", "lite_test", "valgrind", "release", "release_481",
|
||||
"clang_release", "punit", "clang_analyze", "code_cov",
|
||||
"java_build", "no_compression", "unity", "ubsan");
|
||||
|
||||
$send_email_template = array(
|
||||
'type' => 'email',
|
||||
'triggers' => array('fail'),
|
||||
'emails' => array($username . '@fb.com'),
|
||||
);
|
||||
|
||||
// Construct a job definition for each test and add it to the master plan.
|
||||
foreach ($tests as $test) {
|
||||
$stepName = "RocksDB diff " . $diffID . " test " . $test;
|
||||
|
||||
if (!$applyDiff) {
|
||||
$stepName = "RocksDB continuous integration test " . $test;
|
||||
}
|
||||
|
||||
$arg[] = array(
|
||||
"name" => $stepName,
|
||||
"report" => array($send_email_template),
|
||||
"steps" => getSteps($applyDiff, $diffID, $username, $test)
|
||||
);
|
||||
}
|
||||
|
||||
// We cannot submit the parallel execution master plan to Sandcastle and
|
||||
// need supply the job plan as a determinator. So we construct a small job
|
||||
// that will spit out the master job plan which Sandcastle will parse and
|
||||
// execute. Why compress the job definitions? Otherwise we run over the max
|
||||
// string size.
|
||||
$cmd = "echo " . base64_encode(json_encode($arg))
|
||||
. (PHP_OS == "Darwin" ?
|
||||
" | gzip -f | base64" :
|
||||
" | gzip -f | base64 -w0");
|
||||
assert(strlen($cmd) > 0);
|
||||
|
||||
$arg_encoded = shell_exec($cmd);
|
||||
assert(strlen($arg_encoded) > 0);
|
||||
|
||||
$runName = "Run diff " . $diffID . "for user " . $username;
|
||||
|
||||
if (!$applyDiff) {
|
||||
$runName = "RocksDB continuous integration build and test run";
|
||||
}
|
||||
|
||||
$command = array(
|
||||
"name" => $runName,
|
||||
"steps" => array()
|
||||
);
|
||||
|
||||
$command["steps"][] = array(
|
||||
"name" => "Generate determinator",
|
||||
"shell" => "echo " . $arg_encoded . " | base64 --decode | gzip -d"
|
||||
. " | base64 --decode",
|
||||
"determinator" => true,
|
||||
"user" => "root"
|
||||
);
|
||||
|
||||
// Submit to Sandcastle.
|
||||
$url = 'https://interngraph.intern.facebook.com/sandcastle/generate?'
|
||||
.'command=SandcastleUniversalCommand'
|
||||
.'&vcs=rocksdb-git&revision=origin%2Fmaster&type=lego'
|
||||
.'&user=' . $username . '&alias=rocksdb-precommit'
|
||||
.'&command-args=' . urlencode(json_encode($command));
|
||||
|
||||
// Fetch the configuration necessary to submit a successful HTTPS request.
|
||||
$sandcastle_config = getSandcastleConfig();
|
||||
|
||||
$app = $sandcastle_config[0];
|
||||
$token = $sandcastle_config[1];
|
||||
|
||||
$cmd = 'https_proxy= HTTPS_PROXY= curl -s -k -F app=' . $app . ' '
|
||||
. '-F token=' . $token . ' "' . $url . '"';
|
||||
|
||||
$output = shell_exec($cmd);
|
||||
assert(strlen($output) > 0);
|
||||
|
||||
// Extract Sandcastle URL from the response.
|
||||
preg_match('/url": "(.+)"/', $output, $sandcastle_url);
|
||||
|
||||
assert(count($sandcastle_url) > 0, "Unable to submit Sandcastle request.");
|
||||
assert(strlen($sandcastle_url[1]) > 0, "Unable to extract Sandcastle URL.");
|
||||
|
||||
if ($applyDiff) {
|
||||
echo "\nSandcastle URL: " . $sandcastle_url[1] . "\n";
|
||||
// Ask Phabricator to display it on the diff UI.
|
||||
postURL($diffID, $sandcastle_url[1]);
|
||||
} else {
|
||||
echo "Continuous integration started Sandcastle tests. You can look at ";
|
||||
echo "the progress at:\n" . $sandcastle_url[1] . "\n";
|
||||
}
|
||||
}
|
||||
|
||||
// Continuous run cript will set the environment variable and based on that
|
||||
// we'll trigger the execution of tests in Sandcastle. In that case we don't
|
||||
// need to apply any diffs and there's no associated workflow either.
|
||||
if (getenv(ENV_POST_RECEIVE_HOOK)) {
|
||||
startTestsInSandcastle(
|
||||
false /* $applyDiff */,
|
||||
NULL /* $workflow */,
|
||||
NULL /* $diffID */);
|
||||
}
|
||||
@@ -44,7 +44,7 @@ abstract class BaseDirectoryScopedFormatLinter extends ArcanistLinter {
|
||||
$futures[$path] = $this->getFormatFuture($path, $changed);
|
||||
}
|
||||
|
||||
foreach (id(new FutureIterator($futures))->limit(8) as $p => $f) {
|
||||
foreach (Futures($futures)->limit(8) as $p => $f) {
|
||||
$this->rawLintOutput[$p] = $f->resolvex();
|
||||
}
|
||||
}
|
||||
|
||||
@@ -88,9 +88,6 @@ class FbcodeCppLinter extends ArcanistLinter {
|
||||
}
|
||||
|
||||
private function getCppLintOutput($path) {
|
||||
if (!array_key_exists($path, $this->rawLintOutput)) {
|
||||
return array();
|
||||
}
|
||||
list($output) = $this->rawLintOutput[$path];
|
||||
|
||||
$msgs = array();
|
||||
|
||||
Vendored
+2
-2
@@ -1,5 +1,5 @@
|
||||
#!/usr/bin/python
|
||||
# Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
# Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
# This source code is licensed under the BSD-style license found in the
|
||||
# LICENSE file in the root directory of this source tree. An additional grant
|
||||
# of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -2714,7 +2714,7 @@ def CheckSpacing(filename, clean_lines, linenum, nesting_state, error):
|
||||
|
||||
# Look for < that is not surrounded by spaces. This is only
|
||||
# triggered if both sides are missing spaces, even though
|
||||
# technically should flag if at least one side is missing a
|
||||
# technically should should flag if at least one side is missing a
|
||||
# space. This is done to avoid some false positives with shifts.
|
||||
match = Search(r'[^\s<]<([^\s=<].*)', reduced_line)
|
||||
if (match and
|
||||
|
||||
@@ -42,12 +42,14 @@ class FacebookFbcodeLintEngine extends ArcanistLintEngine {
|
||||
$python_linter = new ArcanistPEP8Linter();
|
||||
$linters[] = $python_linter;
|
||||
|
||||
$cpp_linters = array();
|
||||
$cpp_linters[] = $linters[] = new ArcanistCpplintLinter();
|
||||
$cpp_linters[] = $linters[] = new FbcodeCppLinter();
|
||||
if (!$this->getCommitHookMode()) {
|
||||
$cpp_linters = array();
|
||||
$cpp_linters[] = $linters[] = new ArcanistCpplintLinter();
|
||||
$cpp_linters[] = $linters[] = new FbcodeCppLinter();
|
||||
|
||||
$clang_format_linter = new FbcodeClangFormatLinter();
|
||||
$linters[] = $clang_format_linter;
|
||||
$clang_format_linter = new FbcodeClangFormatLinter();
|
||||
$linters[] = $clang_format_linter;
|
||||
}
|
||||
|
||||
$spelling_linter = new ArcanistSpellingLinter();
|
||||
$linters[] = $spelling_linter;
|
||||
|
||||
@@ -4,14 +4,18 @@
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
class FacebookFbcodeUnitTestEngine extends ArcanistUnitTestEngine {
|
||||
class FacebookFbcodeUnitTestEngine extends ArcanistBaseUnitTestEngine {
|
||||
|
||||
public function run() {
|
||||
// For a call to `arc call-conduit differential.updateunitresults` to
|
||||
// succeed we need at least one entry here.
|
||||
$result = new ArcanistUnitTestResult();
|
||||
$result->setName("dummy_placeholder_entry");
|
||||
$result->setResult(ArcanistUnitTestResult::RESULT_PASS);
|
||||
return array($result);
|
||||
// Here we create a new unit test "jenkins_async_test" and promise we'll
|
||||
// update the results later.
|
||||
// Jenkins updates the results using `arc call-conduit
|
||||
// differential.updateunitresults` call. If you change the name here, also
|
||||
// make sure to change the name in Jenkins script that updates the test
|
||||
// result -- they have to be the same.
|
||||
$result = new ArcanistUnitTestResult();
|
||||
$result->setName("jenkins_async_test");
|
||||
$result->setResult(ArcanistUnitTestResult::RESULT_POSTPONED);
|
||||
return array($result);
|
||||
}
|
||||
}
|
||||
@@ -1,17 +0,0 @@
|
||||
<?php
|
||||
// Copyright 2004-present Facebook. All Rights Reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
class FacebookFbcodeUnitTestEngine extends ArcanistBaseUnitTestEngine {
|
||||
|
||||
public function run() {
|
||||
// For a call to `arc call-conduit differential.updateunitresults` to
|
||||
// succeed we need at least one entry here.
|
||||
$result = new ArcanistUnitTestResult();
|
||||
$result->setName("dummy_placeholder_entry");
|
||||
$result->setResult(ArcanistUnitTestResult::RESULT_PASS);
|
||||
return array($result);
|
||||
}
|
||||
}
|
||||
@@ -19,21 +19,22 @@
|
||||
#
|
||||
# The PLATFORM_CCFLAGS and PLATFORM_CXXFLAGS might include the following:
|
||||
#
|
||||
# -DROCKSDB_PLATFORM_POSIX if posix-platform based
|
||||
# -DLEVELDB_PLATFORM_POSIX if cstdatomic is present
|
||||
# -DLEVELDB_PLATFORM_NOATOMIC if it is not
|
||||
# -DSNAPPY if the Snappy library is present
|
||||
# -DLZ4 if the LZ4 library is present
|
||||
# -DZSTD if the ZSTD library is present
|
||||
# -DNUMA if the NUMA library is present
|
||||
# -DTBB if the TBB library is present
|
||||
#
|
||||
# Using gflags in rocksdb:
|
||||
# Our project depends on gflags, which requires users to take some extra steps
|
||||
# before they can compile the whole repository:
|
||||
# 1. Install gflags. You may download it from here:
|
||||
# https://gflags.github.io/gflags/ (Mac users can `brew install gflags`)
|
||||
# 2. Once installed, add the include path for gflags to your CPATH env var and
|
||||
# the lib path to LIBRARY_PATH. If installed with default settings, the lib
|
||||
# will be /usr/local/lib and the include path will be /usr/local/include
|
||||
# https://code.google.com/p/gflags/
|
||||
# 2. Once install, add the include path/lib path for gflags to CPATH and
|
||||
# LIBRARY_PATH respectively. If installed with default mode, the
|
||||
# lib and include path will be /usr/local/lib and /usr/local/include
|
||||
# Mac user can do this by having brew installed and running brew install gflags
|
||||
|
||||
OUTPUT=$1
|
||||
if test -z "$OUTPUT"; then
|
||||
@@ -44,7 +45,7 @@ fi
|
||||
# we depend on C++11
|
||||
PLATFORM_CXXFLAGS="-std=c++11"
|
||||
# we currently depend on POSIX platform
|
||||
COMMON_FLAGS="-DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX"
|
||||
COMMON_FLAGS="-DROCKSDB_PLATFORM_POSIX"
|
||||
|
||||
# Default to fbcode gcc on internal fb machines
|
||||
if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
|
||||
@@ -164,7 +165,7 @@ case "$TARGET_OS" in
|
||||
;;
|
||||
OS_ANDROID_CROSSCOMPILE)
|
||||
PLATFORM=OS_ANDROID
|
||||
COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_ANDROID -DROCKSDB_PLATFORM_POSIX"
|
||||
COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_ANDROID -DLEVELDB_PLATFORM_POSIX"
|
||||
PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS " # All pthread features are in the Android C library
|
||||
# PORT_FILES=port/android/android.cc
|
||||
CROSS_COMPILE=true
|
||||
@@ -183,19 +184,17 @@ if [ "$CROSS_COMPILE" = "true" -o "$FBCODE_BUILD" = "true" ]; then
|
||||
# Also don't need any compilation tests if compiling on fbcode
|
||||
true
|
||||
else
|
||||
if ! test $ROCKSDB_DISABLE_FALLOCATE; then
|
||||
# Test whether fallocate is available
|
||||
$CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
#include <fcntl.h>
|
||||
#include <linux/falloc.h>
|
||||
int main() {
|
||||
int fd = open("/dev/null", 0);
|
||||
fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE, 0, 1024);
|
||||
}
|
||||
# Test whether fallocate is available
|
||||
$CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
#include <fcntl.h>
|
||||
#include <linux/falloc.h>
|
||||
int main() {
|
||||
int fd = open("/dev/null", 0);
|
||||
fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE, 0, 1024);
|
||||
}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_FALLOCATE_PRESENT"
|
||||
fi
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_FALLOCATE_PRESENT"
|
||||
fi
|
||||
|
||||
# Test whether Snappy library is installed
|
||||
@@ -282,7 +281,7 @@ EOF
|
||||
# Test whether numa is available
|
||||
$CXX $CFLAGS -x c++ - -o /dev/null -lnuma 2>/dev/null <<EOF
|
||||
#include <numa.h>
|
||||
#include <numaif.h>
|
||||
#inlcude <numaif.h>
|
||||
int main() {}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
@@ -291,27 +290,15 @@ EOF
|
||||
JAVA_LDFLAGS="$JAVA_LDFLAGS -lnuma"
|
||||
fi
|
||||
|
||||
# Test whether tbb is available
|
||||
$CXX $CFLAGS $LDFLAGS -x c++ - -o /dev/null -ltbb 2>/dev/null <<EOF
|
||||
#include <tbb/tbb.h>
|
||||
int main() {}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DTBB"
|
||||
PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ltbb"
|
||||
JAVA_LDFLAGS="$JAVA_LDFLAGS -ltbb"
|
||||
fi
|
||||
|
||||
# Test whether jemalloc is available
|
||||
if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null -ljemalloc \
|
||||
2>/dev/null; then
|
||||
2>/dev/null; then
|
||||
PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ljemalloc"
|
||||
JAVA_LDFLAGS="$JAVA_LDFLAGS -ljemalloc"
|
||||
JEMALLOC=1
|
||||
else
|
||||
# jemalloc is not available. Let's try tcmalloc
|
||||
if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null \
|
||||
-ltcmalloc 2>/dev/null; then
|
||||
-ltcmalloc 2>/dev/null; then
|
||||
PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ltcmalloc"
|
||||
JAVA_LDFLAGS="$JAVA_LDFLAGS -ltcmalloc"
|
||||
fi
|
||||
@@ -328,55 +315,6 @@ EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_MALLOC_USABLE_SIZE"
|
||||
fi
|
||||
|
||||
# Test whether PTHREAD_MUTEX_ADAPTIVE_NP mutex type is available
|
||||
$CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
#include <pthread.h>
|
||||
int main() {
|
||||
int x = PTHREAD_MUTEX_ADAPTIVE_NP;
|
||||
return 0;
|
||||
}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_PTHREAD_ADAPTIVE_MUTEX"
|
||||
fi
|
||||
|
||||
# Test whether backtrace is available
|
||||
$CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
#include <execinfo.h>>
|
||||
int main() {
|
||||
void* frames[1];
|
||||
backtrace_symbols(frames, backtrace(frames, 1));
|
||||
return 0;
|
||||
}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_BACKTRACE"
|
||||
else
|
||||
# Test whether execinfo library is installed
|
||||
$CXX $CFLAGS -lexecinfo -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
#include <execinfo.h>
|
||||
int main() {
|
||||
void* frames[1];
|
||||
backtrace_symbols(frames, backtrace(frames, 1));
|
||||
}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_BACKTRACE"
|
||||
PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lexecinfo"
|
||||
JAVA_LDFLAGS="$JAVA_LDFLAGS -lexecinfo"
|
||||
fi
|
||||
fi
|
||||
|
||||
# Test if -pg is supported
|
||||
$CXX $CFLAGS -pg -x c++ - -o /dev/null 2>/dev/null <<EOF
|
||||
int main() {
|
||||
return 0;
|
||||
}
|
||||
EOF
|
||||
if [ "$?" = 0 ]; then
|
||||
PROFILING_FLAGS=-pg
|
||||
fi
|
||||
fi
|
||||
|
||||
# TODO(tec): Fix -Wshorten-64-to-32 errors on FreeBSD and enable the warning.
|
||||
@@ -416,15 +354,7 @@ if test "$USE_SSE"; then
|
||||
# if Intel SSE instruction set is supported, set USE_SSE=1
|
||||
COMMON_FLAGS="$COMMON_FLAGS -msse -msse4.2 "
|
||||
elif test -z "$PORTABLE"; then
|
||||
if test -n "`echo $TARGET_ARCHITECTURE | grep ^ppc64`"; then
|
||||
# Tune for this POWER processor, treating '+' models as base models
|
||||
POWER=`LD_SHOW_AUXV=1 /bin/true | grep AT_PLATFORM | grep -E -o power[0-9]+`
|
||||
COMMON_FLAGS="$COMMON_FLAGS -mcpu=$POWER -mtune=$POWER "
|
||||
elif test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
|
||||
COMMON_FLAGS="$COMMON_FLAGS -march=z10 "
|
||||
else
|
||||
COMMON_FLAGS="$COMMON_FLAGS -march=native "
|
||||
fi
|
||||
COMMON_FLAGS="$COMMON_FLAGS -march=native "
|
||||
fi
|
||||
|
||||
PLATFORM_CCFLAGS="$PLATFORM_CCFLAGS $COMMON_FLAGS"
|
||||
@@ -457,7 +387,3 @@ echo "ROCKSDB_MINOR=$ROCKSDB_MINOR" >> "$OUTPUT"
|
||||
echo "ROCKSDB_PATCH=$ROCKSDB_PATCH" >> "$OUTPUT"
|
||||
echo "CLANG_SCAN_BUILD=$CLANG_SCAN_BUILD" >> "$OUTPUT"
|
||||
echo "CLANG_ANALYZER=$CLANG_ANALYZER" >> "$OUTPUT"
|
||||
echo "PROFILING_FLAGS=$PROFILING_FLAGS" >> "$OUTPUT"
|
||||
if test -n "$JEMALLOC"; then
|
||||
echo "JEMALLOC=1" >> "$OUTPUT"
|
||||
fi
|
||||
@@ -1,135 +0,0 @@
|
||||
#!/bin/bash
|
||||
#
|
||||
# Copyright (c) 2016, Facebook. All rights reserved.
|
||||
#
|
||||
# Overall wrapper script for RocksDB continuous builds. The implementation is a
|
||||
# trivial pulling scheme. We loop infinitely, check if any new changes have been
|
||||
# committed, if yes then trigger a Sandcastle run, and finally go to sleep again
|
||||
# for a certain interval.
|
||||
#
|
||||
|
||||
SRC_GIT_REPO=/data/git/rocksdb
|
||||
error=0
|
||||
|
||||
function log {
|
||||
DATE=`date +%Y-%m-%d:%H:%M:%S`
|
||||
echo $DATE $@
|
||||
}
|
||||
|
||||
function log_err {
|
||||
log "ERROR: $@ Error code: $error."
|
||||
}
|
||||
|
||||
function update_repo_status {
|
||||
# Update the parent first.
|
||||
pushd $SRC_GIT_REPO
|
||||
|
||||
# This is a fatal error. Something in the environment isn't right and we will
|
||||
# terminate the execution.
|
||||
error=$?
|
||||
if [ ! $error -eq 0 ]; then
|
||||
log_err "Where is $SRC_GIT_REPO?"
|
||||
exit $error
|
||||
fi
|
||||
|
||||
HTTPS_PROXY=fwdproxy:8080 git fetch -f
|
||||
|
||||
error=$?
|
||||
if [ ! $error -eq 0 ]; then
|
||||
log_err "git fetch -f failed."
|
||||
popd
|
||||
return $error
|
||||
fi
|
||||
|
||||
git update-ref refs/heads/master refs/remotes/origin/master
|
||||
|
||||
error=$?
|
||||
if [ ! $error -eq 0 ]; then
|
||||
log_err "git update-ref failed."
|
||||
popd
|
||||
return $error
|
||||
fi
|
||||
|
||||
popd
|
||||
|
||||
# We're back in an instance-specific directory. Get the latest changes.
|
||||
git pull --rebase
|
||||
|
||||
error=$?
|
||||
if [ ! $error -eq 0 ]; then
|
||||
log_err "git pull --rebase failed."
|
||||
return $error
|
||||
fi
|
||||
}
|
||||
|
||||
#
|
||||
# Execution starts here.
|
||||
#
|
||||
|
||||
# Path to the determinator from the root of the RocksDB repo.
|
||||
CONTRUN_DETERMINATOR=./arcanist_util/config/RocksDBCommonHelper.php
|
||||
|
||||
# Value of the previous commit.
|
||||
PREV_COMMIT=
|
||||
|
||||
log "Starting to monitor for new RocksDB changes ..."
|
||||
log "Running under `pwd` as `whoami`."
|
||||
|
||||
# Paranoia. Make sure that we're using the right branch.
|
||||
git checkout master
|
||||
|
||||
error=$?
|
||||
if [ ! $error -eq 0 ]; then
|
||||
log_err "This is not good. Can't checkout master. Bye-bye!"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# We'll run forever and let the execution environment terminate us if we'll
|
||||
# exceed whatever timeout is set for the job.
|
||||
while true;
|
||||
do
|
||||
# Get the latest changes committed.
|
||||
update_repo_status
|
||||
|
||||
error=$?
|
||||
if [ $error -eq 0 ]; then
|
||||
LAST_COMMIT=`git log -1 | head -1 | grep commit | awk '{ print $2; }'`
|
||||
|
||||
log "Last commit is '$LAST_COMMIT', previous commit is '$PREV_COMMIT'."
|
||||
|
||||
if [ "$PREV_COMMIT" == "$LAST_COMMIT" ]; then
|
||||
log "There were no changes since the last time I checked. Going to sleep."
|
||||
else
|
||||
if [ ! -z "$LAST_COMMIT" ]; then
|
||||
log "New code has been committed or previous commit not known. " \
|
||||
"Will trigger the tests."
|
||||
|
||||
PREV_COMMIT=$LAST_COMMIT
|
||||
log "Updated previous commit to '$PREV_COMMIT'."
|
||||
|
||||
#
|
||||
# This is where we'll trigger the Sandcastle run. The values for
|
||||
# HTTPS_APP_VALUE and HTTPS_APP_VALUE will be set in the container we're
|
||||
# running in.
|
||||
#
|
||||
POST_RECEIVE_HOOK=1 php $CONTRUN_DETERMINATOR
|
||||
|
||||
error=$?
|
||||
if [ $error -eq 0 ]; then
|
||||
log "Sandcastle run successfully triggered."
|
||||
else
|
||||
log_err "Failed to trigger Sandcastle run."
|
||||
fi
|
||||
else
|
||||
log_err "Previous commit not updated. Don't know what the last one is."
|
||||
fi
|
||||
fi
|
||||
else
|
||||
log_err "Getting latest changes failed. Will skip running tests for now."
|
||||
fi
|
||||
|
||||
# Always sleep, even if errors happens while trying to determine the latest
|
||||
# commit. This will prevent us terminating in case of transient errors.
|
||||
log "Will go to sleep for 5 minutes."
|
||||
sleep 5m
|
||||
done
|
||||
@@ -1,17 +0,0 @@
|
||||
GCC_BASE=/mnt/gvfs/third-party2/gcc/cf7d14c625ce30bae1a4661c2319c5a283e4dd22/4.8.1/centos6-native/cc6c9dc
|
||||
CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/8598c375b0e94e1448182eb3df034704144a838d/stable/centos6-native/3f16ddd
|
||||
LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/d6e0a7da6faba45f5e5b1638f9edd7afc2f34e7d/4.8.1/gcc-4.8.1-glibc-2.17/8aac7fc
|
||||
GLIBC_BASE=/mnt/gvfs/third-party2/glibc/d282e6e8f3d20f4e40a516834847bdc038e07973/2.17/gcc-4.8.1-glibc-2.17/99df8fc
|
||||
SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/8c38a4c1e52b4c2cc8a9cdc31b9c947ed7dbfcb4/1.1.3/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
ZLIB_BASE=/mnt/gvfs/third-party2/zlib/0882df3713c7a84f15abe368dc004581f20b39d7/1.2.8/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/740325875f6729f42d28deaa2147b0854f3a347e/1.0.6/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
LZ4_BASE=/mnt/gvfs/third-party2/lz4/0e790b441e2d9acd68d51e1d2e028f88c6a79ddf/r131/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
ZSTD_BASE=/mnt/gvfs/third-party2/zstd/9455f75ff7f4831dc9fda02a6a0f8c68922fad8f/1.0.0/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/f001a51b2854957676d07306ef3abf67186b5c8b/2.1.1/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/fc8a13ca1fffa4d0765c716c5a0b49f0c107518f/master/gcc-4.8.1-glibc-2.17/8d31e51
|
||||
NUMA_BASE=/mnt/gvfs/third-party2/numa/17c514c4d102a25ca15f4558be564eeed76f4b6a/2.0.8/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/ad576de2a1ea560c4d3434304f0fc4e079bede42/trunk/gcc-4.8.1-glibc-2.17/675d945
|
||||
TBB_BASE=/mnt/gvfs/third-party2/tbb/9d9a554877d0c5bef330fe818ab7178806dd316a/4.0_update2/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/7c111ff27e0c466235163f00f280a9d617c3d2ec/4.0.9-36_fbk5_2933_gd092e3f/gcc-4.8.1-glibc-2.17/da39a3e
|
||||
BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/b7fd454c4b10c6a81015d4524ed06cdeab558490/2.26/centos6-native/da39a3e
|
||||
VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/d7f4d4d86674a57668e3a96f76f0e17dd0eb8765/3.8.1/gcc-4.8.1-glibc-2.17/c3f970a
|
||||
@@ -1,167 +0,0 @@
|
||||
# Copyright (c) 2016-present, Facebook, Inc. All rights reserved.
|
||||
# This source code is licensed under the BSD-style license found in the
|
||||
# LICENSE file in the root directory of this source tree. An additional grant
|
||||
# of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
'''Filter for error messages in test output:
|
||||
- Receives merged stdout/stderr from test on stdin
|
||||
- Finds patterns of known error messages for test name (first argument)
|
||||
- Prints those error messages to stdout
|
||||
'''
|
||||
|
||||
from __future__ import absolute_import
|
||||
from __future__ import division
|
||||
from __future__ import print_function
|
||||
from __future__ import unicode_literals
|
||||
|
||||
import re
|
||||
import sys
|
||||
|
||||
|
||||
class ErrorParserBase(object):
|
||||
def parse_error(self, line):
|
||||
'''Parses a line of test output. If it contains an error, returns a
|
||||
formatted message describing the error; otherwise, returns None.
|
||||
Subclasses must override this method.
|
||||
'''
|
||||
raise NotImplementedError
|
||||
|
||||
|
||||
class GTestErrorParser(ErrorParserBase):
|
||||
'''A parser that remembers the last test that began running so it can print
|
||||
that test's name upon detecting failure.
|
||||
'''
|
||||
_GTEST_NAME_PATTERN = re.compile(r'\[ RUN \] (\S+)$')
|
||||
# format: '<filename or "unknown file">:<line #>: Failure'
|
||||
_GTEST_FAIL_PATTERN = re.compile(r'(unknown file|\S+:\d+): Failure$')
|
||||
|
||||
def __init__(self):
|
||||
self._last_gtest_name = 'Unknown test'
|
||||
|
||||
def parse_error(self, line):
|
||||
gtest_name_match = self._GTEST_NAME_PATTERN.match(line)
|
||||
if gtest_name_match:
|
||||
self._last_gtest_name = gtest_name_match.group(1)
|
||||
return None
|
||||
gtest_fail_match = self._GTEST_FAIL_PATTERN.match(line)
|
||||
if gtest_fail_match:
|
||||
return '%s failed: %s' % (
|
||||
self._last_gtest_name, gtest_fail_match.group(1))
|
||||
return None
|
||||
|
||||
|
||||
class MatchErrorParser(ErrorParserBase):
|
||||
'''A simple parser that returns the whole line if it matches the pattern.
|
||||
'''
|
||||
def __init__(self, pattern):
|
||||
self._pattern = re.compile(pattern)
|
||||
|
||||
def parse_error(self, line):
|
||||
if self._pattern.match(line):
|
||||
return line
|
||||
return None
|
||||
|
||||
|
||||
class CompilerErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
# format: '<filename>:<line #>:<column #>: error: <error msg>'
|
||||
super(CompilerErrorParser, self).__init__(r'\S+:\d+:\d+: error:')
|
||||
|
||||
|
||||
class ScanBuildErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(ScanBuildErrorParser, self).__init__(
|
||||
r'scan-build: \d+ bugs found.$')
|
||||
|
||||
|
||||
class DbCrashErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(DbCrashErrorParser, self).__init__(r'\*\*\*.*\^$|TEST FAILED.')
|
||||
|
||||
|
||||
class WriteStressErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(WriteStressErrorParser, self).__init__(
|
||||
r'ERROR: write_stress died with exitcode=\d+')
|
||||
|
||||
|
||||
class AsanErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(AsanErrorParser, self).__init__(
|
||||
r'==\d+==ERROR: AddressSanitizer:')
|
||||
|
||||
|
||||
class UbsanErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
# format: '<filename>:<line #>:<column #>: runtime error: <error msg>'
|
||||
super(UbsanErrorParser, self).__init__(r'\S+:\d+:\d+: runtime error:')
|
||||
|
||||
|
||||
class ValgrindErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
# just grab the summary, valgrind doesn't clearly distinguish errors
|
||||
# from other log messages.
|
||||
super(ValgrindErrorParser, self).__init__(r'==\d+== ERROR SUMMARY:')
|
||||
|
||||
|
||||
class CompatErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(CompatErrorParser, self).__init__(r'==== .*[Ee]rror.* ====$')
|
||||
|
||||
|
||||
class TsanErrorParser(MatchErrorParser):
|
||||
def __init__(self):
|
||||
super(TsanErrorParser, self).__init__(r'WARNING: ThreadSanitizer:')
|
||||
|
||||
|
||||
_TEST_NAME_TO_PARSERS = {
|
||||
'punit': [CompilerErrorParser, GTestErrorParser],
|
||||
'unit': [CompilerErrorParser, GTestErrorParser],
|
||||
'release': [CompilerErrorParser, GTestErrorParser],
|
||||
'unit_481': [CompilerErrorParser, GTestErrorParser],
|
||||
'release_481': [CompilerErrorParser, GTestErrorParser],
|
||||
'clang_unit': [CompilerErrorParser, GTestErrorParser],
|
||||
'clang_release': [CompilerErrorParser, GTestErrorParser],
|
||||
'clang_analyze': [CompilerErrorParser, ScanBuildErrorParser],
|
||||
'code_cov': [CompilerErrorParser, GTestErrorParser],
|
||||
'unity': [CompilerErrorParser, GTestErrorParser],
|
||||
'lite': [CompilerErrorParser],
|
||||
'lite_test': [CompilerErrorParser, GTestErrorParser],
|
||||
'stress_crash': [CompilerErrorParser, DbCrashErrorParser],
|
||||
'write_stress': [CompilerErrorParser, WriteStressErrorParser],
|
||||
'asan': [CompilerErrorParser, GTestErrorParser, AsanErrorParser],
|
||||
'asan_crash': [CompilerErrorParser, AsanErrorParser, DbCrashErrorParser],
|
||||
'ubsan': [CompilerErrorParser, GTestErrorParser, UbsanErrorParser],
|
||||
'ubsan_crash': [CompilerErrorParser, UbsanErrorParser, DbCrashErrorParser],
|
||||
'valgrind': [CompilerErrorParser, GTestErrorParser, ValgrindErrorParser],
|
||||
'tsan': [CompilerErrorParser, GTestErrorParser, TsanErrorParser],
|
||||
'format_compatible': [CompilerErrorParser, CompatErrorParser],
|
||||
'run_format_compatible': [CompilerErrorParser, CompatErrorParser],
|
||||
'no_compression': [CompilerErrorParser, GTestErrorParser],
|
||||
'run_no_compression': [CompilerErrorParser, GTestErrorParser],
|
||||
'regression': [CompilerErrorParser],
|
||||
'run_regression': [CompilerErrorParser],
|
||||
}
|
||||
|
||||
|
||||
def main():
|
||||
if len(sys.argv) != 2:
|
||||
return 'Usage: %s <test name>' % sys.argv[0]
|
||||
test_name = sys.argv[1]
|
||||
if test_name not in _TEST_NAME_TO_PARSERS:
|
||||
return 'Unknown test name: %s' % test_name
|
||||
|
||||
error_parsers = []
|
||||
for parser_cls in _TEST_NAME_TO_PARSERS[test_name]:
|
||||
error_parsers.append(parser_cls())
|
||||
|
||||
for line in sys.stdin:
|
||||
line = line.strip()
|
||||
for error_parser in error_parsers:
|
||||
error_msg = error_parser.parse_error(line)
|
||||
if error_msg is not None:
|
||||
print(error_msg)
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
sys.exit(main())
|
||||
@@ -45,14 +45,6 @@ if test -z $PIC_BUILD; then
|
||||
CFLAGS+=" -DLZ4"
|
||||
fi
|
||||
|
||||
ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
|
||||
if test -z $PIC_BUILD; then
|
||||
ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
|
||||
else
|
||||
ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
|
||||
fi
|
||||
CFLAGS+=" -DZSTD"
|
||||
|
||||
# location of gflags headers and libraries
|
||||
GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
|
||||
if test -z $PIC_BUILD; then
|
||||
@@ -91,7 +83,7 @@ export USE_SSE=1
|
||||
BINUTILS="$BINUTILS_BASE/bin"
|
||||
AR="$BINUTILS/ar"
|
||||
|
||||
DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
|
||||
DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
|
||||
|
||||
STDLIBS="-L $GCC_BASE/lib64"
|
||||
|
||||
@@ -135,7 +127,7 @@ CFLAGS+=" $DEPS_INCLUDE"
|
||||
CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_SUPPORT_THREAD_LOCAL -DHAVE_SSE42"
|
||||
CXXFLAGS+=" $CFLAGS"
|
||||
|
||||
EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
|
||||
EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
|
||||
EXEC_LDFLAGS+=" -B$BINUTILS/gold"
|
||||
EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/platform007/lib/ld.so"
|
||||
EXEC_LDFLAGS+=" $LIBUNWIND"
|
||||
@@ -145,7 +137,7 @@ EXEC_LDFLAGS+=" -ldl"
|
||||
|
||||
PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
|
||||
|
||||
EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS"
|
||||
EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $GFLAGS_LIBS $TBB_LIBS"
|
||||
|
||||
VALGRIND_VER="$VALGRIND_BASE/bin/"
|
||||
|
||||
|
||||
@@ -4,82 +4,80 @@
|
||||
# fbcode settings. It uses the latest g++ compiler and also
|
||||
# uses jemalloc
|
||||
|
||||
BASEDIR=`dirname $BASH_SOURCE`
|
||||
source "$BASEDIR/dependencies_4.8.1.sh"
|
||||
|
||||
# location of libgcc
|
||||
LIBGCC_BASE="/mnt/gvfs/third-party2/libgcc/7712e757d7355cb51292454ee0b7b46a467fdfed/4.8.1/gcc-4.8.1-glibc-2.17/8aac7fc"
|
||||
LIBGCC_INCLUDE="$LIBGCC_BASE/include"
|
||||
LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
|
||||
LIBGCC_LIBS=" -L $LIBGCC_BASE/libs"
|
||||
|
||||
# location of glibc
|
||||
GLIBC_INCLUDE="$GLIBC_BASE/include"
|
||||
GLIBC_LIBS=" -L $GLIBC_BASE/lib"
|
||||
GLIBC_REV=6e40560b4e0b6d690fd1cf8c7a43ad7452b04cfa
|
||||
GLIBC_INCLUDE="/mnt/gvfs/third-party2/glibc/$GLIBC_REV/2.17/gcc-4.8.1-glibc-2.17/99df8fc/include"
|
||||
GLIBC_LIBS=" -L /mnt/gvfs/third-party2/glibc/$GLIBC_REV/2.17/gcc-4.8.1-glibc-2.17/99df8fc/lib"
|
||||
|
||||
# location of snappy headers and libraries
|
||||
SNAPPY_INCLUDE=" -I $SNAPPY_BASE/include"
|
||||
SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy.a"
|
||||
SNAPPY_INCLUDE=" -I /mnt/gvfs/third-party2/snappy/aef17f6c0b44b4fe408bd06f67c93701ab0a6ceb/1.0.3/gcc-4.8.1-glibc-2.17/43d84e2/include"
|
||||
SNAPPY_LIBS=" /mnt/gvfs/third-party2/snappy/aef17f6c0b44b4fe408bd06f67c93701ab0a6ceb/1.0.3/gcc-4.8.1-glibc-2.17/43d84e2/lib/libsnappy.a"
|
||||
|
||||
# location of zlib headers and libraries
|
||||
ZLIB_INCLUDE=" -I $ZLIB_BASE/include"
|
||||
ZLIB_LIBS=" $ZLIB_BASE/lib/libz.a"
|
||||
ZLIB_INCLUDE=" -I /mnt/gvfs/third-party2/zlib/25c6216928b4d77b59ddeca0990ff6fe9ac16b81/1.2.5/gcc-4.8.1-glibc-2.17/c3f970a/include"
|
||||
ZLIB_LIBS=" /mnt/gvfs/third-party2/zlib/25c6216928b4d77b59ddeca0990ff6fe9ac16b81/1.2.5/gcc-4.8.1-glibc-2.17/c3f970a/lib/libz.a"
|
||||
|
||||
# location of bzip headers and libraries
|
||||
BZIP2_INCLUDE=" -I $BZIP2_BASE/include/"
|
||||
BZIP2_LIBS=" $BZIP2_BASE/lib/libbz2.a"
|
||||
BZIP_INCLUDE=" -I /mnt/gvfs/third-party2/bzip2/c9ef7629c2aa0024f7a416e87602f06eb88f5eac/1.0.6/gcc-4.8.1-glibc-2.17/c3f970a/include/"
|
||||
BZIP_LIBS=" /mnt/gvfs/third-party2/bzip2/c9ef7629c2aa0024f7a416e87602f06eb88f5eac/1.0.6/gcc-4.8.1-glibc-2.17/c3f970a/lib/libbz2.a"
|
||||
|
||||
LZ4_INCLUDE=" -I $LZ4_BASE/include"
|
||||
LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
|
||||
LZ4_REV=065ec7e38fe83329031f6668c43bef83eff5808b
|
||||
LZ4_INCLUDE=" -I /mnt/gvfs/third-party2/lz4/$LZ4_REV/r108/gcc-4.8.1-glibc-2.17/c3f970a/include"
|
||||
LZ4_LIBS=" /mnt/gvfs/third-party2/lz4/$LZ4_REV/r108/gcc-4.8.1-glibc-2.17/c3f970a/lib/liblz4.a"
|
||||
|
||||
ZSTD_INCLUDE=" -I $ZSTD_BASE/include"
|
||||
ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
|
||||
ZSTD_REV=8df2d01673ae6afcc8c8d16fec862b2d67ecc1e9
|
||||
ZSTD_INCLUDE=" -I /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.1.1/gcc-4.8.1-glibc-2.17/c3f970a/include"
|
||||
ZSTD_LIBS=" /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.1.1/gcc-4.8.1-glibc-2.17/c3f970a/lib/libzstd.a"
|
||||
|
||||
# location of gflags headers and libraries
|
||||
GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
|
||||
GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags.a"
|
||||
GFLAGS_INCLUDE=" -I /mnt/gvfs/third-party2/gflags/1ad047a6e6f6673991918ecadc670868205a243a/1.6/gcc-4.8.1-glibc-2.17/c3f970a/include/"
|
||||
GFLAGS_LIBS=" /mnt/gvfs/third-party2/gflags/1ad047a6e6f6673991918ecadc670868205a243a/1.6/gcc-4.8.1-glibc-2.17/c3f970a/lib/libgflags.a"
|
||||
|
||||
# location of jemalloc
|
||||
JEMALLOC_INCLUDE=" -I $JEMALLOC_BASE/include"
|
||||
JEMALLOC_LIB="$JEMALLOC_BASE/lib/libjemalloc.a"
|
||||
JEMALLOC_INCLUDE=" -I /mnt/gvfs/third-party2/jemalloc/3691c776ac26dd8781e84f8888b6a0fbdbc0a9ed/dev/gcc-4.8.1-glibc-2.17/4d53c6f/include"
|
||||
JEMALLOC_LIB="/mnt/gvfs/third-party2/jemalloc/3691c776ac26dd8781e84f8888b6a0fbdbc0a9ed/dev/gcc-4.8.1-glibc-2.17/4d53c6f/lib/libjemalloc.a"
|
||||
|
||||
# location of numa
|
||||
NUMA_INCLUDE=" -I $NUMA_BASE/include/"
|
||||
NUMA_LIB=" $NUMA_BASE/lib/libnuma.a"
|
||||
NUMA_REV=829d10dac0230f99cd7e1778869d2adf3da24b65
|
||||
NUMA_INCLUDE=" -I /mnt/gvfs/third-party2/numa/$NUMA_REV/2.0.8/gcc-4.8.1-glibc-2.17/c3f970a/include/"
|
||||
NUMA_LIB=" /mnt/gvfs/third-party2/numa/$NUMA_REV/2.0.8/gcc-4.8.1-glibc-2.17/c3f970a/lib/libnuma.a"
|
||||
|
||||
# location of libunwind
|
||||
LIBUNWIND="$LIBUNWIND_BASE/lib/libunwind.a"
|
||||
|
||||
# location of tbb
|
||||
TBB_INCLUDE=" -isystem $TBB_BASE/include/"
|
||||
TBB_LIBS="$TBB_BASE/lib/libtbb.a"
|
||||
LIBUNWIND_REV=2c060e64064559905d46fd194000d61592087bdc
|
||||
LIBUNWIND="/mnt/gvfs/third-party2/libunwind/$LIBUNWIND_REV/1.1/gcc-4.8.1-glibc-2.17/675d945/lib/libunwind.a"
|
||||
|
||||
# use Intel SSE support for checksum calculations
|
||||
export USE_SSE=1
|
||||
|
||||
BINUTILS="$BINUTILS_BASE/bin"
|
||||
BINUTILS="/mnt/gvfs/third-party2/binutils/2aff2e7b474cd3e6ab23495ad1224b7d214b9f8e/2.21.1/centos6-native/da39a3e/bin"
|
||||
AR="$BINUTILS/ar"
|
||||
|
||||
DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP2_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
|
||||
DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE"
|
||||
|
||||
GCC_BASE="/mnt/gvfs/third-party2/gcc/1ec615e23800f0815d474478ba476a0adc3fe788/4.8.1/centos6-native/cc6c9dc"
|
||||
STDLIBS="-L $GCC_BASE/lib64"
|
||||
|
||||
if [ -z "$USE_CLANG" ]; then
|
||||
# gcc
|
||||
CC="$GCC_BASE/bin/gcc"
|
||||
CXX="$GCC_BASE/bin/g++"
|
||||
|
||||
|
||||
CFLAGS="-B$BINUTILS/gold -m64 -mtune=generic"
|
||||
CFLAGS+=" -isystem $GLIBC_INCLUDE"
|
||||
CFLAGS+=" -isystem $LIBGCC_INCLUDE"
|
||||
JEMALLOC=1
|
||||
else
|
||||
# clang
|
||||
CLANG_BIN="$CLANG_BASE/bin"
|
||||
CLANG_LIB="$CLANG_BASE/lib"
|
||||
CLANG_INCLUDE="$CLANG_LIB/clang/*/include"
|
||||
CC="$CLANG_BIN/clang"
|
||||
CXX="$CLANG_BIN/clang++"
|
||||
# clang
|
||||
CLANG_BASE="/mnt/gvfs/third-party2/clang/9ab68376f938992c4eb5946ca68f90c3185cffc8/3.4"
|
||||
CLANG_INCLUDE="$CLANG_BASE/gcc-4.8.1-glibc-2.17/fb0f730/lib/clang/3.4/include"
|
||||
CC="$CLANG_BASE/centos6-native/9cefd8a/bin/clang"
|
||||
CXX="$CLANG_BASE/centos6-native/9cefd8a/bin/clang++"
|
||||
|
||||
KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include/"
|
||||
KERNEL_HEADERS_INCLUDE="/mnt/gvfs/third-party2/kernel-headers/a683ed7135276731065a9d76d3016c9731f4e2f9/3.2.18_70_fbk11_00129_gc8882d0/gcc-4.8.1-glibc-2.17/da39a3e/include/"
|
||||
|
||||
CFLAGS="-B$BINUTILS/gold -nostdinc -nostdlib"
|
||||
CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/4.8.1 "
|
||||
@@ -93,21 +91,20 @@ else
|
||||
fi
|
||||
|
||||
CFLAGS+=" $DEPS_INCLUDE"
|
||||
CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE"
|
||||
CFLAGS+=" -DSNAPPY -DGFLAGS=google -DZLIB -DBZIP2 -DLZ4 -DZSTD -DNUMA -DTBB"
|
||||
CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE"
|
||||
CFLAGS+=" -DSNAPPY -DGFLAGS=google -DZLIB -DBZIP2 -DLZ4 -DZSTD -DNUMA"
|
||||
CXXFLAGS+=" $CFLAGS"
|
||||
|
||||
EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP2_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
|
||||
EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB"
|
||||
EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/gcc-4.8.1-glibc-2.17/lib/ld.so"
|
||||
EXEC_LDFLAGS+=" $LIBUNWIND"
|
||||
EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/gcc-4.8.1-glibc-2.17/lib"
|
||||
# required by libtbb
|
||||
EXEC_LDFLAGS+=" -ldl"
|
||||
|
||||
PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
|
||||
|
||||
EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP2_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS"
|
||||
EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS"
|
||||
|
||||
VALGRIND_VER="$VALGRIND_BASE/bin/"
|
||||
VALGRIND_REV=b2a9f85e4b70cd03abc85a7f3027fbc4cef35bd0
|
||||
VALGRIND_VER="/mnt/gvfs/third-party2/valgrind/$VALGRIND_REV/3.8.1/gcc-4.8.1-glibc-2.17/c3f970a/bin/"
|
||||
|
||||
export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,5 +1,5 @@
|
||||
#!/bin/bash
|
||||
# Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
# Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
# This source code is licensed under the BSD-style license found in the
|
||||
# LICENSE file in the root directory of this source tree. An additional grant
|
||||
# of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,7 +10,7 @@ then
|
||||
GIT="git"
|
||||
fi
|
||||
|
||||
# Print out the colored progress info so that it can be brainlessly
|
||||
# Print out the colored progress info so that it can be brainlessly
|
||||
# distinguished by users.
|
||||
function title() {
|
||||
echo -e "\033[1;32m$*\033[0m"
|
||||
|
||||
@@ -1,208 +0,0 @@
|
||||
#!/usr/local/fbcode/gcc-4.8.1-glibc-2.17-fb/bin/python2.7
|
||||
|
||||
from __future__ import absolute_import
|
||||
from __future__ import division
|
||||
from __future__ import print_function
|
||||
from __future__ import unicode_literals
|
||||
import argparse
|
||||
import commands
|
||||
import subprocess
|
||||
import sys
|
||||
import re
|
||||
import os
|
||||
import time
|
||||
|
||||
|
||||
#
|
||||
# Simple logger
|
||||
#
|
||||
|
||||
class Log:
|
||||
|
||||
def __init__(self, filename):
|
||||
self.filename = filename
|
||||
self.f = open(self.filename, 'w+', 0)
|
||||
|
||||
def caption(self, str):
|
||||
line = "\n##### %s #####\n" % str
|
||||
if self.f:
|
||||
self.f.write("%s \n" % line)
|
||||
else:
|
||||
print(line)
|
||||
|
||||
def error(self, str):
|
||||
data = "\n\n##### ERROR ##### %s" % str
|
||||
if self.f:
|
||||
self.f.write("%s \n" % data)
|
||||
else:
|
||||
print(data)
|
||||
|
||||
def log(self, str):
|
||||
if self.f:
|
||||
self.f.write("%s \n" % str)
|
||||
else:
|
||||
print(str)
|
||||
|
||||
#
|
||||
# Shell Environment
|
||||
#
|
||||
|
||||
|
||||
class Env(object):
|
||||
|
||||
def __init__(self, logfile, tests):
|
||||
self.tests = tests
|
||||
self.log = Log(logfile)
|
||||
|
||||
def shell(self, cmd, path=os.getcwd()):
|
||||
if path:
|
||||
os.chdir(path)
|
||||
|
||||
self.log.log("==== shell session ===========================")
|
||||
self.log.log("%s> %s" % (path, cmd))
|
||||
status = subprocess.call("cd %s; %s" % (path, cmd), shell=True,
|
||||
stdout=self.log.f, stderr=self.log.f)
|
||||
self.log.log("status = %s" % status)
|
||||
self.log.log("============================================== \n\n")
|
||||
return status
|
||||
|
||||
def GetOutput(self, cmd, path=os.getcwd()):
|
||||
if path:
|
||||
os.chdir(path)
|
||||
|
||||
self.log.log("==== shell session ===========================")
|
||||
self.log.log("%s> %s" % (path, cmd))
|
||||
status, out = commands.getstatusoutput(cmd)
|
||||
self.log.log("status = %s" % status)
|
||||
self.log.log("out = %s" % out)
|
||||
self.log.log("============================================== \n\n")
|
||||
return status, out
|
||||
|
||||
#
|
||||
# Pre-commit checker
|
||||
#
|
||||
|
||||
|
||||
class PreCommitChecker(Env):
|
||||
|
||||
def __init__(self, args):
|
||||
Env.__init__(self, args.logfile, args.tests)
|
||||
self.ignore_failure = args.ignore_failure
|
||||
|
||||
#
|
||||
# Get commands for a given job from the determinator file
|
||||
#
|
||||
def get_commands(self, test):
|
||||
status, out = self.GetOutput(
|
||||
"RATIO=1 build_tools/rocksdb-lego-determinator %s" % test, ".")
|
||||
return status, out
|
||||
|
||||
#
|
||||
# Run a specific CI job
|
||||
#
|
||||
def run_test(self, test):
|
||||
self.log.caption("Running test %s locally" % test)
|
||||
|
||||
# get commands for the CI job determinator
|
||||
status, cmds = self.get_commands(test)
|
||||
if status != 0:
|
||||
self.log.error("Error getting commands for test %s" % test)
|
||||
return False
|
||||
|
||||
# Parse the JSON to extract the commands to run
|
||||
cmds = re.findall("'shell':'([^\']*)'", cmds)
|
||||
|
||||
if len(cmds) == 0:
|
||||
self.log.log("No commands found")
|
||||
return False
|
||||
|
||||
# Run commands
|
||||
for cmd in cmds:
|
||||
# Replace J=<..> with the local environment variable
|
||||
if "J" in os.environ:
|
||||
cmd = cmd.replace("J=1", "J=%s" % os.environ["J"])
|
||||
cmd = cmd.replace("make ", "make -j%s " % os.environ["J"])
|
||||
# Run the command
|
||||
status = self.shell(cmd, ".")
|
||||
if status != 0:
|
||||
self.log.error("Error running command %s for test %s"
|
||||
% (cmd, test))
|
||||
return False
|
||||
|
||||
return True
|
||||
|
||||
#
|
||||
# Run specified CI jobs
|
||||
#
|
||||
def run_tests(self):
|
||||
if not self.tests:
|
||||
self.log.error("Invalid args. Please provide tests")
|
||||
return False
|
||||
|
||||
self.print_separator()
|
||||
self.print_row("TEST", "RESULT")
|
||||
self.print_separator()
|
||||
|
||||
result = True
|
||||
for test in self.tests:
|
||||
start_time = time.time()
|
||||
self.print_test(test)
|
||||
result = self.run_test(test)
|
||||
elapsed_min = (time.time() - start_time) / 60
|
||||
if not result:
|
||||
self.log.error("Error running test %s" % test)
|
||||
self.print_result("FAIL (%dm)" % elapsed_min)
|
||||
if not self.ignore_failure:
|
||||
return False
|
||||
result = False
|
||||
else:
|
||||
self.print_result("PASS (%dm)" % elapsed_min)
|
||||
|
||||
self.print_separator()
|
||||
return result
|
||||
|
||||
#
|
||||
# Print a line
|
||||
#
|
||||
def print_separator(self):
|
||||
print("".ljust(60, "-"))
|
||||
|
||||
#
|
||||
# Print two colums
|
||||
#
|
||||
def print_row(self, c0, c1):
|
||||
print("%s%s" % (c0.ljust(40), c1.ljust(20)))
|
||||
|
||||
def print_test(self, test):
|
||||
print(test.ljust(40), end="")
|
||||
sys.stdout.flush()
|
||||
|
||||
def print_result(self, result):
|
||||
print(result.ljust(20))
|
||||
|
||||
#
|
||||
# Main
|
||||
#
|
||||
parser = argparse.ArgumentParser(description='RocksDB pre-commit checker.')
|
||||
|
||||
# --log <logfile>
|
||||
parser.add_argument('--logfile', default='/tmp/precommit-check.log',
|
||||
help='Log file. Default is /tmp/precommit-check.log')
|
||||
# --ignore_failure
|
||||
parser.add_argument('--ignore_failure', action='store_true', default=False,
|
||||
help='Stop when an error occurs')
|
||||
# <test ....>
|
||||
parser.add_argument('tests', nargs='+',
|
||||
help='CI test(s) to run. e.g: unit punit asan tsan ubsan')
|
||||
|
||||
args = parser.parse_args()
|
||||
checker = PreCommitChecker(args)
|
||||
|
||||
print("Please follow log %s" % checker.log.filename)
|
||||
|
||||
if not checker.run_tests():
|
||||
print("Error running tests. Please check log file %s"
|
||||
% checker.log.filename)
|
||||
sys.exit(1)
|
||||
|
||||
sys.exit(0)
|
||||
@@ -243,7 +243,7 @@ make release
|
||||
--bloom_bits=10 \
|
||||
--num=$((NUM / 4)) \
|
||||
--reads=$((NUM / 4)) \
|
||||
--benchmark_write_rate_limit=$(( 110 * 1024 )) \
|
||||
--writes_per_second=1000 \
|
||||
--write_buffer_size=100000000 \
|
||||
--cache_size=6442450944 \
|
||||
--cache_numshardbits=6 \
|
||||
@@ -329,7 +329,7 @@ common_in_mem_args="--db=/dev/shm/rocksdb \
|
||||
--use_existing_db=1 \
|
||||
--duration=600 \
|
||||
--threads=32 \
|
||||
--benchmark_write_rate_limit=9502720 > ${STAT_FILE}.readwhilewriting_in_ram
|
||||
--writes_per_second=81920 > ${STAT_FILE}.readwhilewriting_in_ram
|
||||
|
||||
# Seekrandomwhilewriting
|
||||
./db_bench \
|
||||
@@ -342,7 +342,7 @@ common_in_mem_args="--db=/dev/shm/rocksdb \
|
||||
--use_tailing_iterator=1 \
|
||||
--duration=600 \
|
||||
--threads=32 \
|
||||
--benchmark_write_rate_limit=9502720 > ${STAT_FILE}.seekwhilewriting_in_ram
|
||||
--writes_per_second=81920 > ${STAT_FILE}.seekwhilewriting_in_ram
|
||||
|
||||
# measure fillseq with bunch of column families
|
||||
./db_bench \
|
||||
|
||||
@@ -8,7 +8,7 @@
|
||||
# Input Value
|
||||
# -------------------------------------------------------------------------
|
||||
# EMAIL Email address to report on trigger conditions
|
||||
# ONCALL Email address to raise a task on failure
|
||||
# ONCAL Email address to raise a task on failure
|
||||
# TRIGGER Trigger conditions for email. Valid values are fail, warn, all
|
||||
# SUBSCRIBER Email addresss to add as subscriber for task
|
||||
#
|
||||
@@ -42,9 +42,6 @@ if [ ! -z $ONCALL ]; then
|
||||
},"
|
||||
fi
|
||||
|
||||
# For now, create the tasks using only the dedicated task creation tool.
|
||||
CREATE_TASK=
|
||||
|
||||
REPORT=
|
||||
if [[ ! -z $REPORT_EMAIL || ! -z $CREATE_TASK ]]; then
|
||||
REPORT="'report': [
|
||||
@@ -59,58 +56,18 @@ fi
|
||||
CLEANUP_ENV="
|
||||
{
|
||||
'name':'Cleanup environment',
|
||||
'shell':'rm -rf /dev/shm/rocksdb && mkdir /dev/shm/rocksdb && (chmod +t /dev/shm || true) && make clean',
|
||||
'shell':'rm -rf /dev/shm/rocksdb && mkdir /dev/shm/rocksdb && make clean',
|
||||
'user':'root'
|
||||
}"
|
||||
|
||||
# We will eventually set the RATIO to 1, but we want do this
|
||||
# in steps. RATIO=$(nproc) will make it work as J=1
|
||||
if [ -z $RATIO ]; then
|
||||
RATIO=$(nproc)
|
||||
fi
|
||||
|
||||
if [ -z $PARALLEL_J ]; then
|
||||
PARALLEL_J="J=$(expr $(nproc) / ${RATIO})"
|
||||
fi
|
||||
|
||||
if [ -z $PARALLEL_j ]; then
|
||||
PARALLEL_j="-j$(expr $(nproc) / ${RATIO})"
|
||||
fi
|
||||
|
||||
PARALLELISM="$PARALLEL_J $PARALLEL_j"
|
||||
|
||||
DEBUG="OPT=-g"
|
||||
SHM="TEST_TMPDIR=/dev/shm/rocksdb"
|
||||
GCC_481="ROCKSDB_FBCODE_BUILD_WITH_481=1"
|
||||
ASAN="COMPILE_WITH_ASAN=1"
|
||||
CLANG="USE_CLANG=1"
|
||||
LITE="OPT=\"-DROCKSDB_LITE -g\""
|
||||
LITE="OPT=-DROCKSDB_LITE"
|
||||
TSAN="COMPILE_WITH_TSAN=1"
|
||||
UBSAN="COMPILE_WITH_UBSAN=1"
|
||||
DISABLE_JEMALLOC="DISABLE_JEMALLOC=1"
|
||||
HTTP_PROXY="https_proxy=http://fwdproxy.29.prn1:8080 http_proxy=http://fwdproxy.29.prn1:8080 ftp_proxy=http://fwdproxy.29.prn1:8080"
|
||||
SETUP_JAVA_ENV="export $HTTP_PROXY; export JAVA_HOME=/usr/local/jdk-7u10-64/; export PATH=\$PATH:\$JAVA_HOME/bin"
|
||||
PARSER="'parser':'python build_tools/error_filter.py $1'"
|
||||
|
||||
CONTRUN_NAME="ROCKSDB_CONTRUN_NAME"
|
||||
|
||||
# This code is getting called under various scenarios. What we care about is to
|
||||
# understand when it's called from nightly contruns because in that case we'll
|
||||
# create tasks for any failures. To follow the existing pattern, we'll check
|
||||
# the value of $ONCALL. If it's a diff then just call `false` to make sure
|
||||
# that errors will be properly propagated to the caller.
|
||||
if [ ! -z $ONCALL ]; then
|
||||
TASK_CREATION_TOOL="/usr/local/bin/mysql_mtr_filter --rocksdb"
|
||||
else
|
||||
TASK_CREATION_TOOL="false"
|
||||
fi
|
||||
|
||||
ARTIFACTS=" 'artifacts': [
|
||||
{
|
||||
'name':'database',
|
||||
'paths':[ '/dev/shm/rocksdb' ],
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# A mechanism to disable tests temporarily
|
||||
@@ -142,16 +99,14 @@ PARALLEL_UNIT_TEST_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug version',
|
||||
'shell':'$DEBUG make -j$(nproc) all && $SHM make check > /dev/null 2>&1 || $CONTRUN_NAME=punit_check $TASK_CREATION_TOOL || cat t/log-*',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$DEBUG make -j$(nproc) all && $SHM make check > /dev/null 2>&1 || cat t/log-*',
|
||||
'user':'root'
|
||||
},
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug version under gcc-4.8.1',
|
||||
'shell':'$GCC_481 $DEBUG make -j$(nproc) all && $SHM make check > /dev/null 2>&1 || $CONTRUN_NAME=punit_check_gcc481 $TASK_CREATION_TOOL || cat t/log-*',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$GCC_481 $DEBUG make -j$(nproc) all && $SHM make check > /dev/null 2>&1 || cat t/log-*',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -169,51 +124,8 @@ UNIT_TEST_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug version',
|
||||
'shell':'$SHM $DEBUG make $PARALLELISM check || $CONTRUN_NAME=check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB unit test not under /dev/shm
|
||||
#
|
||||
UNIT_TEST_NON_SHM_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Unit Test',
|
||||
'oncall':'$ONCALL',
|
||||
'timeout': 86400,
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug version',
|
||||
'timeout': 86400,
|
||||
'shell':'$DEBUG make $PARALLELISM check || $CONTRUN_NAME=non_shm_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB release build and unit tests
|
||||
#
|
||||
RELEASE_BUILD_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Release Build',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB release',
|
||||
'shell':'make $PARALLEL_j release || $CONTRUN_NAME=release $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG make J=1 check',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -231,29 +143,8 @@ UNIT_TEST_COMMANDS_481="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug version',
|
||||
'shell':'$SHM $GCC_481 $DEBUG make $PARALLELISM check || $CONTRUN_NAME=unit_gcc_481_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB release build and unit tests
|
||||
#
|
||||
RELEASE_BUILD_COMMANDS_481="[
|
||||
{
|
||||
'name':'Rocksdb Release on GCC 4.8.1',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB release on GCC 4.8.1',
|
||||
'shell':'$GCC_481 make $PARALLEL_j release || $CONTRUN_NAME=release_gcc481 $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $GCC_481 $DEBUG make J=1 check',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -271,29 +162,8 @@ CLANG_UNIT_TEST_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and test RocksDB debug',
|
||||
'shell':'$CLANG $SHM $DEBUG make $PARALLELISM check || $CONTRUN_NAME=clang_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB release build with CLANG
|
||||
#
|
||||
CLANG_RELEASE_BUILD_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb CLANG Release Build',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB release',
|
||||
'shell':'$CLANG make $PARALLEL_j release|| $CONTRUN_NAME=clang_release $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$CLANG $SHM $DEBUG make J=1 check',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -311,9 +181,8 @@ CLANG_ANALYZE_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'RocksDB build and analyze',
|
||||
'shell':'$CLANG $SHM $DEBUG make $PARALLEL_j analyze || $CONTRUN_NAME=clang_analyze $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$CLANG $SHM $DEBUG make J=1 analyze',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -331,9 +200,8 @@ CODE_COV_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build, test and collect code coverage info',
|
||||
'shell':'$SHM $DEBUG make $PARALLELISM coverage || $CONTRUN_NAME=coverage $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG make J=1 coverage',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -351,9 +219,8 @@ UNITY_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build, test unity test',
|
||||
'shell':'$SHM $DEBUG V=1 make J=1 unity_test || $CONTRUN_NAME=unity_test $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG V=1 make J=1 unity_test',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -371,29 +238,8 @@ LITE_BUILD_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB debug version',
|
||||
'shell':'$LITE make J=1 static_lib || $CONTRUN_NAME=lite_static_lib $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB lite tests
|
||||
#
|
||||
LITE_UNIT_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Lite Unit Test',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB debug version',
|
||||
'shell':'$SHM $LITE make J=1 check || $CONTRUN_NAME=lite_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$LITE $DEBUG make J=1 static_lib',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -412,45 +258,20 @@ STRESS_CRASH_TEST_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and run RocksDB debug stress tests',
|
||||
'shell':'$SHM $DEBUG make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG make J=1 db_stress',
|
||||
'user':'root'
|
||||
},
|
||||
{
|
||||
'name':'Build and run RocksDB debug crash tests',
|
||||
'timeout': 86400,
|
||||
'shell':'$SHM $DEBUG make J=1 crash_test || $CONTRUN_NAME=crash_test $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG make J=1 crash_test',
|
||||
'user':'root'
|
||||
}
|
||||
],
|
||||
$ARTIFACTS,
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
# RocksDB write stress test.
|
||||
# We run on disk device on purpose (i.e. no $SHM)
|
||||
# because we want to add some randomness to fsync commands
|
||||
WRITE_STRESS_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Write Stress Test',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and run RocksDB write stress tests',
|
||||
'shell':'make write_stress && python tools/write_stress_runner.py --runtime_sec=3600 --db=/tmp/rocksdb_write_stress || $CONTRUN_NAME=write_stress $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
}
|
||||
],
|
||||
'artifacts': [{'name': 'database', 'paths': ['/tmp/rocksdb_write_stress']}],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
|
||||
#
|
||||
# RocksDB test under address sanitizer
|
||||
#
|
||||
@@ -462,9 +283,8 @@ ASAN_TEST_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Test RocksDB debug under ASAN',
|
||||
'shell':'set -o pipefail && ($SHM $ASAN $DEBUG make $PARALLELISM asan_check || $CONTRUN_NAME=asan_check $TASK_CREATION_TOOL) |& /usr/facebook/ops/scripts/asan_symbolize.py -d',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'set -o pipefail && $SHM $ASAN $DEBUG make J=1 asan_check |& /usr/facebook/ops/scripts/asan_symbolize.py -d',
|
||||
'user':'root'
|
||||
}
|
||||
],
|
||||
$REPORT
|
||||
@@ -484,51 +304,8 @@ ASAN_CRASH_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Build and run RocksDB debug asan_crash_test',
|
||||
'timeout': 86400,
|
||||
'shell':'$SHM $DEBUG make J=1 asan_crash_test || $CONTRUN_NAME=asan_crash_test $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB test under undefined behavior sanitizer
|
||||
#
|
||||
UBSAN_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Unit Test under UBSAN',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Test RocksDB debug under UBSAN',
|
||||
'shell':'set -o pipefail && $SHM $UBSAN $DEBUG make $PARALLELISM ubsan_check || $CONTRUN_NAME=ubsan_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
}
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB crash testing under udnefined behavior sanitizer
|
||||
#
|
||||
UBSAN_CRASH_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb crash test under UBSAN',
|
||||
'oncall':'$ONCALL',
|
||||
'timeout': 86400,
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build and run RocksDB debug ubsan_crash_test',
|
||||
'timeout': 86400,
|
||||
'shell':'$SHM $DEBUG make J=1 ubsan_crash_test || $CONTRUN_NAME=ubsan_crash_test $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$SHM $DEBUG make J=1 asan_crash_test',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -542,15 +319,12 @@ VALGRIND_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Unit Test under valgrind',
|
||||
'oncall':'$ONCALL',
|
||||
'timeout': 86400,
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Run RocksDB debug unit tests',
|
||||
'timeout': 86400,
|
||||
'shell':'$DISABLE_JEMALLOC $SHM $DEBUG make $PARALLELISM valgrind_check || $CONTRUN_NAME=valgrind_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'$DISABLE_JEMALLOC $SHM $DEBUG make valgrind_check',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -564,15 +338,12 @@ TSAN_UNIT_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Unit Test under TSAN',
|
||||
'oncall':'$ONCALL',
|
||||
'timeout': 86400,
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Run RocksDB debug unit test',
|
||||
'timeout': 86400,
|
||||
'shell':'set -o pipefail && $SHM $DEBUG $TSAN make $PARALLELISM check || $CONTRUN_NAME=tsan_check $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'set -o pipefail && $SHM $DEBUG $TSAN make J=1 check',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -592,9 +363,8 @@ TSAN_CRASH_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Compile and run',
|
||||
'timeout': 86400,
|
||||
'shell':'set -o pipefail && $SHM $DEBUG $TSAN CRASH_TEST_KILL_ODD=1887 make J=1 crash_test || $CONTRUN_NAME=tsan_crash_test $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'set -o pipefail && $SHM $DEBUG $TSAN make J=1 crash_test',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -623,9 +393,9 @@ run_format_compatible()
|
||||
if [ -e "build_detect_platform" ]
|
||||
then
|
||||
sed "s/tcmalloc/nothingnothingnothing/g" build_detect_platform > $TEST_TMPDIR/temp_build_file
|
||||
rm -rf build_detect_platform
|
||||
rm -rf build_detect_platform
|
||||
cp $TEST_TMPDIR/temp_build_file build_detect_platform
|
||||
chmod +x build_detect_platform
|
||||
chmod +x build_detect_platform
|
||||
fi
|
||||
|
||||
make ldb -j32
|
||||
@@ -654,9 +424,8 @@ FORMAT_COMPATIBLE_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Run RocksDB debug unit test',
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_format_compatible || $CONTRUN_NAME=run_format_compatible $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_format_compatible',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -687,9 +456,8 @@ NO_COMPRESSION_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Run RocksDB debug unit test',
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_no_compression || $CONTRUN_NAME=run_no_compression $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_no_compression',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
@@ -744,36 +512,14 @@ REGRESSION_COMMANDS="[
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Make and run script',
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_regression || $CONTRUN_NAME=run_regression $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
'shell':'build_tools/rocksdb-lego-determinator run_regression',
|
||||
'user':'root'
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
#
|
||||
# RocksDB Java build
|
||||
#
|
||||
JAVA_BUILD_TEST_COMMANDS="[
|
||||
{
|
||||
'name':'Rocksdb Java Build',
|
||||
'oncall':'$ONCALL',
|
||||
'steps': [
|
||||
$CLEANUP_ENV,
|
||||
{
|
||||
'name':'Build RocksDB for Java',
|
||||
'shell':'$SETUP_JAVA_ENV; $SHM make rocksdbjava || $CONTRUN_NAME=rocksdbjava $TASK_CREATION_TOOL',
|
||||
'user':'root',
|
||||
$PARSER
|
||||
},
|
||||
],
|
||||
$REPORT
|
||||
}
|
||||
]"
|
||||
|
||||
|
||||
case $1 in
|
||||
punit)
|
||||
echo $PARALLEL_UNIT_TEST_COMMANDS
|
||||
@@ -781,24 +527,12 @@ case $1 in
|
||||
unit)
|
||||
echo $UNIT_TEST_COMMANDS
|
||||
;;
|
||||
unit_non_shm)
|
||||
echo $UNIT_TEST_NON_SHM_COMMANDS
|
||||
;;
|
||||
release)
|
||||
echo $RELEASE_BUILD_COMMANDS
|
||||
;;
|
||||
unit_481)
|
||||
echo $UNIT_TEST_COMMANDS_481
|
||||
;;
|
||||
release_481)
|
||||
echo $RELEASE_BUILD_COMMANDS_481
|
||||
;;
|
||||
clang_unit)
|
||||
echo $CLANG_UNIT_TEST_COMMANDS
|
||||
;;
|
||||
clang_release)
|
||||
echo $CLANG_RELEASE_BUILD_COMMANDS
|
||||
;;
|
||||
clang_analyze)
|
||||
echo $CLANG_ANALYZE_COMMANDS
|
||||
;;
|
||||
@@ -811,27 +545,15 @@ case $1 in
|
||||
lite)
|
||||
echo $LITE_BUILD_COMMANDS
|
||||
;;
|
||||
lite_test)
|
||||
echo $LITE_UNIT_TEST_COMMANDS
|
||||
;;
|
||||
stress_crash)
|
||||
echo $STRESS_CRASH_TEST_COMMANDS
|
||||
;;
|
||||
write_stress)
|
||||
echo $WRITE_STRESS_COMMANDS
|
||||
;;
|
||||
asan)
|
||||
echo $ASAN_TEST_COMMANDS
|
||||
;;
|
||||
asan_crash)
|
||||
echo $ASAN_CRASH_TEST_COMMANDS
|
||||
;;
|
||||
ubsan)
|
||||
echo $UBSAN_TEST_COMMANDS
|
||||
;;
|
||||
ubsan_crash)
|
||||
echo $UBSAN_CRASH_TEST_COMMANDS
|
||||
;;
|
||||
valgrind)
|
||||
echo $VALGRIND_TEST_COMMANDS
|
||||
;;
|
||||
@@ -859,9 +581,6 @@ case $1 in
|
||||
run_regression)
|
||||
run_regression
|
||||
;;
|
||||
java_build)
|
||||
echo $JAVA_BUILD_TEST_COMMANDS
|
||||
;;
|
||||
*)
|
||||
echo "Invalid determinator command"
|
||||
;;
|
||||
|
||||
+93
-171
@@ -1,19 +1,14 @@
|
||||
# This script enables you running RocksDB tests by running
|
||||
# All the tests in parallel and utilizing all the cores
|
||||
# All the tests in paralell and utilizing all the cores
|
||||
# For db_test the script first lists and parses the tests
|
||||
# and then fires them up in parallel using async PS Job functionality
|
||||
# Run the script from the enlistment
|
||||
Param(
|
||||
[switch]$EnableJE = $false, # Use je executable
|
||||
[switch]$EnableRerun = $false, # Rerun failed tests sequentially at the end
|
||||
[string]$WorkFolder = "", # Direct tests to use that folder
|
||||
[int]$Limit = -1, # -1 means run all otherwise limit for testing purposes
|
||||
[string]$Exclude = "", # Expect a comma separated list, no spaces
|
||||
[string]$Run = "db_test", # Run db_test|db_test2|tests|testname1,testname2...
|
||||
# Number of async tasks that would run concurrently. Recommend a number below 64.
|
||||
# However, CPU utlization really depends on the storage media. Recommend ram based disk.
|
||||
# a value of 1 will run everything serially
|
||||
[int]$Concurrency = 16
|
||||
[string]$Run = "db_test" # Run db_test|tests
|
||||
)
|
||||
|
||||
# Folders and commands must be fullpath to run assuming
|
||||
@@ -46,25 +41,15 @@ if($WorkFolder -eq "") {
|
||||
$Env:TEST_TMPDIR = $WorkFolder
|
||||
}
|
||||
|
||||
Write-Output "Root: $RootFolder, WorkFolder: $WorkFolder"
|
||||
|
||||
# Use JEMALLOC executables
|
||||
if($Run -ceq "db_test" -or
|
||||
$Run -ceq "db_test2" ) {
|
||||
|
||||
$file_name = $Run
|
||||
|
||||
if($EnableJE) {
|
||||
$file_name += "_je"
|
||||
}
|
||||
|
||||
$file_name += ".exe"
|
||||
|
||||
$db_test = -Join ($BinariesFolder, $file_name)
|
||||
|
||||
Write-Output "Binaries: $BinariesFolder db_test: $db_test"
|
||||
if($EnableJE) {
|
||||
$db_test = -Join ($BinariesFolder, "db_test_je.exe")
|
||||
} else {
|
||||
$db_test = -Join ($BinariesFolder, "db_test.exe")
|
||||
}
|
||||
|
||||
Write-Output "Root: $RootFolder, WorkFolder: $WorkFolder"
|
||||
Write-Output "Binaries: $BinariesFolder exe: $db_test"
|
||||
|
||||
#Exclusions that we do not want to run
|
||||
$ExcludeTests = New-Object System.Collections.Generic.HashSet[string]
|
||||
@@ -72,7 +57,7 @@ $ExcludeTests = New-Object System.Collections.Generic.HashSet[string]
|
||||
|
||||
if($Exclude -ne "") {
|
||||
Write-Host "Exclude: $Exclude"
|
||||
$l = $Exclude -split ' '
|
||||
$l = $Exclude -split ','
|
||||
ForEach($t in $l) { $ExcludeTests.Add($t) | Out-Null }
|
||||
}
|
||||
|
||||
@@ -124,70 +109,38 @@ function Normalize-DbTests($HashTable) {
|
||||
|
||||
$test_log = $test -replace '[\./]','_'
|
||||
$test_log += ".log"
|
||||
$log_path = -join ($LogFolder, $test_log)
|
||||
|
||||
# Add to a hashtable
|
||||
$HashTable.Add($test, $log_path);
|
||||
$HashTable.Add($test, $test_log);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
# The function removes trailing .exe siffix if any,
|
||||
# creates a name for the log file
|
||||
function MakeAndAdd([string]$token, $HashTable) {
|
||||
$test_name = $token -replace '.exe$', ''
|
||||
$log_name = -join ($test_name, ".log")
|
||||
$log_path = -join ($LogFolder, $log_name)
|
||||
if(!$ExcludeTests.Contains($test_name)) {
|
||||
$HashTable.Add($test_name, $log_path)
|
||||
} else {
|
||||
Write-Warning "Test $test_name is excluded"
|
||||
}
|
||||
}
|
||||
|
||||
# The function scans build\Debug folder to discover
|
||||
# Test executables. It then populates a table with
|
||||
# Test executable name -> Log file
|
||||
function Discover-TestBinaries([string]$Pattern, $HashTable) {
|
||||
function Discover-TestBinaries($HashTable) {
|
||||
|
||||
$Exclusions = @("db_test*", "db_sanity_test*")
|
||||
|
||||
$p = -join ($BinariesFolder, $pattern)
|
||||
|
||||
Write-Host "Path: $p"
|
||||
$p = -join ($BinariesFolder, "*_test*.exe")
|
||||
|
||||
dir -Path $p -Exclude $Exclusions | ForEach-Object {
|
||||
MakeAndAdd -token ($_.Name) -HashTable $HashTable
|
||||
$t = ($_.Name) -replace '.exe$', ''
|
||||
$test_log = -join ($t, ".log")
|
||||
$HashTable.Add($t, $test_log)
|
||||
}
|
||||
}
|
||||
|
||||
$TestsToRun = [ordered]@{}
|
||||
$TestToLog = [ordered]@{}
|
||||
|
||||
if($Run -ceq "db_test" -or
|
||||
$Run -ceq "db_test2") {
|
||||
Normalize-DbTests -HashTable $TestsToRun
|
||||
if($Run -ceq "db_test") {
|
||||
Normalize-DbTests -HashTable $TestToLog
|
||||
} elseif($Run -ceq "tests") {
|
||||
if($EnableJE) {
|
||||
$pattern = "*_test_je.exe"
|
||||
} else {
|
||||
$pattern = "*_test.exe"
|
||||
}
|
||||
Discover-TestBinaries -Pattern $pattern -HashTable $TestsToRun
|
||||
} else {
|
||||
|
||||
$test_list = $Run -split ' '
|
||||
|
||||
ForEach($t in $test_list) {
|
||||
MakeAndAdd -token $t -HashTable $TestsToRun
|
||||
}
|
||||
Discover-TestBinaries -HashTable $TestToLog
|
||||
}
|
||||
|
||||
$NumTestsToStart = $TestsToRun.Count
|
||||
if($Limit -ge 0 -and $NumTestsToStart -gt $Limit) {
|
||||
$NumTestsToStart = $Limit
|
||||
}
|
||||
|
||||
Write-Host "Attempting to start: $NumTestsToStart tests"
|
||||
Write-Host "Attempting to start: " ($TestToLog.Count) " tests"
|
||||
|
||||
# Invoke a test with a filter and redirect all output
|
||||
$InvokeTestCase = {
|
||||
@@ -201,124 +154,93 @@ $InvokeTestAsync = {
|
||||
&$exe > $log 2>&1
|
||||
}
|
||||
|
||||
# Hash that contains tests to rerun if any failed
|
||||
# Those tests will be rerun sequentially
|
||||
$Rerun = [ordered]@{}
|
||||
$jobs = @()
|
||||
$JobToLog = @{}
|
||||
# Test limiting factor here
|
||||
$count = 0
|
||||
# Overall status
|
||||
[bool]$success = $true;
|
||||
|
||||
function RunJobs($TestToLog, [int]$ConcurrencyVal, [bool]$AddForRerun)
|
||||
{
|
||||
# Array to wait for any of the running jobs
|
||||
$jobs = @()
|
||||
# Hash JobToLog
|
||||
$JobToLog = @{}
|
||||
ForEach($k in $TestToLog.keys) {
|
||||
|
||||
# Wait for all to finish and get the results
|
||||
while(($JobToLog.Count -gt 0) -or
|
||||
($TestToLog.Count -gt 0)) {
|
||||
Write-Host "Starting $k"
|
||||
$log_path = -join ($LogFolder, ($TestToLog.$k))
|
||||
|
||||
# Make sure we have maximum concurrent jobs running if anything
|
||||
# and the $Limit either not set or allows to proceed
|
||||
while(($JobToLog.Count -lt $ConcurrencyVal) -and
|
||||
(($TestToLog.Count -gt 0) -and
|
||||
(($Limit -lt 0) -or ($count -lt $Limit)))) {
|
||||
if($Run -ceq "db_test") {
|
||||
$job = Start-Job -Name $k -ScriptBlock $InvokeTestCase -ArgumentList @($db_test,$k,$log_path)
|
||||
} else {
|
||||
[string]$Exe = -Join ($BinariesFolder, $k)
|
||||
$job = Start-Job -Name $k -ScriptBlock $InvokeTestAsync -ArgumentList @($exe,$log_path)
|
||||
}
|
||||
|
||||
$JobToLog.Add($job, $log_path)
|
||||
|
||||
# We only need the first key
|
||||
foreach($key in $TestToLog.keys) {
|
||||
$k = $key
|
||||
break
|
||||
}
|
||||
|
||||
Write-Host "Starting $k"
|
||||
$log_path = ($TestToLog.$k)
|
||||
|
||||
if($Run -ceq "db_test" -or
|
||||
$Run -ceq "db_test2") {
|
||||
$job = Start-Job -Name $k -ScriptBlock $InvokeTestCase -ArgumentList @($db_test,$k,$log_path)
|
||||
} else {
|
||||
[string]$Exe = -Join ($BinariesFolder, $k)
|
||||
$job = Start-Job -Name $k -ScriptBlock $InvokeTestAsync -ArgumentList @($exe,$log_path)
|
||||
}
|
||||
|
||||
$JobToLog.Add($job, $log_path)
|
||||
$TestToLog.Remove($k)
|
||||
|
||||
++$count
|
||||
}
|
||||
|
||||
if($JobToLog.Count -lt 1) {
|
||||
break
|
||||
}
|
||||
|
||||
$jobs = @()
|
||||
foreach($k in $JobToLog.Keys) { $jobs += $k }
|
||||
|
||||
$completed = Wait-Job -Job $jobs -Any
|
||||
$log = $JobToLog[$completed]
|
||||
$JobToLog.Remove($completed)
|
||||
|
||||
$message = -join @($completed.Name, " State: ", ($completed.State))
|
||||
|
||||
$log_content = @(Get-Content $log)
|
||||
|
||||
if($completed.State -ne "Completed") {
|
||||
$success = $false
|
||||
Write-Warning $message
|
||||
$log_content | Write-Warning
|
||||
} else {
|
||||
# Scan the log. If we find PASSED and no occurrence of FAILED
|
||||
# then it is a success
|
||||
[bool]$pass_found = $false
|
||||
ForEach($l in $log_content) {
|
||||
|
||||
if(($l -match "^\[\s+FAILED") -or
|
||||
($l -match "Assertion failed:")) {
|
||||
$pass_found = $false
|
||||
break
|
||||
}
|
||||
|
||||
if(($l -match "^\[\s+PASSED") -or
|
||||
($l -match " : PASSED$") -or
|
||||
($l -match "^PASS$") -or # Special c_test case
|
||||
($l -match "Passed all tests!") ) {
|
||||
$pass_found = $true
|
||||
}
|
||||
}
|
||||
|
||||
if(!$pass_found) {
|
||||
$success = $false;
|
||||
Write-Warning $message
|
||||
$log_content | Write-Warning
|
||||
if($AddForRerun) {
|
||||
$Rerun.Add($completed.Name, $log)
|
||||
}
|
||||
} else {
|
||||
Write-Host $message
|
||||
}
|
||||
}
|
||||
|
||||
# Remove cached job info from the system
|
||||
# Should be no output
|
||||
Receive-Job -Job $completed | Out-Null
|
||||
# Limiting trial runs
|
||||
if(($Limit -gt 0) -and (++$count -ge $Limit)) {
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
RunJobs -TestToLog $TestsToRun -ConcurrencyVal $Concurrency -AddForRerun $EnableRerun
|
||||
[bool]$success = $true;
|
||||
|
||||
if($Rerun.Count -gt 0) {
|
||||
Write-Host "Rerunning " ($Rerun.Count) " tests sequentially"
|
||||
$success = $true
|
||||
$count = 0
|
||||
RunJobs -TestToLog $Rerun -ConcurrencyVal 1 -AddForRerun $false
|
||||
# Wait for all to finish and get the results
|
||||
while($JobToLog.Count -gt 0) {
|
||||
|
||||
$jobs = @()
|
||||
foreach($k in $JobToLog.Keys) { $jobs += $k }
|
||||
|
||||
<#
|
||||
if(!$success) {
|
||||
break
|
||||
}
|
||||
#>
|
||||
|
||||
$completed = Wait-Job -Job $jobs -Any
|
||||
$log = $JobToLog[$completed]
|
||||
$JobToLog.Remove($completed)
|
||||
|
||||
$message = -join @($completed.Name, " State: ", ($completed.State))
|
||||
|
||||
$log_content = @(Get-Content $log)
|
||||
|
||||
if($completed.State -ne "Completed") {
|
||||
$success = $false
|
||||
Write-Warning $message
|
||||
$log_content | Write-Warning
|
||||
} else {
|
||||
# Scan the log. If we find PASSED and no occurence of FAILED
|
||||
# then it is a success
|
||||
[bool]$pass_found = $false
|
||||
ForEach($l in $log_content) {
|
||||
|
||||
if(($l -match "^\[\s+FAILED") -or
|
||||
($l -match "Assertion failed:")) {
|
||||
$pass_found = $false
|
||||
break
|
||||
}
|
||||
|
||||
if(($l -match "^\[\s+PASSED") -or
|
||||
($l -match " : PASSED$") -or
|
||||
($l -match "^PASSED") -or
|
||||
($l -match "Passed all tests!") ) {
|
||||
$pass_found = $true
|
||||
}
|
||||
}
|
||||
|
||||
if(!$pass_found) {
|
||||
$success = $false;
|
||||
Write-Warning $message
|
||||
$log_content | Write-Warning
|
||||
} else {
|
||||
Write-Host $message
|
||||
}
|
||||
}
|
||||
|
||||
# Remove cached job info from the system
|
||||
# Should be no output
|
||||
Receive-Job -Job $completed | Out-Null
|
||||
}
|
||||
|
||||
Get-Date
|
||||
|
||||
|
||||
if(!$success) {
|
||||
# This does not succeed killing off jobs quick
|
||||
# So we simply exit
|
||||
@@ -327,4 +249,4 @@ if(!$success) {
|
||||
exit 12345
|
||||
}
|
||||
|
||||
|
||||
|
||||
@@ -1,129 +0,0 @@
|
||||
#!/bin/sh
|
||||
#
|
||||
# Update dependencies.sh file with the latest avaliable versions
|
||||
|
||||
BASEDIR=$(dirname $0)
|
||||
OUTPUT=""
|
||||
|
||||
function log_variable()
|
||||
{
|
||||
echo "$1=${!1}" >> "$OUTPUT"
|
||||
}
|
||||
|
||||
|
||||
TP2_LATEST="/mnt/vol/engshare/fbcode/third-party2"
|
||||
## $1 => lib name
|
||||
## $2 => lib version (if not provided, will try to pick latest)
|
||||
## $3 => platform (if not provided, will try to pick latest gcc)
|
||||
##
|
||||
## get_lib_base will set a variable named ${LIB_NAME}_BASE to the lib location
|
||||
function get_lib_base()
|
||||
{
|
||||
local lib_name=$1
|
||||
local lib_version=$2
|
||||
local lib_platform=$3
|
||||
|
||||
local result="$TP2_LATEST/$lib_name/"
|
||||
|
||||
# Lib Version
|
||||
if [ -z "$lib_version" ] || [ "$lib_version" = "LATEST" ]; then
|
||||
# version is not provided, use latest
|
||||
result=`ls -dr1v $result/*/ | head -n1`
|
||||
else
|
||||
result="$result/$lib_version/"
|
||||
fi
|
||||
|
||||
# Lib Platform
|
||||
if [ -z "$lib_platform" ]; then
|
||||
# platform is not provided, use latest gcc
|
||||
result=`ls -dr1v $result/gcc-*[^fb]/ | head -n1`
|
||||
else
|
||||
result="$result/$lib_platform/"
|
||||
fi
|
||||
|
||||
result=`ls -1d $result/*/ | head -n1`
|
||||
|
||||
# lib_name => LIB_NAME_BASE
|
||||
local __res_var=${lib_name^^}"_BASE"
|
||||
__res_var=`echo $__res_var | tr - _`
|
||||
# LIB_NAME_BASE=$result
|
||||
eval $__res_var=`readlink -f $result`
|
||||
|
||||
log_variable $__res_var
|
||||
}
|
||||
|
||||
###########################################################
|
||||
# 4.9.x dependencies #
|
||||
###########################################################
|
||||
|
||||
OUTPUT="$BASEDIR/dependencies.sh"
|
||||
|
||||
rm -f "$OUTPUT"
|
||||
touch "$OUTPUT"
|
||||
|
||||
echo "Writing dependencies to $OUTPUT"
|
||||
|
||||
# Compilers locations
|
||||
GCC_BASE=`readlink -f $TP2_LATEST/gcc/4.9.x/centos6-native/*/`
|
||||
CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos6-native/*/`
|
||||
|
||||
log_variable GCC_BASE
|
||||
log_variable CLANG_BASE
|
||||
|
||||
# Libraries locations
|
||||
get_lib_base libgcc 4.9.x
|
||||
get_lib_base glibc 2.20
|
||||
get_lib_base snappy LATEST gcc-4.9-glibc-2.20
|
||||
get_lib_base zlib LATEST
|
||||
get_lib_base bzip2 LATEST
|
||||
get_lib_base lz4 LATEST
|
||||
get_lib_base zstd LATEST
|
||||
get_lib_base gflags LATEST
|
||||
get_lib_base jemalloc LATEST
|
||||
get_lib_base numa LATEST
|
||||
get_lib_base libunwind LATEST
|
||||
get_lib_base tbb 4.0_update2 gcc-4.9-glibc-2.20
|
||||
|
||||
get_lib_base kernel-headers LATEST
|
||||
get_lib_base binutils LATEST centos6-native
|
||||
get_lib_base valgrind 3.10.0 gcc-4.9-glibc-2.20
|
||||
|
||||
git diff $OUTPUT
|
||||
|
||||
###########################################################
|
||||
# 4.8.1 dependencies #
|
||||
###########################################################
|
||||
|
||||
OUTPUT="$BASEDIR/dependencies_4.8.1.sh"
|
||||
|
||||
rm -f "$OUTPUT"
|
||||
touch "$OUTPUT"
|
||||
|
||||
echo "Writing 4.8.1 dependencies to $OUTPUT"
|
||||
|
||||
# Compilers locations
|
||||
GCC_BASE=`readlink -f $TP2_LATEST/gcc/4.8.1/centos6-native/*/`
|
||||
CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos6-native/*/`
|
||||
|
||||
log_variable GCC_BASE
|
||||
log_variable CLANG_BASE
|
||||
|
||||
# Libraries locations
|
||||
get_lib_base libgcc 4.8.1 gcc-4.8.1-glibc-2.17
|
||||
get_lib_base glibc 2.17 gcc-4.8.1-glibc-2.17
|
||||
get_lib_base snappy LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base zlib LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base bzip2 LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base lz4 LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base zstd LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base gflags LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base jemalloc LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base numa LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base libunwind LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base tbb 4.0_update2 gcc-4.8.1-glibc-2.17
|
||||
|
||||
get_lib_base kernel-headers LATEST gcc-4.8.1-glibc-2.17
|
||||
get_lib_base binutils LATEST centos6-native
|
||||
get_lib_base valgrind 3.8.1 gcc-4.8.1-glibc-2.17
|
||||
|
||||
git diff $OUTPUT
|
||||
+17
-53
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -15,7 +15,6 @@
|
||||
|
||||
#include "db/compaction_iterator.h"
|
||||
#include "db/dbformat.h"
|
||||
#include "db/event_helpers.h"
|
||||
#include "db/filename.h"
|
||||
#include "db/internal_stats.h"
|
||||
#include "db/merge_helper.h"
|
||||
@@ -27,7 +26,6 @@
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/table.h"
|
||||
#include "table/block_based_table_builder.h"
|
||||
#include "table/internal_iterator.h"
|
||||
#include "util/file_reader_writer.h"
|
||||
#include "util/iostats_context_imp.h"
|
||||
#include "util/stop_watch.h"
|
||||
@@ -42,39 +40,25 @@ TableBuilder* NewTableBuilder(
|
||||
const InternalKeyComparator& internal_comparator,
|
||||
const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
|
||||
int_tbl_prop_collector_factories,
|
||||
uint32_t column_family_id, const std::string& column_family_name,
|
||||
WritableFileWriter* file, const CompressionType compression_type,
|
||||
const CompressionOptions& compression_opts,
|
||||
const std::string* compression_dict, const bool skip_filters) {
|
||||
assert((column_family_id ==
|
||||
TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
|
||||
column_family_name.empty());
|
||||
const CompressionOptions& compression_opts, const bool skip_filters) {
|
||||
return ioptions.table_factory->NewTableBuilder(
|
||||
TableBuilderOptions(ioptions, internal_comparator,
|
||||
int_tbl_prop_collector_factories, compression_type,
|
||||
compression_opts, compression_dict, skip_filters,
|
||||
column_family_name),
|
||||
column_family_id, file);
|
||||
compression_opts, skip_filters),
|
||||
file);
|
||||
}
|
||||
|
||||
Status BuildTable(
|
||||
const std::string& dbname, Env* env, const ImmutableCFOptions& ioptions,
|
||||
const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
|
||||
TableCache* table_cache, InternalIterator* iter, FileMetaData* meta,
|
||||
const InternalKeyComparator& internal_comparator,
|
||||
const EnvOptions& env_options, TableCache* table_cache, Iterator* iter,
|
||||
FileMetaData* meta, const InternalKeyComparator& internal_comparator,
|
||||
const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
|
||||
int_tbl_prop_collector_factories,
|
||||
uint32_t column_family_id, const std::string& column_family_name,
|
||||
std::vector<SequenceNumber> snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot,
|
||||
const CompressionType compression,
|
||||
std::vector<SequenceNumber> snapshots, const CompressionType compression,
|
||||
const CompressionOptions& compression_opts, bool paranoid_file_checks,
|
||||
InternalStats* internal_stats, TableFileCreationReason reason,
|
||||
EventLogger* event_logger, int job_id, const Env::IOPriority io_priority,
|
||||
TableProperties* table_properties, int level) {
|
||||
assert((column_family_id ==
|
||||
TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
|
||||
column_family_name.empty());
|
||||
InternalStats* internal_stats, const Env::IOPriority io_priority,
|
||||
TableProperties* table_properties) {
|
||||
// Reports the IOStats for flush for every following bytes.
|
||||
const size_t kReportFlushIOStatsEvery = 1048576;
|
||||
Status s;
|
||||
@@ -83,22 +67,13 @@ Status BuildTable(
|
||||
|
||||
std::string fname = TableFileName(ioptions.db_paths, meta->fd.GetNumber(),
|
||||
meta->fd.GetPathId());
|
||||
#ifndef ROCKSDB_LITE
|
||||
EventHelpers::NotifyTableFileCreationStarted(
|
||||
ioptions.listeners, dbname, column_family_name, fname, job_id, reason);
|
||||
#endif // !ROCKSDB_LITE
|
||||
TableProperties tp;
|
||||
|
||||
if (iter->Valid()) {
|
||||
TableBuilder* builder;
|
||||
unique_ptr<WritableFileWriter> file_writer;
|
||||
{
|
||||
unique_ptr<WritableFile> file;
|
||||
s = NewWritableFile(env, fname, &file, env_options);
|
||||
s = env->NewWritableFile(fname, &file, env_options);
|
||||
if (!s.ok()) {
|
||||
EventHelpers::LogAndNotifyTableFileCreationFinished(
|
||||
event_logger, ioptions.listeners, dbname, column_family_name, fname,
|
||||
job_id, meta->fd, tp, reason, s);
|
||||
return s;
|
||||
}
|
||||
file->SetIOPriority(io_priority);
|
||||
@@ -107,19 +82,17 @@ Status BuildTable(
|
||||
|
||||
builder = NewTableBuilder(
|
||||
ioptions, internal_comparator, int_tbl_prop_collector_factories,
|
||||
column_family_id, column_family_name, file_writer.get(), compression,
|
||||
compression_opts);
|
||||
file_writer.get(), compression, compression_opts);
|
||||
}
|
||||
|
||||
MergeHelper merge(env, internal_comparator.user_comparator(),
|
||||
ioptions.merge_operator, nullptr, ioptions.info_log,
|
||||
mutable_cf_options.min_partial_merge_operands,
|
||||
ioptions.min_partial_merge_operands,
|
||||
true /* internal key corruption is not ok */,
|
||||
snapshots.empty() ? 0 : snapshots.back());
|
||||
|
||||
CompactionIterator c_iter(iter, internal_comparator.user_comparator(),
|
||||
&merge, kMaxSequenceNumber, &snapshots,
|
||||
earliest_write_conflict_snapshot, env,
|
||||
&merge, kMaxSequenceNumber, &snapshots, env,
|
||||
true /* internal key corruption is not ok */);
|
||||
c_iter.SeekToFirst();
|
||||
for (; c_iter.Valid(); c_iter.Next()) {
|
||||
@@ -146,13 +119,11 @@ Status BuildTable(
|
||||
}
|
||||
|
||||
if (s.ok() && !empty) {
|
||||
uint64_t file_size = builder->FileSize();
|
||||
meta->fd.file_size = file_size;
|
||||
meta->fd.file_size = builder->FileSize();
|
||||
meta->marked_for_compaction = builder->NeedCompact();
|
||||
assert(meta->fd.GetFileSize() > 0);
|
||||
tp = builder->GetTableProperties();
|
||||
if (table_properties) {
|
||||
*table_properties = tp;
|
||||
*table_properties = builder->GetTableProperties();
|
||||
}
|
||||
}
|
||||
delete builder;
|
||||
@@ -168,12 +139,11 @@ Status BuildTable(
|
||||
|
||||
if (s.ok() && !empty) {
|
||||
// Verify that the table is usable
|
||||
std::unique_ptr<InternalIterator> it(table_cache->NewIterator(
|
||||
std::unique_ptr<Iterator> it(table_cache->NewIterator(
|
||||
ReadOptions(), env_options, internal_comparator, meta->fd, nullptr,
|
||||
(internal_stats == nullptr) ? nullptr
|
||||
: internal_stats->GetFileReadHist(0),
|
||||
false /* for_compaction */, nullptr /* arena */,
|
||||
false /* skip_filter */, level));
|
||||
false));
|
||||
s = it->status();
|
||||
if (s.ok() && paranoid_file_checks) {
|
||||
for (it->SeekToFirst(); it->Valid(); it->Next()) {
|
||||
@@ -191,12 +161,6 @@ Status BuildTable(
|
||||
if (!s.ok() || meta->fd.GetFileSize() == 0) {
|
||||
env->DeleteFile(fname);
|
||||
}
|
||||
|
||||
// Output to event logger and fire events.
|
||||
EventHelpers::LogAndNotifyTableFileCreationFinished(
|
||||
event_logger, ioptions.listeners, dbname, column_family_name, fname,
|
||||
job_id, meta->fd, tp, reason, s);
|
||||
|
||||
return s;
|
||||
}
|
||||
|
||||
|
||||
+9
-27
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -12,13 +12,11 @@
|
||||
#include "db/table_properties_collector.h"
|
||||
#include "rocksdb/comparator.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/listener.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/status.h"
|
||||
#include "rocksdb/table_properties.h"
|
||||
#include "rocksdb/types.h"
|
||||
#include "util/cf_options.h"
|
||||
#include "util/event_logger.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/immutable_options.h"
|
||||
#include "rocksdb/table_properties.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -33,22 +31,14 @@ class VersionEdit;
|
||||
class TableBuilder;
|
||||
class WritableFileWriter;
|
||||
class InternalStats;
|
||||
class InternalIterator;
|
||||
|
||||
// @param column_family_name Name of the column family that is also identified
|
||||
// by column_family_id, or empty string if unknown. It must outlive the
|
||||
// TableBuilder returned by this function.
|
||||
// @param compression_dict Data for presetting the compression library's
|
||||
// dictionary, or nullptr.
|
||||
TableBuilder* NewTableBuilder(
|
||||
const ImmutableCFOptions& options,
|
||||
const InternalKeyComparator& internal_comparator,
|
||||
const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
|
||||
int_tbl_prop_collector_factories,
|
||||
uint32_t column_family_id, const std::string& column_family_name,
|
||||
WritableFileWriter* file, const CompressionType compression_type,
|
||||
const CompressionOptions& compression_opts,
|
||||
const std::string* compression_dict = nullptr,
|
||||
const bool skip_filters = false);
|
||||
|
||||
// Build a Table file from the contents of *iter. The generated file
|
||||
@@ -56,24 +46,16 @@ TableBuilder* NewTableBuilder(
|
||||
// *meta will be filled with metadata about the generated table.
|
||||
// If no data is present in *iter, meta->file_size will be set to
|
||||
// zero, and no Table file will be produced.
|
||||
//
|
||||
// @param column_family_name Name of the column family that is also identified
|
||||
// by column_family_id, or empty string if unknown.
|
||||
extern Status BuildTable(
|
||||
const std::string& dbname, Env* env, const ImmutableCFOptions& options,
|
||||
const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
|
||||
TableCache* table_cache, InternalIterator* iter, FileMetaData* meta,
|
||||
const InternalKeyComparator& internal_comparator,
|
||||
const EnvOptions& env_options, TableCache* table_cache, Iterator* iter,
|
||||
FileMetaData* meta, const InternalKeyComparator& internal_comparator,
|
||||
const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
|
||||
int_tbl_prop_collector_factories,
|
||||
uint32_t column_family_id, const std::string& column_family_name,
|
||||
std::vector<SequenceNumber> snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot,
|
||||
const CompressionType compression,
|
||||
std::vector<SequenceNumber> snapshots, const CompressionType compression,
|
||||
const CompressionOptions& compression_opts, bool paranoid_file_checks,
|
||||
InternalStats* internal_stats, TableFileCreationReason reason,
|
||||
EventLogger* event_logger = nullptr, int job_id = 0,
|
||||
InternalStats* internal_stats,
|
||||
const Env::IOPriority io_priority = Env::IO_HIGH,
|
||||
TableProperties* table_properties = nullptr, int level = -1);
|
||||
TableProperties* table_properties = nullptr);
|
||||
|
||||
} // namespace rocksdb
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -43,7 +43,6 @@ using rocksdb::CompactionFilterContext;
|
||||
using rocksdb::CompactionOptionsFIFO;
|
||||
using rocksdb::Comparator;
|
||||
using rocksdb::CompressionType;
|
||||
using rocksdb::WALRecoveryMode;
|
||||
using rocksdb::DB;
|
||||
using rocksdb::DBOptions;
|
||||
using rocksdb::Env;
|
||||
@@ -124,7 +123,6 @@ struct rocksdb_compactionfilter_t : public CompactionFilter {
|
||||
char** new_value, size_t *new_value_length,
|
||||
unsigned char* value_changed);
|
||||
const char* (*name_)(void*);
|
||||
unsigned char ignore_snapshots_;
|
||||
|
||||
virtual ~rocksdb_compactionfilter_t() {
|
||||
(*destructor_)(state_);
|
||||
@@ -150,8 +148,6 @@ struct rocksdb_compactionfilter_t : public CompactionFilter {
|
||||
}
|
||||
|
||||
virtual const char* Name() const override { return (*name_)(state_); }
|
||||
|
||||
virtual bool IgnoreSnapshots() const override { return ignore_snapshots_; }
|
||||
};
|
||||
|
||||
struct rocksdb_compactionfilterfactory_t : public CompactionFilterFactory {
|
||||
@@ -273,31 +269,33 @@ struct rocksdb_mergeoperator_t : public MergeOperator {
|
||||
|
||||
virtual const char* Name() const override { return (*name_)(state_); }
|
||||
|
||||
virtual bool FullMergeV2(const MergeOperationInput& merge_in,
|
||||
MergeOperationOutput* merge_out) const override {
|
||||
size_t n = merge_in.operand_list.size();
|
||||
virtual bool FullMerge(const Slice& key, const Slice* existing_value,
|
||||
const std::deque<std::string>& operand_list,
|
||||
std::string* new_value,
|
||||
Logger* logger) const override {
|
||||
size_t n = operand_list.size();
|
||||
std::vector<const char*> operand_pointers(n);
|
||||
std::vector<size_t> operand_sizes(n);
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
Slice operand(merge_in.operand_list[i]);
|
||||
Slice operand(operand_list[i]);
|
||||
operand_pointers[i] = operand.data();
|
||||
operand_sizes[i] = operand.size();
|
||||
}
|
||||
|
||||
const char* existing_value_data = nullptr;
|
||||
size_t existing_value_len = 0;
|
||||
if (merge_in.existing_value != nullptr) {
|
||||
existing_value_data = merge_in.existing_value->data();
|
||||
existing_value_len = merge_in.existing_value->size();
|
||||
if (existing_value != nullptr) {
|
||||
existing_value_data = existing_value->data();
|
||||
existing_value_len = existing_value->size();
|
||||
}
|
||||
|
||||
unsigned char success;
|
||||
size_t new_value_len;
|
||||
char* tmp_new_value = (*full_merge_)(
|
||||
state_, merge_in.key.data(), merge_in.key.size(), existing_value_data,
|
||||
existing_value_len, &operand_pointers[0], &operand_sizes[0],
|
||||
static_cast<int>(n), &success, &new_value_len);
|
||||
merge_out->new_value.assign(tmp_new_value, new_value_len);
|
||||
state_, key.data(), key.size(), existing_value_data, existing_value_len,
|
||||
&operand_pointers[0], &operand_sizes[0], static_cast<int>(n), &success,
|
||||
&new_value_len);
|
||||
new_value->assign(tmp_new_value, new_value_len);
|
||||
|
||||
if (delete_value_ != nullptr) {
|
||||
(*delete_value_)(state_, tmp_new_value, new_value_len);
|
||||
@@ -448,12 +446,6 @@ void rocksdb_backup_engine_create_new_backup(rocksdb_backup_engine_t* be,
|
||||
SaveError(errptr, be->rep->CreateNewBackup(db->rep));
|
||||
}
|
||||
|
||||
void rocksdb_backup_engine_purge_old_backups(rocksdb_backup_engine_t* be,
|
||||
uint32_t num_backups_to_keep,
|
||||
char** errptr) {
|
||||
SaveError(errptr, be->rep->PurgeOldBackups(num_backups_to_keep));
|
||||
}
|
||||
|
||||
rocksdb_restore_options_t* rocksdb_restore_options_create() {
|
||||
return new rocksdb_restore_options_t;
|
||||
}
|
||||
@@ -833,31 +825,6 @@ rocksdb_iterator_t* rocksdb_create_iterator_cf(
|
||||
return result;
|
||||
}
|
||||
|
||||
void rocksdb_create_iterators(
|
||||
rocksdb_t *db,
|
||||
rocksdb_readoptions_t* opts,
|
||||
rocksdb_column_family_handle_t** column_families,
|
||||
rocksdb_iterator_t** iterators,
|
||||
size_t size,
|
||||
char** errptr) {
|
||||
std::vector<ColumnFamilyHandle*> column_families_vec;
|
||||
for (size_t i = 0; i < size; i++) {
|
||||
column_families_vec.push_back(column_families[i]->rep);
|
||||
}
|
||||
|
||||
std::vector<Iterator*> res;
|
||||
Status status = db->rep->NewIterators(opts->rep, column_families_vec, &res);
|
||||
assert(res.size() == size);
|
||||
if (SaveError(errptr, status)) {
|
||||
return;
|
||||
}
|
||||
|
||||
for (size_t i = 0; i < size; i++) {
|
||||
iterators[i] = new rocksdb_iterator_t;
|
||||
iterators[i]->rep = res[i];
|
||||
}
|
||||
}
|
||||
|
||||
const rocksdb_snapshot_t* rocksdb_create_snapshot(
|
||||
rocksdb_t* db) {
|
||||
rocksdb_snapshot_t* result = new rocksdb_snapshot_t;
|
||||
@@ -1321,16 +1288,6 @@ void rocksdb_block_based_options_set_cache_index_and_filter_blocks(
|
||||
options->rep.cache_index_and_filter_blocks = v;
|
||||
}
|
||||
|
||||
void rocksdb_block_based_options_set_pin_l0_filter_and_index_blocks_in_cache(
|
||||
rocksdb_block_based_table_options_t* options, unsigned char v) {
|
||||
options->rep.pin_l0_filter_and_index_blocks_in_cache = v;
|
||||
}
|
||||
|
||||
void rocksdb_block_based_options_set_skip_table_builder_flush(
|
||||
rocksdb_block_based_table_options_t* options, unsigned char v) {
|
||||
options->rep.skip_table_builder_flush = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_block_based_table_factory(
|
||||
rocksdb_options_t *opt,
|
||||
rocksdb_block_based_table_options_t* table_options) {
|
||||
@@ -1426,11 +1383,6 @@ void rocksdb_options_set_compaction_filter_factory(
|
||||
std::shared_ptr<CompactionFilterFactory>(factory);
|
||||
}
|
||||
|
||||
void rocksdb_options_compaction_readahead_size(
|
||||
rocksdb_options_t* opt, size_t s) {
|
||||
opt->rep.compaction_readahead_size = s;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_comparator(
|
||||
rocksdb_options_t* opt,
|
||||
rocksdb_comparator_t* cmp) {
|
||||
@@ -1516,9 +1468,14 @@ void rocksdb_options_set_max_bytes_for_level_multiplier(
|
||||
opt->rep.max_bytes_for_level_multiplier = n;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_max_compaction_bytes(rocksdb_options_t* opt,
|
||||
uint64_t n) {
|
||||
opt->rep.max_compaction_bytes = n;
|
||||
void rocksdb_options_set_expanded_compaction_factor(
|
||||
rocksdb_options_t* opt, int n) {
|
||||
opt->rep.expanded_compaction_factor = n;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_max_grandparent_overlap_factor(
|
||||
rocksdb_options_t* opt, int n) {
|
||||
opt->rep.max_grandparent_overlap_factor = n;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_max_bytes_for_level_multiplier_additional(
|
||||
@@ -1555,10 +1512,6 @@ void rocksdb_options_set_level0_stop_writes_trigger(
|
||||
void rocksdb_options_set_max_mem_compaction_level(rocksdb_options_t* opt,
|
||||
int n) {}
|
||||
|
||||
void rocksdb_options_set_wal_recovery_mode(rocksdb_options_t* opt,int mode) {
|
||||
opt->rep.wal_recovery_mode = static_cast<WALRecoveryMode>(mode);
|
||||
}
|
||||
|
||||
void rocksdb_options_set_compression(rocksdb_options_t* opt, int t) {
|
||||
opt->rep.compression = static_cast<CompressionType>(t);
|
||||
}
|
||||
@@ -1573,13 +1526,11 @@ void rocksdb_options_set_compression_per_level(rocksdb_options_t* opt,
|
||||
}
|
||||
}
|
||||
|
||||
void rocksdb_options_set_compression_options(rocksdb_options_t* opt, int w_bits,
|
||||
int level, int strategy,
|
||||
int max_dict_bytes) {
|
||||
void rocksdb_options_set_compression_options(
|
||||
rocksdb_options_t* opt, int w_bits, int level, int strategy) {
|
||||
opt->rep.compression_opts.window_bits = w_bits;
|
||||
opt->rep.compression_opts.level = level;
|
||||
opt->rep.compression_opts.strategy = strategy;
|
||||
opt->rep.compression_opts.max_dict_bytes = max_dict_bytes;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_prefix_extractor(
|
||||
@@ -1693,6 +1644,11 @@ void rocksdb_options_set_verify_checksums_in_compaction(
|
||||
opt->rep.verify_checksums_in_compaction = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_filter_deletes(
|
||||
rocksdb_options_t* opt, unsigned char v) {
|
||||
opt->rep.filter_deletes = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_max_sequential_skip_in_iterations(
|
||||
rocksdb_options_t* opt, uint64_t v) {
|
||||
opt->rep.max_sequential_skip_in_iterations = v;
|
||||
@@ -1731,11 +1687,6 @@ void rocksdb_options_set_keep_log_file_num(rocksdb_options_t* opt, size_t v) {
|
||||
opt->rep.keep_log_file_num = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_recycle_log_file_num(rocksdb_options_t* opt,
|
||||
size_t v) {
|
||||
opt->rep.recycle_log_file_num = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_soft_rate_limit(rocksdb_options_t* opt, double v) {
|
||||
opt->rep.soft_rate_limit = v;
|
||||
}
|
||||
@@ -1778,6 +1729,11 @@ void rocksdb_options_set_delete_obsolete_files_period_micros(
|
||||
opt->rep.delete_obsolete_files_period_micros = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_source_compaction_factor(
|
||||
rocksdb_options_t* opt, int n) {
|
||||
opt->rep.expanded_compaction_factor = n;
|
||||
}
|
||||
|
||||
void rocksdb_options_prepare_for_bulk_load(rocksdb_options_t* opt) {
|
||||
opt->rep.PrepareForBulkLoad();
|
||||
}
|
||||
@@ -1790,14 +1746,14 @@ void rocksdb_options_set_memtable_vector_rep(rocksdb_options_t *opt) {
|
||||
opt->rep.memtable_factory.reset(factory);
|
||||
}
|
||||
|
||||
void rocksdb_options_set_memtable_prefix_bloom_size_ratio(
|
||||
rocksdb_options_t* opt, double v) {
|
||||
opt->rep.memtable_prefix_bloom_size_ratio = v;
|
||||
void rocksdb_options_set_memtable_prefix_bloom_bits(
|
||||
rocksdb_options_t* opt, uint32_t v) {
|
||||
opt->rep.memtable_prefix_bloom_bits = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_memtable_huge_page_size(rocksdb_options_t* opt,
|
||||
size_t v) {
|
||||
opt->rep.memtable_huge_page_size = v;
|
||||
void rocksdb_options_set_memtable_prefix_bloom_probes(
|
||||
rocksdb_options_t* opt, uint32_t v) {
|
||||
opt->rep.memtable_prefix_bloom_probes = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_hash_skip_list_rep(
|
||||
@@ -1861,11 +1817,6 @@ void rocksdb_options_set_inplace_update_num_locks(
|
||||
opt->rep.inplace_update_num_locks = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_report_bg_io_stats(
|
||||
rocksdb_options_t* opt, int v) {
|
||||
opt->rep.report_bg_io_stats = v;
|
||||
}
|
||||
|
||||
void rocksdb_options_set_compaction_style(rocksdb_options_t *opt, int style) {
|
||||
opt->rep.compaction_style = static_cast<rocksdb::CompactionStyle>(style);
|
||||
}
|
||||
@@ -1917,17 +1868,10 @@ rocksdb_compactionfilter_t* rocksdb_compactionfilter_create(
|
||||
result->state_ = state;
|
||||
result->destructor_ = destructor;
|
||||
result->filter_ = filter;
|
||||
result->ignore_snapshots_ = false;
|
||||
result->name_ = name;
|
||||
return result;
|
||||
}
|
||||
|
||||
void rocksdb_compactionfilter_set_ignore_snapshots(
|
||||
rocksdb_compactionfilter_t* filter,
|
||||
unsigned char whether_ignore) {
|
||||
filter->ignore_snapshots_ = whether_ignore;
|
||||
}
|
||||
|
||||
void rocksdb_compactionfilter_destroy(rocksdb_compactionfilter_t* filter) {
|
||||
delete filter;
|
||||
}
|
||||
@@ -2011,7 +1955,7 @@ void rocksdb_filterpolicy_destroy(rocksdb_filterpolicy_t* filter) {
|
||||
delete filter;
|
||||
}
|
||||
|
||||
rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom_format(int bits_per_key, bool original_format) {
|
||||
rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom(int bits_per_key) {
|
||||
// Make a rocksdb_filterpolicy_t, but override all of its methods so
|
||||
// they delegate to a NewBloomFilterPolicy() instead of user
|
||||
// supplied C functions.
|
||||
@@ -2029,21 +1973,13 @@ rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom_format(int bits_per_ke
|
||||
static void DoNothing(void*) { }
|
||||
};
|
||||
Wrapper* wrapper = new Wrapper;
|
||||
wrapper->rep_ = NewBloomFilterPolicy(bits_per_key, original_format);
|
||||
wrapper->rep_ = NewBloomFilterPolicy(bits_per_key);
|
||||
wrapper->state_ = nullptr;
|
||||
wrapper->delete_filter_ = nullptr;
|
||||
wrapper->destructor_ = &Wrapper::DoNothing;
|
||||
return wrapper;
|
||||
}
|
||||
|
||||
rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom_full(int bits_per_key) {
|
||||
return rocksdb_filterpolicy_create_bloom_format(bits_per_key, false);
|
||||
}
|
||||
|
||||
rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom(int bits_per_key) {
|
||||
return rocksdb_filterpolicy_create_bloom_format(bits_per_key, true);
|
||||
}
|
||||
|
||||
rocksdb_mergeoperator_t* rocksdb_mergeoperator_create(
|
||||
void* state, void (*destructor)(void*),
|
||||
char* (*full_merge)(void*, const char* key, size_t key_length,
|
||||
@@ -2120,11 +2056,6 @@ void rocksdb_readoptions_set_tailing(
|
||||
opt->rep.tailing = v;
|
||||
}
|
||||
|
||||
void rocksdb_readoptions_set_readahead_size(
|
||||
rocksdb_readoptions_t* opt, size_t v) {
|
||||
opt->rep.readahead_size = v;
|
||||
}
|
||||
|
||||
rocksdb_writeoptions_t* rocksdb_writeoptions_create() {
|
||||
return new rocksdb_writeoptions_t;
|
||||
}
|
||||
@@ -2166,10 +2097,6 @@ void rocksdb_cache_destroy(rocksdb_cache_t* cache) {
|
||||
delete cache;
|
||||
}
|
||||
|
||||
void rocksdb_cache_set_capacity(rocksdb_cache_t* cache, size_t capacity) {
|
||||
cache->rep->SetCapacity(capacity);
|
||||
}
|
||||
|
||||
rocksdb_env_t* rocksdb_create_default_env() {
|
||||
rocksdb_env_t* result = new rocksdb_env_t;
|
||||
result->rep = Env::Default();
|
||||
@@ -2177,13 +2104,6 @@ rocksdb_env_t* rocksdb_create_default_env() {
|
||||
return result;
|
||||
}
|
||||
|
||||
rocksdb_env_t* rocksdb_create_mem_env() {
|
||||
rocksdb_env_t* result = new rocksdb_env_t;
|
||||
result->rep = rocksdb::NewMemEnv(Env::Default());
|
||||
result->is_default = false;
|
||||
return result;
|
||||
}
|
||||
|
||||
void rocksdb_env_set_background_threads(rocksdb_env_t* env, int n) {
|
||||
env->rep->SetBackgroundThreads(n);
|
||||
}
|
||||
@@ -2395,31 +2315,6 @@ void rocksdb_get_options_from_string(const rocksdb_options_t* base_options,
|
||||
&new_options->rep));
|
||||
}
|
||||
|
||||
void rocksdb_delete_file_in_range(rocksdb_t* db, const char* start_key,
|
||||
size_t start_key_len, const char* limit_key,
|
||||
size_t limit_key_len, char** errptr) {
|
||||
Slice a, b;
|
||||
SaveError(
|
||||
errptr,
|
||||
DeleteFilesInRange(
|
||||
db->rep, db->rep->DefaultColumnFamily(),
|
||||
(start_key ? (a = Slice(start_key, start_key_len), &a) : nullptr),
|
||||
(limit_key ? (b = Slice(limit_key, limit_key_len), &b) : nullptr)));
|
||||
}
|
||||
|
||||
void rocksdb_delete_file_in_range_cf(
|
||||
rocksdb_t* db, rocksdb_column_family_handle_t* column_family,
|
||||
const char* start_key, size_t start_key_len, const char* limit_key,
|
||||
size_t limit_key_len, char** errptr) {
|
||||
Slice a, b;
|
||||
SaveError(
|
||||
errptr,
|
||||
DeleteFilesInRange(
|
||||
db->rep, column_family->rep,
|
||||
(start_key ? (a = Slice(start_key, start_key_len), &a) : nullptr),
|
||||
(limit_key ? (b = Slice(limit_key, limit_key_len), &b) : nullptr)));
|
||||
}
|
||||
|
||||
void rocksdb_free(void* ptr) { free(ptr); }
|
||||
|
||||
} // end extern "C"
|
||||
|
||||
+1
-44
@@ -322,7 +322,7 @@ int main(int argc, char** argv) {
|
||||
rocksdb_options_set_block_based_table_factory(options, table_options);
|
||||
|
||||
rocksdb_options_set_compression(options, rocksdb_no_compression);
|
||||
rocksdb_options_set_compression_options(options, -14, -1, 0, 0);
|
||||
rocksdb_options_set_compression_options(options, -14, -1, 0);
|
||||
int compression_levels[] = {rocksdb_no_compression, rocksdb_no_compression,
|
||||
rocksdb_no_compression, rocksdb_no_compression};
|
||||
rocksdb_options_set_compression_per_level(options, compression_levels, 4);
|
||||
@@ -365,24 +365,6 @@ int main(int argc, char** argv) {
|
||||
rocksdb_backup_engine_create_new_backup(be, db, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
// need a change to trigger a new backup
|
||||
rocksdb_delete(db, woptions, "does-not-exist", 14, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
rocksdb_backup_engine_create_new_backup(be, db, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
const rocksdb_backup_engine_info_t* bei = rocksdb_backup_engine_get_backup_info(be);
|
||||
CheckCondition(rocksdb_backup_engine_info_count(bei) > 1);
|
||||
rocksdb_backup_engine_info_destroy(bei);
|
||||
|
||||
rocksdb_backup_engine_purge_old_backups(be, 1, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
bei = rocksdb_backup_engine_get_backup_info(be);
|
||||
CheckCondition(rocksdb_backup_engine_info_count(bei) == 1);
|
||||
rocksdb_backup_engine_info_destroy(bei);
|
||||
|
||||
rocksdb_delete(db, woptions, "foo", 3, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
@@ -575,7 +557,6 @@ int main(int argc, char** argv) {
|
||||
rocksdb_close(db);
|
||||
rocksdb_options_set_create_if_missing(options, 0);
|
||||
rocksdb_options_set_error_if_exists(options, 0);
|
||||
rocksdb_options_set_wal_recovery_mode(options, 2);
|
||||
rocksdb_repair_db(options, dbname, &err);
|
||||
CheckNoError(err);
|
||||
db = rocksdb_open(options, dbname, &err);
|
||||
@@ -787,30 +768,6 @@ int main(int argc, char** argv) {
|
||||
CheckNoError(err);
|
||||
rocksdb_iter_destroy(iter);
|
||||
|
||||
rocksdb_column_family_handle_t* iters_cf_handles[2] = { handles[0], handles[1] };
|
||||
rocksdb_iterator_t* iters_handles[2];
|
||||
rocksdb_create_iterators(db, roptions, iters_cf_handles, iters_handles, 2, &err);
|
||||
CheckNoError(err);
|
||||
|
||||
iter = iters_handles[0];
|
||||
CheckCondition(!rocksdb_iter_valid(iter));
|
||||
rocksdb_iter_seek_to_first(iter);
|
||||
CheckCondition(!rocksdb_iter_valid(iter));
|
||||
rocksdb_iter_destroy(iter);
|
||||
|
||||
iter = iters_handles[1];
|
||||
CheckCondition(!rocksdb_iter_valid(iter));
|
||||
rocksdb_iter_seek_to_first(iter);
|
||||
CheckCondition(rocksdb_iter_valid(iter));
|
||||
|
||||
for (i = 0; rocksdb_iter_valid(iter) != 0; rocksdb_iter_next(iter)) {
|
||||
i++;
|
||||
}
|
||||
CheckCondition(i == 1);
|
||||
rocksdb_iter_get_error(iter, &err);
|
||||
CheckNoError(err);
|
||||
rocksdb_iter_destroy(iter);
|
||||
|
||||
rocksdb_drop_column_family(db, handles[1], &err);
|
||||
CheckNoError(err);
|
||||
for (i = 0; i < 2; i++) {
|
||||
|
||||
+54
-222
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -26,9 +26,10 @@
|
||||
#include "db/table_properties_collector.h"
|
||||
#include "db/version_set.h"
|
||||
#include "db/write_controller.h"
|
||||
#include "memtable/hash_skiplist_rep.h"
|
||||
#include "db/writebuffer.h"
|
||||
#include "util/autovector.h"
|
||||
#include "util/compression.h"
|
||||
#include "util/hash_skiplist_rep.h"
|
||||
#include "util/options_helper.h"
|
||||
#include "util/thread_status_util.h"
|
||||
#include "util/xfunc.h"
|
||||
@@ -67,21 +68,7 @@ const std::string& ColumnFamilyHandleImpl::GetName() const {
|
||||
return cfd()->GetName();
|
||||
}
|
||||
|
||||
Status ColumnFamilyHandleImpl::GetDescriptor(ColumnFamilyDescriptor* desc) {
|
||||
#ifndef ROCKSDB_LITE
|
||||
// accessing mutable cf-options requires db mutex.
|
||||
InstrumentedMutexLock l(mutex_);
|
||||
*desc = ColumnFamilyDescriptor(
|
||||
cfd()->GetName(),
|
||||
BuildColumnFamilyOptions(*cfd()->options(),
|
||||
*cfd()->GetLatestMutableCFOptions()));
|
||||
return Status::OK();
|
||||
#else
|
||||
return Status::NotSupported();
|
||||
#endif // !ROCKSDB_LITE
|
||||
}
|
||||
|
||||
const Comparator* ColumnFamilyHandleImpl::GetComparator() const {
|
||||
const Comparator* ColumnFamilyHandleImpl::user_comparator() const {
|
||||
return cfd()->user_comparator();
|
||||
}
|
||||
|
||||
@@ -123,28 +110,18 @@ Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
Status CheckConcurrentWritesSupported(const ColumnFamilyOptions& cf_options) {
|
||||
if (cf_options.inplace_update_support) {
|
||||
return Status::InvalidArgument(
|
||||
"In-place memtable updates (inplace_update_support) is not compatible "
|
||||
"with concurrent writes (allow_concurrent_memtable_write)");
|
||||
}
|
||||
if (!cf_options.memtable_factory->IsInsertConcurrentlySupported()) {
|
||||
return Status::InvalidArgument(
|
||||
"Memtable doesn't concurrent writes (allow_concurrent_memtable_write)");
|
||||
}
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
const InternalKeyComparator* icmp,
|
||||
const ColumnFamilyOptions& src) {
|
||||
ColumnFamilyOptions result = src;
|
||||
result.comparator = icmp;
|
||||
size_t clamp_max = std::conditional<
|
||||
sizeof(size_t) == 4, std::integral_constant<size_t, 0xffffffff>,
|
||||
std::integral_constant<uint64_t, 64ull << 30>>::type::value;
|
||||
ClipToRange(&result.write_buffer_size, ((size_t)64) << 10, clamp_max);
|
||||
#ifdef OS_MACOSX
|
||||
// TODO(icanadi) make write_buffer_size uint64_t instead of size_t
|
||||
ClipToRange(&result.write_buffer_size, ((size_t)64) << 10, ((size_t)1) << 30);
|
||||
#else
|
||||
ClipToRange(&result.write_buffer_size,
|
||||
((size_t)64) << 10, ((size_t)64) << 30);
|
||||
#endif
|
||||
// if user sets arena_block_size, we trust user to use this value. Otherwise,
|
||||
// calculate a proper value from writer_buffer_size;
|
||||
if (result.arena_block_size <= 0) {
|
||||
@@ -171,12 +148,6 @@ ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
if (result.max_write_buffer_number_to_maintain < 0) {
|
||||
result.max_write_buffer_number_to_maintain = result.max_write_buffer_number;
|
||||
}
|
||||
// bloom filter size shouldn't exceed 1/4 of memtable size.
|
||||
if (result.memtable_prefix_bloom_size_ratio > 0.25) {
|
||||
result.memtable_prefix_bloom_size_ratio = 0.25;
|
||||
} else if (result.memtable_prefix_bloom_size_ratio < 0) {
|
||||
result.memtable_prefix_bloom_size_ratio = 0;
|
||||
}
|
||||
XFUNC_TEST("memtablelist_history", "transaction_xftest_SanitizeOptions",
|
||||
xf_transaction_set_memtable_history1,
|
||||
xf_transaction_set_memtable_history,
|
||||
@@ -204,16 +175,6 @@ ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
result.level0_stop_writes_trigger = std::numeric_limits<int>::max();
|
||||
}
|
||||
|
||||
if (result.max_bytes_for_level_multiplier <= 0) {
|
||||
result.max_bytes_for_level_multiplier = 1;
|
||||
}
|
||||
|
||||
if (result.level0_file_num_compaction_trigger == 0) {
|
||||
Warn(db_options.info_log.get(),
|
||||
"level0_file_num_compaction_trigger cannot be 0");
|
||||
result.level0_file_num_compaction_trigger = 1;
|
||||
}
|
||||
|
||||
if (result.level0_stop_writes_trigger <
|
||||
result.level0_slowdown_writes_trigger ||
|
||||
result.level0_slowdown_writes_trigger <
|
||||
@@ -244,17 +205,6 @@ ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
result.level0_slowdown_writes_trigger,
|
||||
result.level0_file_num_compaction_trigger);
|
||||
}
|
||||
|
||||
if (result.soft_pending_compaction_bytes_limit == 0) {
|
||||
result.soft_pending_compaction_bytes_limit =
|
||||
result.hard_pending_compaction_bytes_limit;
|
||||
} else if (result.hard_pending_compaction_bytes_limit > 0 &&
|
||||
result.soft_pending_compaction_bytes_limit >
|
||||
result.hard_pending_compaction_bytes_limit) {
|
||||
result.soft_pending_compaction_bytes_limit =
|
||||
result.hard_pending_compaction_bytes_limit;
|
||||
}
|
||||
|
||||
if (result.level_compaction_dynamic_level_bytes) {
|
||||
if (result.compaction_style != kCompactionStyleLevel ||
|
||||
db_options.db_paths.size() > 1U) {
|
||||
@@ -266,10 +216,6 @@ ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
}
|
||||
}
|
||||
|
||||
if (result.max_compaction_bytes == 0) {
|
||||
result.max_compaction_bytes = result.target_file_size_base * 25;
|
||||
}
|
||||
|
||||
return result;
|
||||
}
|
||||
|
||||
@@ -337,7 +283,7 @@ void SuperVersionUnrefHandle(void* ptr) {
|
||||
|
||||
ColumnFamilyData::ColumnFamilyData(
|
||||
uint32_t id, const std::string& name, Version* _dummy_versions,
|
||||
Cache* _table_cache, WriteBufferManager* write_buffer_manager,
|
||||
Cache* _table_cache, WriteBuffer* write_buffer,
|
||||
const ColumnFamilyOptions& cf_options, const DBOptions* db_options,
|
||||
const EnvOptions& env_options, ColumnFamilySet* column_family_set)
|
||||
: id_(id),
|
||||
@@ -351,7 +297,7 @@ ColumnFamilyData::ColumnFamilyData(
|
||||
SanitizeOptions(*db_options, &internal_comparator_, cf_options)),
|
||||
ioptions_(options_),
|
||||
mutable_cf_options_(options_, ioptions_),
|
||||
write_buffer_manager_(write_buffer_manager),
|
||||
write_buffer_(write_buffer),
|
||||
mem_(nullptr),
|
||||
imm_(options_.min_write_buffer_number_to_merge,
|
||||
options_.max_write_buffer_number_to_maintain),
|
||||
@@ -363,8 +309,7 @@ ColumnFamilyData::ColumnFamilyData(
|
||||
log_number_(0),
|
||||
column_family_set_(column_family_set),
|
||||
pending_flush_(false),
|
||||
pending_compaction_(false),
|
||||
prev_compaction_needed_bytes_(0) {
|
||||
pending_compaction_(false) {
|
||||
Ref();
|
||||
|
||||
// Convert user defined table properties collector factories to internal ones.
|
||||
@@ -482,81 +427,13 @@ void ColumnFamilyData::SetDropped() {
|
||||
column_family_set_->RemoveColumnFamily(this);
|
||||
}
|
||||
|
||||
const double kSlowdownRatio = 1.2;
|
||||
|
||||
namespace {
|
||||
std::unique_ptr<WriteControllerToken> SetupDelay(
|
||||
uint64_t max_write_rate, WriteController* write_controller,
|
||||
uint64_t compaction_needed_bytes, uint64_t prev_compaction_neeed_bytes,
|
||||
bool auto_comapctions_disabled) {
|
||||
const uint64_t kMinWriteRate = 1024u; // Minimum write rate 1KB/s.
|
||||
|
||||
uint64_t write_rate = write_controller->delayed_write_rate();
|
||||
|
||||
if (auto_comapctions_disabled) {
|
||||
// When auto compaction is disabled, always use the value user gave.
|
||||
write_rate = max_write_rate;
|
||||
} else if (write_controller->NeedsDelay() && max_write_rate > kMinWriteRate) {
|
||||
// If user gives rate less than kMinWriteRate, don't adjust it.
|
||||
//
|
||||
// If already delayed, need to adjust based on previous compaction debt.
|
||||
// When there are two or more column families require delay, we always
|
||||
// increase or reduce write rate based on information for one single
|
||||
// column family. It is likely to be OK but we can improve if there is a
|
||||
// problem.
|
||||
// Ignore compaction_needed_bytes = 0 case because compaction_needed_bytes
|
||||
// is only available in level-based compaction
|
||||
//
|
||||
// If the compaction debt stays the same as previously, we also further slow
|
||||
// down. It usually means a mem table is full. It's mainly for the case
|
||||
// where both of flush and compaction are much slower than the speed we
|
||||
// insert to mem tables, so we need to actively slow down before we get
|
||||
// feedback signal from compaction and flushes to avoid the full stop
|
||||
// because of hitting the max write buffer number.
|
||||
if (prev_compaction_neeed_bytes > 0 &&
|
||||
prev_compaction_neeed_bytes <= compaction_needed_bytes) {
|
||||
write_rate = static_cast<uint64_t>(static_cast<double>(write_rate) /
|
||||
kSlowdownRatio);
|
||||
if (write_rate < kMinWriteRate) {
|
||||
write_rate = kMinWriteRate;
|
||||
}
|
||||
} else if (prev_compaction_neeed_bytes > compaction_needed_bytes) {
|
||||
// We are speeding up by ratio of kSlowdownRatio when we have paid
|
||||
// compaction debt. But we'll never speed up to faster than the write rate
|
||||
// given by users.
|
||||
write_rate = static_cast<uint64_t>(static_cast<double>(write_rate) *
|
||||
kSlowdownRatio);
|
||||
if (write_rate > max_write_rate) {
|
||||
write_rate = max_write_rate;
|
||||
}
|
||||
}
|
||||
}
|
||||
return write_controller->GetDelayToken(write_rate);
|
||||
}
|
||||
|
||||
int GetL0ThresholdSpeedupCompaction(int level0_file_num_compaction_trigger,
|
||||
int level0_slowdown_writes_trigger) {
|
||||
// SanitizeOptions() ensures it.
|
||||
assert(level0_file_num_compaction_trigger <= level0_slowdown_writes_trigger);
|
||||
|
||||
// 1/4 of the way between L0 compaction trigger threshold and slowdown
|
||||
// condition.
|
||||
// Or twice as compaction trigger, if it is smaller.
|
||||
return std::min(level0_file_num_compaction_trigger * 2,
|
||||
level0_file_num_compaction_trigger +
|
||||
(level0_slowdown_writes_trigger -
|
||||
level0_file_num_compaction_trigger) /
|
||||
4);
|
||||
}
|
||||
} // namespace
|
||||
|
||||
void ColumnFamilyData::RecalculateWriteStallConditions(
|
||||
const MutableCFOptions& mutable_cf_options) {
|
||||
if (current_ != nullptr) {
|
||||
auto* vstorage = current_->storage_info();
|
||||
const double score = vstorage->max_compaction_score();
|
||||
const int max_level = vstorage->max_compaction_score_level();
|
||||
auto write_controller = column_family_set_->write_controller_;
|
||||
uint64_t compaction_needed_bytes =
|
||||
vstorage->estimated_compaction_needed_bytes();
|
||||
|
||||
if (imm()->NumNotFlushed() >= mutable_cf_options.max_write_buffer_number) {
|
||||
write_controller_token_ = write_controller->GetStopToken();
|
||||
@@ -566,9 +443,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
|
||||
"(waiting for flush), max_write_buffer_number is set to %d",
|
||||
name_.c_str(), imm()->NumNotFlushed(),
|
||||
mutable_cf_options.max_write_buffer_number);
|
||||
} else if (!mutable_cf_options.disable_auto_compactions &&
|
||||
vstorage->l0_delay_trigger_count() >=
|
||||
mutable_cf_options.level0_stop_writes_trigger) {
|
||||
} else if (vstorage->l0_delay_trigger_count() >=
|
||||
mutable_cf_options.level0_stop_writes_trigger) {
|
||||
write_controller_token_ = write_controller->GetStopToken();
|
||||
internal_stats_->AddCFStats(InternalStats::LEVEL0_NUM_FILES_TOTAL, 1);
|
||||
if (compaction_picker_->IsLevel0CompactionInProgress()) {
|
||||
@@ -578,92 +454,38 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stopping writes because we have %d level-0 files",
|
||||
name_.c_str(), vstorage->l0_delay_trigger_count());
|
||||
} else if (!mutable_cf_options.disable_auto_compactions &&
|
||||
mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
|
||||
compaction_needed_bytes >=
|
||||
} else if (mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
|
||||
vstorage->estimated_compaction_needed_bytes() >=
|
||||
mutable_cf_options.hard_pending_compaction_bytes_limit) {
|
||||
write_controller_token_ = write_controller->GetStopToken();
|
||||
internal_stats_->AddCFStats(
|
||||
InternalStats::HARD_PENDING_COMPACTION_BYTES_LIMIT, 1);
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stopping writes because of estimated pending compaction "
|
||||
"bytes %" PRIu64,
|
||||
name_.c_str(), compaction_needed_bytes);
|
||||
} else if (mutable_cf_options.max_write_buffer_number > 3 &&
|
||||
imm()->NumNotFlushed() >=
|
||||
mutable_cf_options.max_write_buffer_number - 1) {
|
||||
write_controller_token_ =
|
||||
SetupDelay(ioptions_.delayed_write_rate, write_controller,
|
||||
compaction_needed_bytes, prev_compaction_needed_bytes_,
|
||||
mutable_cf_options.disable_auto_compactions);
|
||||
internal_stats_->AddCFStats(InternalStats::MEMTABLE_SLOWDOWN, 1);
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stalling writes because we have %d immutable memtables "
|
||||
"(waiting for flush), max_write_buffer_number is set to %d "
|
||||
"rate %" PRIu64,
|
||||
name_.c_str(), imm()->NumNotFlushed(),
|
||||
mutable_cf_options.max_write_buffer_number,
|
||||
write_controller->delayed_write_rate());
|
||||
} else if (!mutable_cf_options.disable_auto_compactions &&
|
||||
mutable_cf_options.level0_slowdown_writes_trigger >= 0 &&
|
||||
"[%s] Stopping writes because estimated pending compaction "
|
||||
"bytes exceed %" PRIu64,
|
||||
name_.c_str(), vstorage->estimated_compaction_needed_bytes());
|
||||
} else if (mutable_cf_options.level0_slowdown_writes_trigger >= 0 &&
|
||||
vstorage->l0_delay_trigger_count() >=
|
||||
mutable_cf_options.level0_slowdown_writes_trigger) {
|
||||
write_controller_token_ =
|
||||
SetupDelay(ioptions_.delayed_write_rate, write_controller,
|
||||
compaction_needed_bytes, prev_compaction_needed_bytes_,
|
||||
mutable_cf_options.disable_auto_compactions);
|
||||
write_controller_token_ = write_controller->GetDelayToken();
|
||||
internal_stats_->AddCFStats(InternalStats::LEVEL0_SLOWDOWN_TOTAL, 1);
|
||||
if (compaction_picker_->IsLevel0CompactionInProgress()) {
|
||||
internal_stats_->AddCFStats(
|
||||
InternalStats::LEVEL0_SLOWDOWN_WITH_COMPACTION, 1);
|
||||
}
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stalling writes because we have %d level-0 files "
|
||||
"rate %" PRIu64,
|
||||
name_.c_str(), vstorage->l0_delay_trigger_count(),
|
||||
write_controller->delayed_write_rate());
|
||||
} else if (!mutable_cf_options.disable_auto_compactions &&
|
||||
mutable_cf_options.soft_pending_compaction_bytes_limit > 0 &&
|
||||
vstorage->estimated_compaction_needed_bytes() >=
|
||||
mutable_cf_options.soft_pending_compaction_bytes_limit) {
|
||||
write_controller_token_ =
|
||||
SetupDelay(ioptions_.delayed_write_rate, write_controller,
|
||||
compaction_needed_bytes, prev_compaction_needed_bytes_,
|
||||
mutable_cf_options.disable_auto_compactions);
|
||||
internal_stats_->AddCFStats(
|
||||
InternalStats::SOFT_PENDING_COMPACTION_BYTES_LIMIT, 1);
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stalling writes because of estimated pending compaction "
|
||||
"bytes %" PRIu64 " rate %" PRIu64,
|
||||
name_.c_str(), vstorage->estimated_compaction_needed_bytes(),
|
||||
write_controller->delayed_write_rate());
|
||||
} else if (vstorage->l0_delay_trigger_count() >=
|
||||
GetL0ThresholdSpeedupCompaction(
|
||||
mutable_cf_options.level0_file_num_compaction_trigger,
|
||||
mutable_cf_options.level0_slowdown_writes_trigger)) {
|
||||
write_controller_token_ = write_controller->GetCompactionPressureToken();
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Increasing compaction threads because we have %d level-0 "
|
||||
"files ",
|
||||
"[%s] Stalling writes because we have %d level-0 files",
|
||||
name_.c_str(), vstorage->l0_delay_trigger_count());
|
||||
} else if (vstorage->estimated_compaction_needed_bytes() >=
|
||||
mutable_cf_options.soft_pending_compaction_bytes_limit / 4) {
|
||||
// Increase compaction threads if bytes needed for compaction exceeds
|
||||
// 1/4 of threshold for slowing down.
|
||||
// If soft pending compaction byte limit is not set, always speed up
|
||||
// compaction.
|
||||
write_controller_token_ = write_controller->GetCompactionPressureToken();
|
||||
if (mutable_cf_options.soft_pending_compaction_bytes_limit > 0) {
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Increasing compaction threads because of estimated pending "
|
||||
"compaction "
|
||||
"bytes %" PRIu64,
|
||||
name_.c_str(), vstorage->estimated_compaction_needed_bytes());
|
||||
}
|
||||
} else if (mutable_cf_options.soft_rate_limit > 0.0 &&
|
||||
score > mutable_cf_options.soft_rate_limit) {
|
||||
write_controller_token_ = write_controller->GetDelayToken();
|
||||
internal_stats_->RecordLevelNSlowdown(max_level, true);
|
||||
Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
|
||||
"[%s] Stalling writes because we hit soft limit on level %d",
|
||||
name_.c_str(), max_level);
|
||||
} else {
|
||||
write_controller_token_.reset();
|
||||
}
|
||||
prev_compaction_needed_bytes_ = compaction_needed_bytes;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -687,7 +509,7 @@ MemTable* ColumnFamilyData::ConstructNewMemtable(
|
||||
const MutableCFOptions& mutable_cf_options, SequenceNumber earliest_seq) {
|
||||
assert(current_ != nullptr);
|
||||
return new MemTable(internal_comparator_, ioptions_, mutable_cf_options,
|
||||
write_buffer_manager_, earliest_seq);
|
||||
write_buffer_, earliest_seq);
|
||||
}
|
||||
|
||||
void ColumnFamilyData::CreateNewMemtable(
|
||||
@@ -717,12 +539,13 @@ const int ColumnFamilyData::kCompactAllLevels = -1;
|
||||
const int ColumnFamilyData::kCompactToBaseLevel = -2;
|
||||
|
||||
Compaction* ColumnFamilyData::CompactRange(
|
||||
const MutableCFOptions& mutable_cf_options, int input_level,
|
||||
int output_level, uint32_t output_path_id, const InternalKey* begin,
|
||||
const InternalKey* end, InternalKey** compaction_end, bool* conflict) {
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
int input_level, int output_level, uint32_t output_path_id,
|
||||
const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end) {
|
||||
auto* result = compaction_picker_->CompactRange(
|
||||
GetName(), mutable_cf_options, current_->storage_info(), input_level,
|
||||
output_level, output_path_id, begin, end, compaction_end, conflict);
|
||||
output_level, output_path_id, begin, end, compaction_end);
|
||||
if (result != nullptr) {
|
||||
result->SetInputVersion(current_);
|
||||
}
|
||||
@@ -866,7 +689,7 @@ ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
|
||||
const DBOptions* db_options,
|
||||
const EnvOptions& env_options,
|
||||
Cache* table_cache,
|
||||
WriteBufferManager* write_buffer_manager,
|
||||
WriteBuffer* write_buffer,
|
||||
WriteController* write_controller)
|
||||
: max_column_family_(0),
|
||||
dummy_cfd_(new ColumnFamilyData(0, "", nullptr, nullptr, nullptr,
|
||||
@@ -877,7 +700,7 @@ ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
|
||||
db_options_(db_options),
|
||||
env_options_(env_options),
|
||||
table_cache_(table_cache),
|
||||
write_buffer_manager_(write_buffer_manager),
|
||||
write_buffer_(write_buffer),
|
||||
write_controller_(write_controller) {
|
||||
// initialize linked list
|
||||
dummy_cfd_->prev_ = dummy_cfd_;
|
||||
@@ -940,9 +763,10 @@ ColumnFamilyData* ColumnFamilySet::CreateColumnFamily(
|
||||
const std::string& name, uint32_t id, Version* dummy_versions,
|
||||
const ColumnFamilyOptions& options) {
|
||||
assert(column_families_.find(name) == column_families_.end());
|
||||
ColumnFamilyData* new_cfd = new ColumnFamilyData(
|
||||
id, name, dummy_versions, table_cache_, write_buffer_manager_, options,
|
||||
db_options_, env_options_, this);
|
||||
ColumnFamilyData* new_cfd =
|
||||
new ColumnFamilyData(id, name, dummy_versions, table_cache_,
|
||||
write_buffer_, options, db_options_,
|
||||
env_options_, this);
|
||||
column_families_.insert({name, id});
|
||||
column_family_data_.insert({id, new_cfd});
|
||||
max_column_family_ = std::max(max_column_family_, id);
|
||||
@@ -1007,6 +831,13 @@ ColumnFamilyHandle* ColumnFamilyMemTablesImpl::GetColumnFamilyHandle() {
|
||||
return &handle_;
|
||||
}
|
||||
|
||||
void ColumnFamilyMemTablesImpl::CheckMemtableFull() {
|
||||
if (current_ != nullptr && current_->mem()->ShouldScheduleFlush()) {
|
||||
flush_scheduler_->ScheduleFlush(current_);
|
||||
current_->mem()->MarkFlushScheduled();
|
||||
}
|
||||
}
|
||||
|
||||
uint32_t GetColumnFamilyID(ColumnFamilyHandle* column_family) {
|
||||
uint32_t column_family_id = 0;
|
||||
if (column_family != nullptr) {
|
||||
@@ -1019,7 +850,8 @@ uint32_t GetColumnFamilyID(ColumnFamilyHandle* column_family) {
|
||||
const Comparator* GetColumnFamilyUserComparator(
|
||||
ColumnFamilyHandle* column_family) {
|
||||
if (column_family != nullptr) {
|
||||
return column_family->GetComparator();
|
||||
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
|
||||
return cfh->user_comparator();
|
||||
}
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
+44
-61
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -15,15 +15,17 @@
|
||||
#include <atomic>
|
||||
|
||||
#include "db/memtable_list.h"
|
||||
#include "db/table_cache.h"
|
||||
#include "db/table_properties_collector.h"
|
||||
#include "db/write_batch_internal.h"
|
||||
#include "db/write_controller.h"
|
||||
#include "db/table_cache.h"
|
||||
#include "db/table_properties_collector.h"
|
||||
#include "db/flush_scheduler.h"
|
||||
#include "rocksdb/compaction_job_stats.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "util/cf_options.h"
|
||||
#include "util/instrumented_mutex.h"
|
||||
#include "util/mutable_cf_options.h"
|
||||
#include "util/thread_local.h"
|
||||
|
||||
namespace rocksdb {
|
||||
@@ -42,8 +44,6 @@ class LogBuffer;
|
||||
class InstrumentedMutex;
|
||||
class InstrumentedMutexLock;
|
||||
|
||||
extern const double kSlowdownRatio;
|
||||
|
||||
// ColumnFamilyHandleImpl is the class that clients use to access different
|
||||
// column families. It has non-trivial destructor, which gets called when client
|
||||
// is done using the column family
|
||||
@@ -55,11 +55,10 @@ class ColumnFamilyHandleImpl : public ColumnFamilyHandle {
|
||||
// destroy without mutex
|
||||
virtual ~ColumnFamilyHandleImpl();
|
||||
virtual ColumnFamilyData* cfd() const { return cfd_; }
|
||||
virtual const Comparator* user_comparator() const;
|
||||
|
||||
virtual uint32_t GetID() const override;
|
||||
virtual const std::string& GetName() const override;
|
||||
virtual Status GetDescriptor(ColumnFamilyDescriptor* desc) override;
|
||||
virtual const Comparator* GetComparator() const override;
|
||||
|
||||
private:
|
||||
ColumnFamilyData* cfd_;
|
||||
@@ -133,9 +132,6 @@ struct SuperVersion {
|
||||
|
||||
extern Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options);
|
||||
|
||||
extern Status CheckConcurrentWritesSupported(
|
||||
const ColumnFamilyOptions& cf_options);
|
||||
|
||||
extern ColumnFamilyOptions SanitizeOptions(const DBOptions& db_options,
|
||||
const InternalKeyComparator* icmp,
|
||||
const ColumnFamilyOptions& src);
|
||||
@@ -160,16 +156,14 @@ class ColumnFamilyData {
|
||||
// thread-safe
|
||||
const std::string& GetName() const { return name_; }
|
||||
|
||||
// Ref() can only be called from a context where the caller can guarantee
|
||||
// that ColumnFamilyData is alive (while holding a non-zero ref already,
|
||||
// holding a DB mutex, or as the leader in a write batch group).
|
||||
// Ref() can only be called whily holding a DB mutex or during a
|
||||
// single-threaded write.
|
||||
void Ref() { refs_.fetch_add(1, std::memory_order_relaxed); }
|
||||
|
||||
// Unref decreases the reference count, but does not handle deletion
|
||||
// when the count goes to 0. If this method returns true then the
|
||||
// caller should delete the instance immediately, or later, by calling
|
||||
// FreeDeadColumnFamilies(). Unref() can only be called while holding
|
||||
// a DB mutex, or during single-threaded recovery.
|
||||
// will just decrease reference count to 0, but will not delete it. returns
|
||||
// true if the ref count was decreased to zero. in that case, it can be
|
||||
// deleted by the caller immediately, or later, by calling
|
||||
// FreeDeadColumnFamilies()
|
||||
// Unref() can only be called while holding a DB mutex
|
||||
bool Unref() {
|
||||
int old_refs = refs_.fetch_sub(1, std::memory_order_relaxed);
|
||||
assert(old_refs > 0);
|
||||
@@ -209,7 +203,7 @@ class ColumnFamilyData {
|
||||
const ImmutableCFOptions* ioptions() const { return &ioptions_; }
|
||||
// REQUIRES: DB mutex held
|
||||
// This returns the MutableCFOptions used by current SuperVersion
|
||||
// You should use this API to reference MutableCFOptions most of the time.
|
||||
// You shoul use this API to reference MutableCFOptions most of the time.
|
||||
const MutableCFOptions* GetCurrentMutableCFOptions() const {
|
||||
return &(super_version_->mutable_cf_options);
|
||||
}
|
||||
@@ -230,7 +224,7 @@ class ColumnFamilyData {
|
||||
MemTable* mem() { return mem_; }
|
||||
Version* current() { return current_; }
|
||||
Version* dummy_versions() { return dummy_versions_; }
|
||||
void SetCurrent(Version* _current);
|
||||
void SetCurrent(Version* current);
|
||||
uint64_t GetNumLiveVersions() const; // REQUIRE: DB mutex held
|
||||
uint64_t GetTotalSstFilesSize() const; // REQUIRE: DB mutex held
|
||||
void SetMemtable(MemTable* new_mem) { mem_ = new_mem; }
|
||||
@@ -255,11 +249,11 @@ class ColumnFamilyData {
|
||||
// A flag to tell a manual compaction's output is base level.
|
||||
static const int kCompactToBaseLevel;
|
||||
// REQUIRES: DB mutex held
|
||||
Compaction* CompactRange(const MutableCFOptions& mutable_cf_options,
|
||||
int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin,
|
||||
const InternalKey* end, InternalKey** compaction_end,
|
||||
bool* manual_conflict);
|
||||
Compaction* CompactRange(
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
int input_level, int output_level, uint32_t output_path_id,
|
||||
const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end);
|
||||
|
||||
CompactionPicker* compaction_picker() { return compaction_picker_.get(); }
|
||||
// thread-safe
|
||||
@@ -311,6 +305,15 @@ class ColumnFamilyData {
|
||||
bool pending_flush() { return pending_flush_; }
|
||||
bool pending_compaction() { return pending_compaction_; }
|
||||
|
||||
private:
|
||||
friend class ColumnFamilySet;
|
||||
ColumnFamilyData(uint32_t id, const std::string& name,
|
||||
Version* dummy_versions, Cache* table_cache,
|
||||
WriteBuffer* write_buffer,
|
||||
const ColumnFamilyOptions& options,
|
||||
const DBOptions* db_options, const EnvOptions& env_options,
|
||||
ColumnFamilySet* column_family_set);
|
||||
|
||||
// Recalculate some small conditions, which are changed only during
|
||||
// compaction, adding new memtable and/or
|
||||
// recalculation of compaction score. These values are used in
|
||||
@@ -319,15 +322,6 @@ class ColumnFamilyData {
|
||||
void RecalculateWriteStallConditions(
|
||||
const MutableCFOptions& mutable_cf_options);
|
||||
|
||||
private:
|
||||
friend class ColumnFamilySet;
|
||||
ColumnFamilyData(uint32_t id, const std::string& name,
|
||||
Version* dummy_versions, Cache* table_cache,
|
||||
WriteBufferManager* write_buffer_manager,
|
||||
const ColumnFamilyOptions& options,
|
||||
const DBOptions* db_options, const EnvOptions& env_options,
|
||||
ColumnFamilySet* column_family_set);
|
||||
|
||||
uint32_t id_;
|
||||
const std::string name_;
|
||||
Version* dummy_versions_; // Head of circular doubly-linked list of versions.
|
||||
@@ -348,7 +342,7 @@ class ColumnFamilyData {
|
||||
|
||||
std::unique_ptr<InternalStats> internal_stats_;
|
||||
|
||||
WriteBufferManager* write_buffer_manager_;
|
||||
WriteBuffer* write_buffer_;
|
||||
|
||||
MemTable* mem_;
|
||||
MemTableList imm_;
|
||||
@@ -388,8 +382,6 @@ class ColumnFamilyData {
|
||||
// If true --> this ColumnFamily is currently present in
|
||||
// DBImpl::compaction_queue_
|
||||
bool pending_compaction_;
|
||||
|
||||
uint64_t prev_compaction_needed_bytes_;
|
||||
};
|
||||
|
||||
// ColumnFamilySet has interesting thread-safety requirements
|
||||
@@ -438,8 +430,7 @@ class ColumnFamilySet {
|
||||
|
||||
ColumnFamilySet(const std::string& dbname, const DBOptions* db_options,
|
||||
const EnvOptions& env_options, Cache* table_cache,
|
||||
WriteBufferManager* write_buffer_manager,
|
||||
WriteController* write_controller);
|
||||
WriteBuffer* write_buffer, WriteController* write_controller);
|
||||
~ColumnFamilySet();
|
||||
|
||||
ColumnFamilyData* GetDefault() const;
|
||||
@@ -466,8 +457,6 @@ class ColumnFamilySet {
|
||||
// Don't call while iterating over ColumnFamilySet
|
||||
void FreeDeadColumnFamilies();
|
||||
|
||||
Cache* get_table_cache() { return table_cache_; }
|
||||
|
||||
private:
|
||||
friend class ColumnFamilyData;
|
||||
// helper function that gets called from cfd destructor
|
||||
@@ -496,7 +485,7 @@ class ColumnFamilySet {
|
||||
const DBOptions* const db_options_;
|
||||
const EnvOptions env_options_;
|
||||
Cache* table_cache_;
|
||||
WriteBufferManager* write_buffer_manager_;
|
||||
WriteBuffer* write_buffer_;
|
||||
WriteController* write_controller_;
|
||||
};
|
||||
|
||||
@@ -504,18 +493,15 @@ class ColumnFamilySet {
|
||||
// memtables of different column families (specified by ID in the write batch)
|
||||
class ColumnFamilyMemTablesImpl : public ColumnFamilyMemTables {
|
||||
public:
|
||||
explicit ColumnFamilyMemTablesImpl(ColumnFamilySet* column_family_set)
|
||||
: column_family_set_(column_family_set), current_(nullptr) {}
|
||||
|
||||
// Constructs a ColumnFamilyMemTablesImpl equivalent to one constructed
|
||||
// with the arguments used to construct *orig.
|
||||
explicit ColumnFamilyMemTablesImpl(ColumnFamilyMemTablesImpl* orig)
|
||||
: column_family_set_(orig->column_family_set_), current_(nullptr) {}
|
||||
explicit ColumnFamilyMemTablesImpl(ColumnFamilySet* column_family_set,
|
||||
FlushScheduler* flush_scheduler)
|
||||
: column_family_set_(column_family_set),
|
||||
current_(nullptr),
|
||||
flush_scheduler_(flush_scheduler) {}
|
||||
|
||||
// sets current_ to ColumnFamilyData with column_family_id
|
||||
// returns false if column family doesn't exist
|
||||
// REQUIRES: use this function of DBImpl::column_family_memtables_ should be
|
||||
// under a DB mutex OR from a write thread
|
||||
// REQUIRES: under a DB mutex OR from a write thread
|
||||
bool Seek(uint32_t column_family_id) override;
|
||||
|
||||
// Returns log number of the selected column family
|
||||
@@ -523,23 +509,20 @@ class ColumnFamilyMemTablesImpl : public ColumnFamilyMemTables {
|
||||
uint64_t GetLogNumber() const override;
|
||||
|
||||
// REQUIRES: Seek() called first
|
||||
// REQUIRES: use this function of DBImpl::column_family_memtables_ should be
|
||||
// under a DB mutex OR from a write thread
|
||||
// REQUIRES: under a DB mutex OR from a write thread
|
||||
virtual MemTable* GetMemTable() const override;
|
||||
|
||||
// Returns column family handle for the selected column family
|
||||
// REQUIRES: use this function of DBImpl::column_family_memtables_ should be
|
||||
// under a DB mutex OR from a write thread
|
||||
// REQUIRES: under a DB mutex OR from a write thread
|
||||
virtual ColumnFamilyHandle* GetColumnFamilyHandle() override;
|
||||
|
||||
// Cannot be called while another thread is calling Seek().
|
||||
// REQUIRES: use this function of DBImpl::column_family_memtables_ should be
|
||||
// under a DB mutex OR from a write thread
|
||||
virtual ColumnFamilyData* current() override { return current_; }
|
||||
// REQUIRES: under a DB mutex OR from a write thread
|
||||
virtual void CheckMemtableFull() override;
|
||||
|
||||
private:
|
||||
ColumnFamilySet* column_family_set_;
|
||||
ColumnFamilyData* current_;
|
||||
FlushScheduler* flush_scheduler_;
|
||||
ColumnFamilyHandleInternal handle_;
|
||||
};
|
||||
|
||||
|
||||
+47
-1944
File diff suppressed because it is too large.
Load diff
+8
-162
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -7,14 +7,11 @@
|
||||
|
||||
#include <mutex>
|
||||
#include <string>
|
||||
#include <thread>
|
||||
#include <vector>
|
||||
|
||||
#include "db/db_impl.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "util/string_util.h"
|
||||
#include "util/sync_point.h"
|
||||
#include "util/testharness.h"
|
||||
|
||||
namespace rocksdb {
|
||||
@@ -50,79 +47,22 @@ class FlushedFileCollector : public EventListener {
|
||||
}
|
||||
return result;
|
||||
}
|
||||
void ClearFlushedFiles() { flushed_files_.clear(); }
|
||||
|
||||
private:
|
||||
std::vector<std::string> flushed_files_;
|
||||
std::mutex mutex_;
|
||||
};
|
||||
|
||||
TEST_F(CompactFilesTest, L0ConflictsFiles) {
|
||||
Options options;
|
||||
// to trigger compaction more easily
|
||||
const int kWriteBufferSize = 10000;
|
||||
const int kLevel0Trigger = 2;
|
||||
options.create_if_missing = true;
|
||||
options.compaction_style = kCompactionStyleLevel;
|
||||
// Small slowdown and stop trigger for experimental purpose.
|
||||
options.level0_slowdown_writes_trigger = 20;
|
||||
options.level0_stop_writes_trigger = 20;
|
||||
options.level0_stop_writes_trigger = 20;
|
||||
options.write_buffer_size = kWriteBufferSize;
|
||||
options.level0_file_num_compaction_trigger = kLevel0Trigger;
|
||||
options.compression = kNoCompression;
|
||||
|
||||
DB* db = nullptr;
|
||||
DestroyDB(db_name_, options);
|
||||
Status s = DB::Open(options, db_name_, &db);
|
||||
assert(s.ok());
|
||||
assert(db);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency({
|
||||
{"CompactFilesImpl:0", "BackgroundCallCompaction:0"},
|
||||
{"BackgroundCallCompaction:1", "CompactFilesImpl:1"},
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
// create couple files
|
||||
// Background compaction starts and waits in BackgroundCallCompaction:0
|
||||
for (int i = 0; i < kLevel0Trigger * 4; ++i) {
|
||||
db->Put(WriteOptions(), ToString(i), "");
|
||||
db->Put(WriteOptions(), ToString(100 - i), "");
|
||||
db->Flush(FlushOptions());
|
||||
}
|
||||
|
||||
rocksdb::ColumnFamilyMetaData meta;
|
||||
db->GetColumnFamilyMetaData(&meta);
|
||||
std::string file1;
|
||||
for (auto& file : meta.levels[0].files) {
|
||||
ASSERT_EQ(0, meta.levels[0].level);
|
||||
if (file1 == "") {
|
||||
file1 = file.db_path + "/" + file.name;
|
||||
} else {
|
||||
std::string file2 = file.db_path + "/" + file.name;
|
||||
// Another thread starts a compact files and creates an L0 compaction
|
||||
// The background compaction then notices that there is an L0 compaction
|
||||
// already in progress and doesn't do an L0 compaction
|
||||
// Once the background compaction finishes, the compact files finishes
|
||||
ASSERT_OK(
|
||||
db->CompactFiles(rocksdb::CompactionOptions(), {file1, file2}, 0));
|
||||
break;
|
||||
}
|
||||
}
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
delete db;
|
||||
}
|
||||
|
||||
TEST_F(CompactFilesTest, ObsoleteFiles) {
|
||||
Options options;
|
||||
// to trigger compaction more easily
|
||||
const int kWriteBufferSize = 65536;
|
||||
const int kWriteBufferSize = 10000;
|
||||
options.create_if_missing = true;
|
||||
// Disable RocksDB background compaction.
|
||||
options.compaction_style = kCompactionStyleNone;
|
||||
options.level0_slowdown_writes_trigger = (1 << 30);
|
||||
options.level0_stop_writes_trigger = (1 << 30);
|
||||
// Small slowdown and stop trigger for experimental purpose.
|
||||
options.level0_slowdown_writes_trigger = 20;
|
||||
options.level0_stop_writes_trigger = 20;
|
||||
options.write_buffer_size = kWriteBufferSize;
|
||||
options.max_write_buffer_number = 2;
|
||||
options.compression = kNoCompression;
|
||||
@@ -144,8 +84,10 @@ TEST_F(CompactFilesTest, ObsoleteFiles) {
|
||||
}
|
||||
|
||||
auto l0_files = collector->GetFlushedFiles();
|
||||
CompactionOptions compact_opt;
|
||||
compact_opt.compression = kNoCompression;
|
||||
compact_opt.output_file_size_limit = kWriteBufferSize * 5;
|
||||
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files, 1));
|
||||
reinterpret_cast<DBImpl*>(db)->TEST_WaitForCompact();
|
||||
|
||||
// verify all compaction input files are deleted
|
||||
for (auto fname : l0_files) {
|
||||
@@ -154,102 +96,6 @@ TEST_F(CompactFilesTest, ObsoleteFiles) {
|
||||
delete db;
|
||||
}
|
||||
|
||||
TEST_F(CompactFilesTest, NotCutOutputOnLevel0) {
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
// Disable RocksDB background compaction.
|
||||
options.compaction_style = kCompactionStyleNone;
|
||||
options.level0_slowdown_writes_trigger = 1000;
|
||||
options.level0_stop_writes_trigger = 1000;
|
||||
options.write_buffer_size = 65536;
|
||||
options.max_write_buffer_number = 2;
|
||||
options.compression = kNoCompression;
|
||||
options.max_compaction_bytes = 5000;
|
||||
|
||||
// Add listener
|
||||
FlushedFileCollector* collector = new FlushedFileCollector();
|
||||
options.listeners.emplace_back(collector);
|
||||
|
||||
DB* db = nullptr;
|
||||
DestroyDB(db_name_, options);
|
||||
Status s = DB::Open(options, db_name_, &db);
|
||||
assert(s.ok());
|
||||
assert(db);
|
||||
|
||||
// create couple files
|
||||
for (int i = 0; i < 500; ++i) {
|
||||
db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
|
||||
}
|
||||
reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
|
||||
auto l0_files_1 = collector->GetFlushedFiles();
|
||||
collector->ClearFlushedFiles();
|
||||
for (int i = 0; i < 500; ++i) {
|
||||
db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
|
||||
}
|
||||
reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
|
||||
auto l0_files_2 = collector->GetFlushedFiles();
|
||||
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_1, 0));
|
||||
ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_2, 0));
|
||||
// no assertion failure
|
||||
delete db;
|
||||
}
|
||||
|
||||
TEST_F(CompactFilesTest, CapturingPendingFiles) {
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
// Disable RocksDB background compaction.
|
||||
options.compaction_style = kCompactionStyleNone;
|
||||
// Always do full scans for obsolete files (needed to reproduce the issue).
|
||||
options.delete_obsolete_files_period_micros = 0;
|
||||
|
||||
// Add listener.
|
||||
FlushedFileCollector* collector = new FlushedFileCollector();
|
||||
options.listeners.emplace_back(collector);
|
||||
|
||||
DB* db = nullptr;
|
||||
DestroyDB(db_name_, options);
|
||||
Status s = DB::Open(options, db_name_, &db);
|
||||
assert(s.ok());
|
||||
assert(db);
|
||||
|
||||
// Create 5 files.
|
||||
for (int i = 0; i < 5; ++i) {
|
||||
db->Put(WriteOptions(), "key" + ToString(i), "value");
|
||||
db->Flush(FlushOptions());
|
||||
}
|
||||
|
||||
auto l0_files = collector->GetFlushedFiles();
|
||||
EXPECT_EQ(5, l0_files.size());
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency({
|
||||
{"CompactFilesImpl:2", "CompactFilesTest.CapturingPendingFiles:0"},
|
||||
{"CompactFilesTest.CapturingPendingFiles:1", "CompactFilesImpl:3"},
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
// Start compacting files.
|
||||
std::thread compaction_thread(
|
||||
[&] { EXPECT_OK(db->CompactFiles(CompactionOptions(), l0_files, 1)); });
|
||||
|
||||
// In the meantime flush another file.
|
||||
TEST_SYNC_POINT("CompactFilesTest.CapturingPendingFiles:0");
|
||||
db->Put(WriteOptions(), "key5", "value");
|
||||
db->Flush(FlushOptions());
|
||||
TEST_SYNC_POINT("CompactFilesTest.CapturingPendingFiles:1");
|
||||
|
||||
compaction_thread.join();
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
|
||||
delete db;
|
||||
|
||||
// Make sure we can reopen the DB.
|
||||
s = DB::Open(options, db_name_, &db);
|
||||
ASSERT_TRUE(s.ok());
|
||||
assert(db);
|
||||
delete db;
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -93,7 +93,7 @@ Status CompactedDBImpl::Init(const Options& options) {
|
||||
mutex_.Lock();
|
||||
ColumnFamilyDescriptor cf(kDefaultColumnFamilyName,
|
||||
ColumnFamilyOptions(options));
|
||||
Status s = Recover({cf}, true /* read only */, false, true);
|
||||
Status s = Recover({ cf }, true /* read only */, false);
|
||||
if (s.ok()) {
|
||||
cfd_ = reinterpret_cast<ColumnFamilyHandleImpl*>(
|
||||
DefaultColumnFamily())->cfd();
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
+53
-32
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -16,8 +16,8 @@
|
||||
#include <inttypes.h>
|
||||
#include <vector>
|
||||
|
||||
#include "db/column_family.h"
|
||||
#include "rocksdb/compaction_filter.h"
|
||||
#include "db/column_family.h"
|
||||
#include "util/logging.h"
|
||||
#include "util/sync_point.h"
|
||||
|
||||
@@ -46,7 +46,7 @@ void Compaction::GetBoundaryKeys(
|
||||
Slice* largest_user_key) {
|
||||
bool initialized = false;
|
||||
const Comparator* ucmp = vstorage->InternalComparator()->user_comparator();
|
||||
for (size_t i = 0; i < inputs.size(); ++i) {
|
||||
for (uint32_t i = 0; i < inputs.size(); ++i) {
|
||||
if (inputs[i].files.empty()) {
|
||||
continue;
|
||||
}
|
||||
@@ -128,8 +128,8 @@ bool Compaction::TEST_IsBottommostLevel(
|
||||
bool Compaction::IsFullCompaction(
|
||||
VersionStorageInfo* vstorage,
|
||||
const std::vector<CompactionInputFiles>& inputs) {
|
||||
size_t num_files_in_compaction = 0;
|
||||
size_t total_num_files = 0;
|
||||
int num_files_in_compaction = 0;
|
||||
int total_num_files = 0;
|
||||
for (int l = 0; l < vstorage->num_levels(); l++) {
|
||||
total_num_files += vstorage->NumLevelFiles(l);
|
||||
}
|
||||
@@ -143,16 +143,15 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
|
||||
const MutableCFOptions& _mutable_cf_options,
|
||||
std::vector<CompactionInputFiles> _inputs,
|
||||
int _output_level, uint64_t _target_file_size,
|
||||
uint64_t _max_compaction_bytes, uint32_t _output_path_id,
|
||||
CompressionType _compression,
|
||||
uint64_t _max_grandparent_overlap_bytes,
|
||||
uint32_t _output_path_id, CompressionType _compression,
|
||||
std::vector<FileMetaData*> _grandparents,
|
||||
bool _manual_compaction, double _score,
|
||||
bool _deletion_compaction,
|
||||
CompactionReason _compaction_reason)
|
||||
bool _deletion_compaction)
|
||||
: start_level_(_inputs[0].level),
|
||||
output_level_(_output_level),
|
||||
max_output_file_size_(_target_file_size),
|
||||
max_compaction_bytes_(_max_compaction_bytes),
|
||||
max_grandparent_overlap_bytes_(_max_grandparent_overlap_bytes),
|
||||
mutable_cf_options_(_mutable_cf_options),
|
||||
input_version_(nullptr),
|
||||
number_levels_(vstorage->num_levels()),
|
||||
@@ -162,15 +161,14 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
|
||||
deletion_compaction_(_deletion_compaction),
|
||||
inputs_(std::move(_inputs)),
|
||||
grandparents_(std::move(_grandparents)),
|
||||
grandparent_index_(0),
|
||||
seen_key_(false),
|
||||
overlapped_bytes_(0),
|
||||
score_(_score),
|
||||
bottommost_level_(IsBottommostLevel(output_level_, vstorage, inputs_)),
|
||||
is_full_compaction_(IsFullCompaction(vstorage, inputs_)),
|
||||
is_manual_compaction_(_manual_compaction),
|
||||
compaction_reason_(_compaction_reason) {
|
||||
is_manual_compaction_(_manual_compaction) {
|
||||
MarkFilesBeingCompacted(true);
|
||||
if (is_manual_compaction_) {
|
||||
compaction_reason_ = CompactionReason::kManualCompaction;
|
||||
}
|
||||
|
||||
#ifndef NDEBUG
|
||||
for (size_t i = 1; i < inputs_.size(); ++i) {
|
||||
@@ -186,8 +184,6 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
|
||||
&arena_);
|
||||
}
|
||||
}
|
||||
|
||||
GetBoundaryKeys(vstorage, inputs_, &smallest_user_key_, &largest_user_key_);
|
||||
}
|
||||
|
||||
Compaction::~Compaction() {
|
||||
@@ -202,11 +198,9 @@ Compaction::~Compaction() {
|
||||
}
|
||||
|
||||
bool Compaction::InputCompressionMatchesOutput() const {
|
||||
VersionStorageInfo* vstorage = input_version_->storage_info();
|
||||
int base_level = vstorage->base_level();
|
||||
bool matches =
|
||||
(GetCompressionType(*cfd_->ioptions(), vstorage, mutable_cf_options_,
|
||||
start_level_, base_level) == output_compression_);
|
||||
int base_level = input_version_->storage_info()->base_level();
|
||||
bool matches = (GetCompressionType(*cfd_->ioptions(), start_level_,
|
||||
base_level) == output_compression_);
|
||||
if (matches) {
|
||||
TEST_SYNC_POINT("Compaction::InputCompressionMatchesOutput:Matches");
|
||||
return true;
|
||||
@@ -247,7 +241,7 @@ bool Compaction::IsTrivialMove() const {
|
||||
return (start_level_ != output_level_ && num_input_levels() == 1 &&
|
||||
input(0, 0)->fd.GetPathId() == output_path_id() &&
|
||||
InputCompressionMatchesOutput() &&
|
||||
TotalFileSize(grandparents_) <= max_compaction_bytes_);
|
||||
TotalFileSize(grandparents_) <= max_grandparent_overlap_bytes_);
|
||||
}
|
||||
|
||||
void Compaction::AddInputDeletions(VersionEdit* out_edit) {
|
||||
@@ -288,12 +282,38 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
|
||||
return true;
|
||||
}
|
||||
|
||||
bool Compaction::ShouldStopBefore(const Slice& internal_key) {
|
||||
// Scan to find earliest grandparent file that contains key.
|
||||
const InternalKeyComparator* icmp = &cfd_->internal_comparator();
|
||||
while (grandparent_index_ < grandparents_.size() &&
|
||||
icmp->Compare(internal_key,
|
||||
grandparents_[grandparent_index_]->largest.Encode()) > 0) {
|
||||
if (seen_key_) {
|
||||
overlapped_bytes_ += grandparents_[grandparent_index_]->fd.GetFileSize();
|
||||
}
|
||||
assert(grandparent_index_ + 1 >= grandparents_.size() ||
|
||||
icmp->Compare(grandparents_[grandparent_index_]->largest.Encode(),
|
||||
grandparents_[grandparent_index_+1]->smallest.Encode())
|
||||
< 0);
|
||||
grandparent_index_++;
|
||||
}
|
||||
seen_key_ = true;
|
||||
|
||||
if (overlapped_bytes_ > max_grandparent_overlap_bytes_) {
|
||||
// Too much overlap for current output; start new output
|
||||
overlapped_bytes_ = 0;
|
||||
return true;
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
// Mark (or clear) each file that is being compacted
|
||||
void Compaction::MarkFilesBeingCompacted(bool mark_as_compacted) {
|
||||
for (size_t i = 0; i < num_input_levels(); i++) {
|
||||
for (size_t j = 0; j < inputs_[i].size(); j++) {
|
||||
assert(mark_as_compacted ? !inputs_[i][j]->being_compacted
|
||||
: inputs_[i][j]->being_compacted);
|
||||
for (unsigned int j = 0; j < inputs_[i].size(); j++) {
|
||||
assert(mark_as_compacted ? !inputs_[i][j]->being_compacted :
|
||||
inputs_[i][j]->being_compacted);
|
||||
inputs_[i][j]->being_compacted = mark_as_compacted;
|
||||
}
|
||||
}
|
||||
@@ -351,7 +371,7 @@ int InputSummary(const std::vector<FileMetaData*>& files, char* output,
|
||||
int len) {
|
||||
*output = '\0';
|
||||
int write = 0;
|
||||
for (size_t i = 0; i < files.size(); i++) {
|
||||
for (unsigned int i = 0; i < files.size(); i++) {
|
||||
int sz = len - write;
|
||||
int ret;
|
||||
char sztxt[16];
|
||||
@@ -368,8 +388,10 @@ int InputSummary(const std::vector<FileMetaData*>& files, char* output,
|
||||
|
||||
void Compaction::Summary(char* output, int len) {
|
||||
int write =
|
||||
snprintf(output, len, "Base version %" PRIu64 " Base level %d, inputs: [",
|
||||
input_version_->GetVersionNumber(), start_level_);
|
||||
snprintf(output, len, "Base version %" PRIu64
|
||||
" Base level %d, inputs: [",
|
||||
input_version_->GetVersionNumber(),
|
||||
start_level_);
|
||||
if (write < 0 || write >= len) {
|
||||
return;
|
||||
}
|
||||
@@ -391,7 +413,7 @@ void Compaction::Summary(char* output, int len) {
|
||||
snprintf(output + write, len - write, "]");
|
||||
}
|
||||
|
||||
uint64_t Compaction::OutputFilePreallocationSize() const {
|
||||
uint64_t Compaction::OutputFilePreallocationSize() {
|
||||
uint64_t preallocation_size = 0;
|
||||
|
||||
if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel ||
|
||||
@@ -406,7 +428,7 @@ uint64_t Compaction::OutputFilePreallocationSize() const {
|
||||
}
|
||||
// Over-estimate slightly so we don't end up just barely crossing
|
||||
// the threshold
|
||||
return preallocation_size + (preallocation_size / 10);
|
||||
return preallocation_size * 1.1;
|
||||
}
|
||||
|
||||
std::unique_ptr<CompactionFilter> Compaction::CreateCompactionFilter() const {
|
||||
@@ -417,7 +439,6 @@ std::unique_ptr<CompactionFilter> Compaction::CreateCompactionFilter() const {
|
||||
CompactionFilter::Context context;
|
||||
context.is_full_compaction = is_full_compaction_;
|
||||
context.is_manual_compaction = is_manual_compaction_;
|
||||
context.column_family_id = cfd_->GetID();
|
||||
return cfd_->ioptions()->compaction_filter_factory->CreateCompactionFilter(
|
||||
context);
|
||||
}
|
||||
|
||||
+22
-51
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -8,10 +8,10 @@
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#pragma once
|
||||
#include "db/version_set.h"
|
||||
#include "util/arena.h"
|
||||
#include "util/autovector.h"
|
||||
#include "util/cf_options.h"
|
||||
#include "util/mutable_cf_options.h"
|
||||
#include "db/version_set.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -37,12 +37,11 @@ class Compaction {
|
||||
Compaction(VersionStorageInfo* input_version,
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
std::vector<CompactionInputFiles> inputs, int output_level,
|
||||
uint64_t target_file_size, uint64_t max_compaction_bytes,
|
||||
uint64_t target_file_size, uint64_t max_grandparent_overlap_bytes,
|
||||
uint32_t output_path_id, CompressionType compression,
|
||||
std::vector<FileMetaData*> grandparents,
|
||||
bool manual_compaction = false, double score = -1,
|
||||
bool deletion_compaction = false,
|
||||
CompactionReason compaction_reason = CompactionReason::kUnknown);
|
||||
bool deletion_compaction = false);
|
||||
|
||||
// No copying allowed
|
||||
Compaction(const Compaction&) = delete;
|
||||
@@ -104,7 +103,7 @@ class Compaction {
|
||||
}
|
||||
|
||||
// Returns the LevelFilesBrief of the specified compaction input level.
|
||||
const LevelFilesBrief* input_levels(size_t compaction_input_level) const {
|
||||
LevelFilesBrief* input_levels(size_t compaction_input_level) {
|
||||
return &input_levels_[compaction_input_level];
|
||||
}
|
||||
|
||||
@@ -132,10 +131,12 @@ class Compaction {
|
||||
bool KeyNotExistsBeyondOutputLevel(const Slice& user_key,
|
||||
std::vector<size_t>* level_ptrs) const;
|
||||
|
||||
// Returns true iff we should stop building the current output
|
||||
// before processing "internal_key".
|
||||
bool ShouldStopBefore(const Slice& internal_key);
|
||||
|
||||
// Clear all files to indicate that they are not being compacted
|
||||
// Delete this compaction from the list of running compactions.
|
||||
//
|
||||
// Requirement: DB mutex held
|
||||
void ReleaseCompactionFiles(Status status);
|
||||
|
||||
// Returns the summary of the compaction in "output" with maximum "len"
|
||||
@@ -147,13 +148,13 @@ class Compaction {
|
||||
double score() const { return score_; }
|
||||
|
||||
// Is this compaction creating a file in the bottom most level?
|
||||
bool bottommost_level() const { return bottommost_level_; }
|
||||
bool bottommost_level() { return bottommost_level_; }
|
||||
|
||||
// Does this compaction include all sst files?
|
||||
bool is_full_compaction() const { return is_full_compaction_; }
|
||||
bool is_full_compaction() { return is_full_compaction_; }
|
||||
|
||||
// Was this compaction triggered manually by the client?
|
||||
bool is_manual_compaction() const { return is_manual_compaction_; }
|
||||
bool is_manual_compaction() { return is_manual_compaction_; }
|
||||
|
||||
// Used when allow_trivial_move option is set in
|
||||
// Universal compaction. If all the input files are
|
||||
@@ -166,21 +167,19 @@ class Compaction {
|
||||
// Used when allow_trivial_move option is set in
|
||||
// Universal compaction. Returns true, if the input files
|
||||
// are non-overlapping and can be trivially moved.
|
||||
bool is_trivial_move() const { return is_trivial_move_; }
|
||||
bool is_trivial_move() { return is_trivial_move_; }
|
||||
|
||||
// How many total levels are there?
|
||||
int number_levels() const { return number_levels_; }
|
||||
|
||||
// Return the MutableCFOptions that should be used throughout the compaction
|
||||
// procedure
|
||||
const MutableCFOptions* mutable_cf_options() const {
|
||||
return &mutable_cf_options_;
|
||||
}
|
||||
const MutableCFOptions* mutable_cf_options() { return &mutable_cf_options_; }
|
||||
|
||||
// Returns the size in bytes that the output file should be preallocated to.
|
||||
// In level compaction, that is max_file_size_. In universal compaction, that
|
||||
// is the sum of all input file sizes.
|
||||
uint64_t OutputFilePreallocationSize() const;
|
||||
uint64_t OutputFilePreallocationSize();
|
||||
|
||||
void SetInputVersion(Version* input_version);
|
||||
|
||||
@@ -211,26 +210,6 @@ class Compaction {
|
||||
int output_level, VersionStorageInfo* vstorage,
|
||||
const std::vector<CompactionInputFiles>& inputs);
|
||||
|
||||
TablePropertiesCollection GetOutputTableProperties() const {
|
||||
return output_table_properties_;
|
||||
}
|
||||
|
||||
void SetOutputTableProperties(TablePropertiesCollection tp) {
|
||||
output_table_properties_ = std::move(tp);
|
||||
}
|
||||
|
||||
Slice GetSmallestUserKey() const { return smallest_user_key_; }
|
||||
|
||||
Slice GetLargestUserKey() const { return largest_user_key_; }
|
||||
|
||||
CompactionReason compaction_reason() { return compaction_reason_; }
|
||||
|
||||
const std::vector<FileMetaData*>& grandparents() const {
|
||||
return grandparents_;
|
||||
}
|
||||
|
||||
uint64_t max_compaction_bytes() const { return max_compaction_bytes_; }
|
||||
|
||||
private:
|
||||
// mark (or clear) all files that are being compacted
|
||||
void MarkFilesBeingCompacted(bool mark_as_compacted);
|
||||
@@ -252,7 +231,7 @@ class Compaction {
|
||||
const int start_level_; // the lowest level to be compacted
|
||||
const int output_level_; // levels to which output files are stored
|
||||
uint64_t max_output_file_size_;
|
||||
uint64_t max_compaction_bytes_;
|
||||
uint64_t max_grandparent_overlap_bytes_;
|
||||
MutableCFOptions mutable_cf_options_;
|
||||
Version* input_version_;
|
||||
VersionEdit edit_;
|
||||
@@ -271,9 +250,13 @@ class Compaction {
|
||||
// A copy of inputs_, organized more closely in memory
|
||||
autovector<LevelFilesBrief, 2> input_levels_;
|
||||
|
||||
// State used to check for number of overlapping grandparent files
|
||||
// State used to check for number of of overlapping grandparent files
|
||||
// (grandparent == "output_level_ + 1")
|
||||
std::vector<FileMetaData*> grandparents_;
|
||||
size_t grandparent_index_; // Index in grandparent_starts_
|
||||
bool seen_key_; // Some output key has been seen
|
||||
uint64_t overlapped_bytes_; // Bytes of overlap between current output
|
||||
// and grandparent files
|
||||
const double score_; // score that was used to pick this compaction.
|
||||
|
||||
// Is this compaction creating a file in the bottom most level?
|
||||
@@ -290,18 +273,6 @@ class Compaction {
|
||||
|
||||
// Does input compression match the output compression?
|
||||
bool InputCompressionMatchesOutput() const;
|
||||
|
||||
// table properties of output files
|
||||
TablePropertiesCollection output_table_properties_;
|
||||
|
||||
// smallest user keys in compaction
|
||||
Slice smallest_user_key_;
|
||||
|
||||
// largest user keys in compaction
|
||||
Slice largest_user_key_;
|
||||
|
||||
// Reason for compaction
|
||||
CompactionReason compaction_reason_;
|
||||
};
|
||||
|
||||
// Utility function
|
||||
|
||||
+28
-148
@@ -1,26 +1,23 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
#include "db/compaction_iterator.h"
|
||||
#include "table/internal_iterator.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
CompactionIterator::CompactionIterator(
|
||||
InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
|
||||
Iterator* input, const Comparator* cmp, MergeHelper* merge_helper,
|
||||
SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot, Env* env,
|
||||
bool expect_valid_internal_key, const Compaction* compaction,
|
||||
Env* env, bool expect_valid_internal_key, Compaction* compaction,
|
||||
const CompactionFilter* compaction_filter, LogBuffer* log_buffer)
|
||||
: input_(input),
|
||||
cmp_(cmp),
|
||||
merge_helper_(merge_helper),
|
||||
snapshots_(snapshots),
|
||||
earliest_write_conflict_snapshot_(earliest_write_conflict_snapshot),
|
||||
env_(env),
|
||||
expect_valid_internal_key_(expect_valid_internal_key),
|
||||
compaction_(compaction),
|
||||
@@ -44,17 +41,6 @@ CompactionIterator::CompactionIterator(
|
||||
earliest_snapshot_ = snapshots_->at(0);
|
||||
latest_snapshot_ = snapshots_->back();
|
||||
}
|
||||
if (compaction_filter_ != nullptr && compaction_filter_->IgnoreSnapshots()) {
|
||||
ignore_snapshots_ = true;
|
||||
} else {
|
||||
ignore_snapshots_ = false;
|
||||
}
|
||||
input_->SetPinnedItersMgr(&pinned_iters_mgr_);
|
||||
}
|
||||
|
||||
CompactionIterator::~CompactionIterator() {
|
||||
// input_ Iteartor lifetime is longer than pinned_iters_mgr_ lifetime
|
||||
input_->SetPinnedItersMgr(nullptr);
|
||||
}
|
||||
|
||||
void CompactionIterator::ResetRecordCounts() {
|
||||
@@ -89,8 +75,6 @@ void CompactionIterator::Next() {
|
||||
ikey_.user_key = current_key_.GetUserKey();
|
||||
valid_ = true;
|
||||
} else {
|
||||
// We consumed all pinned merge operands, release pinned iterators
|
||||
pinned_iters_mgr_.ReleasePinnedData();
|
||||
// MergeHelper moves the iterator to the first record after the merged
|
||||
// records, so even though we reached the end of the merge output, we do
|
||||
// not want to advance the iterator.
|
||||
@@ -105,11 +89,6 @@ void CompactionIterator::Next() {
|
||||
NextFromInput();
|
||||
}
|
||||
|
||||
if (valid_) {
|
||||
// Record that we've ouputted a record for the current key.
|
||||
has_outputted_key_ = true;
|
||||
}
|
||||
|
||||
PrepareOutput();
|
||||
}
|
||||
|
||||
@@ -156,14 +135,11 @@ void CompactionIterator::NextFromInput() {
|
||||
key_ = current_key_.SetKey(key_, &ikey_);
|
||||
current_user_key_ = ikey_.user_key;
|
||||
has_current_user_key_ = true;
|
||||
has_outputted_key_ = false;
|
||||
current_user_key_sequence_ = kMaxSequenceNumber;
|
||||
current_user_key_snapshot_ = 0;
|
||||
|
||||
// apply the compaction filter to the first occurrence of the user key
|
||||
if (compaction_filter_ != nullptr && ikey_.type == kTypeValue &&
|
||||
(visible_at_tip_ || ikey_.sequence > latest_snapshot_ ||
|
||||
ignore_snapshots_)) {
|
||||
(visible_at_tip_ || ikey_.sequence > latest_snapshot_)) {
|
||||
// If the user has specified a compaction filter and the sequence
|
||||
// number is greater than any external snapshot, then invoke the
|
||||
// filter. If the return value of the compaction filter is true,
|
||||
@@ -193,9 +169,6 @@ void CompactionIterator::NextFromInput() {
|
||||
} else {
|
||||
// Update the current key to reflect the new sequence number/type without
|
||||
// copying the user key.
|
||||
// TODO(rven): Compaction filter does not process keys in this path
|
||||
// Need to have the compaction filter process multiple versions
|
||||
// if we have versions on both sides of a snapshot
|
||||
current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
|
||||
key_ = current_key_.GetKey();
|
||||
ikey_.user_key = current_key_.GetUserKey();
|
||||
@@ -213,106 +186,31 @@ void CompactionIterator::NextFromInput() {
|
||||
visible_at_tip_ ? visible_at_tip_ : findEarliestVisibleSnapshot(
|
||||
ikey_.sequence, &prev_snapshot);
|
||||
|
||||
if (clear_and_output_next_key_) {
|
||||
// In the previous iteration we encountered a single delete that we could
|
||||
// not compact out. We will keep this Put, but can drop it's data.
|
||||
// (See Optimization 3, below.)
|
||||
assert(ikey_.type == kTypeValue);
|
||||
assert(current_user_key_snapshot_ == last_snapshot);
|
||||
|
||||
value_.clear();
|
||||
valid_ = true;
|
||||
clear_and_output_next_key_ = false;
|
||||
} else if (ikey_.type == kTypeSingleDeletion) {
|
||||
// We can compact out a SingleDelete if:
|
||||
// 1) We encounter the corresponding PUT -OR- we know that this key
|
||||
// doesn't appear past this output level
|
||||
// =AND=
|
||||
// 2) We've already returned a record in this snapshot -OR-
|
||||
// there are no earlier earliest_write_conflict_snapshot.
|
||||
//
|
||||
// Rule 1 is needed for SingleDelete correctness. Rule 2 is needed to
|
||||
// allow Transactions to do write-conflict checking (if we compacted away
|
||||
// all keys, then we wouldn't know that a write happened in this
|
||||
// snapshot). If there is no earlier snapshot, then we know that there
|
||||
// are no active transactions that need to know about any writes.
|
||||
//
|
||||
// Optimization 3:
|
||||
// If we encounter a SingleDelete followed by a PUT and Rule 2 is NOT
|
||||
// true, then we must output a SingleDelete. In this case, we will decide
|
||||
// to also output the PUT. While we are compacting less by outputting the
|
||||
// PUT now, hopefully this will lead to better compaction in the future
|
||||
// when Rule 2 is later true (Ie, We are hoping we can later compact out
|
||||
// both the SingleDelete and the Put, while we couldn't if we only
|
||||
// outputted the SingleDelete now).
|
||||
// In this case, we can save space by removing the PUT's value as it will
|
||||
// never be read.
|
||||
//
|
||||
// Deletes and Merges are not supported on the same key that has a
|
||||
// SingleDelete as it is not possible to correctly do any partial
|
||||
// compaction of such a combination of operations. The result of mixing
|
||||
// those operations for a given key is documented as being undefined. So
|
||||
// we can choose how to handle such a combinations of operations. We will
|
||||
// try to compact out as much as we can in these cases.
|
||||
// We will report counts on these anomalous cases.
|
||||
|
||||
// The easiest way to process a SingleDelete during iteration is to peek
|
||||
// ahead at the next key.
|
||||
if (ikey_.type == kTypeSingleDeletion) {
|
||||
ParsedInternalKey next_ikey;
|
||||
input_->Next();
|
||||
|
||||
// Check whether the next key exists, is not corrupt, and is the same key
|
||||
// Check whether the current key is valid, not corrupt and the same
|
||||
// as the single delete.
|
||||
if (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
|
||||
cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
|
||||
// Check whether the next key belongs to the same snapshot as the
|
||||
// SingleDelete.
|
||||
// Mixing single deletes and merges is not supported. Consecutive
|
||||
// single deletes are not valid.
|
||||
if (next_ikey.type != kTypeValue) {
|
||||
assert(false);
|
||||
status_ =
|
||||
Status::InvalidArgument("Put expected after single delete.");
|
||||
break;
|
||||
}
|
||||
|
||||
// Check whether the current key belongs to the same snapshot as the
|
||||
// single delete.
|
||||
if (prev_snapshot == 0 || next_ikey.sequence > prev_snapshot) {
|
||||
if (next_ikey.type == kTypeSingleDeletion) {
|
||||
// We encountered two SingleDeletes in a row. This could be due to
|
||||
// unexpected user input.
|
||||
// Skip the first SingleDelete and let the next iteration decide how
|
||||
// to handle the second SingleDelete
|
||||
|
||||
// First SingleDelete has been skipped since we already called
|
||||
// input_->Next().
|
||||
++iter_stats_.num_record_drop_obsolete;
|
||||
++iter_stats_.num_single_del_mismatch;
|
||||
} else if ((ikey_.sequence <= earliest_write_conflict_snapshot_) ||
|
||||
has_outputted_key_) {
|
||||
// Found a matching value, we can drop the single delete and the
|
||||
// value. It is safe to drop both records since we've already
|
||||
// outputted a key in this snapshot, or there is no earlier
|
||||
// snapshot (Rule 2 above).
|
||||
|
||||
// Note: it doesn't matter whether the second key is a Put or if it
|
||||
// is an unexpected Merge or Delete. We will compact it out
|
||||
// either way. We will maintain counts of how many mismatches
|
||||
// happened
|
||||
if (next_ikey.type != kTypeValue) {
|
||||
++iter_stats_.num_single_del_mismatch;
|
||||
}
|
||||
|
||||
++iter_stats_.num_record_drop_hidden;
|
||||
++iter_stats_.num_record_drop_obsolete;
|
||||
// Already called input_->Next() once. Call it a second time to
|
||||
// skip past the second key.
|
||||
input_->Next();
|
||||
} else {
|
||||
// Found a matching value, but we cannot drop both keys since
|
||||
// there is an earlier snapshot and we need to leave behind a record
|
||||
// to know that a write happened in this snapshot (Rule 2 above).
|
||||
// Clear the value and output the SingleDelete. (The value will be
|
||||
// outputted on the next iteration.)
|
||||
++iter_stats_.num_record_drop_hidden;
|
||||
|
||||
// Setting valid_ to true will output the current SingleDelete
|
||||
valid_ = true;
|
||||
|
||||
// Set up the Put to be outputted in the next iteration.
|
||||
// (Optimization 3).
|
||||
clear_and_output_next_key_ = true;
|
||||
}
|
||||
// Found the matching value, we can drop the single delete and the
|
||||
// value.
|
||||
++iter_stats_.num_record_drop_hidden;
|
||||
++iter_stats_.num_record_drop_obsolete;
|
||||
input_->Next();
|
||||
} else {
|
||||
// We hit the next snapshot without hitting a put, so the iterator
|
||||
// returns the single delete.
|
||||
@@ -327,15 +225,11 @@ void CompactionIterator::NextFromInput() {
|
||||
// iteration. If the next key is corrupt, we return before the
|
||||
// comparison, so the value of has_current_user_key does not matter.
|
||||
has_current_user_key_ = false;
|
||||
if (compaction_ != nullptr && ikey_.sequence <= earliest_snapshot_ &&
|
||||
if (compaction_ != nullptr &&
|
||||
compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
|
||||
&level_ptrs_)) {
|
||||
// Key doesn't exist outside of this range.
|
||||
// Can compact out this SingleDelete.
|
||||
++iter_stats_.num_record_drop_obsolete;
|
||||
++iter_stats_.num_single_del_fallthru;
|
||||
} else {
|
||||
// Output SingleDelete
|
||||
valid_ = true;
|
||||
}
|
||||
}
|
||||
@@ -349,10 +243,6 @@ void CompactionIterator::NextFromInput() {
|
||||
// same key, then this kv is not visible in any snapshot.
|
||||
// Hidden by an newer entry for same user key
|
||||
// TODO: why not > ?
|
||||
//
|
||||
// Note: Dropping this key will not affect TransactionDB write-conflict
|
||||
// checking since there has already been a record returned for this key
|
||||
// in this snapshot.
|
||||
assert(last_sequence >= current_user_key_sequence_);
|
||||
++iter_stats_.num_record_drop_hidden; // (A)
|
||||
input_->Next();
|
||||
@@ -371,9 +261,6 @@ void CompactionIterator::NextFromInput() {
|
||||
// smaller sequence numbers will be dropped in the next
|
||||
// few iterations of this loop (by rule (A) above).
|
||||
// Therefore this deletion marker is obsolete and can be dropped.
|
||||
//
|
||||
// Note: Dropping this Delete will not affect TransactionDB
|
||||
// write-conflict checking since it is earlier than any snapshot.
|
||||
++iter_stats_.num_record_drop_obsolete;
|
||||
input_->Next();
|
||||
} else if (ikey_.type == kTypeMerge) {
|
||||
@@ -384,7 +271,6 @@ void CompactionIterator::NextFromInput() {
|
||||
return;
|
||||
}
|
||||
|
||||
pinned_iters_mgr_.StartPinning();
|
||||
// We know the merge type entry is not hidden, otherwise we would
|
||||
// have hit (A)
|
||||
// We encapsulate the merge related state machine in a different
|
||||
@@ -412,7 +298,6 @@ void CompactionIterator::NextFromInput() {
|
||||
// batch consumed by the merge operator should not shadow any keys
|
||||
// coming after the merges
|
||||
has_current_user_key_ = false;
|
||||
pinned_iters_mgr_.ReleasePinnedData();
|
||||
}
|
||||
} else {
|
||||
valid_ = true;
|
||||
@@ -424,14 +309,9 @@ void CompactionIterator::PrepareOutput() {
|
||||
// Zeroing out the sequence number leads to better compression.
|
||||
// If this is the bottommost level (no files in lower levels)
|
||||
// and the earliest snapshot is larger than this seqno
|
||||
// and the userkey differs from the last userkey in compaction
|
||||
// then we can squash the seqno to zero.
|
||||
|
||||
// This is safe for TransactionDB write-conflict checking since transactions
|
||||
// only care about sequence number larger than any active snapshots.
|
||||
if (bottommost_level_ && valid_ && ikey_.sequence < earliest_snapshot_ &&
|
||||
ikey_.type != kTypeMerge &&
|
||||
!cmp_->Equal(compaction_->GetLargestUserKey(), ikey_.user_key)) {
|
||||
ikey_.type != kTypeMerge) {
|
||||
assert(ikey_.type != kTypeDeletion && ikey_.type != kTypeSingleDeletion);
|
||||
ikey_.sequence = 0;
|
||||
current_key_.UpdateInternalKey(0, ikey_.type);
|
||||
@@ -441,15 +321,15 @@ void CompactionIterator::PrepareOutput() {
|
||||
inline SequenceNumber CompactionIterator::findEarliestVisibleSnapshot(
|
||||
SequenceNumber in, SequenceNumber* prev_snapshot) {
|
||||
assert(snapshots_->size());
|
||||
SequenceNumber prev __attribute__((__unused__)) = kMaxSequenceNumber;
|
||||
SequenceNumber prev __attribute__((unused)) = 0;
|
||||
for (const auto cur : *snapshots_) {
|
||||
assert(prev == kMaxSequenceNumber || prev <= cur);
|
||||
assert(prev <= cur);
|
||||
if (cur >= in) {
|
||||
*prev_snapshot = prev == kMaxSequenceNumber ? 0 : prev;
|
||||
*prev_snapshot = prev;
|
||||
return cur;
|
||||
}
|
||||
prev = cur;
|
||||
assert(prev < kMaxSequenceNumber);
|
||||
assert(prev);
|
||||
}
|
||||
*prev_snapshot = prev;
|
||||
return kMaxSequenceNumber;
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -13,7 +13,6 @@
|
||||
|
||||
#include "db/compaction.h"
|
||||
#include "db/merge_helper.h"
|
||||
#include "db/pinned_iterators_manager.h"
|
||||
#include "rocksdb/compaction_filter.h"
|
||||
#include "util/log_buffer.h"
|
||||
|
||||
@@ -34,25 +33,18 @@ struct CompactionIteratorStats {
|
||||
uint64_t num_input_corrupt_records = 0;
|
||||
uint64_t total_input_raw_key_bytes = 0;
|
||||
uint64_t total_input_raw_value_bytes = 0;
|
||||
|
||||
// Single-Delete diagnostics for exceptional situations
|
||||
uint64_t num_single_del_fallthru = 0;
|
||||
uint64_t num_single_del_mismatch = 0;
|
||||
};
|
||||
|
||||
class CompactionIterator {
|
||||
public:
|
||||
CompactionIterator(InternalIterator* input, const Comparator* cmp,
|
||||
CompactionIterator(Iterator* input, const Comparator* cmp,
|
||||
MergeHelper* merge_helper, SequenceNumber last_sequence,
|
||||
std::vector<SequenceNumber>* snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot, Env* env,
|
||||
std::vector<SequenceNumber>* snapshots, Env* env,
|
||||
bool expect_valid_internal_key,
|
||||
const Compaction* compaction = nullptr,
|
||||
Compaction* compaction = nullptr,
|
||||
const CompactionFilter* compaction_filter = nullptr,
|
||||
LogBuffer* log_buffer = nullptr);
|
||||
|
||||
~CompactionIterator();
|
||||
|
||||
void ResetRecordCounts();
|
||||
|
||||
// Seek to the beginning of the compaction iterator output.
|
||||
@@ -92,14 +84,13 @@ class CompactionIterator {
|
||||
inline SequenceNumber findEarliestVisibleSnapshot(
|
||||
SequenceNumber in, SequenceNumber* prev_snapshot);
|
||||
|
||||
InternalIterator* input_;
|
||||
Iterator* input_;
|
||||
const Comparator* cmp_;
|
||||
MergeHelper* merge_helper_;
|
||||
const std::vector<SequenceNumber>* snapshots_;
|
||||
const SequenceNumber earliest_write_conflict_snapshot_;
|
||||
Env* env_;
|
||||
bool expect_valid_internal_key_;
|
||||
const Compaction* compaction_;
|
||||
Compaction* compaction_;
|
||||
const CompactionFilter* compaction_filter_;
|
||||
LogBuffer* log_buffer_;
|
||||
bool bottommost_level_;
|
||||
@@ -107,7 +98,6 @@ class CompactionIterator {
|
||||
SequenceNumber visible_at_tip_;
|
||||
SequenceNumber earliest_snapshot_;
|
||||
SequenceNumber latest_snapshot_;
|
||||
bool ignore_snapshots_;
|
||||
|
||||
// State
|
||||
//
|
||||
@@ -134,18 +124,7 @@ class CompactionIterator {
|
||||
Slice current_user_key_;
|
||||
SequenceNumber current_user_key_sequence_;
|
||||
SequenceNumber current_user_key_snapshot_;
|
||||
|
||||
// True if the iterator has already returned a record for the current key.
|
||||
bool has_outputted_key_ = false;
|
||||
|
||||
// truncated the value of the next key and output it without applying any
|
||||
// compaction rules. This is used for outputting a put after a single delete.
|
||||
bool clear_and_output_next_key_ = false;
|
||||
|
||||
MergeOutputIterator merge_out_iter_;
|
||||
// PinnedIteratorsManager used to pin input_ Iterator blocks while reading
|
||||
// merge operands and then releasing them after consuming them.
|
||||
PinnedIteratorsManager pinned_iters_mgr_;
|
||||
std::string compaction_filter_value_;
|
||||
// "level_ptrs" holds indices that remember which file of an associated
|
||||
// level we were last checking during the last call to compaction->
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -20,9 +20,9 @@ class CompactionIteratorTest : public testing::Test {
|
||||
nullptr, 0U, false, 0));
|
||||
iter_.reset(new test::VectorIterator(ks, vs));
|
||||
iter_->SeekToFirst();
|
||||
c_iter_.reset(new CompactionIterator(
|
||||
iter_.get(), cmp_, merge_helper_.get(), last_sequence, &snapshots_,
|
||||
kMaxSequenceNumber, Env::Default(), false));
|
||||
c_iter_.reset(new CompactionIterator(iter_.get(), cmp_, merge_helper_.get(),
|
||||
last_sequence, &snapshots_,
|
||||
Env::Default(), false));
|
||||
}
|
||||
|
||||
const Comparator* cmp_;
|
||||
|
||||
+69
-278
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -16,13 +16,12 @@
|
||||
#include <inttypes.h>
|
||||
#include <algorithm>
|
||||
#include <functional>
|
||||
#include <list>
|
||||
#include <vector>
|
||||
#include <memory>
|
||||
#include <random>
|
||||
#include <list>
|
||||
#include <set>
|
||||
#include <thread>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
#include "db/builder.h"
|
||||
#include "db/db_iter.h"
|
||||
@@ -52,7 +51,6 @@
|
||||
#include "util/iostats_context_imp.h"
|
||||
#include "util/log_buffer.h"
|
||||
#include "util/logging.h"
|
||||
#include "util/sst_file_manager_impl.h"
|
||||
#include "util/mutexlock.h"
|
||||
#include "util/perf_context_imp.h"
|
||||
#include "util/stop_watch.h"
|
||||
@@ -64,7 +62,7 @@ namespace rocksdb {
|
||||
|
||||
// Maintains state for each sub-compaction
|
||||
struct CompactionJob::SubcompactionState {
|
||||
const Compaction* compaction;
|
||||
Compaction* compaction;
|
||||
std::unique_ptr<CompactionIterator> c_iter;
|
||||
|
||||
// The boundaries of the key-range this compaction is interested in. No two
|
||||
@@ -79,7 +77,6 @@ struct CompactionJob::SubcompactionState {
|
||||
struct Output {
|
||||
FileMetaData meta;
|
||||
bool finished;
|
||||
std::shared_ptr<const TableProperties> table_properties;
|
||||
};
|
||||
|
||||
// State kept for output being generated
|
||||
@@ -99,22 +96,12 @@ struct CompactionJob::SubcompactionState {
|
||||
}
|
||||
}
|
||||
|
||||
uint64_t current_output_file_size;
|
||||
|
||||
// State during the subcompaction
|
||||
uint64_t total_bytes;
|
||||
uint64_t num_input_records;
|
||||
uint64_t num_output_records;
|
||||
CompactionJobStats compaction_job_stats;
|
||||
uint64_t approx_size;
|
||||
// An index that used to speed up ShouldStopBefore().
|
||||
size_t grandparent_index = 0;
|
||||
// The number of bytes overlapping between the current output and
|
||||
// grandparent files used in ShouldStopBefore().
|
||||
uint64_t overlapped_bytes = 0;
|
||||
// A flag determine whether the key has been seen in ShouldStopBefore()
|
||||
bool seen_key = false;
|
||||
std::string compression_dict;
|
||||
|
||||
SubcompactionState(Compaction* c, Slice* _start, Slice* _end,
|
||||
uint64_t size = 0)
|
||||
@@ -123,15 +110,10 @@ struct CompactionJob::SubcompactionState {
|
||||
end(_end),
|
||||
outfile(nullptr),
|
||||
builder(nullptr),
|
||||
current_output_file_size(0),
|
||||
total_bytes(0),
|
||||
num_input_records(0),
|
||||
num_output_records(0),
|
||||
approx_size(size),
|
||||
grandparent_index(0),
|
||||
overlapped_bytes(0),
|
||||
seen_key(false),
|
||||
compression_dict() {
|
||||
approx_size(size) {
|
||||
assert(compaction != nullptr);
|
||||
}
|
||||
|
||||
@@ -145,16 +127,11 @@ struct CompactionJob::SubcompactionState {
|
||||
outputs = std::move(o.outputs);
|
||||
outfile = std::move(o.outfile);
|
||||
builder = std::move(o.builder);
|
||||
current_output_file_size = std::move(o.current_output_file_size);
|
||||
total_bytes = std::move(o.total_bytes);
|
||||
num_input_records = std::move(o.num_input_records);
|
||||
num_output_records = std::move(o.num_output_records);
|
||||
compaction_job_stats = std::move(o.compaction_job_stats);
|
||||
approx_size = std::move(o.approx_size);
|
||||
grandparent_index = std::move(o.grandparent_index);
|
||||
overlapped_bytes = std::move(o.overlapped_bytes);
|
||||
seen_key = std::move(o.seen_key);
|
||||
compression_dict = std::move(o.compression_dict);
|
||||
return *this;
|
||||
}
|
||||
|
||||
@@ -162,39 +139,6 @@ struct CompactionJob::SubcompactionState {
|
||||
SubcompactionState(const SubcompactionState&) = delete;
|
||||
|
||||
SubcompactionState& operator=(const SubcompactionState&) = delete;
|
||||
|
||||
// Returns true iff we should stop building the current output
|
||||
// before processing "internal_key".
|
||||
bool ShouldStopBefore(const Slice& internal_key, uint64_t curr_file_size) {
|
||||
const InternalKeyComparator* icmp =
|
||||
&compaction->column_family_data()->internal_comparator();
|
||||
const std::vector<FileMetaData*>& grandparents = compaction->grandparents();
|
||||
|
||||
// Scan to find earliest grandparent file that contains key.
|
||||
while (grandparent_index < grandparents.size() &&
|
||||
icmp->Compare(internal_key,
|
||||
grandparents[grandparent_index]->largest.Encode()) >
|
||||
0) {
|
||||
if (seen_key) {
|
||||
overlapped_bytes += grandparents[grandparent_index]->fd.GetFileSize();
|
||||
}
|
||||
assert(grandparent_index + 1 >= grandparents.size() ||
|
||||
icmp->Compare(
|
||||
grandparents[grandparent_index]->largest.Encode(),
|
||||
grandparents[grandparent_index + 1]->smallest.Encode()) <= 0);
|
||||
grandparent_index++;
|
||||
}
|
||||
seen_key = true;
|
||||
|
||||
if (overlapped_bytes + curr_file_size >
|
||||
compaction->max_compaction_bytes()) {
|
||||
// Too much overlap for current output; start new output
|
||||
overlapped_bytes = 0;
|
||||
return true;
|
||||
}
|
||||
|
||||
return false;
|
||||
}
|
||||
};
|
||||
|
||||
// Maintains state for the entire compaction
|
||||
@@ -266,9 +210,7 @@ CompactionJob::CompactionJob(
|
||||
const EnvOptions& env_options, VersionSet* versions,
|
||||
std::atomic<bool>* shutting_down, LogBuffer* log_buffer,
|
||||
Directory* db_directory, Directory* output_directory, Statistics* stats,
|
||||
InstrumentedMutex* db_mutex, Status* db_bg_error,
|
||||
std::vector<SequenceNumber> existing_snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot,
|
||||
std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
|
||||
bool paranoid_file_checks, bool measure_io_stats, const std::string& dbname,
|
||||
CompactionJobStats* compaction_job_stats)
|
||||
@@ -286,18 +228,13 @@ CompactionJob::CompactionJob(
|
||||
db_directory_(db_directory),
|
||||
output_directory_(output_directory),
|
||||
stats_(stats),
|
||||
db_mutex_(db_mutex),
|
||||
db_bg_error_(db_bg_error),
|
||||
existing_snapshots_(std::move(existing_snapshots)),
|
||||
earliest_write_conflict_snapshot_(earliest_write_conflict_snapshot),
|
||||
table_cache_(std::move(table_cache)),
|
||||
event_logger_(event_logger),
|
||||
paranoid_file_checks_(paranoid_file_checks),
|
||||
measure_io_stats_(measure_io_stats) {
|
||||
assert(log_buffer_ != nullptr);
|
||||
const auto* cfd = compact_->compaction->column_family_data();
|
||||
ThreadStatusUtil::SetColumnFamily(cfd, cfd->ioptions()->env,
|
||||
cfd->options()->enable_thread_tracking);
|
||||
ThreadStatusUtil::SetColumnFamily(compact_->compaction->column_family_data());
|
||||
ThreadStatusUtil::SetThreadOperation(ThreadStatus::OP_COMPACTION);
|
||||
ReportStartedCompaction(compaction);
|
||||
}
|
||||
@@ -309,9 +246,8 @@ CompactionJob::~CompactionJob() {
|
||||
|
||||
void CompactionJob::ReportStartedCompaction(
|
||||
Compaction* compaction) {
|
||||
const auto* cfd = compact_->compaction->column_family_data();
|
||||
ThreadStatusUtil::SetColumnFamily(cfd, cfd->ioptions()->env,
|
||||
cfd->options()->enable_thread_tracking);
|
||||
ThreadStatusUtil::SetColumnFamily(
|
||||
compact_->compaction->column_family_data());
|
||||
|
||||
ThreadStatusUtil::SetThreadOperationProperty(
|
||||
ThreadStatus::COMPACTION_JOB_ID,
|
||||
@@ -395,6 +331,11 @@ struct RangeWithSize {
|
||||
: range(a, b), size(s) {}
|
||||
};
|
||||
|
||||
bool SliceCompare(const Comparator* cmp, const Slice& a, const Slice& b) {
|
||||
// Returns true if a < b
|
||||
return cmp->Compare(ExtractUserKey(a), ExtractUserKey(b)) < 0;
|
||||
}
|
||||
|
||||
// Generates a histogram representing potential divisions of key ranges from
|
||||
// the input. It adds the starting and/or ending keys of certain input files
|
||||
// to the working set and then finds the approximate size of data in between
|
||||
@@ -403,13 +344,14 @@ struct RangeWithSize {
|
||||
void CompactionJob::GenSubcompactionBoundaries() {
|
||||
auto* c = compact_->compaction;
|
||||
auto* cfd = c->column_family_data();
|
||||
const Comparator* cfd_comparator = cfd->user_comparator();
|
||||
std::vector<Slice> bounds;
|
||||
std::set<Slice, std::function<bool(const Slice& a, const Slice& b)> > bounds(
|
||||
std::bind(&SliceCompare, cfd->user_comparator(), std::placeholders::_1,
|
||||
std::placeholders::_2));
|
||||
int start_lvl = c->start_level();
|
||||
int out_lvl = c->output_level();
|
||||
|
||||
// Add the starting and/or ending key of certain input files as a potential
|
||||
// boundary
|
||||
// boundary (because we're inserting into a set, it avoids duplicates)
|
||||
for (size_t lvl_idx = 0; lvl_idx < c->num_input_levels(); lvl_idx++) {
|
||||
int lvl = c->level(lvl_idx);
|
||||
if (lvl >= start_lvl && lvl <= out_lvl) {
|
||||
@@ -417,44 +359,34 @@ void CompactionJob::GenSubcompactionBoundaries() {
|
||||
size_t num_files = flevel->num_files;
|
||||
|
||||
if (num_files == 0) {
|
||||
continue;
|
||||
break;
|
||||
}
|
||||
|
||||
if (lvl == 0) {
|
||||
// For level 0 add the starting and ending key of each file since the
|
||||
// files may have greatly differing key ranges (not range-partitioned)
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
bounds.emplace_back(flevel->files[i].smallest_key);
|
||||
bounds.emplace_back(flevel->files[i].largest_key);
|
||||
bounds.emplace(flevel->files[i].smallest_key);
|
||||
bounds.emplace(flevel->files[i].largest_key);
|
||||
}
|
||||
} else {
|
||||
// For all other levels add the smallest/largest key in the level to
|
||||
// encompass the range covered by that level
|
||||
bounds.emplace_back(flevel->files[0].smallest_key);
|
||||
bounds.emplace_back(flevel->files[num_files - 1].largest_key);
|
||||
bounds.emplace(flevel->files[0].smallest_key);
|
||||
bounds.emplace(flevel->files[num_files - 1].largest_key);
|
||||
if (lvl == out_lvl) {
|
||||
// For the last level include the starting keys of all files since
|
||||
// the last level is the largest and probably has the widest key
|
||||
// range. Since it's range partitioned, the ending key of one file
|
||||
// and the starting key of the next are very close (or identical).
|
||||
for (size_t i = 1; i < num_files; i++) {
|
||||
bounds.emplace_back(flevel->files[i].smallest_key);
|
||||
bounds.emplace(flevel->files[i].smallest_key);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
std::sort(bounds.begin(), bounds.end(),
|
||||
[cfd_comparator] (const Slice& a, const Slice& b) -> bool {
|
||||
return cfd_comparator->Compare(ExtractUserKey(a), ExtractUserKey(b)) < 0;
|
||||
});
|
||||
// Remove duplicated entries from bounds
|
||||
bounds.erase(std::unique(bounds.begin(), bounds.end(),
|
||||
[cfd_comparator] (const Slice& a, const Slice& b) -> bool {
|
||||
return cfd_comparator->Compare(ExtractUserKey(a), ExtractUserKey(b)) == 0;
|
||||
}), bounds.end());
|
||||
|
||||
// Combine consecutive pairs of boundaries into ranges with an approximate
|
||||
// size of data covered by keys in that range
|
||||
uint64_t sum = 0;
|
||||
@@ -476,9 +408,9 @@ void CompactionJob::GenSubcompactionBoundaries() {
|
||||
|
||||
// Group the ranges into subcompactions
|
||||
const double min_file_fill_percent = 4.0 / 5;
|
||||
uint64_t max_output_files = static_cast<uint64_t>(
|
||||
std::ceil(sum / min_file_fill_percent /
|
||||
c->mutable_cf_options()->MaxFileSizeForLevel(out_lvl)));
|
||||
uint64_t max_output_files = std::ceil(
|
||||
sum / min_file_fill_percent /
|
||||
cfd->GetCurrentMutableCFOptions()->MaxFileSizeForLevel(out_lvl));
|
||||
uint64_t subcompactions =
|
||||
std::min({static_cast<uint64_t>(ranges.size()),
|
||||
static_cast<uint64_t>(db_options_.max_subcompactions),
|
||||
@@ -491,12 +423,12 @@ void CompactionJob::GenSubcompactionBoundaries() {
|
||||
// sizes becomes >= the expected mean size of a subcompaction
|
||||
sum = 0;
|
||||
for (size_t i = 0; i < ranges.size() - 1; i++) {
|
||||
sum += ranges[i].size;
|
||||
if (subcompactions == 1) {
|
||||
// If there's only one left to schedule then it goes to the end so no
|
||||
// need to put an end boundary
|
||||
continue;
|
||||
break;
|
||||
}
|
||||
sum += ranges[i].size;
|
||||
if (sum >= mean) {
|
||||
boundaries_.emplace_back(ExtractUserKey(ranges[i].range.limit));
|
||||
sizes_.emplace_back(sum);
|
||||
@@ -555,16 +487,6 @@ Status CompactionJob::Run() {
|
||||
}
|
||||
}
|
||||
|
||||
TablePropertiesCollection tp;
|
||||
for (const auto& state : compact_->sub_compact_states) {
|
||||
for (const auto& output : state.outputs) {
|
||||
auto fn = TableFileName(db_options_.db_paths, output.meta.fd.GetNumber(),
|
||||
output.meta.fd.GetPathId());
|
||||
tp[fn] = output.table_properties;
|
||||
}
|
||||
}
|
||||
compact_->compaction->SetOutputTableProperties(std::move(tp));
|
||||
|
||||
// Finish up all book-keeping to unify the subcompaction results
|
||||
AggregateStatistics();
|
||||
UpdateCompactionStats();
|
||||
@@ -576,17 +498,18 @@ Status CompactionJob::Run() {
|
||||
return status;
|
||||
}
|
||||
|
||||
Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
|
||||
Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options,
|
||||
InstrumentedMutex* db_mutex) {
|
||||
AutoThreadOperationStageUpdater stage_updater(
|
||||
ThreadStatus::STAGE_COMPACTION_INSTALL);
|
||||
db_mutex_->AssertHeld();
|
||||
db_mutex->AssertHeld();
|
||||
Status status = compact_->status;
|
||||
ColumnFamilyData* cfd = compact_->compaction->column_family_data();
|
||||
cfd->internal_stats()->AddCompactionStats(
|
||||
compact_->compaction->output_level(), compaction_stats_);
|
||||
|
||||
if (status.ok()) {
|
||||
status = InstallCompactionResults(mutable_cf_options);
|
||||
status = InstallCompactionResults(mutable_cf_options, db_mutex);
|
||||
}
|
||||
VersionStorageInfo::LevelSummaryStorage tmp;
|
||||
auto vstorage = cfd->current()->storage_info();
|
||||
@@ -629,13 +552,6 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
|
||||
<< "num_output_records" << compact_->num_output_records
|
||||
<< "num_subcompactions" << compact_->sub_compact_states.size();
|
||||
|
||||
if (compaction_job_stats_ != nullptr) {
|
||||
stream << "num_single_delete_mismatches"
|
||||
<< compaction_job_stats_->num_single_del_mismatch;
|
||||
stream << "num_single_delete_fallthrough"
|
||||
<< compaction_job_stats_->num_single_del_fallthru;
|
||||
}
|
||||
|
||||
if (measure_io_stats_ && compaction_job_stats_ != nullptr) {
|
||||
stream << "file_write_nanos" << compaction_job_stats_->file_write_nanos;
|
||||
stream << "file_range_sync_nanos"
|
||||
@@ -658,7 +574,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
|
||||
|
||||
void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
assert(sub_compact != nullptr);
|
||||
std::unique_ptr<InternalIterator> input(
|
||||
std::unique_ptr<Iterator> input(
|
||||
versions_->MakeInputIterator(sub_compact->compaction));
|
||||
|
||||
AutoThreadOperationStageUpdater stage_updater(
|
||||
@@ -674,38 +590,13 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
if (measure_io_stats_) {
|
||||
prev_perf_level = GetPerfLevel();
|
||||
SetPerfLevel(PerfLevel::kEnableTime);
|
||||
prev_write_nanos = IOSTATS(write_nanos);
|
||||
prev_fsync_nanos = IOSTATS(fsync_nanos);
|
||||
prev_range_sync_nanos = IOSTATS(range_sync_nanos);
|
||||
prev_prepare_write_nanos = IOSTATS(prepare_write_nanos);
|
||||
prev_write_nanos = iostats_context.write_nanos;
|
||||
prev_fsync_nanos = iostats_context.fsync_nanos;
|
||||
prev_range_sync_nanos = iostats_context.range_sync_nanos;
|
||||
prev_prepare_write_nanos = iostats_context.prepare_write_nanos;
|
||||
}
|
||||
|
||||
ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
|
||||
const MutableCFOptions* mutable_cf_options =
|
||||
sub_compact->compaction->mutable_cf_options();
|
||||
|
||||
// To build compression dictionary, we sample the first output file, assuming
|
||||
// it'll reach the maximum length, and then use the dictionary for compressing
|
||||
// subsequent output files. The dictionary may be less than max_dict_bytes if
|
||||
// the first output file's length is less than the maximum.
|
||||
const int kSampleLenShift = 6; // 2^6 = 64-byte samples
|
||||
std::set<size_t> sample_begin_offsets;
|
||||
if (bottommost_level_ &&
|
||||
cfd->ioptions()->compression_opts.max_dict_bytes > 0) {
|
||||
const size_t kMaxSamples =
|
||||
cfd->ioptions()->compression_opts.max_dict_bytes >> kSampleLenShift;
|
||||
const size_t kOutFileLen = mutable_cf_options->MaxFileSizeForLevel(
|
||||
compact_->compaction->output_level());
|
||||
if (kOutFileLen != port::kMaxSizet) {
|
||||
const size_t kOutFileNumSamples = kOutFileLen >> kSampleLenShift;
|
||||
Random64 generator{versions_->NewFileNumber()};
|
||||
for (size_t i = 0; i < kMaxSamples; ++i) {
|
||||
sample_begin_offsets.insert(generator.Uniform(kOutFileNumSamples)
|
||||
<< kSampleLenShift);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
auto compaction_filter = cfd->ioptions()->compaction_filter;
|
||||
std::unique_ptr<CompactionFilter> compaction_filter_from_factory = nullptr;
|
||||
if (compaction_filter == nullptr) {
|
||||
@@ -716,7 +607,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
MergeHelper merge(
|
||||
env_, cfd->user_comparator(), cfd->ioptions()->merge_operator,
|
||||
compaction_filter, db_options_.info_log.get(),
|
||||
mutable_cf_options->min_partial_merge_operands,
|
||||
cfd->ioptions()->min_partial_merge_operands,
|
||||
false /* internal key corruption is expected */,
|
||||
existing_snapshots_.empty() ? 0 : existing_snapshots_.back(),
|
||||
compact_->compaction->level(), db_options_.statistics.get());
|
||||
@@ -736,18 +627,11 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
Status status;
|
||||
sub_compact->c_iter.reset(new CompactionIterator(
|
||||
input.get(), cfd->user_comparator(), &merge, versions_->LastSequence(),
|
||||
&existing_snapshots_, earliest_write_conflict_snapshot_, env_, false,
|
||||
sub_compact->compaction, compaction_filter));
|
||||
&existing_snapshots_, env_, false, sub_compact->compaction,
|
||||
compaction_filter));
|
||||
auto c_iter = sub_compact->c_iter.get();
|
||||
c_iter->SeekToFirst();
|
||||
const auto& c_iter_stats = c_iter->iter_stats();
|
||||
auto sample_begin_offset_iter = sample_begin_offsets.cbegin();
|
||||
// data_begin_offset and compression_dict are only valid while generating
|
||||
// dictionary from the first output file.
|
||||
size_t data_begin_offset = 0;
|
||||
std::string compression_dict;
|
||||
compression_dict.reserve(cfd->ioptions()->compression_opts.max_dict_bytes);
|
||||
|
||||
// TODO(noetzli): check whether we could check !shutting_down_->... only
|
||||
// only occasionally (see diff D42687)
|
||||
while (status.ok() && !shutting_down_->load(std::memory_order_acquire) &&
|
||||
@@ -762,9 +646,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
if (end != nullptr &&
|
||||
cfd->user_comparator()->Compare(c_iter->user_key(), *end) >= 0) {
|
||||
break;
|
||||
} else if (sub_compact->compaction->output_level() != 0 &&
|
||||
sub_compact->ShouldStopBefore(
|
||||
key, sub_compact->current_output_file_size) &&
|
||||
} else if (sub_compact->compaction->ShouldStopBefore(key) &&
|
||||
sub_compact->builder != nullptr) {
|
||||
status = FinishCompactionOutputFile(input->status(), sub_compact);
|
||||
if (!status.ok()) {
|
||||
@@ -789,74 +671,20 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
assert(sub_compact->builder != nullptr);
|
||||
assert(sub_compact->current_output() != nullptr);
|
||||
sub_compact->builder->Add(key, value);
|
||||
sub_compact->current_output_file_size = sub_compact->builder->FileSize();
|
||||
sub_compact->current_output()->meta.UpdateBoundaries(
|
||||
key, c_iter->ikey().sequence);
|
||||
sub_compact->num_output_records++;
|
||||
|
||||
if (sub_compact->outputs.size() == 1) { // first output file
|
||||
// Check if this key/value overlaps any sample intervals; if so, appends
|
||||
// overlapping portions to the dictionary.
|
||||
for (const auto& data_elmt : {key, value}) {
|
||||
size_t data_end_offset = data_begin_offset + data_elmt.size();
|
||||
while (sample_begin_offset_iter != sample_begin_offsets.cend() &&
|
||||
*sample_begin_offset_iter < data_end_offset) {
|
||||
size_t sample_end_offset =
|
||||
*sample_begin_offset_iter + (1 << kSampleLenShift);
|
||||
// Invariant: Because we advance sample iterator while processing the
|
||||
// data_elmt containing the sample's last byte, the current sample
|
||||
// cannot end before the current data_elmt.
|
||||
assert(data_begin_offset < sample_end_offset);
|
||||
|
||||
size_t data_elmt_copy_offset, data_elmt_copy_len;
|
||||
if (*sample_begin_offset_iter <= data_begin_offset) {
|
||||
// The sample starts before data_elmt starts, so take bytes starting
|
||||
// at the beginning of data_elmt.
|
||||
data_elmt_copy_offset = 0;
|
||||
} else {
|
||||
// data_elmt starts before the sample starts, so take bytes starting
|
||||
// at the below offset into data_elmt.
|
||||
data_elmt_copy_offset =
|
||||
*sample_begin_offset_iter - data_begin_offset;
|
||||
}
|
||||
if (sample_end_offset <= data_end_offset) {
|
||||
// The sample ends before data_elmt ends, so take as many bytes as
|
||||
// needed.
|
||||
data_elmt_copy_len =
|
||||
sample_end_offset - (data_begin_offset + data_elmt_copy_offset);
|
||||
} else {
|
||||
// data_elmt ends before the sample ends, so take all remaining
|
||||
// bytes in data_elmt.
|
||||
data_elmt_copy_len =
|
||||
data_end_offset - (data_begin_offset + data_elmt_copy_offset);
|
||||
}
|
||||
compression_dict.append(&data_elmt.data()[data_elmt_copy_offset],
|
||||
data_elmt_copy_len);
|
||||
if (sample_end_offset > data_end_offset) {
|
||||
// Didn't finish sample. Try to finish it with the next data_elmt.
|
||||
break;
|
||||
}
|
||||
// Next sample may require bytes from same data_elmt.
|
||||
sample_begin_offset_iter++;
|
||||
}
|
||||
data_begin_offset = data_end_offset;
|
||||
}
|
||||
}
|
||||
|
||||
// Close output file if it is big enough
|
||||
// TODO(aekmekji): determine if file should be closed earlier than this
|
||||
// during subcompactions (i.e. if output size, estimated by input size, is
|
||||
// going to be 1.2MB and max_output_file_size = 1MB, prefer to have 0.6MB
|
||||
// and 0.6MB instead of 1MB and 0.2MB)
|
||||
if (sub_compact->current_output_file_size >=
|
||||
if (sub_compact->builder->FileSize() >=
|
||||
sub_compact->compaction->max_output_file_size()) {
|
||||
status = FinishCompactionOutputFile(input->status(), sub_compact);
|
||||
if (sub_compact->outputs.size() == 1) {
|
||||
// Use dictionary from first output file for compression of subsequent
|
||||
// files.
|
||||
sub_compact->compression_dict = std::move(compression_dict);
|
||||
}
|
||||
}
|
||||
|
||||
c_iter->Next();
|
||||
}
|
||||
|
||||
@@ -865,10 +693,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
c_iter_stats.num_input_deletion_records;
|
||||
sub_compact->compaction_job_stats.num_corrupt_keys =
|
||||
c_iter_stats.num_input_corrupt_records;
|
||||
sub_compact->compaction_job_stats.num_single_del_fallthru =
|
||||
c_iter_stats.num_single_del_fallthru;
|
||||
sub_compact->compaction_job_stats.num_single_del_mismatch =
|
||||
c_iter_stats.num_single_del_mismatch;
|
||||
sub_compact->compaction_job_stats.total_input_raw_key_bytes +=
|
||||
c_iter_stats.total_input_raw_key_bytes;
|
||||
sub_compact->compaction_job_stats.total_input_raw_value_bytes +=
|
||||
@@ -893,13 +717,13 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
|
||||
|
||||
if (measure_io_stats_) {
|
||||
sub_compact->compaction_job_stats.file_write_nanos +=
|
||||
IOSTATS(write_nanos) - prev_write_nanos;
|
||||
iostats_context.write_nanos - prev_write_nanos;
|
||||
sub_compact->compaction_job_stats.file_fsync_nanos +=
|
||||
IOSTATS(fsync_nanos) - prev_fsync_nanos;
|
||||
iostats_context.fsync_nanos - prev_fsync_nanos;
|
||||
sub_compact->compaction_job_stats.file_range_sync_nanos +=
|
||||
IOSTATS(range_sync_nanos) - prev_range_sync_nanos;
|
||||
iostats_context.range_sync_nanos - prev_range_sync_nanos;
|
||||
sub_compact->compaction_job_stats.file_prepare_write_nanos +=
|
||||
IOSTATS(prepare_write_nanos) - prev_prepare_write_nanos;
|
||||
iostats_context.prepare_write_nanos - prev_prepare_write_nanos;
|
||||
if (prev_perf_level != PerfLevel::kEnableTime) {
|
||||
SetPerfLevel(prev_perf_level);
|
||||
}
|
||||
@@ -973,11 +797,10 @@ Status CompactionJob::FinishCompactionOutputFile(
|
||||
}
|
||||
sub_compact->outfile.reset();
|
||||
|
||||
ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
|
||||
TableProperties tp;
|
||||
if (s.ok() && current_entries > 0) {
|
||||
// Verify that the table is usable
|
||||
InternalIterator* iter = cfd->table_cache()->NewIterator(
|
||||
ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
|
||||
Iterator* iter = cfd->table_cache()->NewIterator(
|
||||
ReadOptions(), env_options_, cfd->internal_comparator(), meta->fd,
|
||||
nullptr, cfd->internal_stats()->GetFileReadHist(
|
||||
compact_->compaction->output_level()),
|
||||
@@ -990,52 +813,32 @@ Status CompactionJob::FinishCompactionOutputFile(
|
||||
}
|
||||
|
||||
delete iter;
|
||||
|
||||
// Output to event logger and fire events.
|
||||
if (s.ok()) {
|
||||
tp = sub_compact->builder->GetTableProperties();
|
||||
sub_compact->current_output()->table_properties =
|
||||
std::make_shared<TableProperties>(tp);
|
||||
TableFileCreationInfo info(sub_compact->builder->GetTableProperties());
|
||||
info.db_name = dbname_;
|
||||
info.cf_name = cfd->GetName();
|
||||
info.file_path =
|
||||
TableFileName(cfd->ioptions()->db_paths, meta->fd.GetNumber(),
|
||||
meta->fd.GetPathId());
|
||||
info.file_size = meta->fd.GetFileSize();
|
||||
info.job_id = job_id_;
|
||||
Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
|
||||
"[%s] [JOB %d] Generated table #%" PRIu64 ": %" PRIu64
|
||||
" keys, %" PRIu64 " bytes%s",
|
||||
cfd->GetName().c_str(), job_id_, output_number, current_entries,
|
||||
current_bytes,
|
||||
meta->marked_for_compaction ? " (need compaction)" : "");
|
||||
EventHelpers::LogAndNotifyTableFileCreation(
|
||||
event_logger_, cfd->ioptions()->listeners, meta->fd, info);
|
||||
}
|
||||
}
|
||||
std::string fname = TableFileName(db_options_.db_paths, meta->fd.GetNumber(),
|
||||
meta->fd.GetPathId());
|
||||
EventHelpers::LogAndNotifyTableFileCreationFinished(
|
||||
event_logger_, cfd->ioptions()->listeners, dbname_, cfd->GetName(), fname,
|
||||
job_id_, meta->fd, tp, TableFileCreationReason::kCompaction, s);
|
||||
|
||||
// Report new file to SstFileManagerImpl
|
||||
auto sfm =
|
||||
static_cast<SstFileManagerImpl*>(db_options_.sst_file_manager.get());
|
||||
if (sfm && meta->fd.GetPathId() == 0) {
|
||||
auto fn = TableFileName(cfd->ioptions()->db_paths, meta->fd.GetNumber(),
|
||||
meta->fd.GetPathId());
|
||||
sfm->OnAddFile(fn);
|
||||
if (sfm->IsMaxAllowedSpaceReached()) {
|
||||
InstrumentedMutexLock l(db_mutex_);
|
||||
if (db_bg_error_->ok()) {
|
||||
s = Status::IOError("Max allowed space was reached");
|
||||
*db_bg_error_ = s;
|
||||
TEST_SYNC_POINT(
|
||||
"CompactionJob::FinishCompactionOutputFile:MaxAllowedSpaceReached");
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
sub_compact->builder.reset();
|
||||
sub_compact->current_output_file_size = 0;
|
||||
return s;
|
||||
}
|
||||
|
||||
Status CompactionJob::InstallCompactionResults(
|
||||
const MutableCFOptions& mutable_cf_options) {
|
||||
db_mutex_->AssertHeld();
|
||||
const MutableCFOptions& mutable_cf_options, InstrumentedMutex* db_mutex) {
|
||||
db_mutex->AssertHeld();
|
||||
|
||||
auto* compaction = compact_->compaction;
|
||||
// paranoia: verify that the files that we started with
|
||||
@@ -1070,7 +873,7 @@ Status CompactionJob::InstallCompactionResults(
|
||||
}
|
||||
return versions_->LogAndApply(compaction->column_family_data(),
|
||||
mutable_cf_options, compaction->edit(),
|
||||
db_mutex_, db_directory_);
|
||||
db_mutex, db_directory_);
|
||||
}
|
||||
|
||||
void CompactionJob::RecordCompactionIOStats() {
|
||||
@@ -1090,18 +893,11 @@ Status CompactionJob::OpenCompactionOutputFile(
|
||||
assert(sub_compact->builder == nullptr);
|
||||
// no need to lock because VersionSet::next_file_number_ is atomic
|
||||
uint64_t file_number = versions_->NewFileNumber();
|
||||
std::string fname = TableFileName(db_options_.db_paths, file_number,
|
||||
sub_compact->compaction->output_path_id());
|
||||
// Fire events.
|
||||
ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
|
||||
#ifndef ROCKSDB_LITE
|
||||
EventHelpers::NotifyTableFileCreationStarted(
|
||||
cfd->ioptions()->listeners, dbname_, cfd->GetName(), fname, job_id_,
|
||||
TableFileCreationReason::kCompaction);
|
||||
#endif // !ROCKSDB_LITE
|
||||
// Make the output file
|
||||
unique_ptr<WritableFile> writable_file;
|
||||
Status s = NewWritableFile(env_, fname, &writable_file, env_options_);
|
||||
std::string fname = TableFileName(db_options_.db_paths, file_number,
|
||||
sub_compact->compaction->output_path_id());
|
||||
Status s = env_->NewWritableFile(fname, &writable_file, env_options_);
|
||||
if (!s.ok()) {
|
||||
Log(InfoLogLevel::ERROR_LEVEL, db_options_.info_log,
|
||||
"[%s] [JOB %d] OpenCompactionOutputFiles for table #%" PRIu64
|
||||
@@ -1109,13 +905,8 @@ Status CompactionJob::OpenCompactionOutputFile(
|
||||
sub_compact->compaction->column_family_data()->GetName().c_str(),
|
||||
job_id_, file_number, s.ToString().c_str());
|
||||
LogFlush(db_options_.info_log);
|
||||
EventHelpers::LogAndNotifyTableFileCreationFinished(
|
||||
event_logger_, cfd->ioptions()->listeners, dbname_, cfd->GetName(),
|
||||
fname, job_id_, FileDescriptor(), TableProperties(),
|
||||
TableFileCreationReason::kCompaction, s);
|
||||
return s;
|
||||
}
|
||||
|
||||
SubcompactionState::Output out;
|
||||
out.meta.fd =
|
||||
FileDescriptor(file_number, sub_compact->compaction->output_path_id(), 0);
|
||||
@@ -1128,6 +919,7 @@ Status CompactionJob::OpenCompactionOutputFile(
|
||||
sub_compact->outfile.reset(
|
||||
new WritableFileWriter(std::move(writable_file), env_options_));
|
||||
|
||||
ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
|
||||
// If the Column family flag is to only optimize filters for hits,
|
||||
// we can skip creating filters if this is the bottommost_level where
|
||||
// data is going to be found
|
||||
@@ -1135,10 +927,9 @@ Status CompactionJob::OpenCompactionOutputFile(
|
||||
cfd->ioptions()->optimize_filters_for_hits && bottommost_level_;
|
||||
sub_compact->builder.reset(NewTableBuilder(
|
||||
*cfd->ioptions(), cfd->internal_comparator(),
|
||||
cfd->int_tbl_prop_collector_factories(), cfd->GetID(), cfd->GetName(),
|
||||
sub_compact->outfile.get(), sub_compact->compaction->output_compression(),
|
||||
cfd->ioptions()->compression_opts, &sub_compact->compression_dict,
|
||||
skip_filters));
|
||||
cfd->int_tbl_prop_collector_factories(), sub_compact->outfile.get(),
|
||||
sub_compact->compaction->output_compression(),
|
||||
cfd->ioptions()->compression_opts, skip_filters));
|
||||
LogFlush(db_options_.info_log);
|
||||
return s;
|
||||
}
|
||||
|
||||
+7
-16
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -35,9 +35,9 @@
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/memtablerep.h"
|
||||
#include "rocksdb/transaction_log.h"
|
||||
#include "table/scoped_arena_iterator.h"
|
||||
#include "util/autovector.h"
|
||||
#include "util/event_logger.h"
|
||||
#include "util/scoped_arena_iterator.h"
|
||||
#include "util/stop_watch.h"
|
||||
#include "util/thread_local.h"
|
||||
|
||||
@@ -56,10 +56,8 @@ class CompactionJob {
|
||||
const EnvOptions& env_options, VersionSet* versions,
|
||||
std::atomic<bool>* shutting_down, LogBuffer* log_buffer,
|
||||
Directory* db_directory, Directory* output_directory,
|
||||
Statistics* stats, InstrumentedMutex* db_mutex,
|
||||
Status* db_bg_error,
|
||||
Statistics* stats,
|
||||
std::vector<SequenceNumber> existing_snapshots,
|
||||
SequenceNumber earliest_write_conflict_snapshot,
|
||||
std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
|
||||
bool paranoid_file_checks, bool measure_io_stats,
|
||||
const std::string& dbname,
|
||||
@@ -78,7 +76,8 @@ class CompactionJob {
|
||||
Status Run();
|
||||
|
||||
// REQUIRED: mutex held
|
||||
Status Install(const MutableCFOptions& mutable_cf_options);
|
||||
Status Install(const MutableCFOptions& mutable_cf_options,
|
||||
InstrumentedMutex* db_mutex);
|
||||
|
||||
private:
|
||||
struct SubcompactionState;
|
||||
@@ -95,7 +94,8 @@ class CompactionJob {
|
||||
|
||||
Status FinishCompactionOutputFile(const Status& input_status,
|
||||
SubcompactionState* sub_compact);
|
||||
Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options);
|
||||
Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options,
|
||||
InstrumentedMutex* db_mutex);
|
||||
void RecordCompactionIOStats();
|
||||
Status OpenCompactionOutputFile(SubcompactionState* sub_compact);
|
||||
void CleanupCompaction();
|
||||
@@ -122,7 +122,6 @@ class CompactionJob {
|
||||
const std::string& dbname_;
|
||||
const DBOptions& db_options_;
|
||||
const EnvOptions& env_options_;
|
||||
|
||||
Env* env_;
|
||||
VersionSet* versions_;
|
||||
std::atomic<bool>* shutting_down_;
|
||||
@@ -130,19 +129,11 @@ class CompactionJob {
|
||||
Directory* db_directory_;
|
||||
Directory* output_directory_;
|
||||
Statistics* stats_;
|
||||
InstrumentedMutex* db_mutex_;
|
||||
Status* db_bg_error_;
|
||||
// If there were two snapshots with seq numbers s1 and
|
||||
// s2 and s1 < s2, and if we find two instances of a key k1 then lies
|
||||
// entirely within s1 and s2, then the earlier version of k1 can be safely
|
||||
// deleted because that version is not visible in any snapshot.
|
||||
std::vector<SequenceNumber> existing_snapshots_;
|
||||
|
||||
// This is the earliest snapshot that could be used for write-conflict
|
||||
// checking by a transaction. For any user-key newer than this snapshot, we
|
||||
// should make sure not to remove evidence that a write occurred.
|
||||
SequenceNumber earliest_write_conflict_snapshot_;
|
||||
|
||||
std::shared_ptr<Cache> table_cache_;
|
||||
|
||||
EventLogger* event_logger_;
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -27,7 +27,6 @@
|
||||
#include "db/job_context.h"
|
||||
#include "db/version_set.h"
|
||||
#include "db/write_batch_internal.h"
|
||||
#include "memtable/hash_linklist_rep.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "rocksdb/cache.h"
|
||||
#include "rocksdb/compaction_filter.h"
|
||||
@@ -48,13 +47,14 @@
|
||||
#include "table/block_based_table_factory.h"
|
||||
#include "table/mock_table.h"
|
||||
#include "table/plain_table_factory.h"
|
||||
#include "table/scoped_arena_iterator.h"
|
||||
#include "util/compression.h"
|
||||
#include "util/hash.h"
|
||||
#include "util/hash_linklist_rep.h"
|
||||
#include "util/logging.h"
|
||||
#include "util/mock_env.h"
|
||||
#include "util/mutexlock.h"
|
||||
#include "util/rate_limiter.h"
|
||||
#include "util/scoped_arena_iterator.h"
|
||||
#include "util/statistics.h"
|
||||
#include "util/string_util.h"
|
||||
#include "util/sync_point.h"
|
||||
@@ -64,7 +64,7 @@
|
||||
#include "util/xfunc.h"
|
||||
#include "utilities/merge_operators.h"
|
||||
|
||||
#if !defined(IOS_CROSS_COMPILE)
|
||||
#if !defined(IOS_CROSS_COMPILE) && (!defined(NDEBUG) || !defined(OS_WIN))
|
||||
#ifndef ROCKSDB_LITE
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -551,9 +551,8 @@ uint64_t EstimatedFileSize(
|
||||
const size_t kFooterSize = 512;
|
||||
|
||||
uint64_t data_size =
|
||||
static_cast<uint64_t>(
|
||||
num_records * (key_size + value_size * compression_ratio +
|
||||
kPerKeyOverhead));
|
||||
kPerKeyOverhead);
|
||||
|
||||
return data_size + kFooterSize
|
||||
+ num_records * bloom_bits_per_key / 8 // filter block
|
||||
@@ -624,10 +623,7 @@ CompressionType GetAnyCompression() {
|
||||
return kBZip2Compression;
|
||||
} else if (LZ4_Supported()) {
|
||||
return kLZ4Compression;
|
||||
} else if (XPRESS_Supported()) {
|
||||
return kXpressCompression;
|
||||
}
|
||||
|
||||
return kNoCompression;
|
||||
}
|
||||
|
||||
@@ -663,7 +659,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
|
||||
options.max_subcompactions = max_subcompactions_;
|
||||
options.bytes_per_sync = 512 * 1024;
|
||||
|
||||
options.report_bg_io_stats = true;
|
||||
options.compaction_measure_io_stats = true;
|
||||
for (int test = 0; test < 2; ++test) {
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
@@ -957,7 +953,7 @@ TEST_P(CompactionJobStatsTest, UniversalCompactionTest) {
|
||||
uint64_t key_base = 100000000l;
|
||||
// Note: key_base must be multiple of num_keys_per_L0_file
|
||||
int num_keys_per_table = 100;
|
||||
const uint32_t kTestScale = 6;
|
||||
const uint32_t kTestScale = 8;
|
||||
const int kKeySize = 10;
|
||||
const int kValueSize = 900;
|
||||
double compression_ratio = 1.0;
|
||||
@@ -1008,9 +1004,8 @@ TEST_P(CompactionJobStatsTest, UniversalCompactionTest) {
|
||||
num_input_units,
|
||||
num_keys_per_table * num_input_units,
|
||||
1.0, 0, false));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
}
|
||||
ASSERT_EQ(stats_checker->NumberOfUnverifiedStats(), 3U);
|
||||
ASSERT_EQ(stats_checker->NumberOfUnverifiedStats(), 4U);
|
||||
|
||||
for (uint64_t start_key = key_base;
|
||||
start_key <= key_base * kTestScale;
|
||||
|
||||
+52
-284
@@ -1,22 +1,20 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
#include <algorithm>
|
||||
#include <map>
|
||||
#include <string>
|
||||
#include <tuple>
|
||||
|
||||
#include "db/column_family.h"
|
||||
#include "db/compaction_job.h"
|
||||
#include "db/column_family.h"
|
||||
#include "db/version_set.h"
|
||||
#include "db/writebuffer.h"
|
||||
#include "rocksdb/cache.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/write_buffer_manager.h"
|
||||
#include "table/mock_table.h"
|
||||
#include "util/file_reader_writer.h"
|
||||
#include "util/string_util.h"
|
||||
@@ -70,9 +68,9 @@ class CompactionJobTest : public testing::Test {
|
||||
dbname_(test::TmpDir() + "/compaction_job_test"),
|
||||
mutable_cf_options_(Options(), ImmutableCFOptions(Options())),
|
||||
table_cache_(NewLRUCache(50000, 16)),
|
||||
write_buffer_manager_(db_options_.db_write_buffer_size),
|
||||
write_buffer_(db_options_.db_write_buffer_size),
|
||||
versions_(new VersionSet(dbname_, &db_options_, env_options_,
|
||||
table_cache_.get(), &write_buffer_manager_,
|
||||
table_cache_.get(), &write_buffer_,
|
||||
&write_controller_)),
|
||||
shutting_down_(false),
|
||||
mock_table_factory_(new mock::MockTableFactory()) {
|
||||
@@ -162,12 +160,9 @@ class CompactionJobTest : public testing::Test {
|
||||
auto key = ToString(i * kMatchingKeys + k);
|
||||
auto value = ToString(i * kKeysPerFile + k);
|
||||
InternalKey internal_key(key, ++sequence_number, kTypeValue);
|
||||
|
||||
// This is how the key will look like once it's written in bottommost
|
||||
// file
|
||||
InternalKey bottommost_internal_key(
|
||||
key, (key == "9999") ? sequence_number : 0, kTypeValue);
|
||||
|
||||
InternalKey bottommost_internal_key(key, 0, kTypeValue);
|
||||
if (corrupt_id(k)) {
|
||||
test::CorruptKeyType(&internal_key);
|
||||
test::CorruptKeyType(&bottommost_internal_key);
|
||||
@@ -201,7 +196,7 @@ class CompactionJobTest : public testing::Test {
|
||||
unique_ptr<WritableFileWriter> file_writer(
|
||||
new WritableFileWriter(std::move(file), env_options_));
|
||||
{
|
||||
log::Writer log(std::move(file_writer), 0, false);
|
||||
log::Writer log(std::move(file_writer));
|
||||
std::string record;
|
||||
new_db.EncodeTo(&record);
|
||||
s = log.AddRecord(record);
|
||||
@@ -220,11 +215,9 @@ class CompactionJobTest : public testing::Test {
|
||||
cfd_ = versions_->GetColumnFamilySet()->GetDefault();
|
||||
}
|
||||
|
||||
void RunCompaction(
|
||||
const std::vector<std::vector<FileMetaData*>>& input_files,
|
||||
const stl_wrappers::KVMap& expected_results,
|
||||
const std::vector<SequenceNumber>& snapshots = {},
|
||||
SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
|
||||
void RunCompaction(const std::vector<std::vector<FileMetaData*>>& input_files,
|
||||
const stl_wrappers::KVMap& expected_results,
|
||||
const std::vector<SequenceNumber>& snapshots = {}) {
|
||||
auto cfd = versions_->GetColumnFamilySet()->GetDefault();
|
||||
|
||||
size_t num_input_files = 0;
|
||||
@@ -241,18 +234,18 @@ class CompactionJobTest : public testing::Test {
|
||||
|
||||
Compaction compaction(cfd->current()->storage_info(),
|
||||
*cfd->GetLatestMutableCFOptions(),
|
||||
compaction_input_files, 1, 1024 * 1024,
|
||||
10 * 1024 * 1024, 0, kNoCompression, {}, true);
|
||||
compaction_input_files, 1, 1024 * 1024, 10, 0,
|
||||
kNoCompression, {}, true);
|
||||
compaction.SetInputVersion(cfd->current());
|
||||
|
||||
LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, db_options_.info_log.get());
|
||||
mutex_.Lock();
|
||||
EventLogger event_logger(db_options_.info_log.get());
|
||||
CompactionJob compaction_job(
|
||||
0, &compaction, db_options_, env_options_, versions_.get(),
|
||||
&shutting_down_, &log_buffer, nullptr, nullptr, nullptr, &mutex_,
|
||||
&bg_error_, snapshots, earliest_write_conflict_snapshot, table_cache_,
|
||||
&event_logger, false, false, dbname_, &compaction_job_stats_);
|
||||
CompactionJob compaction_job(0, &compaction, db_options_, env_options_,
|
||||
versions_.get(), &shutting_down_, &log_buffer,
|
||||
nullptr, nullptr, nullptr, snapshots,
|
||||
table_cache_, &event_logger, false, false,
|
||||
dbname_, &compaction_job_stats_);
|
||||
|
||||
VerifyInitializationOfCompactionJobStats(compaction_job_stats_);
|
||||
|
||||
@@ -262,7 +255,8 @@ class CompactionJobTest : public testing::Test {
|
||||
s = compaction_job.Run();
|
||||
ASSERT_OK(s);
|
||||
mutex_.Lock();
|
||||
ASSERT_OK(compaction_job.Install(*cfd->GetLatestMutableCFOptions()));
|
||||
ASSERT_OK(compaction_job.Install(*cfd->GetLatestMutableCFOptions(),
|
||||
&mutex_));
|
||||
mutex_.Unlock();
|
||||
|
||||
if (expected_results.size() == 0) {
|
||||
@@ -285,7 +279,7 @@ class CompactionJobTest : public testing::Test {
|
||||
WriteController write_controller_;
|
||||
DBOptions db_options_;
|
||||
ColumnFamilyOptions cf_options_;
|
||||
WriteBufferManager write_buffer_manager_;
|
||||
WriteBuffer write_buffer_;
|
||||
std::unique_ptr<VersionSet> versions_;
|
||||
InstrumentedMutex mutex_;
|
||||
std::atomic<bool> shutting_down_;
|
||||
@@ -294,7 +288,6 @@ class CompactionJobTest : public testing::Test {
|
||||
ColumnFamilyData* cfd_;
|
||||
std::unique_ptr<CompactionFilter> compaction_filter_;
|
||||
std::shared_ptr<MergeOperator> merge_op_;
|
||||
Status bg_error_;
|
||||
};
|
||||
|
||||
TEST_F(CompactionJobTest, Simple) {
|
||||
@@ -351,7 +344,7 @@ TEST_F(CompactionJobTest, SimpleOverwrite) {
|
||||
|
||||
auto expected_results =
|
||||
mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "val2"},
|
||||
{KeyStr("b", 4U, kTypeValue), "val3"}});
|
||||
{KeyStr("b", 0U, kTypeValue), "val3"}});
|
||||
|
||||
SetLastSequence(4U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
@@ -404,7 +397,7 @@ TEST_F(CompactionJobTest, SimpleMerge) {
|
||||
|
||||
auto expected_results =
|
||||
mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "3,4,5"},
|
||||
{KeyStr("b", 2U, kTypeValue), "1,2"}});
|
||||
{KeyStr("b", 0U, kTypeValue), "1,2"}});
|
||||
|
||||
SetLastSequence(5U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
@@ -564,144 +557,34 @@ TEST_F(CompactionJobTest, SimpleSingleDelete) {
|
||||
TEST_F(CompactionJobTest, SingleDeleteSnapshots) {
|
||||
NewDB();
|
||||
|
||||
auto file1 = mock::MakeMockFile({
|
||||
{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("b", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("c", 22U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("d", 9U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("f", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("j", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("j", 9U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("k", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("k", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("l", 3U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("l", 2U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
auto file1 = mock::MakeMockFile({{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("b", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("c", 22U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("d", 9U, kTypeSingleDeletion), ""}});
|
||||
AddMockFile(file1);
|
||||
|
||||
auto file2 = mock::MakeMockFile({
|
||||
{KeyStr("0", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 11U, kTypeValue), "val1"},
|
||||
{KeyStr("b", 11U, kTypeValue), "val2"},
|
||||
{KeyStr("c", 21U, kTypeValue), "val3"},
|
||||
{KeyStr("d", 8U, kTypeValue), "val4"},
|
||||
{KeyStr("e", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("f", 1U, kTypeValue), "val1"},
|
||||
{KeyStr("g", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("h", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("m", 12U, kTypeValue), "val1"},
|
||||
{KeyStr("m", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("m", 8U, kTypeValue), "val2"},
|
||||
});
|
||||
auto file2 = mock::MakeMockFile({{KeyStr("0", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 11U, kTypeValue), "val1"},
|
||||
{KeyStr("b", 11U, kTypeValue), "val2"},
|
||||
{KeyStr("c", 21U, kTypeValue), "val3"},
|
||||
{KeyStr("d", 8U, kTypeValue), "val4"},
|
||||
{KeyStr("e", 2U, kTypeSingleDeletion), ""}});
|
||||
AddMockFile(file2);
|
||||
|
||||
auto file3 = mock::MakeMockFile({
|
||||
{KeyStr("A", 1U, kTypeValue), "val"},
|
||||
{KeyStr("e", 1U, kTypeValue), "val"},
|
||||
});
|
||||
auto file3 = mock::MakeMockFile({{KeyStr("A", 1U, kTypeValue), "val"},
|
||||
{KeyStr("e", 1U, kTypeValue), "val"}});
|
||||
AddMockFile(file3, 2);
|
||||
|
||||
auto expected_results = mock::MakeMockFile({
|
||||
{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("a", 11U, kTypeValue), ""},
|
||||
{KeyStr("b", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("b", 11U, kTypeValue), "val2"},
|
||||
{KeyStr("c", 22U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("c", 21U, kTypeValue), ""},
|
||||
{KeyStr("e", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("f", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("f", 1U, kTypeValue), "val1"},
|
||||
{KeyStr("g", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("j", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("k", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("m", 12U, kTypeValue), "val1"},
|
||||
{KeyStr("m", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("m", 8U, kTypeValue), "val2"},
|
||||
});
|
||||
auto expected_results =
|
||||
mock::MakeMockFile({{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("b", 21U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("b", 11U, kTypeValue), "val2"},
|
||||
{KeyStr("e", 2U, kTypeSingleDeletion), ""}});
|
||||
|
||||
SetLastSequence(22U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
RunCompaction({files}, expected_results, {10U, 20U}, 10U);
|
||||
}
|
||||
|
||||
TEST_F(CompactionJobTest, EarliestWriteConflictSnapshot) {
|
||||
NewDB();
|
||||
|
||||
// Test multiple snapshots where the earliest snapshot is not a
|
||||
// write-conflic-snapshot.
|
||||
|
||||
auto file1 = mock::MakeMockFile({
|
||||
{KeyStr("A", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 23U, kTypeValue), "val"},
|
||||
{KeyStr("B", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("B", 23U, kTypeValue), "val"},
|
||||
{KeyStr("D", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 32U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 31U, kTypeValue), "val"},
|
||||
{KeyStr("G", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 23U, kTypeValue), "val2"},
|
||||
{KeyStr("H", 31U, kTypeValue), "val"},
|
||||
{KeyStr("H", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("H", 23U, kTypeValue), "val"},
|
||||
{KeyStr("I", 35U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 34U, kTypeValue), "val2"},
|
||||
{KeyStr("I", 33U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 32U, kTypeValue), "val3"},
|
||||
{KeyStr("I", 31U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 34U, kTypeValue), "val"},
|
||||
{KeyStr("J", 33U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 25U, kTypeValue), "val2"},
|
||||
{KeyStr("J", 24U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
AddMockFile(file1);
|
||||
|
||||
auto file2 = mock::MakeMockFile({
|
||||
{KeyStr("A", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 13U, kTypeValue), "val2"},
|
||||
{KeyStr("C", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("C", 13U, kTypeValue), "val"},
|
||||
{KeyStr("E", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("F", 4U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("F", 3U, kTypeValue), "val"},
|
||||
{KeyStr("G", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 13U, kTypeValue), "val3"},
|
||||
{KeyStr("H", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("H", 13U, kTypeValue), "val2"},
|
||||
{KeyStr("I", 13U, kTypeValue), "val4"},
|
||||
{KeyStr("I", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 11U, kTypeValue), "val5"},
|
||||
{KeyStr("J", 15U, kTypeValue), "val3"},
|
||||
{KeyStr("J", 14U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
AddMockFile(file2);
|
||||
|
||||
auto expected_results = mock::MakeMockFile({
|
||||
{KeyStr("A", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 23U, kTypeValue), ""},
|
||||
{KeyStr("B", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("B", 23U, kTypeValue), ""},
|
||||
{KeyStr("D", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 32U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 31U, kTypeValue), ""},
|
||||
{KeyStr("H", 31U, kTypeValue), "val"},
|
||||
{KeyStr("I", 35U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 34U, kTypeValue), ""},
|
||||
{KeyStr("I", 31U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 13U, kTypeValue), "val4"},
|
||||
{KeyStr("J", 34U, kTypeValue), "val"},
|
||||
{KeyStr("J", 33U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 25U, kTypeValue), "val2"},
|
||||
{KeyStr("J", 24U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 15U, kTypeValue), "val3"},
|
||||
{KeyStr("J", 14U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
|
||||
SetLastSequence(24U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
RunCompaction({files}, expected_results, {10U, 20U, 30U}, 20U);
|
||||
RunCompaction({files}, expected_results, {10U, 20U});
|
||||
}
|
||||
|
||||
TEST_F(CompactionJobTest, SingleDeleteZeroSeq) {
|
||||
@@ -719,7 +602,7 @@ TEST_F(CompactionJobTest, SingleDeleteZeroSeq) {
|
||||
AddMockFile(file2);
|
||||
|
||||
auto expected_results = mock::MakeMockFile({
|
||||
{KeyStr("dummy", 5U, kTypeValue), "val2"},
|
||||
{KeyStr("dummy", 0U, kTypeValue), "val2"},
|
||||
});
|
||||
|
||||
SetLastSequence(22U);
|
||||
@@ -731,22 +614,9 @@ TEST_F(CompactionJobTest, MultiSingleDelete) {
|
||||
// Tests three scenarios involving multiple single delete/put pairs:
|
||||
//
|
||||
// A: Put Snapshot SDel Put SDel -> Put Snapshot SDel
|
||||
// B: Snapshot Put SDel Put SDel Snapshot -> Snapshot SDel Snapshot
|
||||
// B: Put SDel Put SDel -> (Removed)
|
||||
// C: SDel Put SDel Snapshot Put -> Snapshot Put
|
||||
// D: (Put) SDel Snapshot Put SDel -> (Put) SDel Snapshot SDel
|
||||
// E: Put SDel Snapshot Put SDel -> Snapshot SDel
|
||||
// F: Put SDel Put Sdel Snapshot -> removed
|
||||
// G: Snapshot SDel Put SDel Put -> Snapshot Put SDel
|
||||
// H: (Put) Put SDel Put Sdel Snapshot -> Removed
|
||||
// I: (Put) Snapshot Put SDel Put SDel -> SDel
|
||||
// J: Put Put SDel Put SDel SDel Snapshot Put Put SDel SDel Put
|
||||
// -> Snapshot Put
|
||||
// K: SDel SDel Put SDel Put Put Snapshot SDel Put SDel SDel Put SDel
|
||||
// -> Snapshot Put Snapshot SDel
|
||||
// L: SDel Put Del Put SDel Snapshot Del Put Del SDel Put SDel
|
||||
// -> Snapshot SDel
|
||||
// M: (Put) SDel Put Del Put SDel Snapshot Put Del SDel Put SDel Del
|
||||
// -> SDel Snapshot Del
|
||||
// D: (Put) SDel Snapshot Put SDel -> (Put) SDel Snapshot
|
||||
NewDB();
|
||||
|
||||
auto file1 = mock::MakeMockFile({
|
||||
@@ -758,34 +628,6 @@ TEST_F(CompactionJobTest, MultiSingleDelete) {
|
||||
{KeyStr("C", 14U, kTypeValue), "val3"},
|
||||
{KeyStr("D", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("D", 11U, kTypeValue), "val4"},
|
||||
{KeyStr("G", 15U, kTypeValue), "val"},
|
||||
{KeyStr("G", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("G", 13U, kTypeValue), "val"},
|
||||
{KeyStr("I", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 13U, kTypeValue), "val"},
|
||||
{KeyStr("J", 15U, kTypeValue), "val"},
|
||||
{KeyStr("J", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 13U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 12U, kTypeValue), "val"},
|
||||
{KeyStr("J", 11U, kTypeValue), "val"},
|
||||
{KeyStr("K", 16U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 15U, kTypeValue), "val1"},
|
||||
{KeyStr("K", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 13U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 12U, kTypeValue), "val2"},
|
||||
{KeyStr("K", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 16U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 15U, kTypeValue), "val"},
|
||||
{KeyStr("L", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 13U, kTypeDeletion), ""},
|
||||
{KeyStr("L", 12U, kTypeValue), "val"},
|
||||
{KeyStr("L", 11U, kTypeDeletion), ""},
|
||||
{KeyStr("M", 16U, kTypeDeletion), ""},
|
||||
{KeyStr("M", 15U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("M", 14U, kTypeValue), "val"},
|
||||
{KeyStr("M", 13U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("M", 12U, kTypeDeletion), ""},
|
||||
{KeyStr("M", 11U, kTypeValue), "val"},
|
||||
});
|
||||
AddMockFile(file1);
|
||||
|
||||
@@ -797,87 +639,24 @@ TEST_F(CompactionJobTest, MultiSingleDelete) {
|
||||
{KeyStr("C", 9U, kTypeValue), "val6"},
|
||||
{KeyStr("C", 8U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("D", 10U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 11U, kTypeValue), "val"},
|
||||
{KeyStr("E", 5U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 4U, kTypeValue), "val"},
|
||||
{KeyStr("F", 6U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("F", 5U, kTypeValue), "val"},
|
||||
{KeyStr("F", 4U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("F", 3U, kTypeValue), "val"},
|
||||
{KeyStr("G", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("H", 6U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("H", 5U, kTypeValue), "val"},
|
||||
{KeyStr("H", 4U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("H", 3U, kTypeValue), "val"},
|
||||
{KeyStr("I", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 11U, kTypeValue), "val"},
|
||||
{KeyStr("J", 6U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 5U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 4U, kTypeValue), "val"},
|
||||
{KeyStr("J", 3U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("J", 2U, kTypeValue), "val"},
|
||||
{KeyStr("K", 8U, kTypeValue), "val3"},
|
||||
{KeyStr("K", 7U, kTypeValue), "val4"},
|
||||
{KeyStr("K", 6U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 5U, kTypeValue), "val5"},
|
||||
{KeyStr("K", 2U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 1U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 5U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 4U, kTypeValue), "val"},
|
||||
{KeyStr("L", 3U, kTypeDeletion), ""},
|
||||
{KeyStr("L", 2U, kTypeValue), "val"},
|
||||
{KeyStr("L", 1U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("M", 10U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("M", 7U, kTypeValue), "val"},
|
||||
{KeyStr("M", 5U, kTypeDeletion), ""},
|
||||
{KeyStr("M", 4U, kTypeValue), "val"},
|
||||
{KeyStr("M", 3U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
AddMockFile(file2);
|
||||
|
||||
auto file3 = mock::MakeMockFile({
|
||||
{KeyStr("D", 1U, kTypeValue), "val"},
|
||||
{KeyStr("H", 1U, kTypeValue), "val"},
|
||||
{KeyStr("I", 2U, kTypeValue), "val"},
|
||||
{KeyStr("D", 11U, kTypeValue), "val"},
|
||||
});
|
||||
AddMockFile(file3, 2);
|
||||
|
||||
auto file4 = mock::MakeMockFile({
|
||||
{KeyStr("M", 1U, kTypeValue), "val"},
|
||||
auto expected_results = mock::MakeMockFile({
|
||||
{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 10U, kTypeValue), "val"},
|
||||
{KeyStr("C", 14U, kTypeValue), "val3"},
|
||||
{KeyStr("D", 10U, kTypeSingleDeletion), ""},
|
||||
});
|
||||
AddMockFile(file4, 2);
|
||||
|
||||
auto expected_results =
|
||||
mock::MakeMockFile({{KeyStr("A", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 13U, kTypeValue), ""},
|
||||
{KeyStr("A", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("A", 10U, kTypeValue), "val"},
|
||||
{KeyStr("B", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("B", 13U, kTypeValue), ""},
|
||||
{KeyStr("C", 14U, kTypeValue), "val3"},
|
||||
{KeyStr("D", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("D", 11U, kTypeValue), ""},
|
||||
{KeyStr("D", 10U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("E", 11U, kTypeValue), ""},
|
||||
{KeyStr("G", 15U, kTypeValue), "val"},
|
||||
{KeyStr("G", 12U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 14U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("I", 13U, kTypeValue), ""},
|
||||
{KeyStr("J", 15U, kTypeValue), "val"},
|
||||
{KeyStr("K", 16U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 15U, kTypeValue), ""},
|
||||
{KeyStr("K", 11U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("K", 8U, kTypeValue), "val3"},
|
||||
{KeyStr("L", 16U, kTypeSingleDeletion), ""},
|
||||
{KeyStr("L", 15U, kTypeValue), ""},
|
||||
{KeyStr("M", 16U, kTypeDeletion), ""},
|
||||
{KeyStr("M", 3U, kTypeSingleDeletion), ""}});
|
||||
|
||||
SetLastSequence(22U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
RunCompaction({files}, expected_results, {10U}, 10U);
|
||||
RunCompaction({files}, expected_results, {10U});
|
||||
}
|
||||
|
||||
// This test documents the behavior where a corrupt key follows a deletion or a
|
||||
@@ -903,7 +682,7 @@ TEST_F(CompactionJobTest, CorruptionAfterDeletion) {
|
||||
mock::MakeMockFile({{test::KeyStr("A", 0U, kTypeValue), "val3"},
|
||||
{test::KeyStr("a", 0U, kTypeValue, true), "val"},
|
||||
{test::KeyStr("b", 0U, kTypeValue, true), "val"},
|
||||
{test::KeyStr("c", 1U, kTypeValue), "val2"}});
|
||||
{test::KeyStr("c", 0U, kTypeValue), "val2"}});
|
||||
|
||||
SetLastSequence(6U);
|
||||
auto files = cfd_->current()->storage_info()->LevelFiles(0);
|
||||
@@ -916,14 +695,3 @@ int main(int argc, char** argv) {
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
|
||||
#else
|
||||
#include <stdio.h>
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
fprintf(stderr,
|
||||
"SKIPPED as CompactionJobStats is not supported in ROCKSDB_LITE\n");
|
||||
return 0;
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
+144
-251
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -70,8 +70,7 @@ struct UserKeyComparator {
|
||||
};
|
||||
|
||||
typedef std::priority_queue<InputFileInfo, std::vector<InputFileInfo>,
|
||||
UserKeyComparator>
|
||||
SmallestKeyHeap;
|
||||
UserKeyComparator> SmallestKeyHeap;
|
||||
|
||||
// This function creates the heap that is used to find if the files are
|
||||
// overlapping during universal compaction when the allow_trivial_move
|
||||
@@ -110,22 +109,13 @@ SmallestKeyHeap create_level_heap(Compaction* c, const Comparator* ucmp) {
|
||||
// matter what the values of the other two parameters are.
|
||||
// Otherwise, the compression type is determined based on options and level.
|
||||
CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
|
||||
const VersionStorageInfo* vstorage,
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
int level, int base_level,
|
||||
const bool enable_compression) {
|
||||
if (!enable_compression) {
|
||||
// disable compression
|
||||
return kNoCompression;
|
||||
}
|
||||
|
||||
// If bottommost_compression is set and we are compacting to the
|
||||
// bottommost level then we should use it.
|
||||
if (ioptions.bottommost_compression != kDisableCompressionOption &&
|
||||
level > base_level && level >= (vstorage->num_non_empty_levels() - 1)) {
|
||||
return ioptions.bottommost_compression;
|
||||
}
|
||||
// If the user has specified a different compression level for each level,
|
||||
// If the use has specified a different compression level for each level,
|
||||
// then pick the compression for that level.
|
||||
if (!ioptions.compression_per_level.empty()) {
|
||||
assert(level == 0 || level >= base_level);
|
||||
@@ -139,7 +129,7 @@ CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
|
||||
// specified compression levels, use the last value.
|
||||
return ioptions.compression_per_level[std::max(0, std::min(idx, n))];
|
||||
} else {
|
||||
return mutable_cf_options.compression;
|
||||
return ioptions.compression;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -151,8 +141,7 @@ CompactionPicker::~CompactionPicker() {}
|
||||
|
||||
// Delete this compaction from the list of running compactions.
|
||||
void CompactionPicker::ReleaseCompactionFiles(Compaction* c, Status status) {
|
||||
if (c->start_level() == 0 ||
|
||||
ioptions_.compaction_style == kCompactionStyleUniversal) {
|
||||
if (c->start_level() == 0) {
|
||||
level0_compactions_in_progress_.erase(c);
|
||||
}
|
||||
if (!status.ok()) {
|
||||
@@ -200,9 +189,10 @@ void CompactionPicker::GetRange(const CompactionInputFiles& inputs1,
|
||||
InternalKey smallest1, smallest2, largest1, largest2;
|
||||
GetRange(inputs1, &smallest1, &largest1);
|
||||
GetRange(inputs2, &smallest2, &largest2);
|
||||
*smallest =
|
||||
icmp_->Compare(smallest1, smallest2) < 0 ? smallest1 : smallest2;
|
||||
*largest = icmp_->Compare(largest1, largest2) < 0 ? largest2 : largest1;
|
||||
*smallest = icmp_->Compare(smallest1, smallest2) < 0 ?
|
||||
smallest1 : smallest2;
|
||||
*largest = icmp_->Compare(largest1, largest2) < 0 ?
|
||||
largest2 : largest1;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -253,7 +243,7 @@ bool CompactionPicker::ExpandWhileOverlapping(const std::string& cf_name,
|
||||
// Returns true if any one of specified files are being compacted
|
||||
bool CompactionPicker::FilesInCompaction(
|
||||
const std::vector<FileMetaData*>& files) {
|
||||
for (size_t i = 0; i < files.size(); i++) {
|
||||
for (unsigned int i = 0; i < files.size(); i++) {
|
||||
if (files[i]->being_compacted) {
|
||||
return true;
|
||||
}
|
||||
@@ -265,32 +255,22 @@ Compaction* CompactionPicker::FormCompaction(
|
||||
const CompactionOptions& compact_options,
|
||||
const std::vector<CompactionInputFiles>& input_files, int output_level,
|
||||
VersionStorageInfo* vstorage, const MutableCFOptions& mutable_cf_options,
|
||||
uint32_t output_path_id) {
|
||||
uint32_t output_path_id) const {
|
||||
uint64_t max_grandparent_overlap_bytes =
|
||||
output_level + 1 < vstorage->num_levels() ?
|
||||
mutable_cf_options.MaxGrandParentOverlapBytes(output_level + 1) :
|
||||
std::numeric_limits<uint64_t>::max();
|
||||
assert(input_files.size());
|
||||
|
||||
// TODO(rven ): we might be able to run concurrent level 0 compaction
|
||||
// if the key ranges of the two compactions do not overlap, but for now
|
||||
// we do not allow it.
|
||||
if ((input_files[0].level == 0) && !level0_compactions_in_progress_.empty()) {
|
||||
return nullptr;
|
||||
}
|
||||
auto c =
|
||||
new Compaction(vstorage, mutable_cf_options, input_files, output_level,
|
||||
compact_options.output_file_size_limit,
|
||||
mutable_cf_options.max_compaction_bytes, output_path_id,
|
||||
compact_options.compression, /* grandparents */ {}, true);
|
||||
|
||||
// If it's level 0 compaction, make sure we don't execute any other level 0
|
||||
// compactions in parallel
|
||||
if ((c != nullptr) && (input_files[0].level == 0)) {
|
||||
level0_compactions_in_progress_.insert(c);
|
||||
}
|
||||
return c;
|
||||
return new Compaction(
|
||||
vstorage, mutable_cf_options, input_files, output_level,
|
||||
compact_options.output_file_size_limit, max_grandparent_overlap_bytes,
|
||||
output_path_id, compact_options.compression, /* grandparents */ {}, true);
|
||||
}
|
||||
|
||||
Status CompactionPicker::GetCompactionInputsFromFileNumbers(
|
||||
std::vector<CompactionInputFiles>* input_files,
|
||||
std::unordered_set<uint64_t>* input_set, const VersionStorageInfo* vstorage,
|
||||
std::unordered_set<uint64_t>* input_set,
|
||||
const VersionStorageInfo* vstorage,
|
||||
const CompactionOptions& compact_options) const {
|
||||
if (input_set->size() == 0U) {
|
||||
return Status::InvalidArgument(
|
||||
@@ -328,8 +308,8 @@ Status CompactionPicker::GetCompactionInputsFromFileNumbers(
|
||||
return Status::InvalidArgument(message);
|
||||
}
|
||||
|
||||
for (int level = first_non_empty_level; level <= last_non_empty_level;
|
||||
++level) {
|
||||
for (int level = first_non_empty_level;
|
||||
level <= last_non_empty_level; ++level) {
|
||||
matched_input_files[level].level = level;
|
||||
input_files->emplace_back(std::move(matched_input_files[level]));
|
||||
}
|
||||
@@ -337,6 +317,8 @@ Status CompactionPicker::GetCompactionInputsFromFileNumbers(
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
|
||||
|
||||
// Returns true if any one of the parent files are being compacted
|
||||
bool CompactionPicker::RangeInCompaction(VersionStorageInfo* vstorage,
|
||||
const InternalKey* smallest,
|
||||
@@ -387,9 +369,6 @@ bool CompactionPicker::SetupOtherInputs(
|
||||
vstorage->GetOverlappingInputs(output_level, &smallest, &largest,
|
||||
&output_level_inputs->files, *parent_index,
|
||||
parent_index);
|
||||
if (!output_level_inputs->empty()) {
|
||||
ExpandWhileOverlapping(cf_name, vstorage, output_level_inputs);
|
||||
}
|
||||
|
||||
if (FilesInCompaction(output_level_inputs->files)) {
|
||||
return false;
|
||||
@@ -413,7 +392,8 @@ bool CompactionPicker::SetupOtherInputs(
|
||||
const uint64_t inputs1_size =
|
||||
TotalCompensatedFileSize(output_level_inputs->files);
|
||||
const uint64_t expanded0_size = TotalCompensatedFileSize(expanded0.files);
|
||||
uint64_t limit = mutable_cf_options.max_compaction_bytes;
|
||||
uint64_t limit =
|
||||
mutable_cf_options.ExpandedCompactionByteSizeLimit(input_level);
|
||||
if (expanded0.size() > inputs->size() &&
|
||||
inputs1_size + expanded0_size < limit &&
|
||||
!FilesInCompaction(expanded0.files) &&
|
||||
@@ -461,7 +441,7 @@ Compaction* CompactionPicker::CompactRange(
|
||||
const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end, bool* manual_conflict) {
|
||||
InternalKey** compaction_end) {
|
||||
// CompactionPickerFIFO has its own implementation of compact range
|
||||
assert(ioptions_.compaction_style != kCompactionStyleFIFO);
|
||||
|
||||
@@ -487,12 +467,6 @@ Compaction* CompactionPicker::CompactRange(
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
if ((start_level == 0) && (!level0_compactions_in_progress_.empty())) {
|
||||
*manual_conflict = true;
|
||||
// Only one level 0 compaction allowed
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
std::vector<CompactionInputFiles> inputs(vstorage->num_levels() -
|
||||
start_level);
|
||||
for (int level = start_level; level < vstorage->num_levels(); level++) {
|
||||
@@ -501,22 +475,13 @@ Compaction* CompactionPicker::CompactRange(
|
||||
for (FileMetaData* f : vstorage->LevelFiles(level)) {
|
||||
files.push_back(f);
|
||||
}
|
||||
if (FilesInCompaction(files)) {
|
||||
*manual_conflict = true;
|
||||
return nullptr;
|
||||
}
|
||||
}
|
||||
Compaction* c = new Compaction(
|
||||
return new Compaction(
|
||||
vstorage, mutable_cf_options, std::move(inputs), output_level,
|
||||
mutable_cf_options.MaxFileSizeForLevel(output_level),
|
||||
/* max_compaction_bytes */ LLONG_MAX, output_path_id,
|
||||
GetCompressionType(ioptions_, vstorage, mutable_cf_options,
|
||||
output_level, 1),
|
||||
/* max_grandparent_overlap_bytes */ LLONG_MAX, output_path_id,
|
||||
GetCompressionType(ioptions_, output_level, 1),
|
||||
/* grandparents */ {}, /* is manual */ true);
|
||||
if (start_level == 0) {
|
||||
level0_compactions_in_progress_.insert(c);
|
||||
}
|
||||
return c;
|
||||
}
|
||||
|
||||
CompactionInputFiles inputs;
|
||||
@@ -535,19 +500,13 @@ Compaction* CompactionPicker::CompactRange(
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
if ((input_level == 0) && (!level0_compactions_in_progress_.empty())) {
|
||||
// Only one level 0 compaction allowed
|
||||
TEST_SYNC_POINT("CompactionPicker::CompactRange:Conflict");
|
||||
*manual_conflict = true;
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
// Avoid compacting too much in one shot in case the range is large.
|
||||
// But we cannot do this for level-0 since level-0 files can overlap
|
||||
// and we must not pick one file and drop another older file if the
|
||||
// two files overlap.
|
||||
if (input_level > 0) {
|
||||
const uint64_t limit = mutable_cf_options.max_compaction_bytes;
|
||||
const uint64_t limit = mutable_cf_options.MaxFileSizeForLevel(input_level) *
|
||||
mutable_cf_options.source_compaction_factor;
|
||||
uint64_t total = 0;
|
||||
for (size_t i = 0; i + 1 < inputs.size(); ++i) {
|
||||
uint64_t s = inputs[i]->compensated_file_size;
|
||||
@@ -563,10 +522,9 @@ Compaction* CompactionPicker::CompactRange(
|
||||
assert(output_path_id < static_cast<uint32_t>(ioptions_.db_paths.size()));
|
||||
|
||||
if (ExpandWhileOverlapping(cf_name, vstorage, &inputs) == false) {
|
||||
// manual compaction is now multi-threaded, so it can
|
||||
// manual compaction is currently single-threaded, so it should never
|
||||
// happen that ExpandWhileOverlapping fails
|
||||
// we handle it higher in RunManualCompaction
|
||||
*manual_conflict = true;
|
||||
assert(false);
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
@@ -585,10 +543,9 @@ Compaction* CompactionPicker::CompactRange(
|
||||
int parent_index = -1;
|
||||
if (!SetupOtherInputs(cf_name, mutable_cf_options, vstorage, &inputs,
|
||||
&output_level_inputs, &parent_index, -1)) {
|
||||
// manual compaction is now multi-threaded, so it can
|
||||
// manual compaction is currently single-threaded, so it should never
|
||||
// happen that SetupOtherInputs fails
|
||||
// we handle it higher in RunManualCompaction
|
||||
*manual_conflict = true;
|
||||
assert(false);
|
||||
return nullptr;
|
||||
}
|
||||
}
|
||||
@@ -597,42 +554,27 @@ Compaction* CompactionPicker::CompactRange(
|
||||
if (!output_level_inputs.empty()) {
|
||||
compaction_inputs.push_back(output_level_inputs);
|
||||
}
|
||||
for (size_t i = 0; i < compaction_inputs.size(); i++) {
|
||||
if (FilesInCompaction(compaction_inputs[i].files)) {
|
||||
*manual_conflict = true;
|
||||
return nullptr;
|
||||
}
|
||||
}
|
||||
|
||||
std::vector<FileMetaData*> grandparents;
|
||||
GetGrandparents(vstorage, inputs, output_level_inputs, &grandparents);
|
||||
Compaction* compaction = new Compaction(
|
||||
vstorage, mutable_cf_options, std::move(compaction_inputs), output_level,
|
||||
mutable_cf_options.MaxFileSizeForLevel(output_level),
|
||||
mutable_cf_options.max_compaction_bytes, output_path_id,
|
||||
GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
|
||||
vstorage->base_level()),
|
||||
mutable_cf_options.MaxGrandParentOverlapBytes(input_level),
|
||||
output_path_id,
|
||||
GetCompressionType(ioptions_, output_level, vstorage->base_level()),
|
||||
std::move(grandparents), /* is manual compaction */ true);
|
||||
|
||||
TEST_SYNC_POINT_CALLBACK("CompactionPicker::CompactRange:Return", compaction);
|
||||
if (input_level == 0) {
|
||||
level0_compactions_in_progress_.insert(compaction);
|
||||
}
|
||||
|
||||
// Creating a compaction influences the compaction score because the score
|
||||
// takes running compactions into account (by skipping files that are already
|
||||
// being compacted). Since we just changed compaction score, we recalculate it
|
||||
// here
|
||||
vstorage->ComputeCompactionScore(mutable_cf_options);
|
||||
|
||||
return compaction;
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
namespace {
|
||||
// Test whether two files have overlapping key-ranges.
|
||||
bool HaveOverlappingKeyRanges(const Comparator* c, const SstFileMetaData& a,
|
||||
const SstFileMetaData& b) {
|
||||
bool HaveOverlappingKeyRanges(
|
||||
const Comparator* c,
|
||||
const SstFileMetaData& a, const SstFileMetaData& b) {
|
||||
if (c->Compare(a.smallestkey, b.smallestkey) >= 0) {
|
||||
if (c->Compare(a.smallestkey, b.largestkey) <= 0) {
|
||||
// b.smallestkey <= a.smallestkey <= b.largestkey
|
||||
@@ -656,8 +598,9 @@ bool HaveOverlappingKeyRanges(const Comparator* c, const SstFileMetaData& a,
|
||||
} // namespace
|
||||
|
||||
Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
|
||||
std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta, const int output_level) const {
|
||||
std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta,
|
||||
const int output_level) const {
|
||||
auto& levels = cf_meta.levels;
|
||||
auto comparator = icmp_->user_comparator();
|
||||
|
||||
@@ -710,17 +653,18 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
|
||||
// has overlapping key-range with other non-compaction input
|
||||
// files in the same level.
|
||||
while (first_included > 0) {
|
||||
if (comparator->Compare(current_files[first_included - 1].largestkey,
|
||||
current_files[first_included].smallestkey) <
|
||||
0) {
|
||||
if (comparator->Compare(
|
||||
current_files[first_included - 1].largestkey,
|
||||
current_files[first_included].smallestkey) < 0) {
|
||||
break;
|
||||
}
|
||||
first_included--;
|
||||
}
|
||||
|
||||
while (last_included < static_cast<int>(current_files.size()) - 1) {
|
||||
if (comparator->Compare(current_files[last_included + 1].smallestkey,
|
||||
current_files[last_included].largestkey) > 0) {
|
||||
if (comparator->Compare(
|
||||
current_files[last_included + 1].smallestkey,
|
||||
current_files[last_included].largestkey) > 0) {
|
||||
break;
|
||||
}
|
||||
last_included++;
|
||||
@@ -730,31 +674,33 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
|
||||
// include all files between the first and the last compaction input files.
|
||||
for (int f = first_included; f <= last_included; ++f) {
|
||||
if (current_files[f].being_compacted) {
|
||||
return Status::Aborted("Necessary compaction input file " +
|
||||
current_files[f].name +
|
||||
" is currently being compacted.");
|
||||
return Status::Aborted(
|
||||
"Necessary compaction input file " + current_files[f].name +
|
||||
" is currently being compacted.");
|
||||
}
|
||||
input_files->insert(TableFileNameToNumber(current_files[f].name));
|
||||
input_files->insert(
|
||||
TableFileNameToNumber(current_files[f].name));
|
||||
}
|
||||
|
||||
// update smallest and largest key
|
||||
if (l == 0) {
|
||||
for (int f = first_included; f <= last_included; ++f) {
|
||||
if (comparator->Compare(smallestkey, current_files[f].smallestkey) >
|
||||
0) {
|
||||
if (comparator->Compare(
|
||||
smallestkey, current_files[f].smallestkey) > 0) {
|
||||
smallestkey = current_files[f].smallestkey;
|
||||
}
|
||||
if (comparator->Compare(largestkey, current_files[f].largestkey) < 0) {
|
||||
if (comparator->Compare(
|
||||
largestkey, current_files[f].largestkey) < 0) {
|
||||
largestkey = current_files[f].largestkey;
|
||||
}
|
||||
}
|
||||
} else {
|
||||
if (comparator->Compare(smallestkey,
|
||||
current_files[first_included].smallestkey) > 0) {
|
||||
if (comparator->Compare(
|
||||
smallestkey, current_files[first_included].smallestkey) > 0) {
|
||||
smallestkey = current_files[first_included].smallestkey;
|
||||
}
|
||||
if (comparator->Compare(largestkey,
|
||||
current_files[last_included].largestkey) < 0) {
|
||||
if (comparator->Compare(
|
||||
largestkey, current_files[last_included].largestkey) < 0) {
|
||||
largestkey = current_files[last_included].largestkey;
|
||||
}
|
||||
}
|
||||
@@ -771,15 +717,16 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
|
||||
// time and not by key
|
||||
for (int m = std::max(l, 1); m <= output_level; ++m) {
|
||||
for (auto& next_lv_file : levels[m].files) {
|
||||
if (HaveOverlappingKeyRanges(comparator, aggregated_file_meta,
|
||||
next_lv_file)) {
|
||||
if (HaveOverlappingKeyRanges(
|
||||
comparator, aggregated_file_meta, next_lv_file)) {
|
||||
if (next_lv_file.being_compacted) {
|
||||
return Status::Aborted(
|
||||
"File " + next_lv_file.name +
|
||||
" that has overlapping key range with one of the compaction "
|
||||
" input file is currently being compacted.");
|
||||
}
|
||||
input_files->insert(TableFileNameToNumber(next_lv_file.name));
|
||||
input_files->insert(
|
||||
TableFileNameToNumber(next_lv_file.name));
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -789,25 +736,28 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
|
||||
|
||||
Status CompactionPicker::SanitizeCompactionInputFiles(
|
||||
std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta, const int output_level) const {
|
||||
const ColumnFamilyMetaData& cf_meta,
|
||||
const int output_level) const {
|
||||
assert(static_cast<int>(cf_meta.levels.size()) - 1 ==
|
||||
cf_meta.levels[cf_meta.levels.size() - 1].level);
|
||||
if (output_level >= static_cast<int>(cf_meta.levels.size())) {
|
||||
return Status::InvalidArgument(
|
||||
"Output level for column family " + cf_meta.name +
|
||||
" must between [0, " +
|
||||
ToString(cf_meta.levels[cf_meta.levels.size() - 1].level) + "].");
|
||||
ToString(cf_meta.levels[cf_meta.levels.size() - 1].level) +
|
||||
"].");
|
||||
}
|
||||
|
||||
if (output_level > MaxOutputLevel()) {
|
||||
return Status::InvalidArgument(
|
||||
"Exceed the maximum output level defined by "
|
||||
"the current compaction algorithm --- " +
|
||||
ToString(MaxOutputLevel()));
|
||||
ToString(MaxOutputLevel()));
|
||||
}
|
||||
|
||||
if (output_level < 0) {
|
||||
return Status::InvalidArgument("Output level cannot be negative.");
|
||||
return Status::InvalidArgument(
|
||||
"Output level cannot be negative.");
|
||||
}
|
||||
|
||||
if (input_files->size() == 0) {
|
||||
@@ -815,8 +765,8 @@ Status CompactionPicker::SanitizeCompactionInputFiles(
|
||||
"A compaction must contain at least one file.");
|
||||
}
|
||||
|
||||
Status s = SanitizeCompactionInputFilesForAllLevels(input_files, cf_meta,
|
||||
output_level);
|
||||
Status s = SanitizeCompactionInputFilesForAllLevels(
|
||||
input_files, cf_meta, output_level);
|
||||
|
||||
if (!s.ok()) {
|
||||
return s;
|
||||
@@ -830,9 +780,10 @@ Status CompactionPicker::SanitizeCompactionInputFiles(
|
||||
for (auto file_meta : level_meta.files) {
|
||||
if (file_num == TableFileNameToNumber(file_meta.name)) {
|
||||
if (file_meta.being_compacted) {
|
||||
return Status::Aborted("Specified compaction input file " +
|
||||
MakeTableFileName("", file_num) +
|
||||
" is already being compacted.");
|
||||
return Status::Aborted(
|
||||
"Specified compaction input file " +
|
||||
MakeTableFileName("", file_num) +
|
||||
" is already being compacted.");
|
||||
}
|
||||
found = true;
|
||||
break;
|
||||
@@ -844,7 +795,8 @@ Status CompactionPicker::SanitizeCompactionInputFiles(
|
||||
}
|
||||
if (!found) {
|
||||
return Status::InvalidArgument(
|
||||
"Specified compaction input file " + MakeTableFileName("", file_num) +
|
||||
"Specified compaction input file " +
|
||||
MakeTableFileName("", file_num) +
|
||||
" does not exist in column family " + cf_meta.name + ".");
|
||||
}
|
||||
}
|
||||
@@ -853,8 +805,8 @@ Status CompactionPicker::SanitizeCompactionInputFiles(
|
||||
}
|
||||
#endif // !ROCKSDB_LITE
|
||||
|
||||
bool LevelCompactionPicker::NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const {
|
||||
bool LevelCompactionPicker::NeedsCompaction(const VersionStorageInfo* vstorage)
|
||||
const {
|
||||
if (!vstorage->FilesMarkedForCompaction().empty()) {
|
||||
return true;
|
||||
}
|
||||
@@ -918,40 +870,22 @@ Compaction* LevelCompactionPicker::PickCompaction(
|
||||
int base_index = -1;
|
||||
CompactionInputFiles inputs;
|
||||
double score = 0;
|
||||
CompactionReason compaction_reason = CompactionReason::kUnknown;
|
||||
|
||||
// Find the compactions by size on all levels.
|
||||
bool skipped_l0 = false;
|
||||
for (int i = 0; i < NumberLevels() - 1; i++) {
|
||||
score = vstorage->CompactionScore(i);
|
||||
level = vstorage->CompactionScoreLevel(i);
|
||||
assert(i == 0 || score <= vstorage->CompactionScore(i - 1));
|
||||
if (score >= 1) {
|
||||
if (skipped_l0 && level == vstorage->base_level()) {
|
||||
// If L0->base_level compaction is pending, don't schedule further
|
||||
// compaction from base level. Otherwise L0->base_level compaction
|
||||
// may starve.
|
||||
continue;
|
||||
}
|
||||
output_level = (level == 0) ? vstorage->base_level() : level + 1;
|
||||
if (PickCompactionBySize(vstorage, level, output_level, &inputs,
|
||||
&parent_index, &base_index) &&
|
||||
ExpandWhileOverlapping(cf_name, vstorage, &inputs)) {
|
||||
// found the compaction!
|
||||
if (level == 0) {
|
||||
// L0 score = `num L0 files` / `level0_file_num_compaction_trigger`
|
||||
compaction_reason = CompactionReason::kLevelL0FilesNum;
|
||||
} else {
|
||||
// L1+ score = `Level files size` / `MaxBytesForLevel`
|
||||
compaction_reason = CompactionReason::kLevelMaxLevelSize;
|
||||
}
|
||||
break;
|
||||
} else {
|
||||
// didn't find the compaction, clear the inputs
|
||||
inputs.clear();
|
||||
if (level == 0) {
|
||||
skipped_l0 = true;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -964,9 +898,6 @@ Compaction* LevelCompactionPicker::PickCompaction(
|
||||
parent_index = base_index = -1;
|
||||
PickFilesMarkedForCompactionExperimental(cf_name, vstorage, &inputs, &level,
|
||||
&output_level);
|
||||
if (!inputs.empty()) {
|
||||
compaction_reason = CompactionReason::kFilesMarkedForCompaction;
|
||||
}
|
||||
}
|
||||
if (inputs.empty()) {
|
||||
return nullptr;
|
||||
@@ -1000,7 +931,7 @@ Compaction* LevelCompactionPicker::PickCompaction(
|
||||
CompactionInputFiles output_level_inputs;
|
||||
output_level_inputs.level = output_level;
|
||||
if (!SetupOtherInputs(cf_name, mutable_cf_options, vstorage, &inputs,
|
||||
&output_level_inputs, &parent_index, base_index)) {
|
||||
&output_level_inputs, &parent_index, base_index)) {
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
@@ -1014,12 +945,10 @@ Compaction* LevelCompactionPicker::PickCompaction(
|
||||
auto c = new Compaction(
|
||||
vstorage, mutable_cf_options, std::move(compaction_inputs), output_level,
|
||||
mutable_cf_options.MaxFileSizeForLevel(output_level),
|
||||
mutable_cf_options.max_compaction_bytes,
|
||||
mutable_cf_options.MaxGrandParentOverlapBytes(level),
|
||||
GetPathId(ioptions_, mutable_cf_options, output_level),
|
||||
GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
|
||||
vstorage->base_level()),
|
||||
std::move(grandparents), is_manual, score,
|
||||
false /* deletion_compaction */, compaction_reason);
|
||||
GetCompressionType(ioptions_, output_level, vstorage->base_level()),
|
||||
std::move(grandparents), is_manual, score);
|
||||
|
||||
// If it's level 0 compaction, make sure we don't execute any other level 0
|
||||
// compactions in parallel
|
||||
@@ -1031,7 +960,11 @@ Compaction* LevelCompactionPicker::PickCompaction(
|
||||
// takes running compactions into account (by skipping files that are already
|
||||
// being compacted). Since we just changed compaction score, we recalculate it
|
||||
// here
|
||||
vstorage->ComputeCompactionScore(mutable_cf_options);
|
||||
{ // this piece of code recomputes compaction score
|
||||
CompactionOptionsFIFO dummy_compaction_options_fifo;
|
||||
vstorage->ComputeCompactionScore(mutable_cf_options,
|
||||
dummy_compaction_options_fifo);
|
||||
}
|
||||
|
||||
TEST_SYNC_POINT_CALLBACK("LevelCompactionPicker::PickCompaction:Return", c);
|
||||
|
||||
@@ -1086,7 +1019,6 @@ bool LevelCompactionPicker::PickCompactionBySize(VersionStorageInfo* vstorage,
|
||||
// could be made better by looking at key-ranges that are
|
||||
// being compacted at level 0.
|
||||
if (level == 0 && !level0_compactions_in_progress_.empty()) {
|
||||
TEST_SYNC_POINT("LevelCompactionPicker::PickCompactionBySize:0");
|
||||
return false;
|
||||
}
|
||||
|
||||
@@ -1163,19 +1095,18 @@ void UniversalCompactionPicker::SortedRun::Dump(char* out_buf,
|
||||
}
|
||||
|
||||
void UniversalCompactionPicker::SortedRun::DumpSizeInfo(
|
||||
char* out_buf, size_t out_buf_size, size_t sorted_run_count) const {
|
||||
char* out_buf, size_t out_buf_size, int sorted_run_count) const {
|
||||
if (level == 0) {
|
||||
assert(file != nullptr);
|
||||
snprintf(out_buf, out_buf_size,
|
||||
"file %" PRIu64 "[%" ROCKSDB_PRIszt
|
||||
"] "
|
||||
"file %" PRIu64
|
||||
"[%d] "
|
||||
"with size %" PRIu64 " (compensated size %" PRIu64 ")",
|
||||
file->fd.GetNumber(), sorted_run_count, file->fd.GetFileSize(),
|
||||
file->compensated_file_size);
|
||||
} else {
|
||||
snprintf(out_buf, out_buf_size,
|
||||
"level %d[%" ROCKSDB_PRIszt
|
||||
"] "
|
||||
"level %d[%d] "
|
||||
"with size %" PRIu64 " (compensated size %" PRIu64 ")",
|
||||
level, sorted_run_count, size, compensated_file_size);
|
||||
}
|
||||
@@ -1303,12 +1234,9 @@ Compaction* UniversalCompactionPicker::PickCompaction(
|
||||
std::vector<SortedRun> sorted_runs =
|
||||
CalculateSortedRuns(*vstorage, ioptions_);
|
||||
|
||||
if (sorted_runs.size() == 0 ||
|
||||
sorted_runs.size() <
|
||||
(unsigned int)mutable_cf_options.level0_file_num_compaction_trigger) {
|
||||
if (sorted_runs.size() <
|
||||
(unsigned int)mutable_cf_options.level0_file_num_compaction_trigger) {
|
||||
LogToBuffer(log_buffer, "[%s] Universal: nothing to do\n", cf_name.c_str());
|
||||
TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
|
||||
nullptr);
|
||||
return nullptr;
|
||||
}
|
||||
VersionStorageInfo::LevelSummaryStorage tmp;
|
||||
@@ -1343,34 +1271,19 @@ Compaction* UniversalCompactionPicker::PickCompaction(
|
||||
assert(sorted_runs.size() >=
|
||||
static_cast<size_t>(
|
||||
mutable_cf_options.level0_file_num_compaction_trigger));
|
||||
// Get the total number of sorted runs that are not being compacted
|
||||
int num_sr_not_compacted = 0;
|
||||
for (size_t i = 0; i < sorted_runs.size(); i++) {
|
||||
if (sorted_runs[i].being_compacted == false) {
|
||||
num_sr_not_compacted++;
|
||||
}
|
||||
}
|
||||
|
||||
// The number of sorted runs that are not being compacted is greater than
|
||||
// the maximum allowed number of sorted runs
|
||||
if (num_sr_not_compacted >
|
||||
mutable_cf_options.level0_file_num_compaction_trigger) {
|
||||
unsigned int num_files =
|
||||
num_sr_not_compacted -
|
||||
mutable_cf_options.level0_file_num_compaction_trigger + 1;
|
||||
if ((c = PickCompactionUniversalReadAmp(
|
||||
cf_name, mutable_cf_options, vstorage, score, UINT_MAX,
|
||||
num_files, sorted_runs, log_buffer)) != nullptr) {
|
||||
LogToBuffer(log_buffer,
|
||||
"[%s] Universal: compacting for file num -- %u\n",
|
||||
cf_name.c_str(), num_files);
|
||||
}
|
||||
unsigned int num_files =
|
||||
static_cast<unsigned int>(sorted_runs.size()) -
|
||||
mutable_cf_options.level0_file_num_compaction_trigger;
|
||||
if ((c = PickCompactionUniversalReadAmp(
|
||||
cf_name, mutable_cf_options, vstorage, score, UINT_MAX,
|
||||
num_files, sorted_runs, log_buffer)) != nullptr) {
|
||||
LogToBuffer(log_buffer,
|
||||
"[%s] Universal: compacting for file num -- %u\n",
|
||||
cf_name.c_str(), num_files);
|
||||
}
|
||||
}
|
||||
}
|
||||
if (c == nullptr) {
|
||||
TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
|
||||
nullptr);
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
@@ -1423,8 +1336,6 @@ Compaction* UniversalCompactionPicker::PickCompaction(
|
||||
|
||||
level0_compactions_in_progress_.insert(c);
|
||||
|
||||
TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
|
||||
c);
|
||||
return c;
|
||||
}
|
||||
|
||||
@@ -1445,9 +1356,8 @@ uint32_t UniversalCompactionPicker::GetPathId(
|
||||
// considered in this algorithm. So the target size can be violated in
|
||||
// that case. We need to improve it.
|
||||
uint64_t accumulated_size = 0;
|
||||
uint64_t future_size =
|
||||
file_size * (100 - ioptions.compaction_options_universal.size_ratio) /
|
||||
100;
|
||||
uint64_t future_size = file_size *
|
||||
(100 - ioptions.compaction_options_universal.size_ratio) / 100;
|
||||
uint32_t p = 0;
|
||||
assert(!ioptions.db_paths.empty());
|
||||
for (; p < ioptions.db_paths.size() - 1; p++) {
|
||||
@@ -1471,27 +1381,22 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
unsigned int max_number_of_files_to_compact,
|
||||
const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer) {
|
||||
unsigned int min_merge_width =
|
||||
ioptions_.compaction_options_universal.min_merge_width;
|
||||
ioptions_.compaction_options_universal.min_merge_width;
|
||||
unsigned int max_merge_width =
|
||||
ioptions_.compaction_options_universal.max_merge_width;
|
||||
ioptions_.compaction_options_universal.max_merge_width;
|
||||
|
||||
const SortedRun* sr = nullptr;
|
||||
bool done = false;
|
||||
size_t start_index = 0;
|
||||
int start_index = 0;
|
||||
unsigned int candidate_count = 0;
|
||||
|
||||
unsigned int max_files_to_compact =
|
||||
std::min(max_merge_width, max_number_of_files_to_compact);
|
||||
unsigned int max_files_to_compact = std::min(max_merge_width,
|
||||
max_number_of_files_to_compact);
|
||||
min_merge_width = std::max(min_merge_width, 2U);
|
||||
|
||||
// Caller checks the size before executing this function. This invariant is
|
||||
// important because otherwise we may have a possible integer underflow when
|
||||
// dealing with unsigned types.
|
||||
assert(sorted_runs.size() > 0);
|
||||
|
||||
// Considers a candidate file only if it is smaller than the
|
||||
// total size accumulated so far.
|
||||
for (size_t loop = 0; loop < sorted_runs.size(); loop++) {
|
||||
for (unsigned int loop = 0; loop < sorted_runs.size(); loop++) {
|
||||
candidate_count = 0;
|
||||
|
||||
// Skip files that are already being compacted
|
||||
@@ -1523,7 +1428,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
}
|
||||
|
||||
// Check if the succeeding files need compaction.
|
||||
for (size_t i = loop + 1;
|
||||
for (unsigned int i = loop + 1;
|
||||
candidate_count < max_files_to_compact && i < sorted_runs.size();
|
||||
i++) {
|
||||
const SortedRun* succeeding_sr = &sorted_runs[i];
|
||||
@@ -1565,7 +1470,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
done = true;
|
||||
break;
|
||||
} else {
|
||||
for (size_t i = loop;
|
||||
for (unsigned int i = loop;
|
||||
i < loop + candidate_count && i < sorted_runs.size(); i++) {
|
||||
const SortedRun* skipping_sr = &sorted_runs[i];
|
||||
char file_num_buf[256];
|
||||
@@ -1578,7 +1483,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
if (!done || candidate_count <= 1) {
|
||||
return nullptr;
|
||||
}
|
||||
size_t first_index_after = start_index + candidate_count;
|
||||
unsigned int first_index_after = start_index + candidate_count;
|
||||
// Compression is enabled if files compacted earlier already reached
|
||||
// size ratio of compression.
|
||||
bool enable_compression = true;
|
||||
@@ -1593,7 +1498,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
uint64_t older_file_size = 0;
|
||||
for (size_t i = sorted_runs.size() - 1; i >= first_index_after; i--) {
|
||||
older_file_size += sorted_runs[i].size;
|
||||
if (older_file_size * 100L >= total_size * (long)ratio_to_compress) {
|
||||
if (older_file_size * 100L >= total_size * (long) ratio_to_compress) {
|
||||
enable_compression = false;
|
||||
break;
|
||||
}
|
||||
@@ -1619,7 +1524,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
for (size_t i = 0; i < inputs.size(); ++i) {
|
||||
inputs[i].level = start_level + static_cast<int>(i);
|
||||
}
|
||||
for (size_t i = start_index; i < first_index_after; i++) {
|
||||
for (unsigned int i = start_index; i < first_index_after; i++) {
|
||||
auto& picking_sr = sorted_runs[i];
|
||||
if (picking_sr.level == 0) {
|
||||
FileMetaData* picking_file = picking_sr.file;
|
||||
@@ -1636,19 +1541,11 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalReadAmp(
|
||||
file_num_buf);
|
||||
}
|
||||
|
||||
CompactionReason compaction_reason;
|
||||
if (max_number_of_files_to_compact == UINT_MAX) {
|
||||
compaction_reason = CompactionReason::kUniversalSortedRunNum;
|
||||
} else {
|
||||
compaction_reason = CompactionReason::kUniversalSizeRatio;
|
||||
}
|
||||
return new Compaction(
|
||||
vstorage, mutable_cf_options, std::move(inputs), output_level,
|
||||
mutable_cf_options.MaxFileSizeForLevel(output_level), LLONG_MAX, path_id,
|
||||
GetCompressionType(ioptions_, vstorage, mutable_cf_options, start_level,
|
||||
1, enable_compression),
|
||||
/* grandparents */ {}, /* is manual */ false, score,
|
||||
false /* deletion_compaction */, compaction_reason);
|
||||
GetCompressionType(ioptions_, start_level, 1, enable_compression),
|
||||
/* grandparents */ {}, /* is manual */ false, score);
|
||||
}
|
||||
|
||||
// Look at overall size amplification. If size amplification
|
||||
@@ -1662,19 +1559,19 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
VersionStorageInfo* vstorage, double score,
|
||||
const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer) {
|
||||
// percentage flexibilty while reducing size amplification
|
||||
uint64_t ratio =
|
||||
ioptions_.compaction_options_universal.max_size_amplification_percent;
|
||||
uint64_t ratio = ioptions_.compaction_options_universal.
|
||||
max_size_amplification_percent;
|
||||
|
||||
unsigned int candidate_count = 0;
|
||||
uint64_t candidate_size = 0;
|
||||
size_t start_index = 0;
|
||||
unsigned int start_index = 0;
|
||||
const SortedRun* sr = nullptr;
|
||||
|
||||
// Skip files that are already being compacted
|
||||
for (size_t loop = 0; loop < sorted_runs.size() - 1; loop++) {
|
||||
for (unsigned int loop = 0; loop < sorted_runs.size() - 1; loop++) {
|
||||
sr = &sorted_runs[loop];
|
||||
if (!sr->being_compacted) {
|
||||
start_index = loop; // Consider this as the first candidate.
|
||||
start_index = loop; // Consider this as the first candidate.
|
||||
break;
|
||||
}
|
||||
char file_num_buf[kFormatFileNumberBufSize];
|
||||
@@ -1686,19 +1583,18 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
}
|
||||
|
||||
if (sr == nullptr) {
|
||||
return nullptr; // no candidate files
|
||||
return nullptr; // no candidate files
|
||||
}
|
||||
{
|
||||
char file_num_buf[kFormatFileNumberBufSize];
|
||||
sr->Dump(file_num_buf, sizeof(file_num_buf), true);
|
||||
LogToBuffer(log_buffer,
|
||||
"[%s] Universal: First candidate %s[%" ROCKSDB_PRIszt "] %s",
|
||||
LogToBuffer(log_buffer, "[%s] Universal: First candidate %s[%d] %s",
|
||||
cf_name.c_str(), file_num_buf, start_index,
|
||||
" to reduce size amp.\n");
|
||||
}
|
||||
|
||||
// keep adding up all the remaining files
|
||||
for (size_t loop = start_index; loop < sorted_runs.size() - 1; loop++) {
|
||||
for (unsigned int loop = start_index; loop < sorted_runs.size() - 1; loop++) {
|
||||
sr = &sorted_runs[loop];
|
||||
if (sr->being_compacted) {
|
||||
char file_num_buf[kFormatFileNumberBufSize];
|
||||
@@ -1724,21 +1620,21 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
LogToBuffer(
|
||||
log_buffer,
|
||||
"[%s] Universal: size amp not needed. newer-files-total-size %" PRIu64
|
||||
" earliest-file-size %" PRIu64,
|
||||
"earliest-file-size %" PRIu64,
|
||||
cf_name.c_str(), candidate_size, earliest_file_size);
|
||||
return nullptr;
|
||||
} else {
|
||||
LogToBuffer(
|
||||
log_buffer,
|
||||
"[%s] Universal: size amp needed. newer-files-total-size %" PRIu64
|
||||
" earliest-file-size %" PRIu64,
|
||||
"earliest-file-size %" PRIu64,
|
||||
cf_name.c_str(), candidate_size, earliest_file_size);
|
||||
}
|
||||
assert(start_index < sorted_runs.size() - 1);
|
||||
|
||||
// Estimate total file size
|
||||
uint64_t estimated_total_size = 0;
|
||||
for (size_t loop = start_index; loop < sorted_runs.size(); loop++) {
|
||||
for (unsigned int loop = start_index; loop < sorted_runs.size(); loop++) {
|
||||
estimated_total_size += sorted_runs[loop].size;
|
||||
}
|
||||
uint32_t path_id = GetPathId(ioptions_, estimated_total_size);
|
||||
@@ -1749,7 +1645,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
inputs[i].level = start_level + static_cast<int>(i);
|
||||
}
|
||||
// We always compact all the files, so always compress.
|
||||
for (size_t loop = start_index; loop < sorted_runs.size(); loop++) {
|
||||
for (unsigned int loop = start_index; loop < sorted_runs.size(); loop++) {
|
||||
auto& picking_sr = sorted_runs[loop];
|
||||
if (picking_sr.level == 0) {
|
||||
FileMetaData* f = picking_sr.file;
|
||||
@@ -1761,7 +1657,7 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
}
|
||||
}
|
||||
char file_num_buf[256];
|
||||
picking_sr.DumpSizeInfo(file_num_buf, sizeof(file_num_buf), loop);
|
||||
sr->DumpSizeInfo(file_num_buf, sizeof(file_num_buf), loop);
|
||||
LogToBuffer(log_buffer, "[%s] Universal: size amp picking %s",
|
||||
cf_name.c_str(), file_num_buf);
|
||||
}
|
||||
@@ -1771,15 +1667,12 @@ Compaction* UniversalCompactionPicker::PickCompactionUniversalSizeAmp(
|
||||
vstorage->num_levels() - 1,
|
||||
mutable_cf_options.MaxFileSizeForLevel(vstorage->num_levels() - 1),
|
||||
/* max_grandparent_overlap_bytes */ LLONG_MAX, path_id,
|
||||
GetCompressionType(ioptions_, vstorage, mutable_cf_options,
|
||||
vstorage->num_levels() - 1, 1),
|
||||
/* grandparents */ {}, /* is manual */ false, score,
|
||||
false /* deletion_compaction */,
|
||||
CompactionReason::kUniversalSizeAmplification);
|
||||
GetCompressionType(ioptions_, vstorage->num_levels() - 1, 1),
|
||||
/* grandparents */ {}, /* is manual */ false, score);
|
||||
}
|
||||
|
||||
bool FIFOCompactionPicker::NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const {
|
||||
bool FIFOCompactionPicker::NeedsCompaction(const VersionStorageInfo* vstorage)
|
||||
const {
|
||||
const int kLevel0 = 0;
|
||||
return vstorage->CompactionScore(kLevel0) >= 1;
|
||||
}
|
||||
@@ -1834,7 +1727,7 @@ Compaction* FIFOCompactionPicker::PickCompaction(
|
||||
Compaction* c = new Compaction(
|
||||
vstorage, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
|
||||
kNoCompression, {}, /* is manual */ false, vstorage->CompactionScore(0),
|
||||
/* is deletion compaction */ true, CompactionReason::kFIFOMaxSize);
|
||||
/* is deletion compaction */ true);
|
||||
level0_compactions_in_progress_.insert(c);
|
||||
return c;
|
||||
}
|
||||
@@ -1843,7 +1736,7 @@ Compaction* FIFOCompactionPicker::CompactRange(
|
||||
const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end, bool* manual_conflict) {
|
||||
InternalKey** compaction_end) {
|
||||
assert(input_level == 0);
|
||||
assert(output_level == 0);
|
||||
*compaction_end = nullptr;
|
||||
|
||||
+46
-45
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -20,7 +20,8 @@
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/status.h"
|
||||
#include "util/cf_options.h"
|
||||
#include "util/mutable_cf_options.h"
|
||||
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -59,28 +60,29 @@ class CompactionPicker {
|
||||
const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end, bool* manual_conflict);
|
||||
InternalKey** compaction_end);
|
||||
|
||||
// The maximum allowed output level. Default value is NumberLevels() - 1.
|
||||
virtual int MaxOutputLevel() const { return NumberLevels() - 1; }
|
||||
virtual int MaxOutputLevel() const {
|
||||
return NumberLevels() - 1;
|
||||
}
|
||||
|
||||
virtual bool NeedsCompaction(const VersionStorageInfo* vstorage) const = 0;
|
||||
|
||||
// Sanitize the input set of compaction input files.
|
||||
// When the input parameters do not describe a valid compaction, the
|
||||
// function will try to fix the input_files by adding necessary
|
||||
// files. If it's not possible to conver an invalid input_files
|
||||
// into a valid one by adding more files, the function will return a
|
||||
// non-ok status with specific reason.
|
||||
// Sanitize the input set of compaction input files.
|
||||
// When the input parameters do not describe a valid compaction, the
|
||||
// function will try to fix the input_files by adding necessary
|
||||
// files. If it's not possible to conver an invalid input_files
|
||||
// into a valid one by adding more files, the function will return a
|
||||
// non-ok status with specific reason.
|
||||
#ifndef ROCKSDB_LITE
|
||||
Status SanitizeCompactionInputFiles(std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta,
|
||||
const int output_level) const;
|
||||
Status SanitizeCompactionInputFiles(
|
||||
std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta,
|
||||
const int output_level) const;
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Free up the files that participated in a compaction
|
||||
//
|
||||
// Requirement: DB mutex held
|
||||
void ReleaseCompactionFiles(Compaction* c, Status status);
|
||||
|
||||
// Returns true if any one of the specified files are being compacted
|
||||
@@ -92,7 +94,7 @@ class CompactionPicker {
|
||||
const CompactionOptions& compact_options,
|
||||
const std::vector<CompactionInputFiles>& input_files, int output_level,
|
||||
VersionStorageInfo* vstorage, const MutableCFOptions& mutable_cf_options,
|
||||
uint32_t output_path_id);
|
||||
uint32_t output_path_id) const;
|
||||
|
||||
// Converts a set of compaction input file numbers into
|
||||
// a list of CompactionInputFiles.
|
||||
@@ -119,15 +121,15 @@ class CompactionPicker {
|
||||
// Stores the minimal range that covers all entries in inputs in
|
||||
// *smallest, *largest.
|
||||
// REQUIRES: inputs is not empty
|
||||
void GetRange(const CompactionInputFiles& inputs, InternalKey* smallest,
|
||||
InternalKey* largest);
|
||||
void GetRange(const CompactionInputFiles& inputs,
|
||||
InternalKey* smallest, InternalKey* largest);
|
||||
|
||||
// Stores the minimal range that covers all entries in inputs1 and inputs2
|
||||
// in *smallest, *largest.
|
||||
// REQUIRES: inputs is not empty
|
||||
void GetRange(const CompactionInputFiles& inputs1,
|
||||
const CompactionInputFiles& inputs2, InternalKey* smallest,
|
||||
InternalKey* largest);
|
||||
const CompactionInputFiles& inputs2,
|
||||
InternalKey* smallest, InternalKey* largest);
|
||||
|
||||
// Add more files to the inputs on "level" to make sure that
|
||||
// no newer version of a key is compacted to "level+1" while leaving an older
|
||||
@@ -162,12 +164,13 @@ class CompactionPicker {
|
||||
|
||||
const ImmutableCFOptions& ioptions_;
|
||||
|
||||
// A helper function to SanitizeCompactionInputFiles() that
|
||||
// sanitizes "input_files" by adding necessary files.
|
||||
// A helper function to SanitizeCompactionInputFiles() that
|
||||
// sanitizes "input_files" by adding necessary files.
|
||||
#ifndef ROCKSDB_LITE
|
||||
virtual Status SanitizeCompactionInputFilesForAllLevels(
|
||||
std::unordered_set<uint64_t>* input_files,
|
||||
const ColumnFamilyMetaData& cf_meta, const int output_level) const;
|
||||
const ColumnFamilyMetaData& cf_meta,
|
||||
const int output_level) const;
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Keeps track of all compactions that are running on Level0.
|
||||
@@ -187,8 +190,8 @@ class LevelCompactionPicker : public CompactionPicker {
|
||||
VersionStorageInfo* vstorage,
|
||||
LogBuffer* log_buffer) override;
|
||||
|
||||
virtual bool NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const override;
|
||||
virtual bool NeedsCompaction(const VersionStorageInfo* vstorage) const
|
||||
override;
|
||||
|
||||
// Pick a path ID to place a newly generated file, with its level
|
||||
static uint32_t GetPathId(const ImmutableCFOptions& ioptions,
|
||||
@@ -227,8 +230,8 @@ class UniversalCompactionPicker : public CompactionPicker {
|
||||
|
||||
virtual int MaxOutputLevel() const override { return NumberLevels() - 1; }
|
||||
|
||||
virtual bool NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const override;
|
||||
virtual bool NeedsCompaction(const VersionStorageInfo* vstorage) const
|
||||
override;
|
||||
|
||||
private:
|
||||
struct SortedRun {
|
||||
@@ -248,7 +251,7 @@ class UniversalCompactionPicker : public CompactionPicker {
|
||||
|
||||
// sorted_run_count is added into the string to print
|
||||
void DumpSizeInfo(char* out_buf, size_t out_buf_size,
|
||||
size_t sorted_run_count) const;
|
||||
int sorted_run_count) const;
|
||||
|
||||
int level;
|
||||
// `file` Will be null for level > 0. For level = 0, the sorted run is
|
||||
@@ -299,20 +302,22 @@ class FIFOCompactionPicker : public CompactionPicker {
|
||||
const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end, bool* manual_conflict) override;
|
||||
InternalKey** compaction_end) override;
|
||||
|
||||
// The maximum allowed output level. Always returns 0.
|
||||
virtual int MaxOutputLevel() const override { return 0; }
|
||||
virtual int MaxOutputLevel() const override {
|
||||
return 0;
|
||||
}
|
||||
|
||||
virtual bool NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const override;
|
||||
virtual bool NeedsCompaction(const VersionStorageInfo* vstorage) const
|
||||
override;
|
||||
};
|
||||
|
||||
class NullCompactionPicker : public CompactionPicker {
|
||||
public:
|
||||
NullCompactionPicker(const ImmutableCFOptions& ioptions,
|
||||
const InternalKeyComparator* icmp)
|
||||
: CompactionPicker(ioptions, icmp) {}
|
||||
const InternalKeyComparator* icmp) :
|
||||
CompactionPicker(ioptions, icmp) {}
|
||||
virtual ~NullCompactionPicker() {}
|
||||
|
||||
// Always return "nullptr"
|
||||
@@ -324,27 +329,23 @@ class NullCompactionPicker : public CompactionPicker {
|
||||
}
|
||||
|
||||
// Always return "nullptr"
|
||||
Compaction* CompactRange(const std::string& cf_name,
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level,
|
||||
int output_level, uint32_t output_path_id,
|
||||
const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end,
|
||||
bool* manual_conflict) override {
|
||||
Compaction* CompactRange(
|
||||
const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
|
||||
VersionStorageInfo* vstorage, int input_level, int output_level,
|
||||
uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
|
||||
InternalKey** compaction_end) override {
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
// Always returns false.
|
||||
virtual bool NeedsCompaction(
|
||||
const VersionStorageInfo* vstorage) const override {
|
||||
virtual bool NeedsCompaction(const VersionStorageInfo* vstorage) const
|
||||
override {
|
||||
return false;
|
||||
}
|
||||
};
|
||||
#endif // !ROCKSDB_LITE
|
||||
|
||||
CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
|
||||
const VersionStorageInfo* vstorage,
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
int level, int base_level,
|
||||
const bool enable_compression = true);
|
||||
|
||||
|
||||
+25
-314
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -79,7 +79,7 @@ class CompactionPickerTest : public testing::Test {
|
||||
}
|
||||
|
||||
void Add(int level, uint32_t file_number, const char* smallest,
|
||||
const char* largest, uint64_t file_size = 1, uint32_t path_id = 0,
|
||||
const char* largest, uint64_t file_size = 0, uint32_t path_id = 0,
|
||||
SequenceNumber smallest_seq = 100,
|
||||
SequenceNumber largest_seq = 100) {
|
||||
assert(level < vstorage_->num_levels());
|
||||
@@ -119,7 +119,7 @@ class CompactionPickerTest : public testing::Test {
|
||||
vstorage_->UpdateNumNonEmptyLevels();
|
||||
vstorage_->GenerateFileIndexer();
|
||||
vstorage_->GenerateLevelFilesBrief();
|
||||
vstorage_->ComputeCompactionScore(mutable_cf_options_);
|
||||
vstorage_->ComputeCompactionScore(mutable_cf_options_, fifo_options_);
|
||||
vstorage_->GenerateLevel0NonOverlapping();
|
||||
vstorage_->SetFinalized();
|
||||
}
|
||||
@@ -195,7 +195,6 @@ TEST_F(CompactionPickerTest, LevelMaxScore) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.target_file_size_base = 10000000;
|
||||
mutable_cf_options_.target_file_size_multiplier = 10;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 10 * 1024 * 1024;
|
||||
Add(0, 1U, "150", "200", 1000000000U);
|
||||
// Level 1 score 1.2
|
||||
Add(1, 66U, "150", "200", 6000000U);
|
||||
@@ -320,7 +319,6 @@ TEST_F(CompactionPickerTest, Level0TriggerDynamic4) {
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 2;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 200;
|
||||
mutable_cf_options_.max_bytes_for_level_multiplier = 10;
|
||||
|
||||
NewVersionStorage(num_levels, kCompactionStyleLevel);
|
||||
Add(0, 1U, "150", "200");
|
||||
Add(0, 2U, "200", "250");
|
||||
@@ -353,7 +351,6 @@ TEST_F(CompactionPickerTest, LevelTriggerDynamic4) {
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 2;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 200;
|
||||
mutable_cf_options_.max_bytes_for_level_multiplier = 10;
|
||||
mutable_cf_options_.compaction_pri = kMinOverlappingRatio;
|
||||
NewVersionStorage(num_levels, kCompactionStyleLevel);
|
||||
Add(0, 1U, "150", "200");
|
||||
Add(num_levels - 1, 3U, "200", "250", 300U);
|
||||
@@ -369,9 +366,11 @@ TEST_F(CompactionPickerTest, LevelTriggerDynamic4) {
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
ASSERT_EQ(5U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(0, compaction->num_input_files(1));
|
||||
ASSERT_EQ(1U, compaction->num_input_levels());
|
||||
ASSERT_EQ(6U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(2U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(3U, compaction->input(1, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(4U, compaction->input(1, 1)->fd.GetNumber());
|
||||
ASSERT_EQ(2U, compaction->num_input_levels());
|
||||
ASSERT_EQ(num_levels - 1, compaction->output_level());
|
||||
}
|
||||
|
||||
@@ -467,6 +466,7 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
|
||||
fifo_options_.max_table_files_size = kMaxSize;
|
||||
ioptions_.compaction_options_fifo = fifo_options_;
|
||||
FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);
|
||||
|
||||
UpdateVersionStorageInfo();
|
||||
// must return false when there's no files.
|
||||
ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()), false);
|
||||
@@ -487,90 +487,6 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.target_file_size_base = 10000000;
|
||||
mutable_cf_options_.target_file_size_multiplier = 10;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 10 * 1024 * 1024;
|
||||
mutable_cf_options_.compaction_pri = kMinOverlappingRatio;
|
||||
|
||||
Add(2, 6U, "150", "179", 50000000U);
|
||||
Add(2, 7U, "180", "220", 50000000U);
|
||||
Add(2, 8U, "321", "400", 50000000U); // File not overlapping
|
||||
Add(2, 9U, "721", "800", 50000000U);
|
||||
|
||||
Add(3, 26U, "150", "170", 260000000U);
|
||||
Add(3, 27U, "171", "179", 260000000U);
|
||||
Add(3, 28U, "191", "220", 260000000U);
|
||||
Add(3, 29U, "221", "300", 260000000U);
|
||||
Add(3, 30U, "750", "900", 260000000U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
// Pick file 8 because it overlaps with 0 files on level 3.
|
||||
ASSERT_EQ(8U, compaction->input(0, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, CompactionPriMinOverlapping2) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.target_file_size_base = 10000000;
|
||||
mutable_cf_options_.target_file_size_multiplier = 10;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 10 * 1024 * 1024;
|
||||
mutable_cf_options_.compaction_pri = kMinOverlappingRatio;
|
||||
|
||||
Add(2, 6U, "150", "175",
|
||||
60000000U); // Overlaps with file 26, 27, total size 521M
|
||||
Add(2, 7U, "176", "200", 60000000U); // Overlaps with file 27, 28, total size
|
||||
// 520M, the smalelst overlapping
|
||||
Add(2, 8U, "201", "300",
|
||||
60000000U); // Overlaps with file 28, 29, total size 521M
|
||||
|
||||
Add(3, 26U, "100", "110", 261000000U);
|
||||
Add(3, 26U, "150", "170", 261000000U);
|
||||
Add(3, 27U, "171", "179", 260000000U);
|
||||
Add(3, 28U, "191", "220", 260000000U);
|
||||
Add(3, 29U, "221", "300", 261000000U);
|
||||
Add(3, 30U, "321", "400", 261000000U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
// Picking file 7 because overlapping ratio is the biggest.
|
||||
ASSERT_EQ(7U, compaction->input(0, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, CompactionPriMinOverlapping3) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.max_bytes_for_level_base = 10000000;
|
||||
mutable_cf_options_.max_bytes_for_level_multiplier = 10;
|
||||
mutable_cf_options_.compaction_pri = kMinOverlappingRatio;
|
||||
|
||||
// file 7 and 8 over lap with the same file, but file 8 is smaller so
|
||||
// it will be picked.
|
||||
Add(2, 6U, "150", "167", 60000000U); // Overlaps with file 26, 27
|
||||
Add(2, 7U, "168", "169", 60000000U); // Overlaps with file 27
|
||||
Add(2, 8U, "201", "300", 61000000U); // Overlaps with file 28, but the file
|
||||
// itself is larger. Should be picked.
|
||||
|
||||
Add(3, 26U, "160", "165", 260000000U);
|
||||
Add(3, 27U, "166", "170", 260000000U);
|
||||
Add(3, 28U, "180", "400", 260000000U);
|
||||
Add(3, 29U, "401", "500", 260000000U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
// Picking file 8 because overlapping ratio is the biggest.
|
||||
ASSERT_EQ(8U, compaction->input(0, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
// This test exhibits the bug where we don't properly reset parent_index in
|
||||
// PickCompaction()
|
||||
TEST_F(CompactionPickerTest, ParentIndexResetBug) {
|
||||
@@ -598,8 +514,6 @@ TEST_F(CompactionPickerTest, ParentIndexResetBug) {
|
||||
// ranges (with different sequence numbers) in the input files.
|
||||
TEST_F(CompactionPickerTest, OverlappingUserKeys) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.compaction_pri = kByCompensatedSize;
|
||||
|
||||
Add(1, 1U, "100", "150", 1U);
|
||||
// Overlapping user keys
|
||||
Add(1, 2U, "200", "400", 1U);
|
||||
@@ -621,9 +535,9 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys2) {
|
||||
// Overlapping user keys on same level and output level
|
||||
Add(1, 1U, "200", "400", 1000000000U);
|
||||
Add(1, 2U, "400", "500", 1U, 0, 0);
|
||||
Add(2, 3U, "000", "100", 1U);
|
||||
Add(2, 4U, "100", "600", 1U, 0, 0);
|
||||
Add(2, 5U, "600", "700", 1U, 0, 0);
|
||||
Add(2, 3U, "400", "600", 1U);
|
||||
// The following file is not in the compaction despite overlapping user keys
|
||||
Add(2, 4U, "600", "700", 1U, 0, 0);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
@@ -631,12 +545,10 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys2) {
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(2U, compaction->num_input_levels());
|
||||
ASSERT_EQ(2U, compaction->num_input_files(0));
|
||||
ASSERT_EQ(3U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(1U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(2U, compaction->input(0, 1)->fd.GetNumber());
|
||||
ASSERT_EQ(3U, compaction->input(1, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(4U, compaction->input(1, 1)->fd.GetNumber());
|
||||
ASSERT_EQ(5U, compaction->input(1, 2)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, OverlappingUserKeys3) {
|
||||
@@ -668,126 +580,6 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys3) {
|
||||
ASSERT_EQ(7U, compaction->input(1, 1)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, OverlappingUserKeys4) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.max_bytes_for_level_base = 1000000;
|
||||
|
||||
Add(1, 1U, "100", "150", 1U);
|
||||
Add(1, 2U, "150", "199", 1U, 0, 0);
|
||||
Add(1, 3U, "200", "250", 1100000U, 0, 0);
|
||||
Add(1, 4U, "251", "300", 1U, 0, 0);
|
||||
Add(1, 5U, "300", "350", 1U, 0, 0);
|
||||
|
||||
Add(2, 6U, "100", "115", 1U);
|
||||
Add(2, 7U, "125", "325", 1U);
|
||||
Add(2, 8U, "350", "400", 1U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(2U, compaction->num_input_levels());
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
ASSERT_EQ(1U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(3U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(7U, compaction->input(1, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri1) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 2;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 900000000U;
|
||||
|
||||
// 6 L0 files, score 3.
|
||||
Add(0, 1U, "000", "400", 1U);
|
||||
Add(0, 2U, "001", "400", 1U, 0, 0);
|
||||
Add(0, 3U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 31U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 32U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 33U, "001", "400", 1000000000U, 0, 0);
|
||||
|
||||
// L1 total size 2GB, score 2.2. If one file being comapcted, score 1.1.
|
||||
Add(1, 4U, "050", "300", 1000000000U, 0, 0);
|
||||
file_map_[4u].first->being_compacted = true;
|
||||
Add(1, 5U, "301", "350", 1000000000U, 0, 0);
|
||||
|
||||
// Output level overlaps with the beginning and the end of the chain
|
||||
Add(2, 6U, "050", "100", 1U);
|
||||
Add(2, 7U, "300", "400", 1U);
|
||||
|
||||
// No compaction should be scheduled, if L0 has higher priority than L1
|
||||
// but L0->L1 compaction is blocked by a file in L1 being compacted.
|
||||
UpdateVersionStorageInfo();
|
||||
ASSERT_EQ(0, vstorage_->CompactionScoreLevel(0));
|
||||
ASSERT_EQ(1, vstorage_->CompactionScoreLevel(1));
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() == nullptr);
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri2) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 2;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 900000000U;
|
||||
|
||||
// 6 L0 files, score 3.
|
||||
Add(0, 1U, "000", "400", 1U);
|
||||
Add(0, 2U, "001", "400", 1U, 0, 0);
|
||||
Add(0, 3U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 31U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 32U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 33U, "001", "400", 1000000000U, 0, 0);
|
||||
|
||||
// L1 total size 2GB, score 2.2. If one file being comapcted, score 1.1.
|
||||
Add(1, 4U, "050", "300", 1000000000U, 0, 0);
|
||||
Add(1, 5U, "301", "350", 1000000000U, 0, 0);
|
||||
|
||||
// Output level overlaps with the beginning and the end of the chain
|
||||
Add(2, 6U, "050", "100", 1U);
|
||||
Add(2, 7U, "300", "400", 1U);
|
||||
|
||||
// If no file in L1 being compacted, L0->L1 compaction will be scheduled.
|
||||
UpdateVersionStorageInfo(); // being_compacted flag is cleared here.
|
||||
ASSERT_EQ(0, vstorage_->CompactionScoreLevel(0));
|
||||
ASSERT_EQ(1, vstorage_->CompactionScoreLevel(1));
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri3) {
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 2;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 900000000U;
|
||||
|
||||
// 6 L0 files, score 3.
|
||||
Add(0, 1U, "000", "400", 1U);
|
||||
Add(0, 2U, "001", "400", 1U, 0, 0);
|
||||
Add(0, 3U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 31U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 32U, "001", "400", 1000000000U, 0, 0);
|
||||
Add(0, 33U, "001", "400", 1000000000U, 0, 0);
|
||||
|
||||
// L1 score more than 6.
|
||||
Add(1, 4U, "050", "300", 1000000000U, 0, 0);
|
||||
file_map_[4u].first->being_compacted = true;
|
||||
Add(1, 5U, "301", "350", 1000000000U, 0, 0);
|
||||
Add(1, 51U, "351", "400", 6000000000U, 0, 0);
|
||||
|
||||
// Output level overlaps with the beginning and the end of the chain
|
||||
Add(2, 6U, "050", "100", 1U);
|
||||
Add(2, 7U, "300", "400", 1U);
|
||||
|
||||
// If score in L1 is larger than L0, L1 compaction goes through despite
|
||||
// there is pending L0 compaction.
|
||||
UpdateVersionStorageInfo();
|
||||
ASSERT_EQ(1, vstorage_->CompactionScoreLevel(0));
|
||||
ASSERT_EQ(0, vstorage_->CompactionScoreLevel(1));
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, EstimateCompactionBytesNeeded1) {
|
||||
int num_levels = ioptions_.num_levels;
|
||||
ioptions_.level_compaction_dynamic_level_bytes = false;
|
||||
@@ -802,22 +594,19 @@ TEST_F(CompactionPickerTest, EstimateCompactionBytesNeeded1) {
|
||||
Add(1, 4U, "400", "500", 600);
|
||||
Add(1, 5U, "600", "700", 600);
|
||||
// Level 2 is less than target 10000 even added size of level 1
|
||||
// Size ratio of L2/L1 is 9600 / 1200 = 8
|
||||
Add(2, 6U, "150", "200", 2500);
|
||||
Add(2, 7U, "201", "210", 2000);
|
||||
Add(2, 8U, "300", "310", 2600);
|
||||
Add(2, 8U, "300", "310", 2500);
|
||||
Add(2, 9U, "400", "500", 2500);
|
||||
// Level 3 exceeds target 100,000 of 1000
|
||||
Add(3, 10U, "400", "500", 101000);
|
||||
// Level 4 exceeds target 1,000,000 by 900 after adding size from level 3
|
||||
// Size ratio L4/L3 is 9.9
|
||||
// After merge from L3, L4 size is 1000900
|
||||
Add(4, 11U, "400", "500", 999900);
|
||||
Add(5, 11U, "400", "500", 8007200);
|
||||
// Level 4 exceeds target 1,000,000 of 500 after adding size from level 3
|
||||
Add(4, 11U, "400", "500", 999500);
|
||||
Add(5, 11U, "400", "500", 8000000);
|
||||
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
ASSERT_EQ(200u * 9u + 10900u + 900u * 9,
|
||||
ASSERT_EQ(2200u + 11000u + 5500u,
|
||||
vstorage_->estimated_compaction_needed_bytes());
|
||||
}
|
||||
|
||||
@@ -833,42 +622,17 @@ TEST_F(CompactionPickerTest, EstimateCompactionBytesNeeded2) {
|
||||
Add(0, 4U, "150", "200", 200);
|
||||
Add(0, 5U, "150", "200", 200);
|
||||
Add(0, 6U, "150", "200", 200);
|
||||
// Level 1 size will be 1400 after merging with L0
|
||||
// Level 1 is over target by
|
||||
Add(1, 7U, "400", "500", 200);
|
||||
Add(1, 8U, "600", "700", 200);
|
||||
// Level 2 is less than target 10000 even added size of level 1
|
||||
Add(2, 9U, "150", "200", 9100);
|
||||
// Level 3 over the target, but since level 4 is empty, we assume it will be
|
||||
// a trivial move.
|
||||
Add(2, 9U, "150", "200", 9500);
|
||||
Add(3, 10U, "400", "500", 101000);
|
||||
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
// estimated L1->L2 merge: 400 * (9100.0 / 1400.0 + 1.0)
|
||||
ASSERT_EQ(1400u + 3000u, vstorage_->estimated_compaction_needed_bytes());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, EstimateCompactionBytesNeeded3) {
|
||||
int num_levels = ioptions_.num_levels;
|
||||
ioptions_.level_compaction_dynamic_level_bytes = false;
|
||||
mutable_cf_options_.level0_file_num_compaction_trigger = 3;
|
||||
mutable_cf_options_.max_bytes_for_level_base = 1000;
|
||||
mutable_cf_options_.max_bytes_for_level_multiplier = 10;
|
||||
NewVersionStorage(num_levels, kCompactionStyleLevel);
|
||||
Add(0, 1U, "150", "200", 2000);
|
||||
Add(0, 2U, "150", "200", 2000);
|
||||
Add(0, 4U, "150", "200", 2000);
|
||||
Add(0, 5U, "150", "200", 2000);
|
||||
Add(0, 6U, "150", "200", 1000);
|
||||
// Level 1 size will be 10000 after merging with L0
|
||||
Add(1, 7U, "400", "500", 500);
|
||||
Add(1, 8U, "600", "700", 500);
|
||||
|
||||
Add(2, 9U, "150", "200", 10000);
|
||||
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
ASSERT_EQ(10000u + 18000u, vstorage_->estimated_compaction_needed_bytes());
|
||||
ASSERT_EQ(1400u + 4400u + 11000u,
|
||||
vstorage_->estimated_compaction_needed_bytes());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, EstimateCompactionBytesNeededDynamicLevel) {
|
||||
@@ -892,14 +656,12 @@ TEST_F(CompactionPickerTest, EstimateCompactionBytesNeededDynamicLevel) {
|
||||
// num_levels - 3 is over target by 100 + 1000
|
||||
Add(num_levels - 3, 7U, "400", "500", 300);
|
||||
Add(num_levels - 3, 8U, "600", "700", 300);
|
||||
// num_levels - 2 is over target by 1100 + 200
|
||||
Add(num_levels - 2, 9U, "150", "200", 5200);
|
||||
// Level 2 is over target by 1100 + 100
|
||||
Add(num_levels - 2, 9U, "150", "200", 5100);
|
||||
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
// Merging to the second last level: (5200 / 1600 + 1) * 1100
|
||||
// Merging to the last level: (50000 / 6300 + 1) * 1300
|
||||
ASSERT_EQ(1600u + 4675u + 11617u,
|
||||
ASSERT_EQ(1600u + 12100u + 13200u,
|
||||
vstorage_->estimated_compaction_needed_bytes());
|
||||
}
|
||||
|
||||
@@ -1061,57 +823,6 @@ TEST_F(CompactionPickerTest, IsBottommostLevelTest) {
|
||||
DeleteVersionStorage();
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, MaxCompactionBytesHit) {
|
||||
mutable_cf_options_.max_bytes_for_level_base = 1000000u;
|
||||
mutable_cf_options_.max_compaction_bytes = 800000u;
|
||||
ioptions_.level_compaction_dynamic_level_bytes = false;
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
// A compaction should be triggered and pick file 2 and 5.
|
||||
// It cannot expand because adding file 1 and 3, the compaction size will
|
||||
// exceed mutable_cf_options_.max_bytes_for_level_base.
|
||||
Add(1, 1U, "100", "150", 300000U);
|
||||
Add(1, 2U, "151", "200", 300001U, 0, 0);
|
||||
Add(1, 3U, "201", "250", 300000U, 0, 0);
|
||||
Add(1, 4U, "251", "300", 300000U, 0, 0);
|
||||
Add(2, 5U, "160", "256", 1U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(2U, compaction->num_input_levels());
|
||||
ASSERT_EQ(1U, compaction->num_input_files(0));
|
||||
ASSERT_EQ(1U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(2U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(5U, compaction->input(1, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
TEST_F(CompactionPickerTest, MaxCompactionBytesNotHit) {
|
||||
mutable_cf_options_.max_bytes_for_level_base = 1000000u;
|
||||
mutable_cf_options_.max_compaction_bytes = 1000000u;
|
||||
ioptions_.level_compaction_dynamic_level_bytes = false;
|
||||
NewVersionStorage(6, kCompactionStyleLevel);
|
||||
// A compaction should be triggered and pick file 2 and 5.
|
||||
// and it expands to file 1 and 3 too.
|
||||
Add(1, 1U, "100", "150", 300000U);
|
||||
Add(1, 2U, "151", "200", 300001U, 0, 0);
|
||||
Add(1, 3U, "201", "250", 300000U, 0, 0);
|
||||
Add(1, 4U, "251", "300", 300000U, 0, 0);
|
||||
Add(2, 5U, "000", "233", 1U);
|
||||
UpdateVersionStorageInfo();
|
||||
|
||||
std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
|
||||
cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
|
||||
ASSERT_TRUE(compaction.get() != nullptr);
|
||||
ASSERT_EQ(2U, compaction->num_input_levels());
|
||||
ASSERT_EQ(3U, compaction->num_input_files(0));
|
||||
ASSERT_EQ(1U, compaction->num_input_files(1));
|
||||
ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
|
||||
ASSERT_EQ(2U, compaction->input(0, 1)->fd.GetNumber());
|
||||
ASSERT_EQ(3U, compaction->input(0, 2)->fd.GetNumber());
|
||||
ASSERT_EQ(5U, compaction->input(1, 0)->fd.GetNumber());
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
|
||||
@@ -1,17 +1,16 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
#include <map>
|
||||
#include <string>
|
||||
|
||||
#include "memtable/stl_wrappers.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "util/hash.h"
|
||||
#include "util/kv_map.h"
|
||||
#include "util/stl_wrappers.h"
|
||||
#include "util/string_util.h"
|
||||
#include "util/testharness.h"
|
||||
#include "util/testutil.h"
|
||||
|
||||
+1
-8
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -18,13 +18,6 @@ namespace rocksdb {
|
||||
void CancelAllBackgroundWork(DB* db, bool wait) {
|
||||
(dynamic_cast<DBImpl*>(db))->CancelAllBackgroundWork(wait);
|
||||
}
|
||||
|
||||
Status DeleteFilesInRange(DB* db, ColumnFamilyHandle* column_family,
|
||||
const Slice* begin, const Slice* end) {
|
||||
return (dynamic_cast<DBImpl*>(db))
|
||||
->DeleteFilesInRange(column_family, begin, end);
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
+9
-20
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -11,7 +11,6 @@
|
||||
|
||||
#include "rocksdb/db.h"
|
||||
|
||||
#include <inttypes.h>
|
||||
#include <errno.h>
|
||||
#include <fcntl.h>
|
||||
#include <sys/stat.h>
|
||||
@@ -42,7 +41,6 @@ class CorruptionTest : public testing::Test {
|
||||
|
||||
CorruptionTest() {
|
||||
tiny_cache_ = NewLRUCache(100);
|
||||
options_.wal_recovery_mode = WALRecoveryMode::kTolerateCorruptedTailRecords;
|
||||
options_.env = &env_;
|
||||
dbname_ = test::TmpDir() + "/corruption_test";
|
||||
DestroyDB(dbname_, options_);
|
||||
@@ -106,8 +104,8 @@ class CorruptionTest : public testing::Test {
|
||||
}
|
||||
|
||||
void Check(int min_expected, int max_expected) {
|
||||
uint64_t next_expected = 0;
|
||||
uint64_t missed = 0;
|
||||
unsigned int next_expected = 0;
|
||||
int missed = 0;
|
||||
int bad_keys = 0;
|
||||
int bad_values = 0;
|
||||
int correct = 0;
|
||||
@@ -128,7 +126,7 @@ class CorruptionTest : public testing::Test {
|
||||
continue;
|
||||
}
|
||||
missed += (key - next_expected);
|
||||
next_expected = key + 1;
|
||||
next_expected = static_cast<unsigned int>(key + 1);
|
||||
if (iter->value() != Value(static_cast<int>(key), &value_space)) {
|
||||
bad_values++;
|
||||
} else {
|
||||
@@ -138,9 +136,8 @@ class CorruptionTest : public testing::Test {
|
||||
delete iter;
|
||||
|
||||
fprintf(stderr,
|
||||
"expected=%d..%d; got=%d; bad_keys=%d; bad_values=%d; missed=%llu\n",
|
||||
min_expected, max_expected, correct, bad_keys, bad_values,
|
||||
static_cast<unsigned long long>(missed));
|
||||
"expected=%d..%d; got=%d; bad_keys=%d; bad_values=%d; missed=%d\n",
|
||||
min_expected, max_expected, correct, bad_keys, bad_values, missed);
|
||||
ASSERT_LE(min_expected, correct);
|
||||
ASSERT_GE(max_expected, correct);
|
||||
}
|
||||
@@ -186,7 +183,7 @@ class CorruptionTest : public testing::Test {
|
||||
FileType type;
|
||||
std::string fname;
|
||||
int picked_number = -1;
|
||||
for (size_t i = 0; i < filenames.size(); i++) {
|
||||
for (unsigned int i = 0; i < filenames.size(); i++) {
|
||||
if (ParseFileName(filenames[i], &number, &type) &&
|
||||
type == filetype &&
|
||||
static_cast<int>(number) > picked_number) { // Pick latest file
|
||||
@@ -235,16 +232,8 @@ class CorruptionTest : public testing::Test {
|
||||
|
||||
// Return the value to associate with the specified key
|
||||
Slice Value(int k, std::string* storage) {
|
||||
if (k == 0) {
|
||||
// Ugh. Random seed of 0 used to produce no entropy. This code
|
||||
// preserves the implementation that was in place when all of the
|
||||
// magic values in this file were picked.
|
||||
*storage = std::string(kValueSize, ' ');
|
||||
return Slice(*storage);
|
||||
} else {
|
||||
Random r(k);
|
||||
return test::RandomString(&r, kValueSize, storage);
|
||||
}
|
||||
Random r(k);
|
||||
return test::RandomString(&r, kValueSize, storage);
|
||||
}
|
||||
};
|
||||
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,579 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
#include <cstdlib>
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/lru_cache.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class DBBlockCacheTest : public DBTestBase {
|
||||
private:
|
||||
size_t miss_count_ = 0;
|
||||
size_t hit_count_ = 0;
|
||||
size_t insert_count_ = 0;
|
||||
size_t failure_count_ = 0;
|
||||
size_t compressed_miss_count_ = 0;
|
||||
size_t compressed_hit_count_ = 0;
|
||||
size_t compressed_insert_count_ = 0;
|
||||
size_t compressed_failure_count_ = 0;
|
||||
|
||||
public:
|
||||
const size_t kNumBlocks = 10;
|
||||
const size_t kValueSize = 100;
|
||||
|
||||
DBBlockCacheTest() : DBTestBase("/db_block_cache_test") {}
|
||||
|
||||
BlockBasedTableOptions GetTableOptions() {
|
||||
BlockBasedTableOptions table_options;
|
||||
// Set a small enough block size so that each key-value get its own block.
|
||||
table_options.block_size = 1;
|
||||
return table_options;
|
||||
}
|
||||
|
||||
Options GetOptions(const BlockBasedTableOptions& table_options) {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.avoid_flush_during_recovery = false;
|
||||
// options.compression = kNoCompression;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
return options;
|
||||
}
|
||||
|
||||
void InitTable(const Options& options) {
|
||||
std::string value(kValueSize, 'a');
|
||||
for (size_t i = 0; i < kNumBlocks; i++) {
|
||||
ASSERT_OK(Put(ToString(i), value.c_str()));
|
||||
}
|
||||
}
|
||||
|
||||
void RecordCacheCounters(const Options& options) {
|
||||
miss_count_ = TestGetTickerCount(options, BLOCK_CACHE_MISS);
|
||||
hit_count_ = TestGetTickerCount(options, BLOCK_CACHE_HIT);
|
||||
insert_count_ = TestGetTickerCount(options, BLOCK_CACHE_ADD);
|
||||
failure_count_ = TestGetTickerCount(options, BLOCK_CACHE_ADD_FAILURES);
|
||||
compressed_miss_count_ =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS);
|
||||
compressed_hit_count_ =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT);
|
||||
compressed_insert_count_ =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD);
|
||||
compressed_failure_count_ =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD_FAILURES);
|
||||
}
|
||||
|
||||
void CheckCacheCounters(const Options& options, size_t expected_misses,
|
||||
size_t expected_hits, size_t expected_inserts,
|
||||
size_t expected_failures) {
|
||||
size_t new_miss_count = TestGetTickerCount(options, BLOCK_CACHE_MISS);
|
||||
size_t new_hit_count = TestGetTickerCount(options, BLOCK_CACHE_HIT);
|
||||
size_t new_insert_count = TestGetTickerCount(options, BLOCK_CACHE_ADD);
|
||||
size_t new_failure_count =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD_FAILURES);
|
||||
ASSERT_EQ(miss_count_ + expected_misses, new_miss_count);
|
||||
ASSERT_EQ(hit_count_ + expected_hits, new_hit_count);
|
||||
ASSERT_EQ(insert_count_ + expected_inserts, new_insert_count);
|
||||
ASSERT_EQ(failure_count_ + expected_failures, new_failure_count);
|
||||
miss_count_ = new_miss_count;
|
||||
hit_count_ = new_hit_count;
|
||||
insert_count_ = new_insert_count;
|
||||
failure_count_ = new_failure_count;
|
||||
}
|
||||
|
||||
void CheckCompressedCacheCounters(const Options& options,
|
||||
size_t expected_misses,
|
||||
size_t expected_hits,
|
||||
size_t expected_inserts,
|
||||
size_t expected_failures) {
|
||||
size_t new_miss_count =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS);
|
||||
size_t new_hit_count =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT);
|
||||
size_t new_insert_count =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD);
|
||||
size_t new_failure_count =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD_FAILURES);
|
||||
ASSERT_EQ(compressed_miss_count_ + expected_misses, new_miss_count);
|
||||
ASSERT_EQ(compressed_hit_count_ + expected_hits, new_hit_count);
|
||||
ASSERT_EQ(compressed_insert_count_ + expected_inserts, new_insert_count);
|
||||
ASSERT_EQ(compressed_failure_count_ + expected_failures, new_failure_count);
|
||||
compressed_miss_count_ = new_miss_count;
|
||||
compressed_hit_count_ = new_hit_count;
|
||||
compressed_insert_count_ = new_insert_count;
|
||||
compressed_failure_count_ = new_failure_count;
|
||||
}
|
||||
};
|
||||
|
||||
TEST_F(DBBlockCacheTest, TestWithoutCompressedBlockCache) {
|
||||
ReadOptions read_options;
|
||||
auto table_options = GetTableOptions();
|
||||
auto options = GetOptions(table_options);
|
||||
InitTable(options);
|
||||
|
||||
std::shared_ptr<Cache> cache = NewLRUCache(0, 0, false);
|
||||
table_options.block_cache = cache;
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
Reopen(options);
|
||||
RecordCacheCounters(options);
|
||||
|
||||
std::vector<std::unique_ptr<Iterator>> iterators(kNumBlocks - 1);
|
||||
Iterator* iter = nullptr;
|
||||
|
||||
// Load blocks into cache.
|
||||
for (size_t i = 0; i < kNumBlocks - 1; i++) {
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(i));
|
||||
ASSERT_OK(iter->status());
|
||||
CheckCacheCounters(options, 1, 0, 1, 0);
|
||||
iterators[i].reset(iter);
|
||||
}
|
||||
size_t usage = cache->GetUsage();
|
||||
ASSERT_LT(0, usage);
|
||||
cache->SetCapacity(usage);
|
||||
ASSERT_EQ(usage, cache->GetPinnedUsage());
|
||||
|
||||
// Test with strict capacity limit.
|
||||
cache->SetStrictCapacityLimit(true);
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(kNumBlocks - 1));
|
||||
ASSERT_TRUE(iter->status().IsIncomplete());
|
||||
CheckCacheCounters(options, 1, 0, 0, 1);
|
||||
delete iter;
|
||||
iter = nullptr;
|
||||
|
||||
// Release interators and access cache again.
|
||||
for (size_t i = 0; i < kNumBlocks - 1; i++) {
|
||||
iterators[i].reset();
|
||||
CheckCacheCounters(options, 0, 0, 0, 0);
|
||||
}
|
||||
ASSERT_EQ(0, cache->GetPinnedUsage());
|
||||
for (size_t i = 0; i < kNumBlocks - 1; i++) {
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(i));
|
||||
ASSERT_OK(iter->status());
|
||||
CheckCacheCounters(options, 0, 1, 0, 0);
|
||||
iterators[i].reset(iter);
|
||||
}
|
||||
}
|
||||
|
||||
#ifdef SNAPPY
|
||||
TEST_F(DBBlockCacheTest, TestWithCompressedBlockCache) {
|
||||
ReadOptions read_options;
|
||||
auto table_options = GetTableOptions();
|
||||
auto options = GetOptions(table_options);
|
||||
options.compression = CompressionType::kSnappyCompression;
|
||||
InitTable(options);
|
||||
|
||||
std::shared_ptr<Cache> cache = NewLRUCache(0, 0, false);
|
||||
std::shared_ptr<Cache> compressed_cache = NewLRUCache(1 << 25, 0, false);
|
||||
table_options.block_cache = cache;
|
||||
table_options.block_cache_compressed = compressed_cache;
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
Reopen(options);
|
||||
RecordCacheCounters(options);
|
||||
|
||||
std::vector<std::unique_ptr<Iterator>> iterators(kNumBlocks - 1);
|
||||
Iterator* iter = nullptr;
|
||||
|
||||
// Load blocks into cache.
|
||||
for (size_t i = 0; i < kNumBlocks - 1; i++) {
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(i));
|
||||
ASSERT_OK(iter->status());
|
||||
CheckCacheCounters(options, 1, 0, 1, 0);
|
||||
CheckCompressedCacheCounters(options, 1, 0, 1, 0);
|
||||
iterators[i].reset(iter);
|
||||
}
|
||||
size_t usage = cache->GetUsage();
|
||||
ASSERT_LT(0, usage);
|
||||
ASSERT_EQ(usage, cache->GetPinnedUsage());
|
||||
size_t compressed_usage = compressed_cache->GetUsage();
|
||||
ASSERT_LT(0, compressed_usage);
|
||||
// Compressed block cache cannot be pinned.
|
||||
ASSERT_EQ(0, compressed_cache->GetPinnedUsage());
|
||||
|
||||
// Set strict capacity limit flag. Now block will only load into compressed
|
||||
// block cache.
|
||||
cache->SetCapacity(usage);
|
||||
cache->SetStrictCapacityLimit(true);
|
||||
ASSERT_EQ(usage, cache->GetPinnedUsage());
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(kNumBlocks - 1));
|
||||
ASSERT_TRUE(iter->status().IsIncomplete());
|
||||
CheckCacheCounters(options, 1, 0, 0, 1);
|
||||
CheckCompressedCacheCounters(options, 1, 0, 1, 0);
|
||||
delete iter;
|
||||
iter = nullptr;
|
||||
|
||||
// Clear strict capacity limit flag. This time we shall hit compressed block
|
||||
// cache.
|
||||
cache->SetStrictCapacityLimit(false);
|
||||
iter = db_->NewIterator(read_options);
|
||||
iter->Seek(ToString(kNumBlocks - 1));
|
||||
ASSERT_OK(iter->status());
|
||||
CheckCacheCounters(options, 1, 0, 1, 0);
|
||||
CheckCompressedCacheCounters(options, 0, 1, 0, 0);
|
||||
delete iter;
|
||||
iter = nullptr;
|
||||
}
|
||||
#endif // SNAPPY
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
// Make sure that when options.block_cache is set, after a new table is
|
||||
// created its index/filter blocks are added to block cache.
|
||||
TEST_F(DBBlockCacheTest, IndexAndFilterBlocksOfNewTableAddedToCache) {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.cache_index_and_filter_blocks = true;
|
||||
table_options.filter_policy.reset(NewBloomFilterPolicy(20));
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
ASSERT_OK(Put(1, "key", "val"));
|
||||
// Create a new table.
|
||||
ASSERT_OK(Flush(1));
|
||||
|
||||
// index/filter blocks added to block cache right after table creation.
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
|
||||
ASSERT_EQ(2, /* only index/filter were added */
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
ASSERT_EQ(0, TestGetTickerCount(options, BLOCK_CACHE_DATA_MISS));
|
||||
uint64_t int_num;
|
||||
ASSERT_TRUE(
|
||||
dbfull()->GetIntProperty("rocksdb.estimate-table-readers-mem", &int_num));
|
||||
ASSERT_EQ(int_num, 0U);
|
||||
|
||||
// Make sure filter block is in cache.
|
||||
std::string value;
|
||||
ReadOptions ropt;
|
||||
db_->KeyMayExist(ReadOptions(), handles_[1], "key", &value);
|
||||
|
||||
// Miss count should remain the same.
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_HIT));
|
||||
|
||||
db_->KeyMayExist(ReadOptions(), handles_[1], "key", &value);
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
|
||||
ASSERT_EQ(2, TestGetTickerCount(options, BLOCK_CACHE_FILTER_HIT));
|
||||
|
||||
// Make sure index block is in cache.
|
||||
auto index_block_hit = TestGetTickerCount(options, BLOCK_CACHE_INDEX_HIT);
|
||||
value = Get(1, "key");
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
|
||||
ASSERT_EQ(index_block_hit + 1,
|
||||
TestGetTickerCount(options, BLOCK_CACHE_INDEX_HIT));
|
||||
|
||||
value = Get(1, "key");
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
|
||||
ASSERT_EQ(index_block_hit + 2,
|
||||
TestGetTickerCount(options, BLOCK_CACHE_INDEX_HIT));
|
||||
}
|
||||
|
||||
TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.cache_index_and_filter_blocks = true;
|
||||
// 200 bytes are enough to hold the first two blocks
|
||||
std::shared_ptr<Cache> cache = NewLRUCache(200, 0, false);
|
||||
table_options.block_cache = cache;
|
||||
table_options.filter_policy.reset(NewBloomFilterPolicy(20));
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
ASSERT_OK(Put(1, "key", "val"));
|
||||
// Create a new table
|
||||
ASSERT_OK(Flush(1));
|
||||
size_t index_bytes_insert =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_INSERT);
|
||||
size_t filter_bytes_insert =
|
||||
TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_INSERT);
|
||||
ASSERT_GT(index_bytes_insert, 0);
|
||||
ASSERT_GT(filter_bytes_insert, 0);
|
||||
ASSERT_EQ(cache->GetUsage(), index_bytes_insert + filter_bytes_insert);
|
||||
// set the cache capacity to the current usage
|
||||
cache->SetCapacity(index_bytes_insert + filter_bytes_insert);
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_EVICT), 0);
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_EVICT), 0);
|
||||
ASSERT_OK(Put(1, "key2", "val"));
|
||||
// Create a new table
|
||||
ASSERT_OK(Flush(1));
|
||||
// cache evicted old index and block entries
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_INSERT),
|
||||
index_bytes_insert);
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_INSERT),
|
||||
filter_bytes_insert);
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_EVICT),
|
||||
index_bytes_insert);
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_EVICT),
|
||||
filter_bytes_insert);
|
||||
}
|
||||
|
||||
namespace {
|
||||
|
||||
// A mock cache wraps LRUCache, and record how many entries have been
|
||||
// inserted for each priority.
|
||||
class MockCache : public LRUCache {
|
||||
public:
|
||||
static uint32_t high_pri_insert_count;
|
||||
static uint32_t low_pri_insert_count;
|
||||
|
||||
MockCache() : LRUCache(1 << 25, 0, false, 0.0) {}
|
||||
|
||||
virtual Status Insert(const Slice& key, void* value, size_t charge,
|
||||
void (*deleter)(const Slice& key, void* value),
|
||||
Handle** handle, Priority priority) override {
|
||||
if (priority == Priority::LOW) {
|
||||
low_pri_insert_count++;
|
||||
} else {
|
||||
high_pri_insert_count++;
|
||||
}
|
||||
return LRUCache::Insert(key, value, charge, deleter, handle, priority);
|
||||
}
|
||||
};
|
||||
|
||||
uint32_t MockCache::high_pri_insert_count = 0;
|
||||
uint32_t MockCache::low_pri_insert_count = 0;
|
||||
|
||||
} // anonymous namespace
|
||||
|
||||
TEST_F(DBBlockCacheTest, IndexAndFilterBlocksCachePriority) {
|
||||
for (auto priority : {Cache::Priority::LOW, Cache::Priority::HIGH}) {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.cache_index_and_filter_blocks = true;
|
||||
table_options.block_cache.reset(new MockCache());
|
||||
table_options.filter_policy.reset(NewBloomFilterPolicy(20));
|
||||
table_options.cache_index_and_filter_blocks_with_high_priority =
|
||||
priority == Cache::Priority::HIGH ? true : false;
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
DestroyAndReopen(options);
|
||||
|
||||
MockCache::high_pri_insert_count = 0;
|
||||
MockCache::low_pri_insert_count = 0;
|
||||
|
||||
// Create a new table.
|
||||
ASSERT_OK(Put("foo", "value"));
|
||||
ASSERT_OK(Put("bar", "value"));
|
||||
ASSERT_OK(Flush());
|
||||
ASSERT_EQ(1, NumTableFilesAtLevel(0));
|
||||
|
||||
// index/filter blocks added to block cache right after table creation.
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
|
||||
ASSERT_EQ(2, /* only index/filter were added */
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
ASSERT_EQ(0, TestGetTickerCount(options, BLOCK_CACHE_DATA_MISS));
|
||||
if (priority == Cache::Priority::LOW) {
|
||||
ASSERT_EQ(0, MockCache::high_pri_insert_count);
|
||||
ASSERT_EQ(2, MockCache::low_pri_insert_count);
|
||||
} else {
|
||||
ASSERT_EQ(2, MockCache::high_pri_insert_count);
|
||||
ASSERT_EQ(0, MockCache::low_pri_insert_count);
|
||||
}
|
||||
|
||||
// Access data block.
|
||||
ASSERT_EQ("value", Get("foo"));
|
||||
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
|
||||
ASSERT_EQ(3, /*adding data block*/
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_DATA_MISS));
|
||||
|
||||
// Data block should be inserted with low priority.
|
||||
if (priority == Cache::Priority::LOW) {
|
||||
ASSERT_EQ(0, MockCache::high_pri_insert_count);
|
||||
ASSERT_EQ(3, MockCache::low_pri_insert_count);
|
||||
} else {
|
||||
ASSERT_EQ(2, MockCache::high_pri_insert_count);
|
||||
ASSERT_EQ(1, MockCache::low_pri_insert_count);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
TEST_F(DBBlockCacheTest, ParanoidFileChecks) {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.paranoid_file_checks = true;
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.cache_index_and_filter_blocks = false;
|
||||
table_options.filter_policy.reset(NewBloomFilterPolicy(20));
|
||||
options.table_factory.reset(new BlockBasedTableFactory(table_options));
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
ASSERT_OK(Put(1, "1_key", "val"));
|
||||
ASSERT_OK(Put(1, "9_key", "val"));
|
||||
// Create a new table.
|
||||
ASSERT_OK(Flush(1));
|
||||
ASSERT_EQ(1, /* read and cache data block */
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
|
||||
ASSERT_OK(Put(1, "1_key2", "val2"));
|
||||
ASSERT_OK(Put(1, "9_key2", "val2"));
|
||||
// Create a new SST file. This will further trigger a compaction
|
||||
// and generate another file.
|
||||
ASSERT_OK(Flush(1));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ(3, /* Totally 3 files created up to now */
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
|
||||
// After disabling options.paranoid_file_checks. NO further block
|
||||
// is added after generating a new file.
|
||||
ASSERT_OK(
|
||||
dbfull()->SetOptions(handles_[1], {{"paranoid_file_checks", "false"}}));
|
||||
|
||||
ASSERT_OK(Put(1, "1_key3", "val3"));
|
||||
ASSERT_OK(Put(1, "9_key3", "val3"));
|
||||
ASSERT_OK(Flush(1));
|
||||
ASSERT_OK(Put(1, "1_key4", "val4"));
|
||||
ASSERT_OK(Put(1, "9_key4", "val4"));
|
||||
ASSERT_OK(Flush(1));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ(3, /* Totally 3 files created up to now */
|
||||
TestGetTickerCount(options, BLOCK_CACHE_ADD));
|
||||
}
|
||||
|
||||
TEST_F(DBBlockCacheTest, CompressedCache) {
|
||||
if (!Snappy_Supported()) {
|
||||
return;
|
||||
}
|
||||
int num_iter = 80;
|
||||
|
||||
// Run this test three iterations.
|
||||
// Iteration 1: only a uncompressed block cache
|
||||
// Iteration 2: only a compressed block cache
|
||||
// Iteration 3: both block cache and compressed cache
|
||||
// Iteration 4: both block cache and compressed cache, but DB is not
|
||||
// compressed
|
||||
for (int iter = 0; iter < 4; iter++) {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 64 * 1024; // small write buffer
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
|
||||
BlockBasedTableOptions table_options;
|
||||
switch (iter) {
|
||||
case 0:
|
||||
// only uncompressed block cache
|
||||
table_options.block_cache = NewLRUCache(8 * 1024);
|
||||
table_options.block_cache_compressed = nullptr;
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
break;
|
||||
case 1:
|
||||
// no block cache, only compressed cache
|
||||
table_options.no_block_cache = true;
|
||||
table_options.block_cache = nullptr;
|
||||
table_options.block_cache_compressed = NewLRUCache(8 * 1024);
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
break;
|
||||
case 2:
|
||||
// both compressed and uncompressed block cache
|
||||
table_options.block_cache = NewLRUCache(1024);
|
||||
table_options.block_cache_compressed = NewLRUCache(8 * 1024);
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
break;
|
||||
case 3:
|
||||
// both block cache and compressed cache, but DB is not compressed
|
||||
// also, make block cache sizes bigger, to trigger block cache hits
|
||||
table_options.block_cache = NewLRUCache(1024 * 1024);
|
||||
table_options.block_cache_compressed = NewLRUCache(8 * 1024 * 1024);
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
options.compression = kNoCompression;
|
||||
break;
|
||||
default:
|
||||
ASSERT_TRUE(false);
|
||||
}
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
// default column family doesn't have block cache
|
||||
Options no_block_cache_opts;
|
||||
no_block_cache_opts.statistics = options.statistics;
|
||||
no_block_cache_opts = CurrentOptions(no_block_cache_opts);
|
||||
BlockBasedTableOptions table_options_no_bc;
|
||||
table_options_no_bc.no_block_cache = true;
|
||||
no_block_cache_opts.table_factory.reset(
|
||||
NewBlockBasedTableFactory(table_options_no_bc));
|
||||
ReopenWithColumnFamilies(
|
||||
{"default", "pikachu"},
|
||||
std::vector<Options>({no_block_cache_opts, options}));
|
||||
|
||||
Random rnd(301);
|
||||
|
||||
// Write 8MB (80 values, each 100K)
|
||||
ASSERT_EQ(NumTableFilesAtLevel(0, 1), 0);
|
||||
std::vector<std::string> values;
|
||||
std::string str;
|
||||
for (int i = 0; i < num_iter; i++) {
|
||||
if (i % 4 == 0) { // high compression ratio
|
||||
str = RandomString(&rnd, 1000);
|
||||
}
|
||||
values.push_back(str);
|
||||
ASSERT_OK(Put(1, Key(i), values[i]));
|
||||
}
|
||||
|
||||
// flush all data from memtable so that reads are from block cache
|
||||
ASSERT_OK(Flush(1));
|
||||
|
||||
for (int i = 0; i < num_iter; i++) {
|
||||
ASSERT_EQ(Get(1, Key(i)), values[i]);
|
||||
}
|
||||
|
||||
// check that we triggered the appropriate code paths in the cache
|
||||
switch (iter) {
|
||||
case 0:
|
||||
// only uncompressed block cache
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
|
||||
break;
|
||||
case 1:
|
||||
// no block cache, only compressed cache
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
|
||||
break;
|
||||
case 2:
|
||||
// both compressed and uncompressed block cache
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
|
||||
break;
|
||||
case 3:
|
||||
// both compressed and uncompressed block cache
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_HIT), 0);
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
|
||||
// compressed doesn't have any hits since blocks are not compressed on
|
||||
// storage
|
||||
ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT), 0);
|
||||
break;
|
||||
default:
|
||||
ASSERT_TRUE(false);
|
||||
}
|
||||
|
||||
options.create_if_missing = true;
|
||||
DestroyAndReopen(options);
|
||||
}
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
File diff suppressed because it is too large.
Load diff
+23
-158
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -7,8 +7,8 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/db_test_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -47,24 +47,6 @@ class DeleteFilter : public CompactionFilter {
|
||||
virtual const char* Name() const override { return "DeleteFilter"; }
|
||||
};
|
||||
|
||||
class DeleteISFilter : public CompactionFilter {
|
||||
public:
|
||||
virtual bool Filter(int level, const Slice& key, const Slice& value,
|
||||
std::string* new_value,
|
||||
bool* value_changed) const override {
|
||||
cfilter_count++;
|
||||
int i = std::stoi(key.ToString());
|
||||
if (i > 5 && i <= 105) {
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
virtual bool IgnoreSnapshots() const override { return true; }
|
||||
|
||||
virtual const char* Name() const override { return "DeleteFilter"; }
|
||||
};
|
||||
|
||||
class DelayFilter : public CompactionFilter {
|
||||
public:
|
||||
explicit DelayFilter(DBTestBase* d) : db_test(d) {}
|
||||
@@ -115,11 +97,8 @@ class ChangeFilter : public CompactionFilter {
|
||||
|
||||
class KeepFilterFactory : public CompactionFilterFactory {
|
||||
public:
|
||||
explicit KeepFilterFactory(bool check_context = false,
|
||||
bool check_context_cf_id = false)
|
||||
: check_context_(check_context),
|
||||
check_context_cf_id_(check_context_cf_id),
|
||||
compaction_filter_created_(false) {}
|
||||
explicit KeepFilterFactory(bool check_context = false)
|
||||
: check_context_(check_context) {}
|
||||
|
||||
virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
|
||||
const CompactionFilter::Context& context) override {
|
||||
@@ -127,22 +106,13 @@ class KeepFilterFactory : public CompactionFilterFactory {
|
||||
EXPECT_EQ(expect_full_compaction_.load(), context.is_full_compaction);
|
||||
EXPECT_EQ(expect_manual_compaction_.load(), context.is_manual_compaction);
|
||||
}
|
||||
if (check_context_cf_id_) {
|
||||
EXPECT_EQ(expect_cf_id_.load(), context.column_family_id);
|
||||
}
|
||||
compaction_filter_created_ = true;
|
||||
return std::unique_ptr<CompactionFilter>(new KeepFilter());
|
||||
}
|
||||
|
||||
bool compaction_filter_created() const { return compaction_filter_created_; }
|
||||
|
||||
virtual const char* Name() const override { return "KeepFilterFactory"; }
|
||||
bool check_context_;
|
||||
bool check_context_cf_id_;
|
||||
std::atomic_bool expect_full_compaction_;
|
||||
std::atomic_bool expect_manual_compaction_;
|
||||
std::atomic<uint32_t> expect_cf_id_;
|
||||
bool compaction_filter_created_;
|
||||
};
|
||||
|
||||
class DeleteFilterFactory : public CompactionFilterFactory {
|
||||
@@ -159,21 +129,6 @@ class DeleteFilterFactory : public CompactionFilterFactory {
|
||||
virtual const char* Name() const override { return "DeleteFilterFactory"; }
|
||||
};
|
||||
|
||||
// Delete Filter Factory which ignores snapshots
|
||||
class DeleteISFilterFactory : public CompactionFilterFactory {
|
||||
public:
|
||||
virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
|
||||
const CompactionFilter::Context& context) override {
|
||||
if (context.is_manual_compaction) {
|
||||
return std::unique_ptr<CompactionFilter>(new DeleteISFilter());
|
||||
} else {
|
||||
return std::unique_ptr<CompactionFilter>(nullptr);
|
||||
}
|
||||
}
|
||||
|
||||
virtual const char* Name() const override { return "DeleteFilterFactory"; }
|
||||
};
|
||||
|
||||
class DelayFilterFactory : public CompactionFilterFactory {
|
||||
public:
|
||||
explicit DelayFilterFactory(DBTestBase* d) : db_test(d) {}
|
||||
@@ -219,7 +174,6 @@ class ChangeFilterFactory : public CompactionFilterFactory {
|
||||
virtual const char* Name() const override { return "ChangeFilterFactory"; }
|
||||
};
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilter) {
|
||||
Options options = CurrentOptions();
|
||||
options.max_open_files = -1;
|
||||
@@ -262,7 +216,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
|
||||
Arena arena;
|
||||
{
|
||||
ScopedArenaIterator iter(
|
||||
dbfull()->NewInternalIterator(&arena, handles_[1]));
|
||||
dbfull()->TEST_NewInternalIterator(&arena, handles_[1]));
|
||||
iter->SeekToFirst();
|
||||
ASSERT_OK(iter->status());
|
||||
while (iter->Valid()) {
|
||||
@@ -350,7 +304,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
|
||||
count = 0;
|
||||
{
|
||||
ScopedArenaIterator iter(
|
||||
dbfull()->NewInternalIterator(&arena, handles_[1]));
|
||||
dbfull()->TEST_NewInternalIterator(&arena, handles_[1]));
|
||||
iter->SeekToFirst();
|
||||
ASSERT_OK(iter->status());
|
||||
while (iter->Valid()) {
|
||||
@@ -368,10 +322,11 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
|
||||
// entries are deleted. The compaction should create bunch of 'DeleteFile'
|
||||
// entries in VersionEdit, but none of the 'AddFile's.
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterDeletesAll) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_filter_factory = std::make_shared<DeleteFilterFactory>();
|
||||
options.disable_auto_compactions = true;
|
||||
options.create_if_missing = true;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// put some data
|
||||
@@ -395,14 +350,14 @@ TEST_F(DBTestCompactionFilter, CompactionFilterDeletesAll) {
|
||||
|
||||
delete itr;
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterWithValueChange) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.num_levels = 3;
|
||||
options.compaction_filter_factory =
|
||||
std::make_shared<ChangeFilterFactory>();
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Write 100K+1 keys, these are written to a few files
|
||||
@@ -465,7 +420,8 @@ TEST_F(DBTestCompactionFilter, CompactionFilterWithMergeOperator) {
|
||||
PutFixed64(&three, 3);
|
||||
PutFixed64(&four, 4);
|
||||
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options = CurrentOptions(options);
|
||||
options.create_if_missing = true;
|
||||
options.merge_operator = MergeOperators::CreateUInt64AddOperator();
|
||||
options.num_levels = 3;
|
||||
@@ -525,9 +481,8 @@ TEST_F(DBTestCompactionFilter, CompactionFilterWithMergeOperator) {
|
||||
ASSERT_EQ(newvalue, four);
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
|
||||
KeepFilterFactory* filter = new KeepFilterFactory(true, true);
|
||||
KeepFilterFactory* filter = new KeepFilterFactory();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
@@ -549,24 +504,22 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
|
||||
// be triggered.
|
||||
num_keys_per_file /= 2;
|
||||
}
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
// Force a manual compaction
|
||||
cfilter_count = 0;
|
||||
filter->expect_manual_compaction_.store(true);
|
||||
filter->expect_full_compaction_.store(true);
|
||||
filter->expect_cf_id_.store(0);
|
||||
filter->expect_full_compaction_.store(false); // Manual compaction always
|
||||
// set this flag.
|
||||
dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
|
||||
ASSERT_EQ(cfilter_count, 700);
|
||||
ASSERT_EQ(NumSortedRuns(0), 1);
|
||||
ASSERT_TRUE(filter->compaction_filter_created());
|
||||
|
||||
// Verify total number of keys is correct after manual compaction.
|
||||
{
|
||||
int count = 0;
|
||||
int total = 0;
|
||||
Arena arena;
|
||||
ScopedArenaIterator iter(dbfull()->NewInternalIterator(&arena));
|
||||
ScopedArenaIterator iter(dbfull()->TEST_NewInternalIterator(&arena));
|
||||
iter->SeekToFirst();
|
||||
ASSERT_OK(iter->status());
|
||||
while (iter->Valid()) {
|
||||
@@ -580,48 +533,18 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
|
||||
iter->Next();
|
||||
}
|
||||
ASSERT_EQ(total, 700);
|
||||
ASSERT_EQ(count, 2);
|
||||
ASSERT_EQ(count, 1);
|
||||
}
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterContextCfId) {
|
||||
KeepFilterFactory* filter = new KeepFilterFactory(false, true);
|
||||
filter->expect_cf_id_.store(1);
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_filter_factory.reset(filter);
|
||||
options.compression = kNoCompression;
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
int num_keys_per_file = 400;
|
||||
for (int j = 0; j < 3; j++) {
|
||||
// Write several keys.
|
||||
const std::string value(10, 'x');
|
||||
for (int i = 0; i < num_keys_per_file; i++) {
|
||||
char key[100];
|
||||
snprintf(key, sizeof(key), "B%08d%02d", i, j);
|
||||
Put(1, key, value);
|
||||
}
|
||||
Flush(1);
|
||||
// Make sure next file is much smaller so automatic compaction will not
|
||||
// be triggered.
|
||||
num_keys_per_file /= 2;
|
||||
}
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
ASSERT_TRUE(filter->compaction_filter_created());
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
// Compaction filters should only be applied to records that are newer than the
|
||||
// latest snapshot. This test inserts records and applies a delete filter.
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterSnapshot) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_filter_factory = std::make_shared<DeleteFilterFactory>();
|
||||
options.disable_auto_compactions = true;
|
||||
options.create_if_missing = true;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Put some data.
|
||||
@@ -650,72 +573,14 @@ TEST_F(DBTestCompactionFilter, CompactionFilterSnapshot) {
|
||||
ASSERT_EQ(0U, CountLiveFiles());
|
||||
}
|
||||
|
||||
// Compaction filters should only be applied to records that are newer than the
|
||||
// latest snapshot. However, if the compaction filter asks to ignore snapshots
|
||||
// records newer than the snapshot will also be processed
|
||||
TEST_F(DBTestCompactionFilter, CompactionFilterIgnoreSnapshot) {
|
||||
std::string five = ToString(5);
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_filter_factory = std::make_shared<DeleteISFilterFactory>();
|
||||
options.disable_auto_compactions = true;
|
||||
options.create_if_missing = true;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Put some data.
|
||||
const Snapshot* snapshot = nullptr;
|
||||
for (int table = 0; table < 4; ++table) {
|
||||
for (int i = 0; i < 10; ++i) {
|
||||
Put(ToString(table * 100 + i), "val");
|
||||
}
|
||||
Flush();
|
||||
|
||||
if (table == 0) {
|
||||
snapshot = db_->GetSnapshot();
|
||||
}
|
||||
}
|
||||
assert(snapshot != nullptr);
|
||||
|
||||
cfilter_count = 0;
|
||||
ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
// The filter should delete 40 records.
|
||||
ASSERT_EQ(40U, cfilter_count);
|
||||
|
||||
{
|
||||
// Scan the entire database as of the snapshot to ensure
|
||||
// that nothing is left
|
||||
ReadOptions read_options;
|
||||
read_options.snapshot = snapshot;
|
||||
std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
|
||||
iter->SeekToFirst();
|
||||
int count = 0;
|
||||
while (iter->Valid()) {
|
||||
count++;
|
||||
iter->Next();
|
||||
}
|
||||
ASSERT_EQ(count, 6);
|
||||
read_options.snapshot = 0;
|
||||
std::unique_ptr<Iterator> iter1(db_->NewIterator(read_options));
|
||||
iter1->SeekToFirst();
|
||||
count = 0;
|
||||
while (iter1->Valid()) {
|
||||
count++;
|
||||
iter1->Next();
|
||||
}
|
||||
// We have deleted 10 keys from 40 using the compaction filter
|
||||
// Keys 6-9 before the snapshot and 100-105 after the snapshot
|
||||
ASSERT_EQ(count, 30);
|
||||
}
|
||||
|
||||
// Release the snapshot and compact again -> now all records should be
|
||||
// removed.
|
||||
db_->ReleaseSnapshot(snapshot);
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
#else
|
||||
return 0;
|
||||
#endif
|
||||
}
|
||||
+78
-777
File diff suppressed because it is too large.
Load diff
+37
-48
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,10 +10,10 @@
|
||||
// Introduction of SyncPoint effectively disabled building and running this test
|
||||
// in Release build.
|
||||
// which is a pity, it is a good test
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/db_test_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
class DBTestDynamicLevel : public DBTestBase {
|
||||
@@ -125,19 +125,19 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.db_write_buffer_size = 204800;
|
||||
options.write_buffer_size = 20480;
|
||||
options.db_write_buffer_size = 2048;
|
||||
options.write_buffer_size = 2048;
|
||||
options.max_write_buffer_number = 2;
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.level0_slowdown_writes_trigger = 9999;
|
||||
options.level0_stop_writes_trigger = 9999;
|
||||
options.target_file_size_base = 9102;
|
||||
options.target_file_size_base = 2048;
|
||||
options.level_compaction_dynamic_level_bytes = true;
|
||||
options.max_bytes_for_level_base = 40960;
|
||||
options.max_bytes_for_level_base = 10240;
|
||||
options.max_bytes_for_level_multiplier = 4;
|
||||
options.max_background_compactions = 2;
|
||||
options.num_levels = 5;
|
||||
options.max_compaction_bytes = 0; // Force not expanding in compactions
|
||||
options.expanded_compaction_factor = 0; // Force not expanding in compactions
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.block_size = 1024;
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
@@ -154,10 +154,10 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
ASSERT_TRUE(db_->GetIntProperty("rocksdb.base-level", &int_prop));
|
||||
ASSERT_EQ(4U, int_prop);
|
||||
|
||||
// Put about 28K to L0
|
||||
// Put about 7K to L0
|
||||
for (int i = 0; i < 70; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 380)));
|
||||
RandomString(&rnd, 80)));
|
||||
}
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "false"},
|
||||
@@ -167,14 +167,14 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
ASSERT_TRUE(db_->GetIntProperty("rocksdb.base-level", &int_prop));
|
||||
ASSERT_EQ(4U, int_prop);
|
||||
|
||||
// Insert extra about 28K to L0. After they are compacted to L4, base level
|
||||
// Insert extra about 3.5K to L0. After they are compacted to L4, base level
|
||||
// should be changed to L3.
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "true"},
|
||||
}));
|
||||
for (int i = 0; i < 70; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 380)));
|
||||
RandomString(&rnd, 80)));
|
||||
}
|
||||
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
@@ -199,10 +199,10 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "true"},
|
||||
}));
|
||||
// Write about 40K more
|
||||
// Write about 10K more
|
||||
for (int i = 0; i < 100; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 380)));
|
||||
RandomString(&rnd, 80)));
|
||||
}
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "false"},
|
||||
@@ -218,16 +218,15 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
|
||||
// Trigger a condition that the compaction changes base level and L0->Lbase
|
||||
// happens at the same time.
|
||||
// We try to make last levels' targets to be 40K, 160K, 640K, add triggers
|
||||
// We try to make last levels' targets to be 10K, 40K, 160K, add triggers
|
||||
// another compaction from 40K->160K.
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "true"},
|
||||
}));
|
||||
// Write about 650K more.
|
||||
// Each file is about 11KB, with 9KB of data.
|
||||
for (int i = 0; i < 1300; i++) {
|
||||
// Write about 150K more
|
||||
for (int i = 0; i < 1350; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 380)));
|
||||
RandomString(&rnd, 80)));
|
||||
}
|
||||
ASSERT_OK(dbfull()->SetOptions({
|
||||
{"disable_auto_compactions", "false"},
|
||||
@@ -237,40 +236,30 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
|
||||
ASSERT_TRUE(db_->GetIntProperty("rocksdb.base-level", &int_prop));
|
||||
ASSERT_EQ(2U, int_prop);
|
||||
|
||||
// A manual compaction will trigger the base level to become L2
|
||||
// Keep Writing data until base level changed 2->1. There will be L0->L2
|
||||
// compaction going on at the same time.
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency({
|
||||
{"CompactionJob::Run():Start", "DynamicLevelMaxBytesBase2:0"},
|
||||
{"DynamicLevelMaxBytesBase2:1", "CompactionJob::Run():End"},
|
||||
{"DynamicLevelMaxBytesBase2:compact_range_finish",
|
||||
"FlushJob::WriteLevel0Table"},
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
for (int attempt = 0; attempt <= 20; attempt++) {
|
||||
// Write about 5K more data with two flushes. It should be flush to level 2
|
||||
// but when it is applied, base level is already 1.
|
||||
for (int i = 0; i < 50; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 80)));
|
||||
}
|
||||
Flush();
|
||||
|
||||
std::thread thread([this] {
|
||||
TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:compact_range_start");
|
||||
ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:compact_range_finish");
|
||||
});
|
||||
|
||||
TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:0");
|
||||
for (int i = 0; i < 2; i++) {
|
||||
ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
|
||||
RandomString(&rnd, 380)));
|
||||
ASSERT_TRUE(db_->GetIntProperty("rocksdb.base-level", &int_prop));
|
||||
if (int_prop == 2U) {
|
||||
env_->SleepForMicroseconds(50000);
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:1");
|
||||
|
||||
Flush();
|
||||
|
||||
thread.join();
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
|
||||
|
||||
env_->SleepForMicroseconds(200000);
|
||||
|
||||
ASSERT_TRUE(db_->GetIntProperty("rocksdb.base-level", &int_prop));
|
||||
ASSERT_EQ(1U, int_prop);
|
||||
}
|
||||
@@ -295,7 +284,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesCompactRange) {
|
||||
options.max_background_compactions = 1;
|
||||
const int kNumLevels = 5;
|
||||
options.num_levels = kNumLevels;
|
||||
options.max_compaction_bytes = 1; // Force not expanding in compactions
|
||||
options.expanded_compaction_factor = 0; // Force not expanding in compactions
|
||||
BlockBasedTableOptions table_options;
|
||||
table_options.block_size = 1024;
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
|
||||
@@ -495,10 +484,10 @@ TEST_F(DBTestDynamicLevel, MigrateToDynamicLevelMaxBytesBase) {
|
||||
}
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // !defined(ROCKSDB_LITE)
|
||||
#endif // !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -14,20 +14,19 @@
|
||||
#endif
|
||||
|
||||
#include <inttypes.h>
|
||||
#include <stdint.h>
|
||||
#include <algorithm>
|
||||
#include <string>
|
||||
#include <stdint.h>
|
||||
#include "db/db_impl.h"
|
||||
#include "db/filename.h"
|
||||
#include "db/job_context.h"
|
||||
#include "db/version_set.h"
|
||||
#include "port/port.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "util/file_util.h"
|
||||
#include "port/port.h"
|
||||
#include "util/mutexlock.h"
|
||||
#include "util/sync_point.h"
|
||||
#include "util/testharness.h"
|
||||
#include "util/file_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -84,6 +83,7 @@ int DBImpl::IsFileDeletionsEnabled() const {
|
||||
Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
|
||||
uint64_t* manifest_file_size,
|
||||
bool flush_memtable) {
|
||||
|
||||
*manifest_file_size = 0;
|
||||
|
||||
mutex_.Lock();
|
||||
@@ -126,7 +126,7 @@ Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
|
||||
}
|
||||
|
||||
ret.clear();
|
||||
ret.reserve(live.size() + 3); // *.sst + CURRENT + MANIFEST + OPTIONS
|
||||
ret.reserve(live.size() + 2); //*.sst + CURRENT + MANIFEST
|
||||
|
||||
// create names of the live files. The names are not absolute
|
||||
// paths, instead they are relative to dbname_;
|
||||
@@ -136,7 +136,6 @@ Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
|
||||
|
||||
ret.push_back(CurrentFileName(""));
|
||||
ret.push_back(DescriptorFileName("", versions_->manifest_file_number()));
|
||||
ret.push_back(OptionsFileName("", versions_->options_file_number()));
|
||||
|
||||
// find length of manifest file while holding the mutex lock
|
||||
*manifest_file_size = versions_->manifest_file_size();
|
||||
|
||||
@@ -1,56 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/sync_point.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class DBFlushTest : public DBTestBase {
|
||||
public:
|
||||
DBFlushTest() : DBTestBase("/db_flush_test") {}
|
||||
};
|
||||
|
||||
// We had issue when two background threads trying to flush at the same time,
|
||||
// only one of them get committed. The test verifies the issue is fixed.
|
||||
TEST_F(DBFlushTest, FlushWhileWritingManifest) {
|
||||
Options options;
|
||||
options.disable_auto_compactions = true;
|
||||
options.max_background_flushes = 2;
|
||||
Reopen(options);
|
||||
FlushOptions no_wait;
|
||||
no_wait.wait = false;
|
||||
|
||||
SyncPoint::GetInstance()->LoadDependency(
|
||||
{{"VersionSet::LogAndApply:WriteManifest",
|
||||
"DBFlushTest::FlushWhileWritingManifest:1"},
|
||||
{"MemTableList::InstallMemtableFlushResults:InProgress",
|
||||
"VersionSet::LogAndApply:WriteManifestDone"}});
|
||||
SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
ASSERT_OK(Put("foo", "v"));
|
||||
ASSERT_OK(dbfull()->Flush(no_wait));
|
||||
TEST_SYNC_POINT("DBFlushTest::FlushWhileWritingManifest:1");
|
||||
ASSERT_OK(Put("bar", "v"));
|
||||
ASSERT_OK(dbfull()->Flush(no_wait));
|
||||
// If the issue is hit we will wait here forever.
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
#ifndef ROCKSDB_LITE
|
||||
ASSERT_EQ(2, TotalTableFiles());
|
||||
#endif // ROCKSDB_LITE
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
+778
-2025
File diff suppressed because it is too large.
Load diff
+61
-344
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,10 +10,8 @@
|
||||
|
||||
#include <atomic>
|
||||
#include <deque>
|
||||
#include <functional>
|
||||
#include <limits>
|
||||
#include <list>
|
||||
#include <queue>
|
||||
#include <set>
|
||||
#include <string>
|
||||
#include <utility>
|
||||
@@ -31,18 +29,18 @@
|
||||
#include "db/wal_manager.h"
|
||||
#include "db/write_controller.h"
|
||||
#include "db/write_thread.h"
|
||||
#include "db/writebuffer.h"
|
||||
#include "memtable_list.h"
|
||||
#include "port/port.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/memtablerep.h"
|
||||
#include "rocksdb/transaction_log.h"
|
||||
#include "rocksdb/write_buffer_manager.h"
|
||||
#include "table/scoped_arena_iterator.h"
|
||||
#include "util/autovector.h"
|
||||
#include "util/event_logger.h"
|
||||
#include "util/hash.h"
|
||||
#include "util/instrumented_mutex.h"
|
||||
#include "util/scoped_arena_iterator.h"
|
||||
#include "util/stop_watch.h"
|
||||
#include "util/thread_local.h"
|
||||
|
||||
@@ -57,7 +55,6 @@ class Arena;
|
||||
class WriteCallback;
|
||||
struct JobContext;
|
||||
struct ExternalSstFileInfo;
|
||||
struct MemTableInfo;
|
||||
|
||||
class DBImpl : public DB {
|
||||
public:
|
||||
@@ -125,9 +122,6 @@ class DBImpl : public DB {
|
||||
using DB::GetIntProperty;
|
||||
virtual bool GetIntProperty(ColumnFamilyHandle* column_family,
|
||||
const Slice& property, uint64_t* value) override;
|
||||
using DB::GetAggregatedIntProperty;
|
||||
virtual bool GetAggregatedIntProperty(const Slice& property,
|
||||
uint64_t* aggregated_value) override;
|
||||
using DB::GetApproximateSizes;
|
||||
virtual void GetApproximateSizes(ColumnFamilyHandle* column_family,
|
||||
const Range* range, int n, uint64_t* sizes,
|
||||
@@ -147,9 +141,6 @@ class DBImpl : public DB {
|
||||
virtual Status PauseBackgroundWork() override;
|
||||
virtual Status ContinueBackgroundWork() override;
|
||||
|
||||
virtual Status EnableAutoCompaction(
|
||||
const std::vector<ColumnFamilyHandle*>& column_family_handles) override;
|
||||
|
||||
using DB::SetOptions;
|
||||
Status SetOptions(
|
||||
ColumnFamilyHandle* column_family,
|
||||
@@ -191,8 +182,6 @@ class DBImpl : public DB {
|
||||
const TransactionLogIterator::ReadOptions&
|
||||
read_options = TransactionLogIterator::ReadOptions()) override;
|
||||
virtual Status DeleteFile(std::string name) override;
|
||||
Status DeleteFilesInRange(ColumnFamilyHandle* column_family,
|
||||
const Slice* begin, const Slice* end);
|
||||
|
||||
virtual void GetLiveFilesMetaData(
|
||||
std::vector<LiveFileMetaData>* metadata) override;
|
||||
@@ -231,48 +220,23 @@ class DBImpl : public DB {
|
||||
bool include_history);
|
||||
|
||||
// For a given key, check to see if there are any records for this key
|
||||
// in the memtables, including memtable history. If cache_only is false,
|
||||
// SST files will also be checked.
|
||||
//
|
||||
// If a key is found, *found_record_for_key will be set to true and
|
||||
// *seq will be set to the stored sequence number for the latest
|
||||
// operation on this key or kMaxSequenceNumber if unknown.
|
||||
// If no key is found, *found_record_for_key will be set to false.
|
||||
//
|
||||
// Note: If cache_only=false, it is possible for *seq to be set to 0 if
|
||||
// the sequence number has been cleared from the record. If the caller is
|
||||
// holding an active db snapshot, we know the missing sequence must be less
|
||||
// than the snapshot's sequence number (sequence numbers are only cleared
|
||||
// when there are no earlier active snapshots).
|
||||
//
|
||||
// If NotFound is returned and found_record_for_key is set to false, then no
|
||||
// record for this key was found. If the caller is holding an active db
|
||||
// snapshot, we know that no key could have existing after this snapshot
|
||||
// (since we do not compact keys that have an earlier snapshot).
|
||||
//
|
||||
// Returns OK or NotFound on success,
|
||||
// other status on unexpected error.
|
||||
// TODO(andrewkr): this API need to be aware of range deletion operations
|
||||
Status GetLatestSequenceForKey(SuperVersion* sv, const Slice& key,
|
||||
bool cache_only, SequenceNumber* seq,
|
||||
bool* found_record_for_key);
|
||||
// in the memtables, including memtable history.
|
||||
|
||||
// On success, *seq will contain the sequence number for the
|
||||
// latest such change or kMaxSequenceNumber if no records were present.
|
||||
// Returns OK on success, other status on error reading memtables.
|
||||
Status GetLatestSequenceForKeyFromMemtable(SuperVersion* sv, const Slice& key,
|
||||
SequenceNumber* seq);
|
||||
|
||||
using DB::AddFile;
|
||||
virtual Status AddFile(ColumnFamilyHandle* column_family,
|
||||
const std::vector<ExternalSstFileInfo>& file_info_list,
|
||||
bool move_file, bool skip_snapshot_check) override;
|
||||
const ExternalSstFileInfo* file_info,
|
||||
bool move_file) override;
|
||||
virtual Status AddFile(ColumnFamilyHandle* column_family,
|
||||
const std::vector<std::string>& file_path_list,
|
||||
bool move_file, bool skip_snapshot_check) override;
|
||||
const std::string& file_path, bool move_file) override;
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Similar to GetSnapshot(), but also lets the db know that this snapshot
|
||||
// will be used for transaction write-conflict checking. The DB can then
|
||||
// make sure not to compact any keys that would prevent a write-conflict from
|
||||
// being detected.
|
||||
const Snapshot* GetSnapshotForWriteConflictBoundary();
|
||||
|
||||
// checks if all live files exist on file system and that their file sizes
|
||||
// match to our in-memory records
|
||||
virtual Status CheckConsistency();
|
||||
@@ -282,16 +246,9 @@ class DBImpl : public DB {
|
||||
Status RunManualCompaction(ColumnFamilyData* cfd, int input_level,
|
||||
int output_level, uint32_t output_path_id,
|
||||
const Slice* begin, const Slice* end,
|
||||
bool exclusive,
|
||||
bool disallow_trivial_move = false);
|
||||
|
||||
// Return an internal iterator over the current state of the database.
|
||||
// The keys of this iterator are internal keys (see format.h).
|
||||
// The returned iterator should be deleted when no longer needed.
|
||||
InternalIterator* NewInternalIterator(
|
||||
Arena* arena, ColumnFamilyHandle* column_family = nullptr);
|
||||
|
||||
#ifndef NDEBUG
|
||||
#ifndef ROCKSDB_LITE
|
||||
// Extra methods (for testing) that are not in the public DB interface
|
||||
// Implemented in db_impl_debug.cc
|
||||
|
||||
@@ -301,8 +258,7 @@ class DBImpl : public DB {
|
||||
bool disallow_trivial_move = false);
|
||||
|
||||
// Force current memtable contents to be flushed.
|
||||
Status TEST_FlushMemTable(bool wait = true,
|
||||
ColumnFamilyHandle* cfh = nullptr);
|
||||
Status TEST_FlushMemTable(bool wait = true);
|
||||
|
||||
// Wait for memtable compaction
|
||||
Status TEST_WaitForFlushMemTable(ColumnFamilyHandle* column_family = nullptr);
|
||||
@@ -310,6 +266,12 @@ class DBImpl : public DB {
|
||||
// Wait for any compaction
|
||||
Status TEST_WaitForCompact();
|
||||
|
||||
// Return an internal iterator over the current state of the database.
|
||||
// The keys of this iterator are internal keys (see format.h).
|
||||
// The returned iterator should be deleted when no longer needed.
|
||||
Iterator* TEST_NewInternalIterator(
|
||||
Arena* arena, ColumnFamilyHandle* column_family = nullptr);
|
||||
|
||||
// Return the maximum overlapping data (in bytes) at next level for any
|
||||
// file at a level >= 1.
|
||||
int64_t TEST_MaxNextLevelOverlappingBytes(ColumnFamilyHandle* column_family =
|
||||
@@ -343,32 +305,7 @@ class DBImpl : public DB {
|
||||
|
||||
uint64_t TEST_LogfileNumber();
|
||||
|
||||
uint64_t TEST_total_log_size() const { return total_log_size_; }
|
||||
|
||||
// Returns column family name to ImmutableCFOptions map.
|
||||
Status TEST_GetAllImmutableCFOptions(
|
||||
std::unordered_map<std::string, const ImmutableCFOptions*>* iopts_map);
|
||||
|
||||
// Return the lastest MutableCFOptions of of a column family
|
||||
Status TEST_GetLatestMutableCFOptions(ColumnFamilyHandle* column_family,
|
||||
MutableCFOptions* mutable_cf_opitons);
|
||||
|
||||
Cache* TEST_table_cache() { return table_cache_.get(); }
|
||||
|
||||
WriteController& TEST_write_controler() { return write_controller_; }
|
||||
|
||||
uint64_t TEST_FindMinLogContainingOutstandingPrep();
|
||||
uint64_t TEST_FindMinPrepLogReferencedByMemTable();
|
||||
|
||||
#endif // NDEBUG
|
||||
|
||||
// Return maximum background compaction allowed to be scheduled based on
|
||||
// compaction status.
|
||||
int BGCompactionsAllowed() const;
|
||||
|
||||
// move logs pending closing from job_context to the DB queue and
|
||||
// schedule a purge
|
||||
void ScheduleBgLogWriterClose(JobContext* job_context);
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Returns the list of live files in 'live' and the list
|
||||
// of all files in the filesystem in 'candidate_files'.
|
||||
@@ -382,10 +319,7 @@ class DBImpl : public DB {
|
||||
// belong to live files are posibly removed. Also, removes all the
|
||||
// files in sst_delete_files and log_delete_files.
|
||||
// It is not necessary to hold the mutex when invoking this method.
|
||||
void PurgeObsoleteFiles(const JobContext& background_contet,
|
||||
bool schedule_only = false);
|
||||
|
||||
void SchedulePurge();
|
||||
void PurgeObsoleteFiles(const JobContext& background_contet);
|
||||
|
||||
ColumnFamilyHandle* DefaultColumnFamily() const override;
|
||||
|
||||
@@ -429,116 +363,24 @@ class DBImpl : public DB {
|
||||
// Same as above, should called without mutex held and not on write thread.
|
||||
ColumnFamilyHandle* GetColumnFamilyHandleUnlocked(uint32_t column_family_id);
|
||||
|
||||
// Returns the number of currently running flushes.
|
||||
// REQUIREMENT: mutex_ must be held when calling this function.
|
||||
int num_running_flushes() {
|
||||
mutex_.AssertHeld();
|
||||
return num_running_flushes_;
|
||||
}
|
||||
|
||||
// Returns the number of currently running compactions.
|
||||
// REQUIREMENT: mutex_ must be held when calling this function.
|
||||
int num_running_compactions() {
|
||||
mutex_.AssertHeld();
|
||||
return num_running_compactions_;
|
||||
}
|
||||
|
||||
// hollow transactions shell used for recovery.
|
||||
// these will then be passed to TransactionDB so that
|
||||
// locks can be reacquired before writing can resume.
|
||||
struct RecoveredTransaction {
|
||||
uint64_t log_number_;
|
||||
std::string name_;
|
||||
WriteBatch* batch_;
|
||||
explicit RecoveredTransaction(const uint64_t log, const std::string& name,
|
||||
WriteBatch* batch)
|
||||
: log_number_(log), name_(name), batch_(batch) {}
|
||||
|
||||
~RecoveredTransaction() { delete batch_; }
|
||||
};
|
||||
|
||||
bool allow_2pc() const { return db_options_.allow_2pc; }
|
||||
|
||||
std::unordered_map<std::string, RecoveredTransaction*>
|
||||
recovered_transactions() {
|
||||
return recovered_transactions_;
|
||||
}
|
||||
|
||||
RecoveredTransaction* GetRecoveredTransaction(const std::string& name) {
|
||||
auto it = recovered_transactions_.find(name);
|
||||
if (it == recovered_transactions_.end()) {
|
||||
return nullptr;
|
||||
} else {
|
||||
return it->second;
|
||||
}
|
||||
}
|
||||
|
||||
void InsertRecoveredTransaction(const uint64_t log, const std::string& name,
|
||||
WriteBatch* batch) {
|
||||
recovered_transactions_[name] = new RecoveredTransaction(log, name, batch);
|
||||
MarkLogAsContainingPrepSection(log);
|
||||
}
|
||||
|
||||
void DeleteRecoveredTransaction(const std::string& name) {
|
||||
auto it = recovered_transactions_.find(name);
|
||||
assert(it != recovered_transactions_.end());
|
||||
auto* trx = it->second;
|
||||
recovered_transactions_.erase(it);
|
||||
MarkLogAsHavingPrepSectionFlushed(trx->log_number_);
|
||||
delete trx;
|
||||
}
|
||||
|
||||
void DeleteAllRecoveredTransactions() {
|
||||
for (auto it = recovered_transactions_.begin();
|
||||
it != recovered_transactions_.end(); it++) {
|
||||
delete it->second;
|
||||
}
|
||||
recovered_transactions_.clear();
|
||||
}
|
||||
|
||||
void MarkLogAsHavingPrepSectionFlushed(uint64_t log);
|
||||
void MarkLogAsContainingPrepSection(uint64_t log);
|
||||
void AddToLogsToFreeQueue(log::Writer* log_writer) {
|
||||
logs_to_free_queue_.push_back(log_writer);
|
||||
}
|
||||
|
||||
Status NewDB();
|
||||
|
||||
protected:
|
||||
Env* const env_;
|
||||
const std::string dbname_;
|
||||
unique_ptr<VersionSet> versions_;
|
||||
const DBOptions db_options_;
|
||||
Statistics* stats_;
|
||||
std::unordered_map<std::string, RecoveredTransaction*>
|
||||
recovered_transactions_;
|
||||
|
||||
InternalIterator* NewInternalIterator(const ReadOptions&,
|
||||
ColumnFamilyData* cfd,
|
||||
SuperVersion* super_version,
|
||||
Arena* arena);
|
||||
|
||||
// Except in DB::Open(), WriteOptionsFile can only be called when:
|
||||
// 1. WriteThread::Writer::EnterUnbatched() is used.
|
||||
// 2. db_mutex is held
|
||||
Status WriteOptionsFile();
|
||||
|
||||
// The following two functions can only be called when:
|
||||
// 1. WriteThread::Writer::EnterUnbatched() is used.
|
||||
// 2. db_mutex is NOT held
|
||||
Status RenameTempFileToOptionsFile(const std::string& file_name);
|
||||
Status DeleteObsoleteOptionsFiles();
|
||||
Iterator* NewInternalIterator(const ReadOptions&, ColumnFamilyData* cfd,
|
||||
SuperVersion* super_version, Arena* arena);
|
||||
|
||||
void NotifyOnFlushCompleted(ColumnFamilyData* cfd, FileMetaData* file_meta,
|
||||
const MutableCFOptions& mutable_cf_options,
|
||||
int job_id, TableProperties prop);
|
||||
int job_id);
|
||||
|
||||
void NotifyOnCompactionCompleted(ColumnFamilyData* cfd,
|
||||
Compaction *c, const Status &st,
|
||||
const CompactionJobStats& job_stats,
|
||||
int job_id);
|
||||
void NotifyOnMemTableSealed(ColumnFamilyData* cfd,
|
||||
const MemTableInfo& mem_table_info);
|
||||
|
||||
void NewThreadStatusCfInfo(ColumnFamilyData* cfd) const;
|
||||
|
||||
@@ -547,17 +389,11 @@ class DBImpl : public DB {
|
||||
void EraseThreadStatusDbInfo() const;
|
||||
|
||||
Status WriteImpl(const WriteOptions& options, WriteBatch* updates,
|
||||
WriteCallback* callback = nullptr,
|
||||
uint64_t* log_used = nullptr, uint64_t log_ref = 0,
|
||||
bool disable_memtable = false);
|
||||
|
||||
uint64_t FindMinLogContainingOutstandingPrep();
|
||||
uint64_t FindMinPrepLogReferencedByMemTable();
|
||||
WriteCallback* callback);
|
||||
|
||||
private:
|
||||
friend class DB;
|
||||
friend class InternalStats;
|
||||
friend class TransactionImpl;
|
||||
#ifndef ROCKSDB_LITE
|
||||
friend class ForwardIterator;
|
||||
#endif
|
||||
@@ -570,14 +406,13 @@ class DBImpl : public DB {
|
||||
|
||||
struct WriteContext;
|
||||
|
||||
struct PurgeFileInfo;
|
||||
Status NewDB();
|
||||
|
||||
// Recover the descriptor from persistent storage. May do a significant
|
||||
// amount of work to recover recently logged updates. Any changes to
|
||||
// be made to the descriptor are added to *edit.
|
||||
Status Recover(const std::vector<ColumnFamilyDescriptor>& column_families,
|
||||
bool read_only = false, bool error_if_log_file_exist = false,
|
||||
bool error_if_data_exists_in_logs = false);
|
||||
bool read_only = false, bool error_if_log_file_exist = false);
|
||||
|
||||
void MaybeIgnoreError(Status* s) const;
|
||||
|
||||
@@ -585,18 +420,13 @@ class DBImpl : public DB {
|
||||
|
||||
// Delete any unneeded files and stale in-memory entries.
|
||||
void DeleteObsoleteFiles();
|
||||
// Delete obsolete files and log status and information of file deletion
|
||||
void DeleteObsoleteFileImpl(Status file_deletion_status, int job_id,
|
||||
const std::string& fname, FileType type,
|
||||
uint64_t number, uint32_t path_id);
|
||||
|
||||
// Background process needs to call
|
||||
// auto x = CaptureCurrentFileNumberInPendingOutputs()
|
||||
// auto file_num = versions_->NewFileNumber();
|
||||
// <do something>
|
||||
// ReleaseFileNumberFromPendingOutputs(x)
|
||||
// This will protect any file with number `file_num` or greater from being
|
||||
// deleted while <do something> is running.
|
||||
// This will protect any temporary files created while <do something> is
|
||||
// executing from being deleted.
|
||||
// -----------
|
||||
// This function will capture current file number and append it to
|
||||
// pending_outputs_. This will prevent any background process to delete any
|
||||
@@ -609,8 +439,6 @@ class DBImpl : public DB {
|
||||
// and blocked by any other pending_outputs_ calls)
|
||||
void ReleaseFileNumberFromPendingOutputs(std::list<uint64_t>::iterator v);
|
||||
|
||||
Status SyncClosedLogs(JobContext* job_context);
|
||||
|
||||
// Flush the in-memory write buffer to storage. Switches to a new
|
||||
// log-file/memtable and writes a new descriptor iff successful.
|
||||
Status FlushMemTableToOutputFile(ColumnFamilyData* cfd,
|
||||
@@ -620,10 +448,10 @@ class DBImpl : public DB {
|
||||
|
||||
// REQUIRES: log_numbers are sorted in ascending order
|
||||
Status RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
|
||||
SequenceNumber* next_sequence, bool read_only);
|
||||
SequenceNumber* max_sequence, bool read_only);
|
||||
|
||||
// The following two methods are used to flush a memtable to
|
||||
// storage. The first one is used at database RecoveryTime (when the
|
||||
// storage. The first one is used atdatabase RecoveryTime (when the
|
||||
// database is opened) and is heavyweight because it holds the mutex
|
||||
// for the entire period. The second method WriteLevel0Table supports
|
||||
// concurrent flush memtables to storage.
|
||||
@@ -644,30 +472,15 @@ class DBImpl : public DB {
|
||||
// Wait for memtable flushed
|
||||
Status WaitForFlushMemTable(ColumnFamilyData* cfd);
|
||||
|
||||
void RecordFlushIOStats();
|
||||
void RecordCompactionIOStats();
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
// Finds the lowest level in the DB that the ingested file can be added to
|
||||
// REQUIRES: mutex_ held
|
||||
int PickLevelForIngestedFile(ColumnFamilyData* cfd,
|
||||
const ExternalSstFileInfo& file_info);
|
||||
|
||||
// Wait for current AddFile() calls to finish.
|
||||
// REQUIRES: mutex_ held
|
||||
void WaitForAddFile();
|
||||
|
||||
Status CompactFilesImpl(const CompactionOptions& compact_options,
|
||||
ColumnFamilyData* cfd, Version* version,
|
||||
const std::vector<std::string>& input_file_names,
|
||||
const int output_level, int output_path_id,
|
||||
JobContext* job_context, LogBuffer* log_buffer);
|
||||
|
||||
Status ReadExternalSstFileInfo(ColumnFamilyHandle* column_family,
|
||||
const std::string& file_path,
|
||||
ExternalSstFileInfo* file_info);
|
||||
|
||||
#else
|
||||
// AddFile is not supported in ROCKSDB_LITE so this function
|
||||
// will be no-op
|
||||
void WaitForAddFile() {}
|
||||
Status CompactFilesImpl(
|
||||
const CompactionOptions& compact_options, ColumnFamilyData* cfd,
|
||||
Version* version, const std::vector<std::string>& input_file_names,
|
||||
const int output_level, int output_path_id, JobContext* job_context,
|
||||
LogBuffer* log_buffer);
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
ColumnFamilyData* GetColumnFamilyDataByName(const std::string& cf_name);
|
||||
@@ -675,17 +488,12 @@ class DBImpl : public DB {
|
||||
void MaybeScheduleFlushOrCompaction();
|
||||
void SchedulePendingFlush(ColumnFamilyData* cfd);
|
||||
void SchedulePendingCompaction(ColumnFamilyData* cfd);
|
||||
void SchedulePendingPurge(std::string fname, FileType type, uint64_t number,
|
||||
uint32_t path_id, int job_id);
|
||||
static void BGWorkCompaction(void* arg);
|
||||
static void BGWorkCompaction(void* db);
|
||||
static void BGWorkFlush(void* db);
|
||||
static void BGWorkPurge(void* arg);
|
||||
static void UnscheduleCallback(void* arg);
|
||||
void BackgroundCallCompaction(void* arg);
|
||||
void BackgroundCallCompaction();
|
||||
void BackgroundCallFlush();
|
||||
void BackgroundCallPurge();
|
||||
Status BackgroundCompaction(bool* madeProgress, JobContext* job_context,
|
||||
LogBuffer* log_buffer, void* m = 0);
|
||||
LogBuffer* log_buffer);
|
||||
Status BackgroundFlush(bool* madeProgress, JobContext* job_context,
|
||||
LogBuffer* log_buffer);
|
||||
|
||||
@@ -713,35 +521,25 @@ class DBImpl : public DB {
|
||||
// helper function to call after some of the logs_ were synced
|
||||
void MarkLogsSynced(uint64_t up_to, bool synced_dir, const Status& status);
|
||||
|
||||
const Snapshot* GetSnapshotImpl(bool is_write_conflict_boundary);
|
||||
|
||||
// table_cache_ provides its own synchronization
|
||||
std::shared_ptr<Cache> table_cache_;
|
||||
|
||||
// Lock over the persistent DB state. Non-nullptr iff successfully acquired.
|
||||
FileLock* db_lock_;
|
||||
|
||||
// The mutex for options file related operations.
|
||||
// NOTE: should never acquire options_file_mutex_ and mutex_ at the
|
||||
// same time.
|
||||
InstrumentedMutex options_files_mutex_;
|
||||
// State below is protected by mutex_
|
||||
InstrumentedMutex mutex_;
|
||||
|
||||
std::atomic<bool> shutting_down_;
|
||||
// This condition variable is signaled on these conditions:
|
||||
// * whenever bg_compaction_scheduled_ goes down to 0
|
||||
// * if AnyManualCompaction, whenever a compaction finishes, even if it hasn't
|
||||
// * if bg_manual_only_ > 0, whenever a compaction finishes, even if it hasn't
|
||||
// made any progress
|
||||
// * whenever a compaction made any progress
|
||||
// * whenever bg_flush_scheduled_ or bg_purge_scheduled_ value decreases
|
||||
// (i.e. whenever a flush is done, even if it didn't make any progress)
|
||||
// * whenever there is an error in background purge, flush or compaction
|
||||
// * whenever num_running_addfile_ goes to 0.
|
||||
// * whenever bg_flush_scheduled_ value decreases (i.e. whenever a flush is
|
||||
// done, even if it didn't make any progress)
|
||||
// * whenever there is an error in background flush or compaction
|
||||
InstrumentedCondVar bg_cv_;
|
||||
uint64_t logfile_number_;
|
||||
std::deque<uint64_t>
|
||||
log_recycle_files; // a list of log files that we can recycle
|
||||
bool log_dir_synced_;
|
||||
bool log_empty_;
|
||||
ColumnFamilyHandleImpl* default_cf_handle_;
|
||||
@@ -830,7 +628,7 @@ class DBImpl : public DB {
|
||||
|
||||
Directories directories_;
|
||||
|
||||
WriteBufferManager* write_buffer_manager_;
|
||||
WriteBuffer write_buffer_;
|
||||
|
||||
WriteThread write_thread_;
|
||||
|
||||
@@ -857,19 +655,6 @@ class DBImpl : public DB {
|
||||
// State is protected with db mutex.
|
||||
std::list<uint64_t> pending_outputs_;
|
||||
|
||||
// PurgeFileInfo is a structure to hold information of files to be deleted in
|
||||
// purge_queue_
|
||||
struct PurgeFileInfo {
|
||||
std::string fname;
|
||||
FileType type;
|
||||
uint64_t number;
|
||||
uint32_t path_id;
|
||||
int job_id;
|
||||
PurgeFileInfo(std::string fn, FileType t, uint64_t num, uint32_t pid,
|
||||
int jid)
|
||||
: fname(fn), type(t), number(num), path_id(pid), job_id(jid) {}
|
||||
};
|
||||
|
||||
// flush_queue_ and compaction_queue_ hold column families that we need to
|
||||
// flush and compact, respectively.
|
||||
// A column family is inserted into flush_queue_ when it satisfies condition
|
||||
@@ -894,55 +679,35 @@ class DBImpl : public DB {
|
||||
// invariant(column family present in compaction_queue_ <==>
|
||||
// ColumnFamilyData::pending_compaction_ == true)
|
||||
std::deque<ColumnFamilyData*> compaction_queue_;
|
||||
|
||||
// A queue to store filenames of the files to be purged
|
||||
std::deque<PurgeFileInfo> purge_queue_;
|
||||
|
||||
// A queue to store log writers to close
|
||||
std::deque<log::Writer*> logs_to_free_queue_;
|
||||
int unscheduled_flushes_;
|
||||
int unscheduled_compactions_;
|
||||
|
||||
// count how many background compactions are running or have been scheduled
|
||||
int bg_compaction_scheduled_;
|
||||
|
||||
// stores the number of compactions are currently running
|
||||
int num_running_compactions_;
|
||||
// If non-zero, MaybeScheduleFlushOrCompaction() will only schedule manual
|
||||
// compactions (if manual_compaction_ is not null). This mechanism enables
|
||||
// manual compactions to wait until all other compactions are finished.
|
||||
int bg_manual_only_;
|
||||
|
||||
// number of background memtable flush jobs, submitted to the HIGH pool
|
||||
int bg_flush_scheduled_;
|
||||
|
||||
// stores the number of flushes are currently running
|
||||
int num_running_flushes_;
|
||||
|
||||
// number of background obsolete file purge jobs, submitted to the HIGH pool
|
||||
int bg_purge_scheduled_;
|
||||
|
||||
// Information for a manual compaction
|
||||
struct ManualCompaction {
|
||||
ColumnFamilyData* cfd;
|
||||
int input_level;
|
||||
int output_level;
|
||||
uint32_t output_path_id;
|
||||
Status status;
|
||||
bool done;
|
||||
Status status;
|
||||
bool in_progress; // compaction request being processed?
|
||||
bool incomplete; // only part of requested range compacted
|
||||
bool exclusive; // current behavior of only one manual
|
||||
bool disallow_trivial_move; // Force actual compaction to run
|
||||
const InternalKey* begin; // nullptr means beginning of key range
|
||||
const InternalKey* end; // nullptr means end of key range
|
||||
InternalKey* manual_end; // how far we are compacting
|
||||
InternalKey tmp_storage; // Used to keep track of compaction progress
|
||||
InternalKey tmp_storage1; // Used to keep track of compaction progress
|
||||
Compaction* compaction;
|
||||
};
|
||||
std::deque<ManualCompaction*> manual_compaction_dequeue_;
|
||||
|
||||
struct CompactionArg {
|
||||
DBImpl* db;
|
||||
ManualCompaction* m;
|
||||
bool disallow_trivial_move; // Force actual compaction to run
|
||||
};
|
||||
ManualCompaction* manual_compaction_;
|
||||
|
||||
// Have we encountered a background error in paranoid mode?
|
||||
Status bg_error_;
|
||||
@@ -965,10 +730,7 @@ class DBImpl : public DB {
|
||||
// they're unique
|
||||
std::atomic<int> next_job_id_;
|
||||
|
||||
// A flag indicating whether the current rocksdb database has any
|
||||
// data that is not yet persisted into either WAL or SST file.
|
||||
// Used when disableWAL is true.
|
||||
bool has_unpersisted_data_;
|
||||
bool flush_on_destroy_; // Used when disableWAL is true.
|
||||
|
||||
static const int KEEP_LOG_FILE_NUM = 1000;
|
||||
// MSVC version 1800 still does not have constexpr for ::max()
|
||||
@@ -979,14 +741,6 @@ class DBImpl : public DB {
|
||||
// The options to access storage files
|
||||
const EnvOptions env_options_;
|
||||
|
||||
// A set of compactions that are running right now
|
||||
// REQUIRES: mutex held
|
||||
std::unordered_set<Compaction*> running_compactions_;
|
||||
|
||||
// Number of running AddFile() calls.
|
||||
// REQUIRES: mutex held
|
||||
int num_running_addfile_;
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
WalManager wal_manager_;
|
||||
#endif // ROCKSDB_LITE
|
||||
@@ -994,40 +748,15 @@ class DBImpl : public DB {
|
||||
// Unified interface for logging events
|
||||
EventLogger event_logger_;
|
||||
|
||||
// A value of > 0 temporarily disables scheduling of background work
|
||||
// A value of >0 temporarily disables scheduling of background work
|
||||
int bg_work_paused_;
|
||||
|
||||
// A value of > 0 temporarily disables scheduling of background compaction
|
||||
int bg_compaction_paused_;
|
||||
|
||||
// Guard against multiple concurrent refitting
|
||||
bool refitting_level_;
|
||||
|
||||
// Indicate DB was opened successfully
|
||||
bool opened_successfully_;
|
||||
|
||||
// minmum log number still containing prepared data.
|
||||
// this is used by FindObsoleteFiles to determine which
|
||||
// flushed logs we must keep around because they still
|
||||
// contain prepared data which has not been flushed or rolled back
|
||||
std::priority_queue<uint64_t, std::vector<uint64_t>, std::greater<uint64_t>>
|
||||
min_log_with_prep_;
|
||||
|
||||
// to be used in conjunction with min_log_with_prep_.
|
||||
// once a transaction with data in log L is committed or rolled back
|
||||
// rather than removing the value from the heap we add that value
|
||||
// to prepared_section_completed_ which maps LOG -> instance_count
|
||||
// since a log could contain multiple prepared sections
|
||||
//
|
||||
// when trying to determine the minmum log still active we first
|
||||
// consult min_log_with_prep_. while that root value maps to
|
||||
// a value > 0 in prepared_section_completed_ we decrement the
|
||||
// instance_count for that log and pop the root value in
|
||||
// min_log_with_prep_. This will work the same as a min_heap
|
||||
// where we are deleteing arbitrary elements and the up heaping.
|
||||
std::unordered_map<uint64_t, uint64_t> prepared_section_completed_;
|
||||
std::mutex prep_heap_mutex_;
|
||||
|
||||
// No copying allowed
|
||||
DBImpl(const DBImpl&);
|
||||
void operator=(const DBImpl&);
|
||||
@@ -1059,10 +788,6 @@ class DBImpl : public DB {
|
||||
virtual Status GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
|
||||
TablePropertiesCollection* props)
|
||||
override;
|
||||
virtual Status GetPropertiesOfTablesInRange(
|
||||
ColumnFamilyHandle* column_family, const Range* range, std::size_t n,
|
||||
TablePropertiesCollection* props) override;
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Function that Get and KeyMayExist call with no_io true or false
|
||||
@@ -1071,17 +796,9 @@ class DBImpl : public DB {
|
||||
const Slice& key, std::string* value,
|
||||
bool* value_found = nullptr);
|
||||
|
||||
bool GetIntPropertyInternal(ColumnFamilyData* cfd,
|
||||
const DBPropertyInfo& property_info,
|
||||
bool is_locked, uint64_t* value);
|
||||
|
||||
bool HasPendingManualCompaction();
|
||||
bool HasExclusiveManualCompaction();
|
||||
void AddManualCompaction(ManualCompaction* m);
|
||||
void RemoveManualCompaction(ManualCompaction* m);
|
||||
bool ShouldntRunManualCompaction(ManualCompaction* m);
|
||||
bool HaveManualCompaction(ColumnFamilyData* cfd);
|
||||
bool MCOverlap(ManualCompaction* m, ManualCompaction* m1);
|
||||
bool GetIntPropertyInternal(ColumnFamilyHandle* column_family,
|
||||
DBPropertyType property_type,
|
||||
bool need_out_of_mutex, uint64_t* value);
|
||||
};
|
||||
|
||||
// Sanitize db options. The caller should delete result.info_log if
|
||||
|
||||
@@ -1,417 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
#include "db/db_impl.h"
|
||||
|
||||
#ifndef __STDC_FORMAT_MACROS
|
||||
#define __STDC_FORMAT_MACROS
|
||||
#endif
|
||||
|
||||
#include <inttypes.h>
|
||||
|
||||
#include "db/builder.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/sst_file_writer.h"
|
||||
#include "table/table_builder.h"
|
||||
#include "util/file_reader_writer.h"
|
||||
#include "util/file_util.h"
|
||||
#include "util/sync_point.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
Status DBImpl::ReadExternalSstFileInfo(ColumnFamilyHandle* column_family,
|
||||
const std::string& file_path,
|
||||
ExternalSstFileInfo* file_info) {
|
||||
Status status;
|
||||
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
|
||||
auto cfd = cfh->cfd();
|
||||
|
||||
file_info->file_path = file_path;
|
||||
status = env_->GetFileSize(file_path, &file_info->file_size);
|
||||
if (!status.ok()) {
|
||||
return status;
|
||||
}
|
||||
|
||||
// Access the file using TableReader to extract
|
||||
// version, number of entries, smallest user key, largest user key
|
||||
std::unique_ptr<RandomAccessFile> sst_file;
|
||||
status = env_->NewRandomAccessFile(file_path, &sst_file, env_options_);
|
||||
if (!status.ok()) {
|
||||
return status;
|
||||
}
|
||||
std::unique_ptr<RandomAccessFileReader> sst_file_reader;
|
||||
sst_file_reader.reset(new RandomAccessFileReader(std::move(sst_file)));
|
||||
|
||||
std::unique_ptr<TableReader> table_reader;
|
||||
status = cfd->ioptions()->table_factory->NewTableReader(
|
||||
TableReaderOptions(*cfd->ioptions(), env_options_,
|
||||
cfd->internal_comparator()),
|
||||
std::move(sst_file_reader), file_info->file_size, &table_reader);
|
||||
if (!status.ok()) {
|
||||
return status;
|
||||
}
|
||||
|
||||
// Get the external sst file version from table properties
|
||||
const UserCollectedProperties& user_collected_properties =
|
||||
table_reader->GetTableProperties()->user_collected_properties;
|
||||
UserCollectedProperties::const_iterator external_sst_file_version_iter =
|
||||
user_collected_properties.find(ExternalSstFilePropertyNames::kVersion);
|
||||
if (external_sst_file_version_iter == user_collected_properties.end()) {
|
||||
return Status::InvalidArgument("Generated table version not found");
|
||||
}
|
||||
|
||||
file_info->version =
|
||||
DecodeFixed32(external_sst_file_version_iter->second.c_str());
|
||||
if (file_info->version == 1) {
|
||||
// version 1 imply that all sequence numbers in table equal 0
|
||||
file_info->sequence_number = 0;
|
||||
} else {
|
||||
return Status::InvalidArgument("Generated table version is not supported");
|
||||
}
|
||||
// Get number of entries in table
|
||||
file_info->num_entries = table_reader->GetTableProperties()->num_entries;
|
||||
|
||||
ParsedInternalKey key;
|
||||
std::unique_ptr<InternalIterator> iter(
|
||||
table_reader->NewIterator(ReadOptions()));
|
||||
|
||||
// Get first (smallest) key from file
|
||||
iter->SeekToFirst();
|
||||
if (!ParseInternalKey(iter->key(), &key)) {
|
||||
return Status::Corruption("Generated table have corrupted keys");
|
||||
}
|
||||
if (key.sequence != 0) {
|
||||
return Status::Corruption("Generated table have non zero sequence number");
|
||||
}
|
||||
file_info->smallest_key = key.user_key.ToString();
|
||||
|
||||
// Get last (largest) key from file
|
||||
iter->SeekToLast();
|
||||
if (!ParseInternalKey(iter->key(), &key)) {
|
||||
return Status::Corruption("Generated table have corrupted keys");
|
||||
}
|
||||
if (key.sequence != 0) {
|
||||
return Status::Corruption("Generated table have non zero sequence number");
|
||||
}
|
||||
file_info->largest_key = key.user_key.ToString();
|
||||
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
Status DBImpl::AddFile(ColumnFamilyHandle* column_family,
|
||||
const std::vector<std::string>& file_path_list,
|
||||
bool move_file, bool skip_snapshot_check) {
|
||||
Status status;
|
||||
auto num_files = file_path_list.size();
|
||||
if (num_files == 0) {
|
||||
return Status::InvalidArgument("The list of files is empty");
|
||||
}
|
||||
|
||||
std::vector<ExternalSstFileInfo> file_info_list(num_files);
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
status = ReadExternalSstFileInfo(column_family, file_path_list[i],
|
||||
&file_info_list[i]);
|
||||
if (!status.ok()) {
|
||||
return status;
|
||||
}
|
||||
}
|
||||
return AddFile(column_family, file_info_list, move_file, skip_snapshot_check);
|
||||
}
|
||||
|
||||
Status DBImpl::AddFile(ColumnFamilyHandle* column_family,
|
||||
const std::vector<ExternalSstFileInfo>& file_info_list,
|
||||
bool move_file, bool skip_snapshot_check) {
|
||||
Status status;
|
||||
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
|
||||
ColumnFamilyData* cfd = cfh->cfd();
|
||||
const Comparator* user_cmp = cfd->internal_comparator().user_comparator();
|
||||
|
||||
auto num_files = file_info_list.size();
|
||||
if (num_files == 0) {
|
||||
return Status::InvalidArgument("The list of files is empty");
|
||||
}
|
||||
|
||||
// Verify that passed files dont have overlapping ranges
|
||||
if (num_files > 1) {
|
||||
std::vector<const ExternalSstFileInfo*> sorted_file_info_list(num_files);
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
sorted_file_info_list[i] = &file_info_list[i];
|
||||
}
|
||||
|
||||
std::sort(sorted_file_info_list.begin(), sorted_file_info_list.end(),
|
||||
[&user_cmp, &file_info_list](const ExternalSstFileInfo* info1,
|
||||
const ExternalSstFileInfo* info2) {
|
||||
return user_cmp->Compare(info1->smallest_key,
|
||||
info2->smallest_key) < 0;
|
||||
});
|
||||
|
||||
for (size_t i = 0; i < num_files - 1; i++) {
|
||||
if (user_cmp->Compare(sorted_file_info_list[i]->largest_key,
|
||||
sorted_file_info_list[i + 1]->smallest_key) >= 0) {
|
||||
return Status::NotSupported("Files have overlapping ranges");
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
std::vector<uint64_t> micro_list(num_files, 0);
|
||||
std::vector<FileMetaData> meta_list(num_files);
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
StopWatch sw(env_, nullptr, 0, µ_list[i], false);
|
||||
if (file_info_list[i].num_entries == 0) {
|
||||
return Status::InvalidArgument("File contain no entries");
|
||||
}
|
||||
if (file_info_list[i].version != 1) {
|
||||
return Status::InvalidArgument(
|
||||
"Generated table version is not supported");
|
||||
}
|
||||
// version 1 imply that file have only Put Operations with Sequence Number =
|
||||
// 0
|
||||
|
||||
meta_list[i].smallest =
|
||||
InternalKey(file_info_list[i].smallest_key,
|
||||
file_info_list[i].sequence_number, ValueType::kTypeValue);
|
||||
meta_list[i].largest =
|
||||
InternalKey(file_info_list[i].largest_key,
|
||||
file_info_list[i].sequence_number, ValueType::kTypeValue);
|
||||
if (!meta_list[i].smallest.Valid() || !meta_list[i].largest.Valid()) {
|
||||
return Status::Corruption("Generated table have corrupted keys");
|
||||
}
|
||||
meta_list[i].smallest_seqno = file_info_list[i].sequence_number;
|
||||
meta_list[i].largest_seqno = file_info_list[i].sequence_number;
|
||||
if (meta_list[i].smallest_seqno != 0 || meta_list[i].largest_seqno != 0) {
|
||||
return Status::InvalidArgument(
|
||||
"Non zero sequence numbers are not supported");
|
||||
}
|
||||
}
|
||||
|
||||
std::vector<std::list<uint64_t>::iterator> pending_outputs_inserted_elem_list(
|
||||
num_files);
|
||||
// Generate locations for the new tables
|
||||
{
|
||||
InstrumentedMutexLock l(&mutex_);
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
StopWatch sw(env_, nullptr, 0, µ_list[i], false);
|
||||
pending_outputs_inserted_elem_list[i] =
|
||||
CaptureCurrentFileNumberInPendingOutputs();
|
||||
meta_list[i].fd = FileDescriptor(versions_->NewFileNumber(), 0,
|
||||
file_info_list[i].file_size);
|
||||
}
|
||||
}
|
||||
|
||||
// Copy/Move external files into DB
|
||||
std::vector<std::string> db_fname_list(num_files);
|
||||
size_t j = 0;
|
||||
for (; j < num_files; j++) {
|
||||
StopWatch sw(env_, nullptr, 0, µ_list[j], false);
|
||||
db_fname_list[j] =
|
||||
TableFileName(db_options_.db_paths, meta_list[j].fd.GetNumber(),
|
||||
meta_list[j].fd.GetPathId());
|
||||
if (move_file) {
|
||||
status = env_->LinkFile(file_info_list[j].file_path, db_fname_list[j]);
|
||||
if (status.IsNotSupported()) {
|
||||
// Original file is on a different FS, use copy instead of hard linking
|
||||
status =
|
||||
CopyFile(env_, file_info_list[j].file_path, db_fname_list[j], 0);
|
||||
}
|
||||
} else {
|
||||
status = CopyFile(env_, file_info_list[j].file_path, db_fname_list[j], 0);
|
||||
}
|
||||
TEST_SYNC_POINT("DBImpl::AddFile:FileCopied");
|
||||
if (!status.ok()) {
|
||||
for (size_t i = 0; i < j; i++) {
|
||||
Status s = env_->DeleteFile(db_fname_list[i]);
|
||||
if (!s.ok()) {
|
||||
Log(InfoLogLevel::WARN_LEVEL, db_options_.info_log,
|
||||
"AddFile() clean up for file %s failed : %s",
|
||||
db_fname_list[i].c_str(), s.ToString().c_str());
|
||||
}
|
||||
}
|
||||
return status;
|
||||
}
|
||||
}
|
||||
|
||||
{
|
||||
InstrumentedMutexLock l(&mutex_);
|
||||
TEST_SYNC_POINT("DBImpl::AddFile:MutexLock");
|
||||
|
||||
const MutableCFOptions mutable_cf_options =
|
||||
*cfd->GetLatestMutableCFOptions();
|
||||
|
||||
WriteThread::Writer w;
|
||||
write_thread_.EnterUnbatched(&w, &mutex_);
|
||||
|
||||
num_running_addfile_++;
|
||||
|
||||
if (!skip_snapshot_check && !snapshots_.empty()) {
|
||||
// Check that no snapshots are being held
|
||||
status =
|
||||
Status::NotSupported("Cannot add a file while holding snapshots");
|
||||
}
|
||||
|
||||
if (status.ok()) {
|
||||
// Verify that added file key range dont overlap with any keys in DB
|
||||
SuperVersion* sv = cfd->GetSuperVersion()->Ref();
|
||||
Arena arena;
|
||||
ReadOptions ro;
|
||||
ro.total_order_seek = true;
|
||||
ScopedArenaIterator iter(NewInternalIterator(ro, cfd, sv, &arena));
|
||||
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
StopWatch sw(env_, nullptr, 0, µ_list[i], false);
|
||||
InternalKey range_start(file_info_list[i].smallest_key,
|
||||
kMaxSequenceNumber, kTypeValue);
|
||||
iter->Seek(range_start.Encode());
|
||||
status = iter->status();
|
||||
|
||||
if (status.ok() && iter->Valid()) {
|
||||
ParsedInternalKey seek_result;
|
||||
if (ParseInternalKey(iter->key(), &seek_result)) {
|
||||
if (user_cmp->Compare(seek_result.user_key,
|
||||
file_info_list[i].largest_key) <= 0) {
|
||||
status = Status::NotSupported("Cannot add overlapping range");
|
||||
break;
|
||||
}
|
||||
} else {
|
||||
status = Status::Corruption("DB have corrupted keys");
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// The levels that the files will be ingested into
|
||||
std::vector<int> target_level_list(num_files, 0);
|
||||
if (status.ok()) {
|
||||
VersionEdit edit;
|
||||
edit.SetColumnFamily(cfd->GetID());
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
StopWatch sw(env_, nullptr, 0, µ_list[i], false);
|
||||
// Add file to the lowest possible level
|
||||
target_level_list[i] = PickLevelForIngestedFile(cfd, file_info_list[i]);
|
||||
edit.AddFile(target_level_list[i], meta_list[i].fd.GetNumber(),
|
||||
meta_list[i].fd.GetPathId(), meta_list[i].fd.GetFileSize(),
|
||||
meta_list[i].smallest, meta_list[i].largest,
|
||||
meta_list[i].smallest_seqno, meta_list[i].largest_seqno,
|
||||
meta_list[i].marked_for_compaction);
|
||||
}
|
||||
status = versions_->LogAndApply(cfd, mutable_cf_options, &edit, &mutex_,
|
||||
directories_.GetDbDir());
|
||||
}
|
||||
write_thread_.ExitUnbatched(&w);
|
||||
|
||||
if (status.ok()) {
|
||||
delete InstallSuperVersionAndScheduleWork(cfd, nullptr,
|
||||
mutable_cf_options);
|
||||
}
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
// Update internal stats
|
||||
InternalStats::CompactionStats stats(1);
|
||||
stats.micros = micro_list[i];
|
||||
stats.bytes_written = meta_list[i].fd.GetFileSize();
|
||||
stats.num_output_files = 1;
|
||||
cfd->internal_stats()->AddCompactionStats(target_level_list[i], stats);
|
||||
cfd->internal_stats()->AddCFStats(InternalStats::BYTES_INGESTED_ADD_FILE,
|
||||
meta_list[i].fd.GetFileSize());
|
||||
ReleaseFileNumberFromPendingOutputs(
|
||||
pending_outputs_inserted_elem_list[i]);
|
||||
}
|
||||
|
||||
num_running_addfile_--;
|
||||
if (num_running_addfile_ == 0) {
|
||||
bg_cv_.SignalAll();
|
||||
}
|
||||
TEST_SYNC_POINT("DBImpl::AddFile:MutexUnlock");
|
||||
} // mutex_ is unlocked here;
|
||||
|
||||
if (!status.ok()) {
|
||||
// We failed to add the files to the database
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
Status s = env_->DeleteFile(db_fname_list[i]);
|
||||
if (!s.ok()) {
|
||||
Log(InfoLogLevel::WARN_LEVEL, db_options_.info_log,
|
||||
"AddFile() clean up for file %s failed : %s",
|
||||
db_fname_list[i].c_str(), s.ToString().c_str());
|
||||
}
|
||||
}
|
||||
} else if (status.ok() && move_file) {
|
||||
// The files were moved and added successfully, remove original file links
|
||||
for (size_t i = 0; i < num_files; i++) {
|
||||
Status s = env_->DeleteFile(file_info_list[i].file_path);
|
||||
if (!s.ok()) {
|
||||
Log(InfoLogLevel::WARN_LEVEL, db_options_.info_log,
|
||||
"%s was added to DB successfully but failed to remove original "
|
||||
"file "
|
||||
"link : %s",
|
||||
file_info_list[i].file_path.c_str(), s.ToString().c_str());
|
||||
}
|
||||
}
|
||||
}
|
||||
return status;
|
||||
}
|
||||
|
||||
// Finds the lowest level in the DB that the ingested file can be added to
|
||||
int DBImpl::PickLevelForIngestedFile(ColumnFamilyData* cfd,
|
||||
const ExternalSstFileInfo& file_info) {
|
||||
mutex_.AssertHeld();
|
||||
|
||||
int target_level = 0;
|
||||
auto* vstorage = cfd->current()->storage_info();
|
||||
auto* ucmp = vstorage->InternalComparator()->user_comparator();
|
||||
|
||||
Slice file_smallest_user_key(file_info.smallest_key);
|
||||
Slice file_largest_user_key(file_info.largest_key);
|
||||
|
||||
for (int lvl = cfd->NumberLevels() - 1; lvl >= vstorage->base_level();
|
||||
lvl--) {
|
||||
if (vstorage->OverlapInLevel(lvl, &file_smallest_user_key,
|
||||
&file_largest_user_key) == false) {
|
||||
// Make sure that the file dont overlap with the output of any
|
||||
// compaction running right now
|
||||
Slice compaction_smallest_user_key;
|
||||
Slice compaction_largest_user_key;
|
||||
bool overlap_with_compaction_output = false;
|
||||
for (Compaction* c : running_compactions_) {
|
||||
if (c->column_family_data()->GetID() != cfd->GetID() ||
|
||||
c->output_level() != lvl) {
|
||||
continue;
|
||||
}
|
||||
|
||||
compaction_smallest_user_key = c->GetSmallestUserKey();
|
||||
compaction_largest_user_key = c->GetLargestUserKey();
|
||||
|
||||
if (ucmp->Compare(file_smallest_user_key,
|
||||
compaction_largest_user_key) <= 0 &&
|
||||
ucmp->Compare(file_largest_user_key,
|
||||
compaction_smallest_user_key) >= 0) {
|
||||
overlap_with_compaction_output = true;
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
if (overlap_with_compaction_output == false) {
|
||||
// Level lvl is the lowest level that dont have any files with key
|
||||
// range overlapping with our file key range and no compactions
|
||||
// planning to add overlapping files in it.
|
||||
target_level = lvl;
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
return target_level;
|
||||
}
|
||||
|
||||
void DBImpl::WaitForAddFile() {
|
||||
mutex_.AssertHeld();
|
||||
while (num_running_addfile_ > 0) {
|
||||
bg_cv_.Wait();
|
||||
}
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
+23
-49
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -7,7 +7,7 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#ifndef NDEBUG
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
#include "db/db_impl.h"
|
||||
#include "util/thread_status_updater.h"
|
||||
@@ -19,6 +19,23 @@ uint64_t DBImpl::TEST_GetLevel0TotalSize() {
|
||||
return default_cf_handle_->cfd()->current()->storage_info()->NumLevelBytes(0);
|
||||
}
|
||||
|
||||
Iterator* DBImpl::TEST_NewInternalIterator(Arena* arena,
|
||||
ColumnFamilyHandle* column_family) {
|
||||
ColumnFamilyData* cfd;
|
||||
if (column_family == nullptr) {
|
||||
cfd = default_cf_handle_->cfd();
|
||||
} else {
|
||||
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
|
||||
cfd = cfh->cfd();
|
||||
}
|
||||
|
||||
mutex_.Lock();
|
||||
SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
|
||||
mutex_.Unlock();
|
||||
ReadOptions roptions;
|
||||
return NewInternalIterator(roptions, cfd, super_version, arena);
|
||||
}
|
||||
|
||||
int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
|
||||
ColumnFamilyHandle* column_family) {
|
||||
ColumnFamilyData* cfd;
|
||||
@@ -70,21 +87,14 @@ Status DBImpl::TEST_CompactRange(int level, const Slice* begin,
|
||||
cfd->ioptions()->compaction_style == kCompactionStyleFIFO)
|
||||
? level
|
||||
: level + 1;
|
||||
return RunManualCompaction(cfd, level, output_level, 0, begin, end, true,
|
||||
return RunManualCompaction(cfd, level, output_level, 0, begin, end,
|
||||
disallow_trivial_move);
|
||||
}
|
||||
|
||||
Status DBImpl::TEST_FlushMemTable(bool wait, ColumnFamilyHandle* cfh) {
|
||||
Status DBImpl::TEST_FlushMemTable(bool wait) {
|
||||
FlushOptions fo;
|
||||
fo.wait = wait;
|
||||
ColumnFamilyData* cfd;
|
||||
if (cfh == nullptr) {
|
||||
cfd = default_cf_handle_->cfd();
|
||||
} else {
|
||||
auto cfhi = reinterpret_cast<ColumnFamilyHandleImpl*>(cfh);
|
||||
cfd = cfhi->cfd();
|
||||
}
|
||||
return FlushMemTable(cfd, fo);
|
||||
return FlushMemTable(default_cf_handle_->cfd(), fo);
|
||||
}
|
||||
|
||||
Status DBImpl::TEST_WaitForFlushMemTable(ColumnFamilyHandle* column_family) {
|
||||
@@ -142,41 +152,5 @@ uint64_t DBImpl::TEST_LogfileNumber() {
|
||||
return logfile_number_;
|
||||
}
|
||||
|
||||
Status DBImpl::TEST_GetAllImmutableCFOptions(
|
||||
std::unordered_map<std::string, const ImmutableCFOptions*>* iopts_map) {
|
||||
std::vector<std::string> cf_names;
|
||||
std::vector<const ImmutableCFOptions*> iopts;
|
||||
{
|
||||
InstrumentedMutexLock l(&mutex_);
|
||||
for (auto cfd : *versions_->GetColumnFamilySet()) {
|
||||
cf_names.push_back(cfd->GetName());
|
||||
iopts.push_back(cfd->ioptions());
|
||||
}
|
||||
}
|
||||
iopts_map->clear();
|
||||
for (size_t i = 0; i < cf_names.size(); ++i) {
|
||||
iopts_map->insert({cf_names[i], iopts[i]});
|
||||
}
|
||||
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
uint64_t DBImpl::TEST_FindMinLogContainingOutstandingPrep() {
|
||||
return FindMinLogContainingOutstandingPrep();
|
||||
}
|
||||
|
||||
uint64_t DBImpl::TEST_FindMinPrepLogReferencedByMemTable() {
|
||||
return FindMinPrepLogReferencedByMemTable();
|
||||
}
|
||||
|
||||
Status DBImpl::TEST_GetLatestMutableCFOptions(
|
||||
ColumnFamilyHandle* column_family, MutableCFOptions* mutable_cf_options) {
|
||||
InstrumentedMutexLock l(&mutex_);
|
||||
|
||||
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
|
||||
*mutable_cf_options = *cfh->cfd()->GetLatestMutableCFOptions();
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
#endif // NDEBUG
|
||||
#endif // ROCKSDB_LITE
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -49,7 +49,8 @@ Status DBImpl::SuggestCompactRange(ColumnFamilyHandle* column_family,
|
||||
}
|
||||
// Since we have some more files to compact, we should also recompute
|
||||
// compaction score
|
||||
vstorage->ComputeCompactionScore(*cfd->GetLatestMutableCFOptions());
|
||||
vstorage->ComputeCompactionScore(*cfd->GetLatestMutableCFOptions(),
|
||||
CompactionOptionsFIFO());
|
||||
SchedulePendingCompaction(cfd);
|
||||
MaybeScheduleFlushOrCompaction();
|
||||
}
|
||||
|
||||
+8
-10
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -54,11 +54,10 @@ Iterator* DBImplReadOnly::NewIterator(const ReadOptions& read_options,
|
||||
auto db_iter = NewArenaWrappedDbIterator(
|
||||
env_, *cfd->ioptions(), cfd->user_comparator(),
|
||||
(read_options.snapshot != nullptr
|
||||
? reinterpret_cast<const SnapshotImpl*>(read_options.snapshot)
|
||||
->number_
|
||||
? reinterpret_cast<const SnapshotImpl*>(
|
||||
read_options.snapshot)->number_
|
||||
: latest_snapshot),
|
||||
super_version->mutable_cf_options.max_sequential_skip_in_iterations,
|
||||
super_version->version_number);
|
||||
super_version->mutable_cf_options.max_sequential_skip_in_iterations);
|
||||
auto internal_iter = NewInternalIterator(
|
||||
read_options, cfd, super_version, db_iter->GetArena());
|
||||
db_iter->SetIterUnderDBIter(internal_iter);
|
||||
@@ -82,11 +81,10 @@ Status DBImplReadOnly::NewIterators(
|
||||
auto* db_iter = NewArenaWrappedDbIterator(
|
||||
env_, *cfd->ioptions(), cfd->user_comparator(),
|
||||
(read_options.snapshot != nullptr
|
||||
? reinterpret_cast<const SnapshotImpl*>(read_options.snapshot)
|
||||
->number_
|
||||
: latest_snapshot),
|
||||
sv->mutable_cf_options.max_sequential_skip_in_iterations,
|
||||
sv->version_number);
|
||||
? reinterpret_cast<const SnapshotImpl*>(
|
||||
read_options.snapshot)->number_
|
||||
: latest_snapshot),
|
||||
sv->mutable_cf_options.max_sequential_skip_in_iterations);
|
||||
auto* internal_iter = NewInternalIterator(
|
||||
read_options, cfd, sv, db_iter->GetArena());
|
||||
db_iter->SetIterUnderDBIter(internal_iter);
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -6,8 +6,8 @@
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/db_test_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -18,11 +18,12 @@ class DBTestInPlaceUpdate : public DBTestBase {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdate) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
options.env = env_;
|
||||
options.write_buffer_size = 100000;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Update key with values of smaller size
|
||||
@@ -40,11 +41,12 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdate) {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdateLargeNewValue) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
options.env = env_;
|
||||
options.write_buffer_size = 100000;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Update key with values of larger size
|
||||
@@ -62,7 +64,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateLargeNewValue) {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerSize) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
|
||||
@@ -70,6 +72,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerSize) {
|
||||
options.write_buffer_size = 100000;
|
||||
options.inplace_callback =
|
||||
rocksdb::DBTestInPlaceUpdate::updateInPlaceSmallerSize;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Update key with values of smaller size
|
||||
@@ -89,7 +92,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerSize) {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerVarintSize) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
|
||||
@@ -97,6 +100,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerVarintSize) {
|
||||
options.write_buffer_size = 100000;
|
||||
options.inplace_callback =
|
||||
rocksdb::DBTestInPlaceUpdate::updateInPlaceSmallerVarintSize;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Update key with values of smaller varint size
|
||||
@@ -116,7 +120,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackSmallerVarintSize) {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackLargeNewValue) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
|
||||
@@ -124,6 +128,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackLargeNewValue) {
|
||||
options.write_buffer_size = 100000;
|
||||
options.inplace_callback =
|
||||
rocksdb::DBTestInPlaceUpdate::updateInPlaceLargerSize;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Update key with values of larger size
|
||||
@@ -141,7 +146,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackLargeNewValue) {
|
||||
|
||||
TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackNoAction) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.inplace_update_support = true;
|
||||
|
||||
@@ -149,6 +154,7 @@ TEST_F(DBTestInPlaceUpdate, InPlaceUpdateCallbackNoAction) {
|
||||
options.write_buffer_size = 100000;
|
||||
options.inplace_callback =
|
||||
rocksdb::DBTestInPlaceUpdate::updateInPlaceNoAction;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Callback function requests no actions from db
|
||||
|
||||
@@ -1,260 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class DBIOFailureTest : public DBTestBase {
|
||||
public:
|
||||
DBIOFailureTest() : DBTestBase("/db_io_failure_test") {}
|
||||
};
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
// Check that number of files does not grow when writes are dropped
|
||||
TEST_F(DBIOFailureTest, DropWrites) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.paranoid_checks = false;
|
||||
Reopen(options);
|
||||
|
||||
ASSERT_OK(Put("foo", "v1"));
|
||||
ASSERT_EQ("v1", Get("foo"));
|
||||
Compact("a", "z");
|
||||
const size_t num_files = CountFiles();
|
||||
// Force out-of-space errors
|
||||
env_->drop_writes_.store(true, std::memory_order_release);
|
||||
env_->sleep_counter_.Reset();
|
||||
env_->no_sleep_ = true;
|
||||
for (int i = 0; i < 5; i++) {
|
||||
if (option_config_ != kUniversalCompactionMultiLevel &&
|
||||
option_config_ != kUniversalSubcompactions) {
|
||||
for (int level = 0; level < dbfull()->NumberLevels(); level++) {
|
||||
if (level > 0 && level == dbfull()->NumberLevels() - 1) {
|
||||
break;
|
||||
}
|
||||
dbfull()->TEST_CompactRange(level, nullptr, nullptr, nullptr,
|
||||
true /* disallow trivial move */);
|
||||
}
|
||||
} else {
|
||||
dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
|
||||
}
|
||||
}
|
||||
|
||||
std::string property_value;
|
||||
ASSERT_TRUE(db_->GetProperty("rocksdb.background-errors", &property_value));
|
||||
ASSERT_EQ("5", property_value);
|
||||
|
||||
env_->drop_writes_.store(false, std::memory_order_release);
|
||||
ASSERT_LT(CountFiles(), num_files + 3);
|
||||
|
||||
// Check that compaction attempts slept after errors
|
||||
// TODO @krad: Figure out why ASSERT_EQ 5 keeps failing in certain compiler
|
||||
// versions
|
||||
ASSERT_GE(env_->sleep_counter_.Read(), 4);
|
||||
} while (ChangeCompactOptions());
|
||||
}
|
||||
|
||||
// Check background error counter bumped on flush failures.
|
||||
TEST_F(DBIOFailureTest, DropWritesFlush) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.max_background_flushes = 1;
|
||||
Reopen(options);
|
||||
|
||||
ASSERT_OK(Put("foo", "v1"));
|
||||
// Force out-of-space errors
|
||||
env_->drop_writes_.store(true, std::memory_order_release);
|
||||
|
||||
std::string property_value;
|
||||
// Background error count is 0 now.
|
||||
ASSERT_TRUE(db_->GetProperty("rocksdb.background-errors", &property_value));
|
||||
ASSERT_EQ("0", property_value);
|
||||
|
||||
dbfull()->TEST_FlushMemTable(true);
|
||||
|
||||
ASSERT_TRUE(db_->GetProperty("rocksdb.background-errors", &property_value));
|
||||
ASSERT_EQ("1", property_value);
|
||||
|
||||
env_->drop_writes_.store(false, std::memory_order_release);
|
||||
} while (ChangeCompactOptions());
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
// Check that CompactRange() returns failure if there is not enough space left
|
||||
// on device
|
||||
TEST_F(DBIOFailureTest, NoSpaceCompactRange) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.disable_auto_compactions = true;
|
||||
Reopen(options);
|
||||
|
||||
// generate 5 tables
|
||||
for (int i = 0; i < 5; ++i) {
|
||||
ASSERT_OK(Put(Key(i), Key(i) + "v"));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
|
||||
// Force out-of-space errors
|
||||
env_->no_space_.store(true, std::memory_order_release);
|
||||
|
||||
Status s = dbfull()->TEST_CompactRange(0, nullptr, nullptr, nullptr,
|
||||
true /* disallow trivial move */);
|
||||
ASSERT_TRUE(s.IsIOError());
|
||||
ASSERT_TRUE(s.IsNoSpace());
|
||||
|
||||
env_->no_space_.store(false, std::memory_order_release);
|
||||
} while (ChangeCompactOptions());
|
||||
}
|
||||
|
||||
TEST_F(DBIOFailureTest, NonWritableFileSystem) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 4096;
|
||||
options.arena_block_size = 4096;
|
||||
options.env = env_;
|
||||
Reopen(options);
|
||||
ASSERT_OK(Put("foo", "v1"));
|
||||
env_->non_writeable_rate_.store(100);
|
||||
std::string big(100000, 'x');
|
||||
int errors = 0;
|
||||
for (int i = 0; i < 20; i++) {
|
||||
if (!Put("foo", big).ok()) {
|
||||
errors++;
|
||||
env_->SleepForMicroseconds(100000);
|
||||
}
|
||||
}
|
||||
ASSERT_GT(errors, 0);
|
||||
env_->non_writeable_rate_.store(0);
|
||||
} while (ChangeCompactOptions());
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBIOFailureTest, ManifestWriteError) {
|
||||
// Test for the following problem:
|
||||
// (a) Compaction produces file F
|
||||
// (b) Log record containing F is written to MANIFEST file, but Sync() fails
|
||||
// (c) GC deletes F
|
||||
// (d) After reopening DB, reads fail since deleted F is named in log record
|
||||
|
||||
// We iterate twice. In the second iteration, everything is the
|
||||
// same except the log record never makes it to the MANIFEST file.
|
||||
for (int iter = 0; iter < 2; iter++) {
|
||||
std::atomic<bool>* error_type = (iter == 0) ? &env_->manifest_sync_error_
|
||||
: &env_->manifest_write_error_;
|
||||
|
||||
// Insert foo=>bar mapping
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.create_if_missing = true;
|
||||
options.error_if_exists = false;
|
||||
options.paranoid_checks = true;
|
||||
DestroyAndReopen(options);
|
||||
ASSERT_OK(Put("foo", "bar"));
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
|
||||
// Memtable compaction (will succeed)
|
||||
Flush();
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
const int last = 2;
|
||||
MoveFilesToLevel(2);
|
||||
ASSERT_EQ(NumTableFilesAtLevel(last), 1); // foo=>bar is now in last level
|
||||
|
||||
// Merging compaction (will fail)
|
||||
error_type->store(true, std::memory_order_release);
|
||||
dbfull()->TEST_CompactRange(last, nullptr, nullptr); // Should fail
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
|
||||
error_type->store(false, std::memory_order_release);
|
||||
|
||||
// Since paranoid_checks=true, writes should fail
|
||||
ASSERT_NOK(Put("foo2", "bar2"));
|
||||
|
||||
// Recovery: should not lose data
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
|
||||
// Try again with paranoid_checks=false
|
||||
Close();
|
||||
options.paranoid_checks = false;
|
||||
Reopen(options);
|
||||
|
||||
// Merging compaction (will fail)
|
||||
error_type->store(true, std::memory_order_release);
|
||||
dbfull()->TEST_CompactRange(last, nullptr, nullptr); // Should fail
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
|
||||
// Recovery: should not lose data
|
||||
error_type->store(false, std::memory_order_release);
|
||||
Reopen(options);
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
|
||||
// Since paranoid_checks=false, writes should succeed
|
||||
ASSERT_OK(Put("foo2", "bar2"));
|
||||
ASSERT_EQ("bar", Get("foo"));
|
||||
ASSERT_EQ("bar2", Get("foo2"));
|
||||
}
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBIOFailureTest, PutFailsParanoid) {
|
||||
// Test the following:
|
||||
// (a) A random put fails in paranoid mode (simulate by sync fail)
|
||||
// (b) All other puts have to fail, even if writes would succeed
|
||||
// (c) All of that should happen ONLY if paranoid_checks = true
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.create_if_missing = true;
|
||||
options.error_if_exists = false;
|
||||
options.paranoid_checks = true;
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
Status s;
|
||||
|
||||
ASSERT_OK(Put(1, "foo", "bar"));
|
||||
ASSERT_OK(Put(1, "foo1", "bar1"));
|
||||
// simulate error
|
||||
env_->log_write_error_.store(true, std::memory_order_release);
|
||||
s = Put(1, "foo2", "bar2");
|
||||
ASSERT_TRUE(!s.ok());
|
||||
env_->log_write_error_.store(false, std::memory_order_release);
|
||||
s = Put(1, "foo3", "bar3");
|
||||
// the next put should fail, too
|
||||
ASSERT_TRUE(!s.ok());
|
||||
// but we're still able to read
|
||||
ASSERT_EQ("bar", Get(1, "foo"));
|
||||
|
||||
// do the same thing with paranoid checks off
|
||||
options.paranoid_checks = false;
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
ASSERT_OK(Put(1, "foo", "bar"));
|
||||
ASSERT_OK(Put(1, "foo1", "bar1"));
|
||||
// simulate error
|
||||
env_->log_write_error_.store(true, std::memory_order_release);
|
||||
s = Put(1, "foo2", "bar2");
|
||||
ASSERT_TRUE(!s.ok());
|
||||
env_->log_write_error_.store(false, std::memory_order_release);
|
||||
s = Put(1, "foo3", "bar3");
|
||||
// the next put should NOT fail
|
||||
ASSERT_TRUE(s.ok());
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
+118
-287
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -13,22 +13,17 @@
|
||||
#include <string>
|
||||
#include <limits>
|
||||
|
||||
#include "db/dbformat.h"
|
||||
#include "db/filename.h"
|
||||
#include "db/merge_context.h"
|
||||
#include "db/merge_helper.h"
|
||||
#include "db/pinned_iterators_manager.h"
|
||||
#include "port/port.h"
|
||||
#include "db/dbformat.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "rocksdb/iterator.h"
|
||||
#include "rocksdb/merge_operator.h"
|
||||
#include "rocksdb/options.h"
|
||||
#include "table/internal_iterator.h"
|
||||
#include "port/port.h"
|
||||
#include "util/arena.h"
|
||||
#include "util/logging.h"
|
||||
#include "util/mutexlock.h"
|
||||
#include "util/perf_context_imp.h"
|
||||
#include "util/string_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -63,91 +58,37 @@ class DBIter: public Iterator {
|
||||
kReverse
|
||||
};
|
||||
|
||||
// LocalStatistics contain Statistics counters that will be aggregated per
|
||||
// each iterator instance and then will be sent to the global statistics when
|
||||
// the iterator is destroyed.
|
||||
//
|
||||
// The purpose of this approach is to avoid perf regression happening
|
||||
// when multiple threads bump the atomic counters from a DBIter::Next().
|
||||
struct LocalStatistics {
|
||||
explicit LocalStatistics() { ResetCounters(); }
|
||||
|
||||
void ResetCounters() {
|
||||
next_count_ = 0;
|
||||
next_found_count_ = 0;
|
||||
prev_count_ = 0;
|
||||
prev_found_count_ = 0;
|
||||
bytes_read_ = 0;
|
||||
}
|
||||
|
||||
void BumpGlobalStatistics(Statistics* global_statistics) {
|
||||
RecordTick(global_statistics, NUMBER_DB_NEXT, next_count_);
|
||||
RecordTick(global_statistics, NUMBER_DB_NEXT_FOUND, next_found_count_);
|
||||
RecordTick(global_statistics, NUMBER_DB_PREV, prev_count_);
|
||||
RecordTick(global_statistics, NUMBER_DB_PREV_FOUND, prev_found_count_);
|
||||
RecordTick(global_statistics, ITER_BYTES_READ, bytes_read_);
|
||||
ResetCounters();
|
||||
}
|
||||
|
||||
// Map to Tickers::NUMBER_DB_NEXT
|
||||
uint64_t next_count_;
|
||||
// Map to Tickers::NUMBER_DB_NEXT_FOUND
|
||||
uint64_t next_found_count_;
|
||||
// Map to Tickers::NUMBER_DB_PREV
|
||||
uint64_t prev_count_;
|
||||
// Map to Tickers::NUMBER_DB_PREV_FOUND
|
||||
uint64_t prev_found_count_;
|
||||
// Map to Tickers::ITER_BYTES_READ
|
||||
uint64_t bytes_read_;
|
||||
};
|
||||
|
||||
DBIter(Env* env, const ImmutableCFOptions& ioptions, const Comparator* cmp,
|
||||
InternalIterator* iter, SequenceNumber s, bool arena_mode,
|
||||
uint64_t max_sequential_skip_in_iterations, uint64_t version_number,
|
||||
const Slice* iterate_upper_bound = nullptr,
|
||||
bool prefix_same_as_start = false, bool pin_data = false)
|
||||
DBIter(Env* env, const ImmutableCFOptions& ioptions,
|
||||
const Comparator* cmp, Iterator* iter, SequenceNumber s,
|
||||
bool arena_mode, uint64_t max_sequential_skip_in_iterations,
|
||||
const Slice* iterate_upper_bound = nullptr)
|
||||
: arena_mode_(arena_mode),
|
||||
env_(env),
|
||||
logger_(ioptions.info_log),
|
||||
user_comparator_(cmp),
|
||||
merge_operator_(ioptions.merge_operator),
|
||||
user_merge_operator_(ioptions.merge_operator),
|
||||
iter_(iter),
|
||||
sequence_(s),
|
||||
direction_(kForward),
|
||||
valid_(false),
|
||||
current_entry_is_merged_(false),
|
||||
statistics_(ioptions.statistics),
|
||||
version_number_(version_number),
|
||||
iterate_upper_bound_(iterate_upper_bound),
|
||||
prefix_same_as_start_(prefix_same_as_start),
|
||||
pin_thru_lifetime_(pin_data) {
|
||||
iterate_upper_bound_(iterate_upper_bound) {
|
||||
RecordTick(statistics_, NO_ITERATORS);
|
||||
prefix_extractor_ = ioptions.prefix_extractor;
|
||||
max_skip_ = max_sequential_skip_in_iterations;
|
||||
if (pin_thru_lifetime_) {
|
||||
pinned_iters_mgr_.StartPinning();
|
||||
}
|
||||
if (iter_) {
|
||||
iter_->SetPinnedItersMgr(&pinned_iters_mgr_);
|
||||
}
|
||||
}
|
||||
virtual ~DBIter() {
|
||||
// Release pinned data if any
|
||||
if (pinned_iters_mgr_.PinningEnabled()) {
|
||||
pinned_iters_mgr_.ReleasePinnedData();
|
||||
}
|
||||
RecordTick(statistics_, NO_ITERATORS, -1);
|
||||
local_stats_.BumpGlobalStatistics(statistics_);
|
||||
if (!arena_mode_) {
|
||||
delete iter_;
|
||||
} else {
|
||||
iter_->~InternalIterator();
|
||||
iter_->~Iterator();
|
||||
}
|
||||
}
|
||||
virtual void SetIter(InternalIterator* iter) {
|
||||
virtual void SetIter(Iterator* iter) {
|
||||
assert(iter_ == nullptr);
|
||||
iter_ = iter;
|
||||
iter_->SetPinnedItersMgr(&pinned_iters_mgr_);
|
||||
}
|
||||
virtual bool Valid() const override { return valid_; }
|
||||
virtual Slice key() const override {
|
||||
@@ -156,15 +97,8 @@ class DBIter: public Iterator {
|
||||
}
|
||||
virtual Slice value() const override {
|
||||
assert(valid_);
|
||||
if (current_entry_is_merged_) {
|
||||
// If pinned_value_ is set then the result of merge operator is one of
|
||||
// the merge operands and we should return it.
|
||||
return pinned_value_.data() ? pinned_value_ : saved_value_;
|
||||
} else if (direction_ == kReverse) {
|
||||
return pinned_value_;
|
||||
} else {
|
||||
return iter_->value();
|
||||
}
|
||||
return (direction_ == kForward && !current_entry_is_merged_) ?
|
||||
iter_->value() : saved_value_;
|
||||
}
|
||||
virtual Status status() const override {
|
||||
if (status_.ok()) {
|
||||
@@ -174,28 +108,6 @@ class DBIter: public Iterator {
|
||||
}
|
||||
}
|
||||
|
||||
virtual Status GetProperty(std::string prop_name,
|
||||
std::string* prop) override {
|
||||
if (prop == nullptr) {
|
||||
return Status::InvalidArgument("prop is nullptr");
|
||||
}
|
||||
if (prop_name == "rocksdb.iterator.super-version-number") {
|
||||
// First try to pass the value returned from inner iterator.
|
||||
if (!iter_->GetProperty(prop_name, prop).ok()) {
|
||||
*prop = ToString(version_number_);
|
||||
}
|
||||
return Status::OK();
|
||||
} else if (prop_name == "rocksdb.iterator.is-key-pinned") {
|
||||
if (valid_) {
|
||||
*prop = (pin_thru_lifetime_ && saved_key_.IsKeyPinned()) ? "1" : "0";
|
||||
} else {
|
||||
*prop = "Iterator is not valid.";
|
||||
}
|
||||
return Status::OK();
|
||||
}
|
||||
return Status::InvalidArgument("Undentified property.");
|
||||
}
|
||||
|
||||
virtual void Next() override;
|
||||
virtual void Prev() override;
|
||||
virtual void Seek(const Slice& target) override;
|
||||
@@ -210,26 +122,11 @@ class DBIter: public Iterator {
|
||||
bool FindValueForCurrentKeyUsingSeek();
|
||||
void FindPrevUserKey();
|
||||
void FindNextUserKey();
|
||||
inline void FindNextUserEntry(bool skipping, bool prefix_check);
|
||||
void FindNextUserEntryInternal(bool skipping, bool prefix_check);
|
||||
inline void FindNextUserEntry(bool skipping);
|
||||
void FindNextUserEntryInternal(bool skipping);
|
||||
bool ParseKey(ParsedInternalKey* key);
|
||||
void MergeValuesNewToOld();
|
||||
|
||||
// Temporarily pin the blocks that we encounter until ReleaseTempPinnedData()
|
||||
// is called
|
||||
void TempPinData() {
|
||||
if (!pin_thru_lifetime_) {
|
||||
pinned_iters_mgr_.StartPinning();
|
||||
}
|
||||
}
|
||||
|
||||
// Release blocks pinned by TempPinData()
|
||||
void ReleaseTempPinnedData() {
|
||||
if (!pin_thru_lifetime_ && pinned_iters_mgr_.PinningEnabled()) {
|
||||
pinned_iters_mgr_.ReleasePinnedData();
|
||||
}
|
||||
}
|
||||
|
||||
inline void ClearSavedValue() {
|
||||
if (saved_value_.capacity() > 1048576) {
|
||||
std::string empty;
|
||||
@@ -244,31 +141,19 @@ class DBIter: public Iterator {
|
||||
Env* const env_;
|
||||
Logger* logger_;
|
||||
const Comparator* const user_comparator_;
|
||||
const MergeOperator* const merge_operator_;
|
||||
InternalIterator* iter_;
|
||||
const MergeOperator* const user_merge_operator_;
|
||||
Iterator* iter_;
|
||||
SequenceNumber const sequence_;
|
||||
|
||||
Status status_;
|
||||
IterKey saved_key_;
|
||||
std::string saved_value_;
|
||||
Slice pinned_value_;
|
||||
Direction direction_;
|
||||
bool valid_;
|
||||
bool current_entry_is_merged_;
|
||||
Statistics* statistics_;
|
||||
uint64_t max_skip_;
|
||||
uint64_t version_number_;
|
||||
const Slice* iterate_upper_bound_;
|
||||
IterKey prefix_start_buf_;
|
||||
Slice prefix_start_key_;
|
||||
const bool prefix_same_as_start_;
|
||||
// Means that we will pin all data blocks we read as long the Iterator
|
||||
// is not deleted, will be true if ReadOptions::pin_data is true
|
||||
const bool pin_thru_lifetime_;
|
||||
// List of operands for merge operator.
|
||||
MergeContext merge_context_;
|
||||
LocalStatistics local_stats_;
|
||||
PinnedIteratorsManager pinned_iters_mgr_;
|
||||
|
||||
// No copying allowed
|
||||
DBIter(const DBIter&);
|
||||
@@ -290,37 +175,26 @@ inline bool DBIter::ParseKey(ParsedInternalKey* ikey) {
|
||||
void DBIter::Next() {
|
||||
assert(valid_);
|
||||
|
||||
// Release temporarily pinned blocks from last operation
|
||||
ReleaseTempPinnedData();
|
||||
if (direction_ == kReverse) {
|
||||
FindNextUserKey();
|
||||
direction_ = kForward;
|
||||
if (!iter_->Valid()) {
|
||||
iter_->SeekToFirst();
|
||||
}
|
||||
} else if (iter_->Valid() && !current_entry_is_merged_) {
|
||||
// If the current value is not a merge, the iter position is the
|
||||
// current key, which is already returned. We can safely issue a
|
||||
// Next() without checking the current key.
|
||||
// If the current key is a merge, very likely iter already points
|
||||
// to the next internal position.
|
||||
iter_->Next();
|
||||
PERF_COUNTER_ADD(internal_key_skipped_count, 1);
|
||||
}
|
||||
|
||||
if (statistics_ != nullptr) {
|
||||
local_stats_.next_count_++;
|
||||
}
|
||||
// Now we point to the next internal position, for both of merge and
|
||||
// not merge cases.
|
||||
// If the current value is merged, we might already hit end of iter_
|
||||
if (!iter_->Valid()) {
|
||||
valid_ = false;
|
||||
return;
|
||||
}
|
||||
FindNextUserEntry(true /* skipping the current user key */, prefix_same_as_start_);
|
||||
if (statistics_ != nullptr && valid_) {
|
||||
local_stats_.next_found_count_++;
|
||||
local_stats_.bytes_read_ += (key().size() + value().size());
|
||||
FindNextUserEntry(true /* skipping the current user key */);
|
||||
if (statistics_ != nullptr) {
|
||||
RecordTick(statistics_, NUMBER_DB_NEXT);
|
||||
if (valid_) {
|
||||
RecordTick(statistics_, NUMBER_DB_NEXT_FOUND);
|
||||
RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -332,18 +206,13 @@ void DBIter::Next() {
|
||||
//
|
||||
// NOTE: In between, saved_key_ can point to a user key that has
|
||||
// a delete marker
|
||||
//
|
||||
// The prefix_check parameter controls whether we check the iterated
|
||||
// keys against the prefix of the seeked key. Set to false when
|
||||
// performing a seek without a key (e.g. SeekToFirst). Set to
|
||||
// prefix_same_as_start_ for other iterations.
|
||||
inline void DBIter::FindNextUserEntry(bool skipping, bool prefix_check) {
|
||||
inline void DBIter::FindNextUserEntry(bool skipping) {
|
||||
PERF_TIMER_GUARD(find_next_user_entry_time);
|
||||
FindNextUserEntryInternal(skipping, prefix_check);
|
||||
FindNextUserEntryInternal(skipping);
|
||||
}
|
||||
|
||||
// Actual implementation of DBIter::FindNextUserEntry()
|
||||
void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
|
||||
void DBIter::FindNextUserEntryInternal(bool skipping) {
|
||||
// Loop until we hit an acceptable entry to yield
|
||||
assert(iter_->Valid());
|
||||
assert(direction_ == kForward);
|
||||
@@ -354,12 +223,7 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
|
||||
|
||||
if (ParseKey(&ikey)) {
|
||||
if (iterate_upper_bound_ != nullptr &&
|
||||
user_comparator_->Compare(ikey.user_key, *iterate_upper_bound_) >= 0) {
|
||||
break;
|
||||
}
|
||||
|
||||
if (prefix_extractor_ && prefix_check &&
|
||||
prefix_extractor_->Transform(ikey.user_key).compare(prefix_start_key_) != 0) {
|
||||
ikey.user_key.compare(*iterate_upper_bound_) >= 0) {
|
||||
break;
|
||||
}
|
||||
|
||||
@@ -374,24 +238,18 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
|
||||
case kTypeSingleDeletion:
|
||||
// Arrange to skip all upcoming entries for this key since
|
||||
// they are hidden by this deletion.
|
||||
saved_key_.SetKey(
|
||||
ikey.user_key,
|
||||
!iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
|
||||
saved_key_.SetKey(ikey.user_key);
|
||||
skipping = true;
|
||||
num_skipped = 0;
|
||||
PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
|
||||
break;
|
||||
case kTypeValue:
|
||||
valid_ = true;
|
||||
saved_key_.SetKey(
|
||||
ikey.user_key,
|
||||
!iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
|
||||
saved_key_.SetKey(ikey.user_key);
|
||||
return;
|
||||
case kTypeMerge:
|
||||
// By now, we are sure the current ikey is going to yield a value
|
||||
saved_key_.SetKey(
|
||||
ikey.user_key,
|
||||
!iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
|
||||
saved_key_.SetKey(ikey.user_key);
|
||||
current_entry_is_merged_ = true;
|
||||
valid_ = true;
|
||||
MergeValuesNewToOld(); // Go to a different state machine
|
||||
@@ -429,20 +287,17 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
|
||||
// POST: saved_value_ has the merged value for the user key
|
||||
// iter_ points to the next entry (or invalid)
|
||||
void DBIter::MergeValuesNewToOld() {
|
||||
if (!merge_operator_) {
|
||||
if (!user_merge_operator_) {
|
||||
Log(InfoLogLevel::ERROR_LEVEL,
|
||||
logger_, "Options::merge_operator is null.");
|
||||
status_ = Status::InvalidArgument("merge_operator_ must be set.");
|
||||
status_ = Status::InvalidArgument("user_merge_operator_ must be set.");
|
||||
valid_ = false;
|
||||
return;
|
||||
}
|
||||
|
||||
// Temporarily pin the blocks that hold merge operands
|
||||
TempPinData();
|
||||
merge_context_.Clear();
|
||||
// Start the merge process by pushing the first operand
|
||||
merge_context_.PushOperand(iter_->value(),
|
||||
iter_->IsValuePinned() /* operand_pinned */);
|
||||
std::deque<std::string> operands;
|
||||
operands.push_front(iter_->value().ToString());
|
||||
|
||||
ParsedInternalKey ikey;
|
||||
for (iter_->Next(); iter_->Valid(); iter_->Next()) {
|
||||
@@ -464,50 +319,53 @@ void DBIter::MergeValuesNewToOld() {
|
||||
// final result in saved_value_. We are done!
|
||||
// ignore corruption if there is any.
|
||||
const Slice val = iter_->value();
|
||||
MergeHelper::TimedFullMerge(merge_operator_, ikey.user_key, &val,
|
||||
merge_context_.GetOperands(), &saved_value_,
|
||||
logger_, statistics_, env_, &pinned_value_);
|
||||
{
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
user_merge_operator_->FullMerge(ikey.user_key, &val, operands,
|
||||
&saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME,
|
||||
timer.ElapsedNanos());
|
||||
}
|
||||
// iter_ is positioned after put
|
||||
iter_->Next();
|
||||
return;
|
||||
} else if (kTypeMerge == ikey.type) {
|
||||
// hit a merge, add the value as an operand and run associative merge.
|
||||
// when complete, add result to operands and continue.
|
||||
merge_context_.PushOperand(iter_->value(),
|
||||
iter_->IsValuePinned() /* operand_pinned */);
|
||||
const Slice& val = iter_->value();
|
||||
operands.push_front(val.ToString());
|
||||
} else {
|
||||
assert(false);
|
||||
}
|
||||
}
|
||||
|
||||
// we either exhausted all internal keys under this user key, or hit
|
||||
// a deletion marker.
|
||||
// feed null as the existing value to the merge operator, such that
|
||||
// client can differentiate this scenario and do things accordingly.
|
||||
MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(), nullptr,
|
||||
merge_context_.GetOperands(), &saved_value_,
|
||||
logger_, statistics_, env_, &pinned_value_);
|
||||
{
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
// we either exhausted all internal keys under this user key, or hit
|
||||
// a deletion marker.
|
||||
// feed null as the existing value to the merge operator, such that
|
||||
// client can differentiate this scenario and do things accordingly.
|
||||
user_merge_operator_->FullMerge(saved_key_.GetKey(), nullptr, operands,
|
||||
&saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME, timer.ElapsedNanos());
|
||||
}
|
||||
}
|
||||
|
||||
void DBIter::Prev() {
|
||||
assert(valid_);
|
||||
ReleaseTempPinnedData();
|
||||
if (direction_ == kForward) {
|
||||
ReverseToBackward();
|
||||
}
|
||||
PrevInternal();
|
||||
if (statistics_ != nullptr) {
|
||||
local_stats_.prev_count_++;
|
||||
RecordTick(statistics_, NUMBER_DB_PREV);
|
||||
if (valid_) {
|
||||
local_stats_.prev_found_count_++;
|
||||
local_stats_.bytes_read_ += (key().size() + value().size());
|
||||
RecordTick(statistics_, NUMBER_DB_PREV_FOUND);
|
||||
RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
|
||||
}
|
||||
}
|
||||
if (valid_ && prefix_extractor_ && prefix_same_as_start_ &&
|
||||
prefix_extractor_->Transform(saved_key_.GetKey())
|
||||
.compare(prefix_start_key_) != 0) {
|
||||
valid_ = false;
|
||||
}
|
||||
}
|
||||
|
||||
void DBIter::ReverseToBackward() {
|
||||
@@ -546,8 +404,7 @@ void DBIter::PrevInternal() {
|
||||
ParsedInternalKey ikey;
|
||||
|
||||
while (iter_->Valid()) {
|
||||
saved_key_.SetKey(ExtractUserKey(iter_->key()),
|
||||
!iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
|
||||
saved_key_.SetKey(ExtractUserKey(iter_->key()));
|
||||
if (FindValueForCurrentKey()) {
|
||||
valid_ = true;
|
||||
if (!iter_->Valid()) {
|
||||
@@ -577,8 +434,8 @@ void DBIter::PrevInternal() {
|
||||
// saved_value_
|
||||
bool DBIter::FindValueForCurrentKey() {
|
||||
assert(iter_->Valid());
|
||||
merge_context_.Clear();
|
||||
current_entry_is_merged_ = false;
|
||||
// Contains operands for merge operator.
|
||||
std::deque<std::string> operands;
|
||||
// last entry before merge (could be kTypeDeletion, kTypeSingleDeletion or
|
||||
// kTypeValue)
|
||||
ValueType last_not_merge_type = kTypeDeletion;
|
||||
@@ -587,9 +444,6 @@ bool DBIter::FindValueForCurrentKey() {
|
||||
ParsedInternalKey ikey;
|
||||
FindParseableKey(&ikey, kReverse);
|
||||
|
||||
// Temporarily pin blocks that hold (merge operands / the value)
|
||||
ReleaseTempPinnedData();
|
||||
TempPinData();
|
||||
size_t num_skipped = 0;
|
||||
while (iter_->Valid() && ikey.sequence <= sequence_ &&
|
||||
user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
|
||||
@@ -601,21 +455,19 @@ bool DBIter::FindValueForCurrentKey() {
|
||||
last_key_entry_type = ikey.type;
|
||||
switch (last_key_entry_type) {
|
||||
case kTypeValue:
|
||||
merge_context_.Clear();
|
||||
assert(iter_->IsValuePinned());
|
||||
pinned_value_ = iter_->value();
|
||||
operands.clear();
|
||||
saved_value_ = iter_->value().ToString();
|
||||
last_not_merge_type = kTypeValue;
|
||||
break;
|
||||
case kTypeDeletion:
|
||||
case kTypeSingleDeletion:
|
||||
merge_context_.Clear();
|
||||
operands.clear();
|
||||
last_not_merge_type = last_key_entry_type;
|
||||
PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
|
||||
break;
|
||||
case kTypeMerge:
|
||||
assert(merge_operator_ != nullptr);
|
||||
merge_context_.PushOperandBack(
|
||||
iter_->value(), iter_->IsValuePinned() /* operand_pinned */);
|
||||
assert(user_merge_operator_ != nullptr);
|
||||
operands.push_back(iter_->value().ToString());
|
||||
break;
|
||||
default:
|
||||
assert(false);
|
||||
@@ -634,18 +486,25 @@ bool DBIter::FindValueForCurrentKey() {
|
||||
valid_ = false;
|
||||
return false;
|
||||
case kTypeMerge:
|
||||
current_entry_is_merged_ = true;
|
||||
if (last_not_merge_type == kTypeDeletion) {
|
||||
MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(),
|
||||
nullptr, merge_context_.GetOperands(),
|
||||
&saved_value_, logger_, statistics_, env_,
|
||||
&pinned_value_);
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
user_merge_operator_->FullMerge(saved_key_.GetKey(), nullptr, operands,
|
||||
&saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME,
|
||||
timer.ElapsedNanos());
|
||||
} else {
|
||||
assert(last_not_merge_type == kTypeValue);
|
||||
MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(),
|
||||
&pinned_value_,
|
||||
merge_context_.GetOperands(), &saved_value_,
|
||||
logger_, statistics_, env_, &pinned_value_);
|
||||
std::string last_put_value = saved_value_;
|
||||
Slice temp_slice(last_put_value);
|
||||
{
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
user_merge_operator_->FullMerge(saved_key_.GetKey(), &temp_slice,
|
||||
operands, &saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME,
|
||||
timer.ElapsedNanos());
|
||||
}
|
||||
}
|
||||
break;
|
||||
case kTypeValue:
|
||||
@@ -662,9 +521,6 @@ bool DBIter::FindValueForCurrentKey() {
|
||||
// This function is used in FindValueForCurrentKey.
|
||||
// We use Seek() function instead of Prev() to find necessary value
|
||||
bool DBIter::FindValueForCurrentKeyUsingSeek() {
|
||||
// FindValueForCurrentKey will enable pinning before calling
|
||||
// FindValueForCurrentKeyUsingSeek()
|
||||
assert(pinned_iters_mgr_.PinningEnabled());
|
||||
std::string last_key;
|
||||
AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetKey(), sequence_,
|
||||
kValueTypeForSeek));
|
||||
@@ -678,8 +534,7 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
|
||||
if (ikey.type == kTypeValue || ikey.type == kTypeDeletion ||
|
||||
ikey.type == kTypeSingleDeletion) {
|
||||
if (ikey.type == kTypeValue) {
|
||||
assert(iter_->IsValuePinned());
|
||||
pinned_value_ = iter_->value();
|
||||
saved_value_ = iter_->value().ToString();
|
||||
valid_ = true;
|
||||
return true;
|
||||
}
|
||||
@@ -689,13 +544,11 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
|
||||
|
||||
// kTypeMerge. We need to collect all kTypeMerge values and save them
|
||||
// in operands
|
||||
current_entry_is_merged_ = true;
|
||||
merge_context_.Clear();
|
||||
std::deque<std::string> operands;
|
||||
while (iter_->Valid() &&
|
||||
user_comparator_->Equal(ikey.user_key, saved_key_.GetKey()) &&
|
||||
ikey.type == kTypeMerge) {
|
||||
merge_context_.PushOperand(iter_->value(),
|
||||
iter_->IsValuePinned() /* operand_pinned */);
|
||||
operands.push_front(iter_->value().ToString());
|
||||
iter_->Next();
|
||||
FindParseableKey(&ikey, kForward);
|
||||
}
|
||||
@@ -703,9 +556,13 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
|
||||
if (!iter_->Valid() ||
|
||||
!user_comparator_->Equal(ikey.user_key, saved_key_.GetKey()) ||
|
||||
ikey.type == kTypeDeletion || ikey.type == kTypeSingleDeletion) {
|
||||
MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(), nullptr,
|
||||
merge_context_.GetOperands(), &saved_value_,
|
||||
logger_, statistics_, env_, &pinned_value_);
|
||||
{
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
user_merge_operator_->FullMerge(saved_key_.GetKey(), nullptr, operands,
|
||||
&saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME, timer.ElapsedNanos());
|
||||
}
|
||||
// Make iter_ valid and point to saved_key_
|
||||
if (!iter_->Valid() ||
|
||||
!user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
|
||||
@@ -717,9 +574,13 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
|
||||
}
|
||||
|
||||
const Slice& val = iter_->value();
|
||||
MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(), &val,
|
||||
merge_context_.GetOperands(), &saved_value_,
|
||||
logger_, statistics_, env_, &pinned_value_);
|
||||
{
|
||||
StopWatchNano timer(env_, statistics_ != nullptr);
|
||||
PERF_TIMER_GUARD(merge_operator_time_nanos);
|
||||
user_merge_operator_->FullMerge(saved_key_.GetKey(), &val, operands,
|
||||
&saved_value_, logger_);
|
||||
RecordTick(statistics_, MERGE_OPERATION_TOTAL_TIME, timer.ElapsedNanos());
|
||||
}
|
||||
valid_ = true;
|
||||
return true;
|
||||
}
|
||||
@@ -783,7 +644,6 @@ void DBIter::FindParseableKey(ParsedInternalKey* ikey, Direction direction) {
|
||||
|
||||
void DBIter::Seek(const Slice& target) {
|
||||
StopWatch sw(env_, statistics_, DB_SEEK);
|
||||
ReleaseTempPinnedData();
|
||||
saved_key_.Clear();
|
||||
// now savved_key is used to store internal key.
|
||||
saved_key_.SetInternalKey(target, sequence_);
|
||||
@@ -795,15 +655,9 @@ void DBIter::Seek(const Slice& target) {
|
||||
|
||||
RecordTick(statistics_, NUMBER_DB_SEEK);
|
||||
if (iter_->Valid()) {
|
||||
if (prefix_extractor_ && prefix_same_as_start_) {
|
||||
prefix_start_key_ = prefix_extractor_->Transform(target);
|
||||
}
|
||||
direction_ = kForward;
|
||||
ClearSavedValue();
|
||||
FindNextUserEntry(false /* not skipping */, prefix_same_as_start_);
|
||||
if (!valid_) {
|
||||
prefix_start_key_.clear();
|
||||
}
|
||||
FindNextUserEntry(false /* not skipping */);
|
||||
if (statistics_ != nullptr) {
|
||||
if (valid_) {
|
||||
RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
|
||||
@@ -813,10 +667,6 @@ void DBIter::Seek(const Slice& target) {
|
||||
} else {
|
||||
valid_ = false;
|
||||
}
|
||||
if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
|
||||
prefix_start_buf_.SetKey(prefix_start_key_);
|
||||
prefix_start_key_ = prefix_start_buf_.GetKey();
|
||||
}
|
||||
}
|
||||
|
||||
void DBIter::SeekToFirst() {
|
||||
@@ -826,7 +676,6 @@ void DBIter::SeekToFirst() {
|
||||
max_skip_ = std::numeric_limits<uint64_t>::max();
|
||||
}
|
||||
direction_ = kForward;
|
||||
ReleaseTempPinnedData();
|
||||
ClearSavedValue();
|
||||
|
||||
{
|
||||
@@ -836,7 +685,7 @@ void DBIter::SeekToFirst() {
|
||||
|
||||
RecordTick(statistics_, NUMBER_DB_SEEK);
|
||||
if (iter_->Valid()) {
|
||||
FindNextUserEntry(false /* not skipping */, false /* no prefix check */);
|
||||
FindNextUserEntry(false /* not skipping */);
|
||||
if (statistics_ != nullptr) {
|
||||
if (valid_) {
|
||||
RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
|
||||
@@ -846,10 +695,6 @@ void DBIter::SeekToFirst() {
|
||||
} else {
|
||||
valid_ = false;
|
||||
}
|
||||
if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
|
||||
prefix_start_buf_.SetKey(prefix_extractor_->Transform(saved_key_.GetKey()));
|
||||
prefix_start_key_ = prefix_start_buf_.GetKey();
|
||||
}
|
||||
}
|
||||
|
||||
void DBIter::SeekToLast() {
|
||||
@@ -859,7 +704,6 @@ void DBIter::SeekToLast() {
|
||||
max_skip_ = std::numeric_limits<uint64_t>::max();
|
||||
}
|
||||
direction_ = kReverse;
|
||||
ReleaseTempPinnedData();
|
||||
ClearSavedValue();
|
||||
|
||||
{
|
||||
@@ -870,7 +714,7 @@ void DBIter::SeekToLast() {
|
||||
// it will seek to the last key before the
|
||||
// ReadOptions.iterate_upper_bound
|
||||
if (iter_->Valid() && iterate_upper_bound_ != nullptr) {
|
||||
saved_key_.SetKey(*iterate_upper_bound_, false /* copy */);
|
||||
saved_key_.SetKey(*iterate_upper_bound_);
|
||||
std::string last_key;
|
||||
AppendInternalKey(&last_key,
|
||||
ParsedInternalKey(saved_key_.GetKey(), kMaxSequenceNumber,
|
||||
@@ -896,32 +740,24 @@ void DBIter::SeekToLast() {
|
||||
RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
|
||||
}
|
||||
}
|
||||
if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
|
||||
prefix_start_buf_.SetKey(prefix_extractor_->Transform(saved_key_.GetKey()));
|
||||
prefix_start_key_ = prefix_start_buf_.GetKey();
|
||||
}
|
||||
}
|
||||
|
||||
Iterator* NewDBIterator(Env* env, const ImmutableCFOptions& ioptions,
|
||||
const Comparator* user_key_comparator,
|
||||
InternalIterator* internal_iter,
|
||||
Iterator* internal_iter,
|
||||
const SequenceNumber& sequence,
|
||||
uint64_t max_sequential_skip_in_iterations,
|
||||
uint64_t version_number,
|
||||
const Slice* iterate_upper_bound,
|
||||
bool prefix_same_as_start, bool pin_data) {
|
||||
DBIter* db_iter =
|
||||
new DBIter(env, ioptions, user_key_comparator, internal_iter, sequence,
|
||||
false, max_sequential_skip_in_iterations, version_number,
|
||||
iterate_upper_bound, prefix_same_as_start, pin_data);
|
||||
return db_iter;
|
||||
const Slice* iterate_upper_bound) {
|
||||
return new DBIter(env, ioptions, user_key_comparator, internal_iter, sequence,
|
||||
false, max_sequential_skip_in_iterations,
|
||||
iterate_upper_bound);
|
||||
}
|
||||
|
||||
ArenaWrappedDBIter::~ArenaWrappedDBIter() { db_iter_->~DBIter(); }
|
||||
|
||||
void ArenaWrappedDBIter::SetDBIter(DBIter* iter) { db_iter_ = iter; }
|
||||
|
||||
void ArenaWrappedDBIter::SetIterUnderDBIter(InternalIterator* iter) {
|
||||
void ArenaWrappedDBIter::SetIterUnderDBIter(Iterator* iter) {
|
||||
static_cast<DBIter*>(db_iter_)->SetIter(iter);
|
||||
}
|
||||
|
||||
@@ -936,10 +772,6 @@ inline void ArenaWrappedDBIter::Prev() { db_iter_->Prev(); }
|
||||
inline Slice ArenaWrappedDBIter::key() const { return db_iter_->key(); }
|
||||
inline Slice ArenaWrappedDBIter::value() const { return db_iter_->value(); }
|
||||
inline Status ArenaWrappedDBIter::status() const { return db_iter_->status(); }
|
||||
inline Status ArenaWrappedDBIter::GetProperty(std::string prop_name,
|
||||
std::string* prop) {
|
||||
return db_iter_->GetProperty(prop_name, prop);
|
||||
}
|
||||
void ArenaWrappedDBIter::RegisterCleanup(CleanupFunction function, void* arg1,
|
||||
void* arg2) {
|
||||
db_iter_->RegisterCleanup(function, arg1, arg2);
|
||||
@@ -947,17 +779,16 @@ void ArenaWrappedDBIter::RegisterCleanup(CleanupFunction function, void* arg1,
|
||||
|
||||
ArenaWrappedDBIter* NewArenaWrappedDbIterator(
|
||||
Env* env, const ImmutableCFOptions& ioptions,
|
||||
const Comparator* user_key_comparator, const SequenceNumber& sequence,
|
||||
uint64_t max_sequential_skip_in_iterations, uint64_t version_number,
|
||||
const Slice* iterate_upper_bound, bool prefix_same_as_start,
|
||||
bool pin_data) {
|
||||
const Comparator* user_key_comparator,
|
||||
const SequenceNumber& sequence,
|
||||
uint64_t max_sequential_skip_in_iterations,
|
||||
const Slice* iterate_upper_bound) {
|
||||
ArenaWrappedDBIter* iter = new ArenaWrappedDBIter();
|
||||
Arena* arena = iter->GetArena();
|
||||
auto mem = arena->AllocateAligned(sizeof(DBIter));
|
||||
DBIter* db_iter =
|
||||
new (mem) DBIter(env, ioptions, user_key_comparator, nullptr, sequence,
|
||||
true, max_sequential_skip_in_iterations, version_number,
|
||||
iterate_upper_bound, prefix_same_as_start, pin_data);
|
||||
DBIter* db_iter = new (mem) DBIter(env, ioptions, user_key_comparator,
|
||||
nullptr, sequence, true, max_sequential_skip_in_iterations,
|
||||
iterate_upper_bound);
|
||||
|
||||
iter->SetDBIter(db_iter);
|
||||
|
||||
|
||||
+13
-18
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -9,29 +9,27 @@
|
||||
|
||||
#pragma once
|
||||
#include <stdint.h>
|
||||
#include <string>
|
||||
#include "db/dbformat.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/iterator.h"
|
||||
#include "db/dbformat.h"
|
||||
#include "util/arena.h"
|
||||
#include "util/autovector.h"
|
||||
#include "util/cf_options.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class Arena;
|
||||
class DBIter;
|
||||
class InternalIterator;
|
||||
|
||||
// Return a new iterator that converts internal keys (yielded by
|
||||
// "*internal_iter") that were live at the specified "sequence" number
|
||||
// into appropriate user keys.
|
||||
extern Iterator* NewDBIterator(
|
||||
Env* env, const ImmutableCFOptions& options,
|
||||
const Comparator* user_key_comparator, InternalIterator* internal_iter,
|
||||
const SequenceNumber& sequence, uint64_t max_sequential_skip_in_iterations,
|
||||
uint64_t version_number, const Slice* iterate_upper_bound = nullptr,
|
||||
bool prefix_same_as_start = false, bool pin_data = false);
|
||||
Env* env,
|
||||
const ImmutableCFOptions& options,
|
||||
const Comparator *user_key_comparator,
|
||||
Iterator* internal_iter,
|
||||
const SequenceNumber& sequence,
|
||||
uint64_t max_sequential_skip_in_iterations,
|
||||
const Slice* iterate_upper_bound = nullptr);
|
||||
|
||||
// A wrapper iterator which wraps DB Iterator and the arena, with which the DB
|
||||
// iterator is supposed be allocated. This class is used as an entry point of
|
||||
@@ -52,7 +50,7 @@ class ArenaWrappedDBIter : public Iterator {
|
||||
|
||||
// Set the internal iterator wrapped inside the DB Iterator. Usually it is
|
||||
// a merging iterator.
|
||||
virtual void SetIterUnderDBIter(InternalIterator* iter);
|
||||
virtual void SetIterUnderDBIter(Iterator* iter);
|
||||
virtual bool Valid() const override;
|
||||
virtual void SeekToFirst() override;
|
||||
virtual void SeekToLast() override;
|
||||
@@ -62,9 +60,7 @@ class ArenaWrappedDBIter : public Iterator {
|
||||
virtual Slice key() const override;
|
||||
virtual Slice value() const override;
|
||||
virtual Status status() const override;
|
||||
|
||||
void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
|
||||
virtual Status GetProperty(std::string prop_name, std::string* prop) override;
|
||||
|
||||
private:
|
||||
DBIter* db_iter_;
|
||||
@@ -74,9 +70,8 @@ class ArenaWrappedDBIter : public Iterator {
|
||||
// Generate the arena wrapped iterator class.
|
||||
extern ArenaWrappedDBIter* NewArenaWrappedDbIterator(
|
||||
Env* env, const ImmutableCFOptions& options,
|
||||
const Comparator* user_key_comparator, const SequenceNumber& sequence,
|
||||
uint64_t max_sequential_skip_in_iterations, uint64_t version_number,
|
||||
const Slice* iterate_upper_bound = nullptr,
|
||||
bool prefix_same_as_start = false, bool pin_data = false);
|
||||
const Comparator* user_key_comparator,
|
||||
const SequenceNumber& sequence, uint64_t max_sequential_skip_in_iterations,
|
||||
const Slice* iterate_upper_bound = nullptr);
|
||||
|
||||
} // namespace rocksdb
|
||||
+103
-98
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -29,7 +29,7 @@ static uint64_t TestGetTickerCount(const Options& options,
|
||||
return options.statistics->getTickerCount(ticker_type);
|
||||
}
|
||||
|
||||
class TestIterator : public InternalIterator {
|
||||
class TestIterator : public Iterator {
|
||||
public:
|
||||
explicit TestIterator(const Comparator* comparator)
|
||||
: initialized_(false),
|
||||
@@ -150,9 +150,6 @@ class TestIterator : public InternalIterator {
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
virtual bool IsKeyPinned() const override { return true; }
|
||||
virtual bool IsValuePinned() const override { return true; }
|
||||
|
||||
private:
|
||||
bool initialized_;
|
||||
bool valid_;
|
||||
@@ -184,9 +181,10 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
internal_iter->AddPut("b", "val_b");
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 10,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -217,7 +215,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
10, options.max_sequential_skip_in_iterations));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -243,8 +241,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
10, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -276,8 +273,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
10, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -312,8 +308,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
10, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
@@ -342,8 +337,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
7, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
7, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
SetPerfLevel(kEnableCount);
|
||||
ASSERT_TRUE(GetPerfLevel() == kEnableCount);
|
||||
@@ -380,8 +374,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
4, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
4, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -406,8 +399,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
10, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
@@ -429,8 +421,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
10, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -465,8 +456,7 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
7, options.max_sequential_skip_in_iterations, 0,
|
||||
ro.iterate_upper_bound));
|
||||
7, options.max_sequential_skip_in_iterations, ro.iterate_upper_bound));
|
||||
|
||||
SetPerfLevel(kEnableCount);
|
||||
ASSERT_TRUE(GetPerfLevel() == kEnableCount);
|
||||
@@ -492,9 +482,10 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
internal_iter->AddPut("b", "val_b");
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 10,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -533,9 +524,10 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
internal_iter->AddPut("b", "val_b");
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
2, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "b");
|
||||
@@ -563,9 +555,10 @@ TEST_F(DBIteratorTest, DBIteratorPrevNext) {
|
||||
internal_iter->AddPut("c", "val_c");
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 10,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "c");
|
||||
@@ -590,9 +583,10 @@ TEST_F(DBIteratorTest, DBIteratorEmpty) {
|
||||
TestIterator* internal_iter = new TestIterator(BytewiseComparator());
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
0, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 0,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
}
|
||||
@@ -601,9 +595,10 @@ TEST_F(DBIteratorTest, DBIteratorEmpty) {
|
||||
TestIterator* internal_iter = new TestIterator(BytewiseComparator());
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
0, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 0,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
}
|
||||
@@ -622,9 +617,10 @@ TEST_F(DBIteratorTest, DBIteratorUseSkipCountSkips) {
|
||||
}
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 2,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "c");
|
||||
@@ -663,8 +659,9 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, i + 2, options.max_sequential_skip_in_iterations, 0));
|
||||
env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, i + 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -698,8 +695,9 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, i + 2, options.max_sequential_skip_in_iterations, 0));
|
||||
env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, i + 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -726,8 +724,9 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, 202, options.max_sequential_skip_in_iterations, 0));
|
||||
env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 202,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -757,9 +756,10 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
}
|
||||
internal_iter->AddPut("c", "200");
|
||||
internal_iter->Finish();
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, i, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, i,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
|
||||
@@ -773,9 +773,10 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
}
|
||||
internal_iter->AddPut("c", "200");
|
||||
internal_iter->Finish();
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
200, options.max_sequential_skip_in_iterations, 0));
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, 200,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "c");
|
||||
@@ -808,8 +809,9 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, i + 2, options.max_sequential_skip_in_iterations, 0));
|
||||
env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, i + 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -842,8 +844,9 @@ TEST_F(DBIteratorTest, DBIteratorUseSkip) {
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, i + 2, options.max_sequential_skip_in_iterations, 0));
|
||||
env_, ImmutableCFOptions(options),
|
||||
BytewiseComparator(), internal_iter, i + 2,
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -884,7 +887,7 @@ TEST_F(DBIteratorTest, DBIterator1) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 1,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -910,7 +913,7 @@ TEST_F(DBIteratorTest, DBIterator2) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 0,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -933,7 +936,7 @@ TEST_F(DBIteratorTest, DBIterator3) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 2,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -955,7 +958,7 @@ TEST_F(DBIteratorTest, DBIterator4) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 4,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -984,7 +987,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
0, options.max_sequential_skip_in_iterations, 0));
|
||||
0, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1006,7 +1009,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
1, options.max_sequential_skip_in_iterations, 0));
|
||||
1, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1028,7 +1031,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
2, options.max_sequential_skip_in_iterations, 0));
|
||||
2, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1050,7 +1053,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
3, options.max_sequential_skip_in_iterations, 0));
|
||||
3, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1072,7 +1075,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
4, options.max_sequential_skip_in_iterations, 0));
|
||||
4, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1094,7 +1097,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
5, options.max_sequential_skip_in_iterations, 0));
|
||||
5, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1116,7 +1119,7 @@ TEST_F(DBIteratorTest, DBIterator5) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
6, options.max_sequential_skip_in_iterations, 0));
|
||||
6, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1142,7 +1145,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
0, options.max_sequential_skip_in_iterations, 0));
|
||||
0, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1164,7 +1167,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
1, options.max_sequential_skip_in_iterations, 0));
|
||||
1, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1186,7 +1189,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
2, options.max_sequential_skip_in_iterations, 0));
|
||||
2, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1208,7 +1211,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
3, options.max_sequential_skip_in_iterations, 0));
|
||||
3, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(!db_iter->Valid());
|
||||
}
|
||||
@@ -1226,7 +1229,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
4, options.max_sequential_skip_in_iterations, 0));
|
||||
4, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1248,7 +1251,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
5, options.max_sequential_skip_in_iterations, 0));
|
||||
5, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1270,7 +1273,7 @@ TEST_F(DBIteratorTest, DBIterator6) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
6, options.max_sequential_skip_in_iterations, 0));
|
||||
6, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1308,7 +1311,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
0, options.max_sequential_skip_in_iterations, 0));
|
||||
0, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1342,7 +1345,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
2, options.max_sequential_skip_in_iterations, 0));
|
||||
2, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1382,7 +1385,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
4, options.max_sequential_skip_in_iterations, 0));
|
||||
4, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1422,7 +1425,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
5, options.max_sequential_skip_in_iterations, 0));
|
||||
5, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1467,7 +1470,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
6, options.max_sequential_skip_in_iterations, 0));
|
||||
6, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1513,7 +1516,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
7, options.max_sequential_skip_in_iterations, 0));
|
||||
7, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1553,7 +1556,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
9, options.max_sequential_skip_in_iterations, 0));
|
||||
9, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1599,7 +1602,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
13, options.max_sequential_skip_in_iterations, 0));
|
||||
13, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1646,7 +1649,7 @@ TEST_F(DBIteratorTest, DBIterator7) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
14, options.max_sequential_skip_in_iterations, 0));
|
||||
14, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
|
||||
@@ -1675,7 +1678,7 @@ TEST_F(DBIteratorTest, DBIterator8) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
10, options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "b");
|
||||
@@ -1704,7 +1707,7 @@ TEST_F(DBIteratorTest, DBIterator9) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
10, options.max_sequential_skip_in_iterations));
|
||||
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -1751,7 +1754,7 @@ TEST_F(DBIteratorTest, DBIterator10) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, options.max_sequential_skip_in_iterations, 0));
|
||||
10, options.max_sequential_skip_in_iterations));
|
||||
|
||||
db_iter->Seek("c");
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
@@ -1775,9 +1778,9 @@ TEST_F(DBIteratorTest, SeekToLastOccurrenceSeq0) {
|
||||
internal_iter->AddPut("b", "2");
|
||||
internal_iter->Finish();
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, 10, 0 /* force seek */, 0));
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter,
|
||||
10, 0 /* force seek */));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1804,7 +1807,7 @@ TEST_F(DBIteratorTest, DBIterator11) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(NewDBIterator(
|
||||
env_, ImmutableCFOptions(options), BytewiseComparator(), internal_iter, 1,
|
||||
options.max_sequential_skip_in_iterations, 0));
|
||||
options.max_sequential_skip_in_iterations));
|
||||
db_iter->SeekToFirst();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "a");
|
||||
@@ -1829,7 +1832,7 @@ TEST_F(DBIteratorTest, DBIterator12) {
|
||||
|
||||
std::unique_ptr<Iterator> db_iter(
|
||||
NewDBIterator(env_, ImmutableCFOptions(options), BytewiseComparator(),
|
||||
internal_iter, 10, 0, 0));
|
||||
internal_iter, 10, 0));
|
||||
db_iter->SeekToLast();
|
||||
ASSERT_TRUE(db_iter->Valid());
|
||||
ASSERT_EQ(db_iter->key().ToString(), "c");
|
||||
@@ -1861,17 +1864,16 @@ class DBIterWithMergeIterTest : public testing::Test {
|
||||
internal_iter2_->Add("d", kTypeValue, "7", 3u);
|
||||
internal_iter2_->Finish();
|
||||
|
||||
std::vector<InternalIterator*> child_iters;
|
||||
std::vector<Iterator*> child_iters;
|
||||
child_iters.push_back(internal_iter1_);
|
||||
child_iters.push_back(internal_iter2_);
|
||||
InternalKeyComparator icomp(BytewiseComparator());
|
||||
InternalIterator* merge_iter =
|
||||
NewMergingIterator(&icomp_, &child_iters[0], 2u);
|
||||
Iterator* merge_iter = NewMergingIterator(&icomp_, &child_iters[0], 2u);
|
||||
|
||||
db_iter_.reset(NewDBIterator(env_, ImmutableCFOptions(options_),
|
||||
BytewiseComparator(), merge_iter,
|
||||
8 /* read data earlier than seqId 8 */,
|
||||
3 /* max iterators before reseek */, 0));
|
||||
3 /* max iterators before reseek */));
|
||||
}
|
||||
|
||||
Env* env_;
|
||||
@@ -1940,6 +1942,8 @@ TEST_F(DBIterWithMergeIterTest, InnerMergeIterator2) {
|
||||
ASSERT_EQ(db_iter_->value().ToString(), "4");
|
||||
}
|
||||
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
TEST_F(DBIterWithMergeIterTest, InnerMergeIteratorDataRace1) {
|
||||
// Test Prev() when one child iterator is at its end but more rows
|
||||
// are added.
|
||||
@@ -2290,6 +2294,7 @@ TEST_F(DBIterWithMergeIterTest, InnerMergeIteratorDataRace8) {
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
#endif // #if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,10 +10,10 @@
|
||||
// Introduction of SyncPoint effectively disabled building and running this test
|
||||
// in Release build.
|
||||
// which is a pity, it is a good test
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/db_test_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -27,7 +27,7 @@ class DBTestXactLogIterator : public DBTestBase {
|
||||
Status status = dbfull()->GetUpdatesSince(seq, &iter);
|
||||
EXPECT_OK(status);
|
||||
EXPECT_TRUE(iter->Valid());
|
||||
return iter;
|
||||
return std::move(iter);
|
||||
}
|
||||
};
|
||||
|
||||
@@ -277,10 +277,10 @@ TEST_F(DBTestXactLogIterator, TransactionLogIteratorBlobs) {
|
||||
}
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // !defined(ROCKSDB_LITE)
|
||||
#endif // !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
|
||||
@@ -1,137 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
#include <limits>
|
||||
#include <string>
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/sync_point.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class DBOptionsTest : public DBTestBase {
|
||||
public:
|
||||
DBOptionsTest() : DBTestBase("/db_options_test") {}
|
||||
};
|
||||
|
||||
// RocksDB lite don't support dynamic options.
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBOptionsTest, EnableAutoCompactionAndTriggerStall) {
|
||||
const std::string kValue(1024, 'v');
|
||||
for (int method_type = 0; method_type < 2; method_type++) {
|
||||
for (int option_type = 0; option_type < 4; option_type++) {
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.disable_auto_compactions = true;
|
||||
options.write_buffer_size = 1024 * 1024 * 10;
|
||||
options.compression = CompressionType::kNoCompression;
|
||||
options.level0_file_num_compaction_trigger = 1;
|
||||
options.level0_stop_writes_trigger = std::numeric_limits<int>::max();
|
||||
options.level0_slowdown_writes_trigger = std::numeric_limits<int>::max();
|
||||
options.hard_pending_compaction_bytes_limit =
|
||||
std::numeric_limits<uint64_t>::max();
|
||||
options.soft_pending_compaction_bytes_limit =
|
||||
std::numeric_limits<uint64_t>::max();
|
||||
|
||||
DestroyAndReopen(options);
|
||||
int i = 0;
|
||||
for (; i < 1024; i++) {
|
||||
Put(Key(i), kValue);
|
||||
}
|
||||
Flush();
|
||||
for (; i < 1024 * 2; i++) {
|
||||
Put(Key(i), kValue);
|
||||
}
|
||||
Flush();
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
ASSERT_EQ(2, NumTableFilesAtLevel(0));
|
||||
uint64_t l0_size = SizeAtLevel(0);
|
||||
|
||||
switch (option_type) {
|
||||
case 0:
|
||||
// test with level0_stop_writes_trigger
|
||||
options.level0_stop_writes_trigger = 2;
|
||||
options.level0_slowdown_writes_trigger = 2;
|
||||
break;
|
||||
case 1:
|
||||
options.level0_slowdown_writes_trigger = 2;
|
||||
break;
|
||||
case 2:
|
||||
options.hard_pending_compaction_bytes_limit = l0_size;
|
||||
options.soft_pending_compaction_bytes_limit = l0_size;
|
||||
break;
|
||||
case 3:
|
||||
options.soft_pending_compaction_bytes_limit = l0_size;
|
||||
break;
|
||||
}
|
||||
Reopen(options);
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().NeedsDelay());
|
||||
|
||||
SyncPoint::GetInstance()->LoadDependency(
|
||||
{{"DBOptionsTest::EnableAutoCompactionAndTriggerStall:1",
|
||||
"BackgroundCallCompaction:0"},
|
||||
{"DBImpl::BackgroundCompaction():BeforePickCompaction",
|
||||
"DBOptionsTest::EnableAutoCompactionAndTriggerStall:2"},
|
||||
{"DBOptionsTest::EnableAutoCompactionAndTriggerStall:3",
|
||||
"DBImpl::BackgroundCompaction():AfterPickCompaction"}});
|
||||
// Block background compaction.
|
||||
SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
switch (method_type) {
|
||||
case 0:
|
||||
ASSERT_OK(
|
||||
dbfull()->SetOptions({{"disable_auto_compactions", "false"}}));
|
||||
break;
|
||||
case 1:
|
||||
ASSERT_OK(dbfull()->EnableAutoCompaction(
|
||||
{dbfull()->DefaultColumnFamily()}));
|
||||
break;
|
||||
}
|
||||
TEST_SYNC_POINT("DBOptionsTest::EnableAutoCompactionAndTriggerStall:1");
|
||||
// Wait for stall condition recalculate.
|
||||
TEST_SYNC_POINT("DBOptionsTest::EnableAutoCompactionAndTriggerStall:2");
|
||||
|
||||
switch (option_type) {
|
||||
case 0:
|
||||
ASSERT_TRUE(dbfull()->TEST_write_controler().IsStopped());
|
||||
break;
|
||||
case 1:
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
|
||||
ASSERT_TRUE(dbfull()->TEST_write_controler().NeedsDelay());
|
||||
break;
|
||||
case 2:
|
||||
ASSERT_TRUE(dbfull()->TEST_write_controler().IsStopped());
|
||||
break;
|
||||
case 3:
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
|
||||
ASSERT_TRUE(dbfull()->TEST_write_controler().NeedsDelay());
|
||||
break;
|
||||
}
|
||||
TEST_SYNC_POINT("DBOptionsTest::EnableAutoCompactionAndTriggerStall:3");
|
||||
|
||||
// Background compaction executed.
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
|
||||
ASSERT_FALSE(dbfull()->TEST_write_controler().NeedsDelay());
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,801 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/port.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "rocksdb/sst_file_manager.h"
|
||||
#include "util/sst_file_manager_impl.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
class DBSSTTest : public DBTestBase {
|
||||
public:
|
||||
DBSSTTest() : DBTestBase("/db_sst_test") {}
|
||||
};
|
||||
|
||||
TEST_F(DBSSTTest, DontDeletePendingOutputs) {
|
||||
Options options;
|
||||
options.env = env_;
|
||||
options.create_if_missing = true;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Every time we write to a table file, call FOF/POF with full DB scan. This
|
||||
// will make sure our pending_outputs_ protection work correctly
|
||||
std::function<void()> purge_obsolete_files_function = [&]() {
|
||||
JobContext job_context(0);
|
||||
dbfull()->TEST_LockMutex();
|
||||
dbfull()->FindObsoleteFiles(&job_context, true /*force*/);
|
||||
dbfull()->TEST_UnlockMutex();
|
||||
dbfull()->PurgeObsoleteFiles(job_context);
|
||||
job_context.Clean();
|
||||
};
|
||||
|
||||
env_->table_write_callback_ = &purge_obsolete_files_function;
|
||||
|
||||
for (int i = 0; i < 2; ++i) {
|
||||
ASSERT_OK(Put("a", "begin"));
|
||||
ASSERT_OK(Put("z", "end"));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
|
||||
// If pending output guard does not work correctly, PurgeObsoleteFiles() will
|
||||
// delete the file that Compaction is trying to create, causing this: error
|
||||
// db/db_test.cc:975: IO error:
|
||||
// /tmp/rocksdbtest-1552237650/db_test/000009.sst: No such file or directory
|
||||
Compact("a", "b");
|
||||
}
|
||||
|
||||
// 1 Create some SST files by inserting K-V pairs into DB
|
||||
// 2 Close DB and change suffix from ".sst" to ".ldb" for every other SST file
|
||||
// 3 Open DB and check if all key can be read
|
||||
TEST_F(DBSSTTest, SSTsWithLdbSuffixHandling) {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 110 << 10; // 110KB
|
||||
options.num_levels = 4;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
int key_id = 0;
|
||||
for (int i = 0; i < 10; ++i) {
|
||||
GenerateNewFile(&rnd, &key_id, false);
|
||||
}
|
||||
Flush();
|
||||
Close();
|
||||
int const num_files = GetSstFileCount(dbname_);
|
||||
ASSERT_GT(num_files, 0);
|
||||
|
||||
std::vector<std::string> filenames;
|
||||
GetSstFiles(dbname_, &filenames);
|
||||
int num_ldb_files = 0;
|
||||
for (size_t i = 0; i < filenames.size(); ++i) {
|
||||
if (i & 1) {
|
||||
continue;
|
||||
}
|
||||
std::string const rdb_name = dbname_ + "/" + filenames[i];
|
||||
std::string const ldb_name = Rocks2LevelTableFileName(rdb_name);
|
||||
ASSERT_TRUE(env_->RenameFile(rdb_name, ldb_name).ok());
|
||||
++num_ldb_files;
|
||||
}
|
||||
ASSERT_GT(num_ldb_files, 0);
|
||||
ASSERT_EQ(num_files, GetSstFileCount(dbname_));
|
||||
|
||||
Reopen(options);
|
||||
for (int k = 0; k < key_id; ++k) {
|
||||
ASSERT_NE("NOT_FOUND", Get(Key(k)));
|
||||
}
|
||||
Destroy(options);
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBSSTTest, DontDeleteMovedFile) {
|
||||
// This test triggers move compaction and verifies that the file is not
|
||||
// deleted when it's part of move compaction
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.create_if_missing = true;
|
||||
options.max_bytes_for_level_base = 1024 * 1024; // 1 MB
|
||||
options.level0_file_num_compaction_trigger =
|
||||
2; // trigger compaction when we have 2 files
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
// Create two 1MB sst files
|
||||
for (int i = 0; i < 2; ++i) {
|
||||
// Create 1MB sst file
|
||||
for (int j = 0; j < 100; ++j) {
|
||||
ASSERT_OK(Put(Key(i * 50 + j), RandomString(&rnd, 10 * 1024)));
|
||||
}
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
// this should execute both L0->L1 and L1->(move)->L2 compactions
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ("0,0,1", FilesPerLevel(0));
|
||||
|
||||
// If the moved file is actually deleted (the move-safeguard in
|
||||
// ~Version::Version() is not there), we get this failure:
|
||||
// Corruption: Can't access /000009.sst
|
||||
Reopen(options);
|
||||
}
|
||||
|
||||
// This reproduces a bug where we don't delete a file because when it was
|
||||
// supposed to be deleted, it was blocked by pending_outputs
|
||||
// Consider:
|
||||
// 1. current file_number is 13
|
||||
// 2. compaction (1) starts, blocks deletion of all files starting with 13
|
||||
// (pending outputs)
|
||||
// 3. file 13 is created by compaction (2)
|
||||
// 4. file 13 is consumed by compaction (3) and file 15 was created. Since file
|
||||
// 13 has no references, it is put into VersionSet::obsolete_files_
|
||||
// 5. FindObsoleteFiles() gets file 13 from VersionSet::obsolete_files_. File 13
|
||||
// is deleted from obsolete_files_ set.
|
||||
// 6. PurgeObsoleteFiles() tries to delete file 13, but this file is blocked by
|
||||
// pending outputs since compaction (1) is still running. It is not deleted and
|
||||
// it is not present in obsolete_files_ anymore. Therefore, we never delete it.
|
||||
TEST_F(DBSSTTest, DeleteObsoleteFilesPendingOutputs) {
|
||||
Options options = CurrentOptions();
|
||||
options.env = env_;
|
||||
options.write_buffer_size = 2 * 1024 * 1024; // 2 MB
|
||||
options.max_bytes_for_level_base = 1024 * 1024; // 1 MB
|
||||
options.level0_file_num_compaction_trigger =
|
||||
2; // trigger compaction when we have 2 files
|
||||
options.max_background_flushes = 2;
|
||||
options.max_background_compactions = 2;
|
||||
|
||||
OnFileDeletionListener* listener = new OnFileDeletionListener();
|
||||
options.listeners.emplace_back(listener);
|
||||
|
||||
Reopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
// Create two 1MB sst files
|
||||
for (int i = 0; i < 2; ++i) {
|
||||
// Create 1MB sst file
|
||||
for (int j = 0; j < 100; ++j) {
|
||||
ASSERT_OK(Put(Key(i * 50 + j), RandomString(&rnd, 10 * 1024)));
|
||||
}
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
// this should execute both L0->L1 and L1->(move)->L2 compactions
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ("0,0,1", FilesPerLevel(0));
|
||||
|
||||
test::SleepingBackgroundTask blocking_thread;
|
||||
port::Mutex mutex_;
|
||||
bool already_blocked(false);
|
||||
|
||||
// block the flush
|
||||
std::function<void()> block_first_time = [&]() {
|
||||
bool blocking = false;
|
||||
{
|
||||
MutexLock l(&mutex_);
|
||||
if (!already_blocked) {
|
||||
blocking = true;
|
||||
already_blocked = true;
|
||||
}
|
||||
}
|
||||
if (blocking) {
|
||||
blocking_thread.DoSleep();
|
||||
}
|
||||
};
|
||||
env_->table_write_callback_ = &block_first_time;
|
||||
// Insert 2.5MB data, which should trigger a flush because we exceed
|
||||
// write_buffer_size. The flush will be blocked with block_first_time
|
||||
// pending_file is protecting all the files created after
|
||||
for (int j = 0; j < 256; ++j) {
|
||||
ASSERT_OK(Put(Key(j), RandomString(&rnd, 10 * 1024)));
|
||||
}
|
||||
blocking_thread.WaitUntilSleeping();
|
||||
|
||||
ASSERT_OK(dbfull()->TEST_CompactRange(2, nullptr, nullptr));
|
||||
|
||||
ASSERT_EQ("0,0,0,1", FilesPerLevel(0));
|
||||
std::vector<LiveFileMetaData> metadata;
|
||||
db_->GetLiveFilesMetaData(&metadata);
|
||||
ASSERT_EQ(metadata.size(), 1U);
|
||||
auto file_on_L2 = metadata[0].name;
|
||||
listener->SetExpectedFileName(dbname_ + file_on_L2);
|
||||
|
||||
ASSERT_OK(dbfull()->TEST_CompactRange(3, nullptr, nullptr, nullptr,
|
||||
true /* disallow trivial move */));
|
||||
ASSERT_EQ("0,0,0,0,1", FilesPerLevel(0));
|
||||
|
||||
// finish the flush!
|
||||
blocking_thread.WakeUp();
|
||||
blocking_thread.WaitUntilDone();
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
ASSERT_EQ("1,0,0,0,1", FilesPerLevel(0));
|
||||
|
||||
metadata.clear();
|
||||
db_->GetLiveFilesMetaData(&metadata);
|
||||
ASSERT_EQ(metadata.size(), 2U);
|
||||
|
||||
// This file should have been deleted during last compaction
|
||||
ASSERT_EQ(Status::NotFound(), env_->FileExists(dbname_ + file_on_L2));
|
||||
listener->VerifyMatchedCount(1);
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBSSTTest, DBWithSstFileManager) {
|
||||
std::shared_ptr<SstFileManager> sst_file_manager(NewSstFileManager(env_));
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(sst_file_manager.get());
|
||||
|
||||
int files_added = 0;
|
||||
int files_deleted = 0;
|
||||
int files_moved = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"SstFileManagerImpl::OnAddFile", [&](void* arg) { files_added++; });
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"SstFileManagerImpl::OnDeleteFile", [&](void* arg) { files_deleted++; });
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"SstFileManagerImpl::OnMoveFile", [&](void* arg) { files_moved++; });
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.sst_file_manager = sst_file_manager;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
for (int i = 0; i < 25; i++) {
|
||||
GenerateNewRandomFile(&rnd);
|
||||
ASSERT_OK(Flush());
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
// Verify that we are tracking all sst files in dbname_
|
||||
ASSERT_EQ(sfm->GetTrackedFiles(), GetAllSSTFiles());
|
||||
}
|
||||
ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
|
||||
auto files_in_db = GetAllSSTFiles();
|
||||
// Verify that we are tracking all sst files in dbname_
|
||||
ASSERT_EQ(sfm->GetTrackedFiles(), files_in_db);
|
||||
// Verify the total files size
|
||||
uint64_t total_files_size = 0;
|
||||
for (auto& file_to_size : files_in_db) {
|
||||
total_files_size += file_to_size.second;
|
||||
}
|
||||
ASSERT_EQ(sfm->GetTotalSize(), total_files_size);
|
||||
// We flushed at least 25 files
|
||||
ASSERT_GE(files_added, 25);
|
||||
// Compaction must have deleted some files
|
||||
ASSERT_GT(files_deleted, 0);
|
||||
// No files were moved
|
||||
ASSERT_EQ(files_moved, 0);
|
||||
|
||||
Close();
|
||||
Reopen(options);
|
||||
ASSERT_EQ(sfm->GetTrackedFiles(), files_in_db);
|
||||
ASSERT_EQ(sfm->GetTotalSize(), total_files_size);
|
||||
|
||||
// Verify that we track all the files again after the DB is closed and opened
|
||||
Close();
|
||||
sst_file_manager.reset(NewSstFileManager(env_));
|
||||
options.sst_file_manager = sst_file_manager;
|
||||
sfm = static_cast<SstFileManagerImpl*>(sst_file_manager.get());
|
||||
|
||||
Reopen(options);
|
||||
ASSERT_EQ(sfm->GetTrackedFiles(), files_in_db);
|
||||
ASSERT_EQ(sfm->GetTotalSize(), total_files_size);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBSSTTest, RateLimitedDelete) {
|
||||
Destroy(last_options_);
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency({
|
||||
{"DBSSTTest::RateLimitedDelete:1",
|
||||
"DeleteScheduler::BackgroundEmptyTrash"},
|
||||
});
|
||||
|
||||
std::vector<uint64_t> penalties;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"DeleteScheduler::BackgroundEmptyTrash:Wait",
|
||||
[&](void* arg) { penalties.push_back(*(static_cast<int*>(arg))); });
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.env = env_;
|
||||
|
||||
std::string trash_dir = test::TmpDir(env_) + "/trash";
|
||||
int64_t rate_bytes_per_sec = 1024 * 10; // 10 Kbs / Sec
|
||||
Status s;
|
||||
options.sst_file_manager.reset(NewSstFileManager(
|
||||
env_, nullptr, trash_dir, rate_bytes_per_sec, false, &s));
|
||||
ASSERT_OK(s);
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());
|
||||
|
||||
ASSERT_OK(TryReopen(options));
|
||||
// Create 4 files in L0
|
||||
for (char v = 'a'; v <= 'd'; v++) {
|
||||
ASSERT_OK(Put("Key2", DummyString(1024, v)));
|
||||
ASSERT_OK(Put("Key3", DummyString(1024, v)));
|
||||
ASSERT_OK(Put("Key4", DummyString(1024, v)));
|
||||
ASSERT_OK(Put("Key1", DummyString(1024, v)));
|
||||
ASSERT_OK(Put("Key4", DummyString(1024, v)));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
// We created 4 sst files in L0
|
||||
ASSERT_EQ("4", FilesPerLevel(0));
|
||||
|
||||
std::vector<LiveFileMetaData> metadata;
|
||||
db_->GetLiveFilesMetaData(&metadata);
|
||||
|
||||
// Compaction will move the 4 files in L0 to trash and create 1 L1 file
|
||||
ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
ASSERT_EQ("0,1", FilesPerLevel(0));
|
||||
|
||||
uint64_t delete_start_time = env_->NowMicros();
|
||||
// Hold BackgroundEmptyTrash
|
||||
TEST_SYNC_POINT("DBSSTTest::RateLimitedDelete:1");
|
||||
sfm->WaitForEmptyTrash();
|
||||
uint64_t time_spent_deleting = env_->NowMicros() - delete_start_time;
|
||||
|
||||
uint64_t total_files_size = 0;
|
||||
uint64_t expected_penlty = 0;
|
||||
ASSERT_EQ(penalties.size(), metadata.size());
|
||||
for (size_t i = 0; i < metadata.size(); i++) {
|
||||
total_files_size += metadata[i].size;
|
||||
expected_penlty = ((total_files_size * 1000000) / rate_bytes_per_sec);
|
||||
ASSERT_EQ(expected_penlty, penalties[i]);
|
||||
}
|
||||
ASSERT_GT(time_spent_deleting, expected_penlty * 0.9);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
// Create a DB with 2 db_paths, and generate multiple files in the 2
|
||||
// db_paths using CompactRangeOptions, make sure that files that were
|
||||
// deleted from first db_path were deleted using DeleteScheduler and
|
||||
// files in the second path were not.
|
||||
TEST_F(DBSSTTest, DeleteSchedulerMultipleDBPaths) {
|
||||
int bg_delete_file = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"DeleteScheduler::DeleteTrashFile:DeleteFile",
|
||||
[&](void* arg) { bg_delete_file++; });
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.db_paths.emplace_back(dbname_, 1024 * 100);
|
||||
options.db_paths.emplace_back(dbname_ + "_2", 1024 * 100);
|
||||
options.env = env_;
|
||||
|
||||
std::string trash_dir = test::TmpDir(env_) + "/trash";
|
||||
int64_t rate_bytes_per_sec = 1024 * 1024; // 1 Mb / Sec
|
||||
Status s;
|
||||
options.sst_file_manager.reset(NewSstFileManager(
|
||||
env_, nullptr, trash_dir, rate_bytes_per_sec, false, &s));
|
||||
ASSERT_OK(s);
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());
|
||||
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Create 4 files in L0
|
||||
for (int i = 0; i < 4; i++) {
|
||||
ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'A')));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
// We created 4 sst files in L0
|
||||
ASSERT_EQ("4", FilesPerLevel(0));
|
||||
// Compaction will delete files from L0 in first db path and generate a new
|
||||
// file in L1 in second db path
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.target_path_id = 1;
|
||||
Slice begin("Key0");
|
||||
Slice end("Key3");
|
||||
ASSERT_OK(db_->CompactRange(compact_options, &begin, &end));
|
||||
ASSERT_EQ("0,1", FilesPerLevel(0));
|
||||
|
||||
// Create 4 files in L0
|
||||
for (int i = 4; i < 8; i++) {
|
||||
ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'B')));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
ASSERT_EQ("4,1", FilesPerLevel(0));
|
||||
|
||||
// Compaction will delete files from L0 in first db path and generate a new
|
||||
// file in L1 in second db path
|
||||
begin = "Key4";
|
||||
end = "Key7";
|
||||
ASSERT_OK(db_->CompactRange(compact_options, &begin, &end));
|
||||
ASSERT_EQ("0,2", FilesPerLevel(0));
|
||||
|
||||
sfm->WaitForEmptyTrash();
|
||||
ASSERT_EQ(bg_delete_file, 8);
|
||||
|
||||
compact_options.bottommost_level_compaction =
|
||||
BottommostLevelCompaction::kForce;
|
||||
ASSERT_OK(db_->CompactRange(compact_options, nullptr, nullptr));
|
||||
ASSERT_EQ("0,1", FilesPerLevel(0));
|
||||
|
||||
sfm->WaitForEmptyTrash();
|
||||
ASSERT_EQ(bg_delete_file, 8);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
TEST_F(DBSSTTest, DestroyDBWithRateLimitedDelete) {
|
||||
int bg_delete_file = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"DeleteScheduler::DeleteTrashFile:DeleteFile",
|
||||
[&](void* arg) { bg_delete_file++; });
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.env = env_;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Create 4 files in L0
|
||||
for (int i = 0; i < 4; i++) {
|
||||
ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'A')));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
// We created 4 sst files in L0
|
||||
ASSERT_EQ("4", FilesPerLevel(0));
|
||||
|
||||
// Close DB and destroy it using DeleteScheduler
|
||||
Close();
|
||||
std::string trash_dir = test::TmpDir(env_) + "/trash";
|
||||
int64_t rate_bytes_per_sec = 1024 * 1024; // 1 Mb / Sec
|
||||
Status s;
|
||||
options.sst_file_manager.reset(NewSstFileManager(
|
||||
env_, nullptr, trash_dir, rate_bytes_per_sec, false, &s));
|
||||
ASSERT_OK(s);
|
||||
ASSERT_OK(DestroyDB(dbname_, options));
|
||||
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());
|
||||
sfm->WaitForEmptyTrash();
|
||||
// We have deleted the 4 sst files in the delete_scheduler
|
||||
ASSERT_EQ(bg_delete_file, 4);
|
||||
}
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
TEST_F(DBSSTTest, DBWithMaxSpaceAllowed) {
|
||||
std::shared_ptr<SstFileManager> sst_file_manager(NewSstFileManager(env_));
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(sst_file_manager.get());
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.sst_file_manager = sst_file_manager;
|
||||
options.disable_auto_compactions = true;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
|
||||
// Generate a file containing 100 keys.
|
||||
for (int i = 0; i < 100; i++) {
|
||||
ASSERT_OK(Put(Key(i), RandomString(&rnd, 50)));
|
||||
}
|
||||
ASSERT_OK(Flush());
|
||||
|
||||
uint64_t first_file_size = 0;
|
||||
auto files_in_db = GetAllSSTFiles(&first_file_size);
|
||||
ASSERT_EQ(sfm->GetTotalSize(), first_file_size);
|
||||
|
||||
// Set the maximum allowed space usage to the current total size
|
||||
sfm->SetMaxAllowedSpaceUsage(first_file_size + 1);
|
||||
|
||||
ASSERT_OK(Put("key1", "val1"));
|
||||
// This flush will cause bg_error_ and will fail
|
||||
ASSERT_NOK(Flush());
|
||||
}
|
||||
|
||||
TEST_F(DBSSTTest, DBWithMaxSpaceAllowedRandomized) {
|
||||
// This test will set a maximum allowed space for the DB, then it will
|
||||
// keep filling the DB until the limit is reached and bg_error_ is set.
|
||||
// When bg_error_ is set we will verify that the DB size is greater
|
||||
// than the limit.
|
||||
|
||||
std::vector<int> max_space_limits_mbs = {1, 2, 4, 8, 10};
|
||||
|
||||
bool bg_error_set = false;
|
||||
uint64_t total_sst_files_size = 0;
|
||||
|
||||
int reached_max_space_on_flush = 0;
|
||||
int reached_max_space_on_compaction = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"DBImpl::FlushMemTableToOutputFile:MaxAllowedSpaceReached",
|
||||
[&](void* arg) {
|
||||
bg_error_set = true;
|
||||
GetAllSSTFiles(&total_sst_files_size);
|
||||
reached_max_space_on_flush++;
|
||||
});
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"CompactionJob::FinishCompactionOutputFile:MaxAllowedSpaceReached",
|
||||
[&](void* arg) {
|
||||
bg_error_set = true;
|
||||
GetAllSSTFiles(&total_sst_files_size);
|
||||
reached_max_space_on_compaction++;
|
||||
});
|
||||
|
||||
for (auto limit_mb : max_space_limits_mbs) {
|
||||
bg_error_set = false;
|
||||
total_sst_files_size = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->ClearTrace();
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
std::shared_ptr<SstFileManager> sst_file_manager(NewSstFileManager(env_));
|
||||
auto sfm = static_cast<SstFileManagerImpl*>(sst_file_manager.get());
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.sst_file_manager = sst_file_manager;
|
||||
options.write_buffer_size = 1024 * 512; // 512 Kb
|
||||
DestroyAndReopen(options);
|
||||
Random rnd(301);
|
||||
|
||||
sfm->SetMaxAllowedSpaceUsage(limit_mb * 1024 * 1024);
|
||||
|
||||
int keys_written = 0;
|
||||
uint64_t estimated_db_size = 0;
|
||||
while (true) {
|
||||
auto s = Put(RandomString(&rnd, 10), RandomString(&rnd, 50));
|
||||
if (!s.ok()) {
|
||||
break;
|
||||
}
|
||||
keys_written++;
|
||||
// Check the estimated db size vs the db limit just to make sure we
|
||||
// dont run into an infinite loop
|
||||
estimated_db_size = keys_written * 60; // ~60 bytes per key
|
||||
ASSERT_LT(estimated_db_size, limit_mb * 1024 * 1024 * 2);
|
||||
}
|
||||
ASSERT_TRUE(bg_error_set);
|
||||
ASSERT_GE(total_sst_files_size, limit_mb * 1024 * 1024);
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
ASSERT_GT(reached_max_space_on_flush, 0);
|
||||
ASSERT_GT(reached_max_space_on_compaction, 0);
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBSSTTest, OpenDBWithInfiniteMaxOpenFiles) {
|
||||
// Open DB with infinite max open files
|
||||
// - First iteration use 1 thread to open files
|
||||
// - Second iteration use 5 threads to open files
|
||||
for (int iter = 0; iter < 2; iter++) {
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.write_buffer_size = 100000;
|
||||
options.disable_auto_compactions = true;
|
||||
options.max_open_files = -1;
|
||||
if (iter == 0) {
|
||||
options.max_file_opening_threads = 1;
|
||||
} else {
|
||||
options.max_file_opening_threads = 5;
|
||||
}
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Create 12 Files in L0 (then move then to L2)
|
||||
for (int i = 0; i < 12; i++) {
|
||||
std::string k = "L2_" + Key(i);
|
||||
ASSERT_OK(Put(k, k + std::string(1000, 'a')));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 2;
|
||||
db_->CompactRange(compact_options, nullptr, nullptr);
|
||||
|
||||
// Create 12 Files in L0
|
||||
for (int i = 0; i < 12; i++) {
|
||||
std::string k = "L0_" + Key(i);
|
||||
ASSERT_OK(Put(k, k + std::string(1000, 'a')));
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
Close();
|
||||
|
||||
// Reopening the DB will load all exisitng files
|
||||
Reopen(options);
|
||||
ASSERT_EQ("12,0,12", FilesPerLevel(0));
|
||||
std::vector<std::vector<FileMetaData>> files;
|
||||
dbfull()->TEST_GetFilesMetaData(db_->DefaultColumnFamily(), &files);
|
||||
|
||||
for (const auto& level : files) {
|
||||
for (const auto& file : level) {
|
||||
ASSERT_TRUE(file.table_reader_handle != nullptr);
|
||||
}
|
||||
}
|
||||
|
||||
for (int i = 0; i < 12; i++) {
|
||||
ASSERT_EQ(Get("L0_" + Key(i)), "L0_" + Key(i) + std::string(1000, 'a'));
|
||||
ASSERT_EQ(Get("L2_" + Key(i)), "L2_" + Key(i) + std::string(1000, 'a'));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
TEST_F(DBSSTTest, GetTotalSstFilesSize) {
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.compression = kNoCompression;
|
||||
DestroyAndReopen(options);
|
||||
// Generate 5 files in L0
|
||||
for (int i = 0; i < 5; i++) {
|
||||
for (int j = 0; j < 10; j++) {
|
||||
std::string val = "val_file_" + ToString(i);
|
||||
ASSERT_OK(Put(Key(j), val));
|
||||
}
|
||||
Flush();
|
||||
}
|
||||
ASSERT_EQ("5", FilesPerLevel(0));
|
||||
|
||||
std::vector<LiveFileMetaData> live_files_meta;
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 5);
|
||||
uint64_t single_file_size = live_files_meta[0].size;
|
||||
|
||||
uint64_t live_sst_files_size = 0;
|
||||
uint64_t total_sst_files_size = 0;
|
||||
for (const auto& file_meta : live_files_meta) {
|
||||
live_sst_files_size += file_meta.size;
|
||||
}
|
||||
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 5
|
||||
// Total SST files = 5
|
||||
ASSERT_EQ(live_sst_files_size, 5 * single_file_size);
|
||||
ASSERT_EQ(total_sst_files_size, 5 * single_file_size);
|
||||
|
||||
// hold current version
|
||||
std::unique_ptr<Iterator> iter1(dbfull()->NewIterator(ReadOptions()));
|
||||
|
||||
// Compact 5 files into 1 file in L0
|
||||
ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
ASSERT_EQ("0,1", FilesPerLevel(0));
|
||||
|
||||
live_files_meta.clear();
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 1);
|
||||
|
||||
live_sst_files_size = 0;
|
||||
total_sst_files_size = 0;
|
||||
for (const auto& file_meta : live_files_meta) {
|
||||
live_sst_files_size += file_meta.size;
|
||||
}
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 1 (compacted file)
|
||||
// Total SST files = 6 (5 original files + compacted file)
|
||||
ASSERT_EQ(live_sst_files_size, 1 * single_file_size);
|
||||
ASSERT_EQ(total_sst_files_size, 6 * single_file_size);
|
||||
|
||||
// hold current version
|
||||
std::unique_ptr<Iterator> iter2(dbfull()->NewIterator(ReadOptions()));
|
||||
|
||||
// Delete all keys and compact, this will delete all live files
|
||||
for (int i = 0; i < 10; i++) {
|
||||
ASSERT_OK(Delete(Key(i)));
|
||||
}
|
||||
Flush();
|
||||
ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
ASSERT_EQ("", FilesPerLevel(0));
|
||||
|
||||
live_files_meta.clear();
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 0);
|
||||
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 0
|
||||
// Total SST files = 6 (5 original files + compacted file)
|
||||
ASSERT_EQ(total_sst_files_size, 6 * single_file_size);
|
||||
|
||||
iter1.reset();
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 0
|
||||
// Total SST files = 1 (compacted file)
|
||||
ASSERT_EQ(total_sst_files_size, 1 * single_file_size);
|
||||
|
||||
iter2.reset();
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 0
|
||||
// Total SST files = 0
|
||||
ASSERT_EQ(total_sst_files_size, 0);
|
||||
}
|
||||
|
||||
TEST_F(DBSSTTest, GetTotalSstFilesSizeVersionsFilesShared) {
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.compression = kNoCompression;
|
||||
DestroyAndReopen(options);
|
||||
// Generate 5 files in L0
|
||||
for (int i = 0; i < 5; i++) {
|
||||
ASSERT_OK(Put(Key(i), "val"));
|
||||
Flush();
|
||||
}
|
||||
ASSERT_EQ("5", FilesPerLevel(0));
|
||||
|
||||
std::vector<LiveFileMetaData> live_files_meta;
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 5);
|
||||
uint64_t single_file_size = live_files_meta[0].size;
|
||||
|
||||
uint64_t live_sst_files_size = 0;
|
||||
uint64_t total_sst_files_size = 0;
|
||||
for (const auto& file_meta : live_files_meta) {
|
||||
live_sst_files_size += file_meta.size;
|
||||
}
|
||||
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
|
||||
// Live SST files = 5
|
||||
// Total SST files = 5
|
||||
ASSERT_EQ(live_sst_files_size, 5 * single_file_size);
|
||||
ASSERT_EQ(total_sst_files_size, 5 * single_file_size);
|
||||
|
||||
// hold current version
|
||||
std::unique_ptr<Iterator> iter1(dbfull()->NewIterator(ReadOptions()));
|
||||
|
||||
// Compaction will do trivial move from L0 to L1
|
||||
ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
ASSERT_EQ("0,5", FilesPerLevel(0));
|
||||
|
||||
live_files_meta.clear();
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 5);
|
||||
|
||||
live_sst_files_size = 0;
|
||||
total_sst_files_size = 0;
|
||||
for (const auto& file_meta : live_files_meta) {
|
||||
live_sst_files_size += file_meta.size;
|
||||
}
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 5
|
||||
// Total SST files = 5 (used in 2 version)
|
||||
ASSERT_EQ(live_sst_files_size, 5 * single_file_size);
|
||||
ASSERT_EQ(total_sst_files_size, 5 * single_file_size);
|
||||
|
||||
// hold current version
|
||||
std::unique_ptr<Iterator> iter2(dbfull()->NewIterator(ReadOptions()));
|
||||
|
||||
// Delete all keys and compact, this will delete all live files
|
||||
for (int i = 0; i < 5; i++) {
|
||||
ASSERT_OK(Delete(Key(i)));
|
||||
}
|
||||
Flush();
|
||||
ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
|
||||
ASSERT_EQ("", FilesPerLevel(0));
|
||||
|
||||
live_files_meta.clear();
|
||||
dbfull()->GetLiveFilesMetaData(&live_files_meta);
|
||||
ASSERT_EQ(live_files_meta.size(), 0);
|
||||
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 0
|
||||
// Total SST files = 5 (used in 2 version)
|
||||
ASSERT_EQ(total_sst_files_size, 5 * single_file_size);
|
||||
|
||||
iter1.reset();
|
||||
iter2.reset();
|
||||
|
||||
ASSERT_TRUE(dbfull()->GetIntProperty("rocksdb.total-sst-files-size",
|
||||
&total_sst_files_size));
|
||||
// Live SST files = 0
|
||||
// Total SST files = 0
|
||||
ASSERT_EQ(total_sst_files_size, 0);
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
@@ -1,260 +0,0 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
//
|
||||
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include <unordered_set>
|
||||
#include <vector>
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "util/testharness.h"
|
||||
#include "util/testutil.h"
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
// A helper function that ensures the table properties returned in
|
||||
// `GetPropertiesOfAllTablesTest` is correct.
|
||||
// This test assumes entries size is different for each of the tables.
|
||||
namespace {
|
||||
|
||||
void VerifyTableProperties(DB* db, uint64_t expected_entries_size) {
|
||||
TablePropertiesCollection props;
|
||||
ASSERT_OK(db->GetPropertiesOfAllTables(&props));
|
||||
|
||||
ASSERT_EQ(4U, props.size());
|
||||
std::unordered_set<uint64_t> unique_entries;
|
||||
|
||||
// Indirect test
|
||||
uint64_t sum = 0;
|
||||
for (const auto& item : props) {
|
||||
unique_entries.insert(item.second->num_entries);
|
||||
sum += item.second->num_entries;
|
||||
}
|
||||
|
||||
ASSERT_EQ(props.size(), unique_entries.size());
|
||||
ASSERT_EQ(expected_entries_size, sum);
|
||||
}
|
||||
} // namespace
|
||||
|
||||
class DBTablePropertiesTest : public DBTestBase {
|
||||
public:
|
||||
DBTablePropertiesTest() : DBTestBase("/db_table_properties_test") {}
|
||||
TablePropertiesCollection TestGetPropertiesOfTablesInRange(
|
||||
std::vector<Range> ranges, std::size_t* num_properties = nullptr,
|
||||
std::size_t* num_files = nullptr);
|
||||
};
|
||||
|
||||
TEST_F(DBTablePropertiesTest, GetPropertiesOfAllTablesTest) {
|
||||
Options options = CurrentOptions();
|
||||
options.level0_file_num_compaction_trigger = 8;
|
||||
Reopen(options);
|
||||
// Create 4 tables
|
||||
for (int table = 0; table < 4; ++table) {
|
||||
for (int i = 0; i < 10 + table; ++i) {
|
||||
db_->Put(WriteOptions(), ToString(table * 100 + i), "val");
|
||||
}
|
||||
db_->Flush(FlushOptions());
|
||||
}
|
||||
|
||||
// 1. Read table properties directly from file
|
||||
Reopen(options);
|
||||
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
|
||||
|
||||
// 2. Put two tables to table cache and
|
||||
Reopen(options);
|
||||
// fetch key from 1st and 2nd table, which will internally place that table to
|
||||
// the table cache.
|
||||
for (int i = 0; i < 2; ++i) {
|
||||
Get(ToString(i * 100 + 0));
|
||||
}
|
||||
|
||||
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
|
||||
|
||||
// 3. Put all tables to table cache
|
||||
Reopen(options);
|
||||
// fetch key from 1st and 2nd table, which will internally place that table to
|
||||
// the table cache.
|
||||
for (int i = 0; i < 4; ++i) {
|
||||
Get(ToString(i * 100 + 0));
|
||||
}
|
||||
VerifyTableProperties(db_, 10 + 11 + 12 + 13);
|
||||
}
|
||||
|
||||
TablePropertiesCollection
|
||||
DBTablePropertiesTest::TestGetPropertiesOfTablesInRange(
|
||||
std::vector<Range> ranges, std::size_t* num_properties,
|
||||
std::size_t* num_files) {
|
||||
|
||||
// Since we deref zero element in the vector it can not be empty
|
||||
// otherwise we pass an address to some random memory
|
||||
EXPECT_GT(ranges.size(), 0U);
|
||||
// run the query
|
||||
TablePropertiesCollection props;
|
||||
EXPECT_OK(db_->GetPropertiesOfTablesInRange(
|
||||
db_->DefaultColumnFamily(), &ranges[0], ranges.size(), &props));
|
||||
|
||||
// Make sure that we've received properties for those and for those files
|
||||
// only which fall within requested ranges
|
||||
std::vector<LiveFileMetaData> vmd;
|
||||
db_->GetLiveFilesMetaData(&vmd);
|
||||
for (auto& md : vmd) {
|
||||
std::string fn = md.db_path + md.name;
|
||||
bool in_range = false;
|
||||
for (auto& r : ranges) {
|
||||
// smallestkey < limit && largestkey >= start
|
||||
if (r.limit.compare(md.smallestkey) >= 0 &&
|
||||
r.start.compare(md.largestkey) <= 0) {
|
||||
in_range = true;
|
||||
EXPECT_GT(props.count(fn), 0);
|
||||
}
|
||||
}
|
||||
if (!in_range) {
|
||||
EXPECT_EQ(props.count(fn), 0);
|
||||
}
|
||||
}
|
||||
|
||||
if (num_properties) {
|
||||
*num_properties = props.size();
|
||||
}
|
||||
|
||||
if (num_files) {
|
||||
*num_files = vmd.size();
|
||||
}
|
||||
return props;
|
||||
}
|
||||
|
||||
TEST_F(DBTablePropertiesTest, GetPropertiesOfTablesInRange) {
|
||||
// Fixed random sead
|
||||
Random rnd(301);
|
||||
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.write_buffer_size = 4096;
|
||||
options.max_write_buffer_number = 3;
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.level0_slowdown_writes_trigger = 2;
|
||||
options.level0_stop_writes_trigger = 4;
|
||||
options.target_file_size_base = 2048;
|
||||
options.max_bytes_for_level_base = 10240;
|
||||
options.max_bytes_for_level_multiplier = 4;
|
||||
options.hard_pending_compaction_bytes_limit = 16 * 1024;
|
||||
options.num_levels = 8;
|
||||
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// build a decent LSM
|
||||
for (int i = 0; i < 10000; i++) {
|
||||
ASSERT_OK(Put(test::RandomKey(&rnd, 5), RandomString(&rnd, 102)));
|
||||
}
|
||||
Flush();
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
if (NumTableFilesAtLevel(0) == 0) {
|
||||
ASSERT_OK(Put(test::RandomKey(&rnd, 5), RandomString(&rnd, 102)));
|
||||
Flush();
|
||||
}
|
||||
|
||||
db_->PauseBackgroundWork();
|
||||
|
||||
// Ensure that we have at least L0, L1 and L2
|
||||
ASSERT_GT(NumTableFilesAtLevel(0), 0);
|
||||
ASSERT_GT(NumTableFilesAtLevel(1), 0);
|
||||
ASSERT_GT(NumTableFilesAtLevel(2), 0);
|
||||
|
||||
// Query the largest range
|
||||
std::size_t num_properties, num_files;
|
||||
TestGetPropertiesOfTablesInRange(
|
||||
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST),
|
||||
test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
|
||||
&num_properties, &num_files);
|
||||
ASSERT_EQ(num_properties, num_files);
|
||||
|
||||
// Query the empty range
|
||||
TestGetPropertiesOfTablesInRange(
|
||||
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST),
|
||||
test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST))},
|
||||
&num_properties, &num_files);
|
||||
ASSERT_GT(num_files, 0);
|
||||
ASSERT_EQ(num_properties, 0);
|
||||
|
||||
// Query the middle rangee
|
||||
TestGetPropertiesOfTablesInRange(
|
||||
{Range(test::RandomKey(&rnd, 5, test::RandomKeyType::MIDDLE),
|
||||
test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
|
||||
&num_properties, &num_files);
|
||||
ASSERT_GT(num_files, 0);
|
||||
ASSERT_GT(num_files, num_properties);
|
||||
ASSERT_GT(num_properties, 0);
|
||||
|
||||
// Query a bunch of random ranges
|
||||
for (int j = 0; j < 100; j++) {
|
||||
// create a bunch of ranges
|
||||
std::vector<std::string> random_keys;
|
||||
// Random returns numbers with zero included
|
||||
// when we pass empty ranges TestGetPropertiesOfTablesInRange()
|
||||
// derefs random memory in the empty ranges[0]
|
||||
// so want to be greater than zero and even since
|
||||
// the below loop requires that random_keys.size() to be even.
|
||||
auto n = 2 * (rnd.Uniform(50) + 1);
|
||||
|
||||
for (uint32_t i = 0; i < n; ++i) {
|
||||
random_keys.push_back(test::RandomKey(&rnd, 5));
|
||||
}
|
||||
|
||||
ASSERT_GT(random_keys.size(), 0U);
|
||||
ASSERT_EQ((random_keys.size() % 2), 0U);
|
||||
|
||||
std::vector<Range> ranges;
|
||||
auto it = random_keys.begin();
|
||||
while (it != random_keys.end()) {
|
||||
ranges.push_back(Range(*it, *(it + 1)));
|
||||
it += 2;
|
||||
}
|
||||
|
||||
TestGetPropertiesOfTablesInRange(std::move(ranges));
|
||||
}
|
||||
}
|
||||
|
||||
TEST_F(DBTablePropertiesTest, GetColumnFamilyNameProperty) {
|
||||
std::string kExtraCfName = "pikachu";
|
||||
CreateAndReopenWithCF({kExtraCfName}, CurrentOptions());
|
||||
|
||||
// Create one table per CF, then verify it was created with the column family
|
||||
// name property.
|
||||
for (int cf = 0; cf < 2; ++cf) {
|
||||
Put(cf, "key", "val");
|
||||
Flush(cf);
|
||||
|
||||
TablePropertiesCollection fname_to_props;
|
||||
ASSERT_OK(db_->GetPropertiesOfAllTables(handles_[cf], &fname_to_props));
|
||||
ASSERT_EQ(1U, fname_to_props.size());
|
||||
|
||||
std::string expected_cf_name;
|
||||
if (cf > 0) {
|
||||
expected_cf_name = kExtraCfName;
|
||||
} else {
|
||||
expected_cf_name = kDefaultColumnFamilyName;
|
||||
}
|
||||
ASSERT_EQ(expected_cf_name,
|
||||
fname_to_props.begin()->second->column_family_name);
|
||||
ASSERT_EQ(cf, static_cast<uint32_t>(
|
||||
fname_to_props.begin()->second->column_family_id));
|
||||
}
|
||||
}
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
}
|
||||
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,11 +10,11 @@
|
||||
// Introduction of SyncPoint effectively disabled building and running this test
|
||||
// in Release build.
|
||||
// which is a pity, it is a good test
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "db/forward_iterator.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/db_test_util.h"
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -140,8 +140,6 @@ TEST_F(DBTestTailingIterator, TailingIteratorTrimSeekToNext) {
|
||||
std::unique_ptr<Iterator> iterh(db_->NewIterator(read_options, handles_[1]));
|
||||
std::string value(1024, 'a');
|
||||
bool file_iters_deleted = false;
|
||||
bool file_iters_renewed_null = false;
|
||||
bool file_iters_renewed_copy = false;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"ForwardIterator::SeekInternal:Return", [&](void* arg) {
|
||||
ForwardIterator* fiter = reinterpret_cast<ForwardIterator*>(arg);
|
||||
@@ -154,12 +152,6 @@ TEST_F(DBTestTailingIterator, TailingIteratorTrimSeekToNext) {
|
||||
ASSERT_TRUE(!file_iters_deleted ||
|
||||
fiter->TEST_CheckDeletedIters(&deleted_iters, &num_iters));
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"ForwardIterator::RenewIterators:Null",
|
||||
[&](void* arg) { file_iters_renewed_null = true; });
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"ForwardIterator::RenewIterators:Copy",
|
||||
[&](void* arg) { file_iters_renewed_copy = true; });
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
const int num_records = 1000;
|
||||
for (int i = 1; i < num_records; ++i) {
|
||||
@@ -168,7 +160,7 @@ TEST_F(DBTestTailingIterator, TailingIteratorTrimSeekToNext) {
|
||||
char buf3[32];
|
||||
char buf4[32];
|
||||
snprintf(buf1, sizeof(buf1), "00a0%016d", i * 5);
|
||||
snprintf(buf3, sizeof(buf3), "00b0%016d", i * 5);
|
||||
snprintf(buf3, sizeof(buf1), "00b0%016d", i * 5);
|
||||
|
||||
Slice key(buf1, 20);
|
||||
ASSERT_OK(Put(1, key, value));
|
||||
@@ -181,7 +173,7 @@ TEST_F(DBTestTailingIterator, TailingIteratorTrimSeekToNext) {
|
||||
if (i == 299) {
|
||||
file_iters_deleted = true;
|
||||
}
|
||||
snprintf(buf4, sizeof(buf4), "00a0%016d", i * 5 / 2);
|
||||
snprintf(buf4, sizeof(buf1), "00a0%016d", i * 5 / 2);
|
||||
Slice target(buf4, 20);
|
||||
iterh->Seek(target);
|
||||
ASSERT_TRUE(iter->Valid());
|
||||
@@ -211,8 +203,6 @@ TEST_F(DBTestTailingIterator, TailingIteratorTrimSeekToNext) {
|
||||
ASSERT_LE(num_iters, 1);
|
||||
file_iters_deleted = false;
|
||||
}
|
||||
ASSERT_TRUE(file_iters_renewed_null);
|
||||
ASSERT_TRUE(file_iters_renewed_copy);
|
||||
iter = 0;
|
||||
itern = 0;
|
||||
iterh = 0;
|
||||
@@ -654,57 +644,12 @@ TEST_F(DBTestTailingIterator, ManagedTailingIteratorSeekToSame) {
|
||||
ASSERT_EQ(found, iter->key().ToString());
|
||||
}
|
||||
|
||||
TEST_F(DBTestTailingIterator, ForwardIteratorVersionProperty) {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 1000;
|
||||
|
||||
ReadOptions read_options;
|
||||
read_options.tailing = true;
|
||||
|
||||
Put("foo", "bar");
|
||||
|
||||
uint64_t v1, v2, v3, v4;
|
||||
{
|
||||
std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
|
||||
iter->Seek("foo");
|
||||
std::string prop_value;
|
||||
ASSERT_OK(iter->GetProperty("rocksdb.iterator.super-version-number",
|
||||
&prop_value));
|
||||
v1 = static_cast<uint64_t>(std::atoi(prop_value.c_str()));
|
||||
|
||||
Put("foo1", "bar1");
|
||||
Flush();
|
||||
|
||||
ASSERT_OK(iter->GetProperty("rocksdb.iterator.super-version-number",
|
||||
&prop_value));
|
||||
v2 = static_cast<uint64_t>(std::atoi(prop_value.c_str()));
|
||||
|
||||
iter->Seek("f");
|
||||
|
||||
ASSERT_OK(iter->GetProperty("rocksdb.iterator.super-version-number",
|
||||
&prop_value));
|
||||
v3 = static_cast<uint64_t>(std::atoi(prop_value.c_str()));
|
||||
|
||||
ASSERT_EQ(v1, v2);
|
||||
ASSERT_GT(v3, v2);
|
||||
}
|
||||
|
||||
{
|
||||
std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
|
||||
iter->Seek("foo");
|
||||
std::string prop_value;
|
||||
ASSERT_OK(iter->GetProperty("rocksdb.iterator.super-version-number",
|
||||
&prop_value));
|
||||
v4 = static_cast<uint64_t>(std::atoi(prop_value.c_str()));
|
||||
}
|
||||
ASSERT_EQ(v3, v4);
|
||||
}
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // !defined(ROCKSDB_LITE)
|
||||
#endif // !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
|
||||
+5231
-1076
File diff suppressed because it is too large.
Load diff
-2075
File diff suppressed because it is too large.
Load diff
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -7,9 +7,9 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#include "util/db_test_util.h"
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
#include "util/sync_point.h"
|
||||
|
||||
namespace rocksdb {
|
||||
@@ -22,16 +22,12 @@ static std::string CompressibleString(Random* rnd, int len) {
|
||||
|
||||
class DBTestUniversalCompactionBase
|
||||
: public DBTestBase,
|
||||
public ::testing::WithParamInterface<std::tuple<int, bool>> {
|
||||
public ::testing::WithParamInterface<int> {
|
||||
public:
|
||||
explicit DBTestUniversalCompactionBase(
|
||||
const std::string& path) : DBTestBase(path) {}
|
||||
virtual void SetUp() override {
|
||||
num_levels_ = std::get<0>(GetParam());
|
||||
exclusive_manual_compaction_ = std::get<1>(GetParam());
|
||||
}
|
||||
virtual void SetUp() override { num_levels_ = GetParam(); }
|
||||
int num_levels_;
|
||||
bool exclusive_manual_compaction_;
|
||||
};
|
||||
|
||||
class DBTestUniversalCompaction : public DBTestUniversalCompactionBase {
|
||||
@@ -116,111 +112,6 @@ class DelayFilterFactory : public CompactionFilterFactory {
|
||||
};
|
||||
} // namespace
|
||||
|
||||
// Make sure we don't trigger a problem if the trigger conditon is given
|
||||
// to be 0, which is invalid.
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionSingleSortedRun) {
|
||||
Options options = CurrentOptions();
|
||||
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = num_levels_;
|
||||
// Config universal compaction to always compact to one single sorted run.
|
||||
options.level0_file_num_compaction_trigger = 0;
|
||||
options.compaction_options_universal.size_ratio = 10;
|
||||
options.compaction_options_universal.min_merge_width = 2;
|
||||
options.compaction_options_universal.max_size_amplification_percent = 0;
|
||||
|
||||
options.write_buffer_size = 105 << 10; // 105KB
|
||||
options.arena_block_size = 4 << 10;
|
||||
options.target_file_size_base = 32 << 10; // 32KB
|
||||
// trigger compaction if there are >= 4 files
|
||||
KeepFilterFactory* filter = new KeepFilterFactory(true);
|
||||
filter->expect_manual_compaction_.store(false);
|
||||
options.compaction_filter_factory.reset(filter);
|
||||
|
||||
DestroyAndReopen(options);
|
||||
ASSERT_EQ(1, db_->GetOptions().level0_file_num_compaction_trigger);
|
||||
|
||||
Random rnd(301);
|
||||
int key_idx = 0;
|
||||
|
||||
filter->expect_full_compaction_.store(true);
|
||||
|
||||
for (int num = 0; num < 16; num++) {
|
||||
// Write 100KB file. And immediately it should be compacted to one file.
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ(NumSortedRuns(0), 1);
|
||||
}
|
||||
ASSERT_OK(Put(Key(key_idx), ""));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
ASSERT_EQ(NumSortedRuns(0), 1);
|
||||
}
|
||||
|
||||
TEST_P(DBTestUniversalCompaction, OptimizeFiltersForHits) {
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.size_ratio = 5;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 105 << 10; // 105KB
|
||||
options.arena_block_size = 4 << 10;
|
||||
options.target_file_size_base = 32 << 10; // 32KB
|
||||
// trigger compaction if there are >= 4 files
|
||||
options.level0_file_num_compaction_trigger = 4;
|
||||
BlockBasedTableOptions bbto;
|
||||
bbto.cache_index_and_filter_blocks = true;
|
||||
bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
|
||||
bbto.whole_key_filtering = true;
|
||||
options.table_factory.reset(NewBlockBasedTableFactory(bbto));
|
||||
options.optimize_filters_for_hits = true;
|
||||
options.statistics = rocksdb::CreateDBStatistics();
|
||||
options.memtable_factory.reset(new SpecialSkipListFactory(3));
|
||||
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// block compaction from happening
|
||||
env_->SetBackgroundThreads(1, Env::LOW);
|
||||
test::SleepingBackgroundTask sleeping_task_low;
|
||||
env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task_low,
|
||||
Env::Priority::LOW);
|
||||
|
||||
for (int num = 0; num < options.level0_file_num_compaction_trigger; num++) {
|
||||
Put(Key(num * 10), "val");
|
||||
if (num) {
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
}
|
||||
Put(Key(30 + num * 10), "val");
|
||||
Put(Key(60 + num * 10), "val");
|
||||
}
|
||||
Put("", "");
|
||||
dbfull()->TEST_WaitForFlushMemTable();
|
||||
|
||||
// Query set of non existing keys
|
||||
for (int i = 5; i < 90; i += 10) {
|
||||
ASSERT_EQ(Get(Key(i)), "NOT_FOUND");
|
||||
}
|
||||
|
||||
// Make sure bloom filter is used at least once.
|
||||
ASSERT_GT(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
|
||||
auto prev_counter = TestGetTickerCount(options, BLOOM_FILTER_USEFUL);
|
||||
|
||||
// Make sure bloom filter is used for all but the last L0 file when looking
|
||||
// up a non-existent key that's in the range of all L0 files.
|
||||
ASSERT_EQ(Get(Key(35)), "NOT_FOUND");
|
||||
ASSERT_EQ(prev_counter + NumTableFilesAtLevel(0) - 1,
|
||||
TestGetTickerCount(options, BLOOM_FILTER_USEFUL));
|
||||
prev_counter = TestGetTickerCount(options, BLOOM_FILTER_USEFUL);
|
||||
|
||||
// Unblock compaction and wait it for happening.
|
||||
sleeping_task_low.WakeUp();
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
// The same queries will not trigger bloom filter
|
||||
for (int i = 5; i < 90; i += 10) {
|
||||
ASSERT_EQ(Get(Key(i)), "NOT_FOUND");
|
||||
}
|
||||
ASSERT_EQ(prev_counter, TestGetTickerCount(options, BLOOM_FILTER_USEFUL));
|
||||
}
|
||||
|
||||
// TODO(kailiu) The tests on UniversalCompaction has some issues:
|
||||
// 1. A lot of magic numbers ("11" or "12").
|
||||
// 2. Made assumption on the memtable flush conditions, which may change from
|
||||
@@ -228,7 +119,6 @@ TEST_P(DBTestUniversalCompaction, OptimizeFiltersForHits) {
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionTrigger) {
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.size_ratio = 5;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 105 << 10; // 105KB
|
||||
options.arena_block_size = 4 << 10;
|
||||
@@ -333,12 +223,13 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrigger) {
|
||||
}
|
||||
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionSizeAmplification) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.target_file_size_base = 32 << 10; // 32KB
|
||||
options.level0_file_num_compaction_trigger = 3;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -383,12 +274,12 @@ TEST_P(DBTestUniversalCompaction, CompactFilesOnUniversalCompaction) {
|
||||
ChangeCompactOptions();
|
||||
Options options;
|
||||
options.create_if_missing = true;
|
||||
options.write_buffer_size = kEntrySize * kEntriesPerBuffer;
|
||||
options.compaction_style = kCompactionStyleLevel;
|
||||
options.num_levels = 1;
|
||||
options.target_file_size_base = options.write_buffer_size;
|
||||
options.compression = kNoCompression;
|
||||
options = CurrentOptions(options);
|
||||
options.write_buffer_size = kEntrySize * kEntriesPerBuffer;
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
ASSERT_EQ(options.compaction_style, kCompactionStyleUniversal);
|
||||
Random rnd(301);
|
||||
@@ -446,11 +337,12 @@ TEST_P(DBTestUniversalCompaction, CompactFilesOnUniversalCompaction) {
|
||||
}
|
||||
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionTargetLevel) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.num_levels = 7;
|
||||
options.disable_auto_compactions = true;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// Generate 3 overlapping files
|
||||
@@ -475,7 +367,6 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTargetLevel) {
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 4;
|
||||
compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
|
||||
db_->CompactRange(compact_options, nullptr, nullptr);
|
||||
ASSERT_EQ("0,0,0,0,1", FilesPerLevel(0));
|
||||
}
|
||||
@@ -490,13 +381,14 @@ class DBTestUniversalCompactionMultiLevels
|
||||
};
|
||||
|
||||
TEST_P(DBTestUniversalCompactionMultiLevels, UniversalCompactionMultiLevels) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.level0_file_num_compaction_trigger = 8;
|
||||
options.max_background_compactions = 3;
|
||||
options.target_file_size_base = 32 * 1024;
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
// Trigger compaction if size amplification exceeds 110%
|
||||
@@ -532,7 +424,7 @@ TEST_P(DBTestUniversalCompactionMultiLevels, UniversalCompactionTrivialMove) {
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.allow_trivial_move = true;
|
||||
options.num_levels = 3;
|
||||
@@ -540,6 +432,7 @@ TEST_P(DBTestUniversalCompactionMultiLevels, UniversalCompactionTrivialMove) {
|
||||
options.level0_file_num_compaction_trigger = 3;
|
||||
options.max_background_compactions = 2;
|
||||
options.target_file_size_base = 32 * 1024;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -566,8 +459,7 @@ TEST_P(DBTestUniversalCompactionMultiLevels, UniversalCompactionTrivialMove) {
|
||||
|
||||
INSTANTIATE_TEST_CASE_P(DBTestUniversalCompactionMultiLevels,
|
||||
DBTestUniversalCompactionMultiLevels,
|
||||
::testing::Combine(::testing::Values(3, 20),
|
||||
::testing::Bool()));
|
||||
::testing::Values(3, 20));
|
||||
|
||||
class DBTestUniversalCompactionParallel :
|
||||
public DBTestUniversalCompactionBase {
|
||||
@@ -578,7 +470,7 @@ class DBTestUniversalCompactionParallel :
|
||||
};
|
||||
|
||||
TEST_P(DBTestUniversalCompactionParallel, UniversalCompactionParallel) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 1 << 10; // 1KB
|
||||
@@ -587,6 +479,7 @@ TEST_P(DBTestUniversalCompactionParallel, UniversalCompactionParallel) {
|
||||
options.max_background_flushes = 3;
|
||||
options.target_file_size_base = 1 * 1024;
|
||||
options.compaction_options_universal.max_size_amplification_percent = 110;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -637,135 +530,12 @@ TEST_P(DBTestUniversalCompactionParallel, UniversalCompactionParallel) {
|
||||
}
|
||||
}
|
||||
|
||||
TEST_P(DBTestUniversalCompactionParallel, PickByFileNumberBug) {
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = num_levels_;
|
||||
options.write_buffer_size = 1 * 1024; // 1KB
|
||||
options.level0_file_num_compaction_trigger = 7;
|
||||
options.max_background_compactions = 2;
|
||||
options.target_file_size_base = 1024 * 1024; // 1MB
|
||||
|
||||
// Disable size amplifiction compaction
|
||||
options.compaction_options_universal.max_size_amplification_percent =
|
||||
UINT_MAX;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency(
|
||||
{{"DBTestUniversalCompactionParallel::PickByFileNumberBug:0",
|
||||
"BackgroundCallCompaction:0"},
|
||||
{"UniversalCompactionPicker::PickCompaction:Return",
|
||||
"DBTestUniversalCompactionParallel::PickByFileNumberBug:1"},
|
||||
{"DBTestUniversalCompactionParallel::PickByFileNumberBug:2",
|
||||
"CompactionJob::Run():Start"}});
|
||||
|
||||
int total_picked_compactions = 0;
|
||||
rocksdb::SyncPoint::GetInstance()->SetCallBack(
|
||||
"UniversalCompactionPicker::PickCompaction:Return", [&](void* arg) {
|
||||
if (arg) {
|
||||
total_picked_compactions++;
|
||||
}
|
||||
});
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
// Write 7 files to trigger compaction
|
||||
int key_idx = 1;
|
||||
for (int i = 1; i <= 70; i++) {
|
||||
std::string k = Key(key_idx++);
|
||||
ASSERT_OK(Put(k, k));
|
||||
if (i % 10 == 0) {
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
}
|
||||
|
||||
// Wait for the 1st background compaction process to start
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:0");
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:1");
|
||||
rocksdb::SyncPoint::GetInstance()->ClearTrace();
|
||||
|
||||
// Write 3 files while 1st compaction is held
|
||||
// These 3 files have different sizes to avoid compacting based on size_ratio
|
||||
int num_keys = 1000;
|
||||
for (int i = 0; i < 3; i++) {
|
||||
for (int j = 1; j <= num_keys; j++) {
|
||||
std::string k = Key(key_idx++);
|
||||
ASSERT_OK(Put(k, k));
|
||||
}
|
||||
ASSERT_OK(Flush());
|
||||
num_keys -= 100;
|
||||
}
|
||||
|
||||
// Wait for the 2nd background compaction process to start
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:0");
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:1");
|
||||
|
||||
// Hold the 1st and 2nd compaction from finishing
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:2");
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
// Although 2 compaction threads started, the second one did not compact
|
||||
// anything because the number of files not being compacted is less than
|
||||
// level0_file_num_compaction_trigger
|
||||
EXPECT_EQ(total_picked_compactions, 1);
|
||||
EXPECT_EQ(TotalTableFiles(), 4);
|
||||
|
||||
// Stop SyncPoint and destroy the DB and reopen it again
|
||||
rocksdb::SyncPoint::GetInstance()->ClearTrace();
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
key_idx = 1;
|
||||
total_picked_compactions = 0;
|
||||
DestroyAndReopen(options);
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
// Write 7 files to trigger compaction
|
||||
for (int i = 1; i <= 70; i++) {
|
||||
std::string k = Key(key_idx++);
|
||||
ASSERT_OK(Put(k, k));
|
||||
if (i % 10 == 0) {
|
||||
ASSERT_OK(Flush());
|
||||
}
|
||||
}
|
||||
|
||||
// Wait for the 1st background compaction process to start
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:0");
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:1");
|
||||
rocksdb::SyncPoint::GetInstance()->ClearTrace();
|
||||
|
||||
// Write 8 files while 1st compaction is held
|
||||
// These 8 files have different sizes to avoid compacting based on size_ratio
|
||||
num_keys = 1000;
|
||||
for (int i = 0; i < 8; i++) {
|
||||
for (int j = 1; j <= num_keys; j++) {
|
||||
std::string k = Key(key_idx++);
|
||||
ASSERT_OK(Put(k, k));
|
||||
}
|
||||
ASSERT_OK(Flush());
|
||||
num_keys -= 100;
|
||||
}
|
||||
|
||||
// Wait for the 2nd background compaction process to start
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:0");
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:1");
|
||||
|
||||
// Hold the 1st and 2nd compaction from finishing
|
||||
TEST_SYNC_POINT("DBTestUniversalCompactionParallel::PickByFileNumberBug:2");
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
// This time we will trigger a compaction because of size ratio and
|
||||
// another compaction because of number of files that are not compacted
|
||||
// greater than 7
|
||||
EXPECT_GE(total_picked_compactions, 2);
|
||||
}
|
||||
|
||||
INSTANTIATE_TEST_CASE_P(DBTestUniversalCompactionParallel,
|
||||
DBTestUniversalCompactionParallel,
|
||||
::testing::Combine(::testing::Values(1, 10),
|
||||
::testing::Values(false)));
|
||||
::testing::Values(1, 10));
|
||||
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionOptions) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.write_buffer_size = 105 << 10; // 105KB
|
||||
options.arena_block_size = 4 << 10; // 4KB
|
||||
@@ -773,6 +543,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionOptions) {
|
||||
options.level0_file_num_compaction_trigger = 4;
|
||||
options.num_levels = num_levels_;
|
||||
options.compaction_options_universal.compression_size_percent = -1;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -885,13 +656,14 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionCompressRatio1) {
|
||||
return;
|
||||
}
|
||||
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.target_file_size_base = 32 << 10; // 32KB
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.num_levels = num_levels_;
|
||||
options.compaction_options_universal.compression_size_percent = 70;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
@@ -952,13 +724,14 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionCompressRatio2) {
|
||||
if (!Snappy_Supported()) {
|
||||
return;
|
||||
}
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.target_file_size_base = 32 << 10; // 32KB
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.num_levels = num_levels_;
|
||||
options.compaction_options_universal.compression_size_percent = 95;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
|
||||
Random rnd(301);
|
||||
@@ -994,7 +767,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrivialMoveTest1) {
|
||||
});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.allow_trivial_move = true;
|
||||
options.num_levels = 2;
|
||||
@@ -1002,6 +775,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrivialMoveTest1) {
|
||||
options.level0_file_num_compaction_trigger = 3;
|
||||
options.max_background_compactions = 1;
|
||||
options.target_file_size_base = 32 * 1024;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -1038,7 +812,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrivialMoveTest2) {
|
||||
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.allow_trivial_move = true;
|
||||
options.num_levels = 15;
|
||||
@@ -1046,6 +820,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrivialMoveTest2) {
|
||||
options.level0_file_num_compaction_trigger = 8;
|
||||
options.max_background_compactions = 4;
|
||||
options.target_file_size_base = 64 * 1024;
|
||||
options = CurrentOptions(options);
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
@@ -1071,19 +846,17 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionTrivialMoveTest2) {
|
||||
}
|
||||
|
||||
TEST_P(DBTestUniversalCompaction, UniversalCompactionFourPaths) {
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.db_paths.emplace_back(dbname_, 300 * 1024);
|
||||
options.db_paths.emplace_back(dbname_ + "_2", 300 * 1024);
|
||||
options.db_paths.emplace_back(dbname_ + "_3", 500 * 1024);
|
||||
options.db_paths.emplace_back(dbname_ + "_4", 1024 * 1024 * 1024);
|
||||
options.memtable_factory.reset(
|
||||
new SpecialSkipListFactory(KNumKeysByGenerateNewFile - 1));
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.size_ratio = 5;
|
||||
options.write_buffer_size = 111 << 10; // 114KB
|
||||
options.write_buffer_size = 110 << 10; // 105KB
|
||||
options.arena_block_size = 4 << 10;
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.num_levels = 1;
|
||||
options = CurrentOptions(options);
|
||||
|
||||
std::vector<std::string> filenames;
|
||||
env_->GetChildren(options.db_paths[1].path, &filenames);
|
||||
@@ -1118,13 +891,13 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionFourPaths) {
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 2, 4) -> (3, 4)
|
||||
// (1, 2, 4)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[2].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 3, 4) -> (8)
|
||||
// (1, 1, 2, 4) -> (8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[3].path));
|
||||
|
||||
@@ -1138,22 +911,22 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionFourPaths) {
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[3].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
|
||||
// (1, 2, 8) -> (3, 8)
|
||||
// (1, 2, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[3].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 3, 8) -> (4, 8)
|
||||
// (1, 1, 2, 8) -> (4, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[2].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[3].path));
|
||||
|
||||
// (1, 4, 8) -> (5, 8)
|
||||
// (1, 4, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[3].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[2].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
for (int i = 0; i < key_idx; i++) {
|
||||
auto v = Get(Key(i));
|
||||
@@ -1195,27 +968,29 @@ TEST_P(DBTestUniversalCompaction, IncreaseUniversalCompactionNumLevels) {
|
||||
int max_key1 = 200;
|
||||
int max_key2 = 600;
|
||||
int max_key3 = 800;
|
||||
const int KNumKeysPerFile = 10;
|
||||
|
||||
// Stage 1: open a DB with universal compaction, num_levels=1
|
||||
Options options = CurrentOptions();
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = 1;
|
||||
options.write_buffer_size = 200 << 10; // 200KB
|
||||
options.write_buffer_size = 100 << 10; // 100KB
|
||||
options.level0_file_num_compaction_trigger = 3;
|
||||
options.memtable_factory.reset(new SpecialSkipListFactory(KNumKeysPerFile));
|
||||
options = CurrentOptions(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
for (int i = 0; i <= max_key1; i++) {
|
||||
// each value is 10K
|
||||
ASSERT_OK(Put(1, Key(i), RandomString(&rnd, 10000)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
}
|
||||
ASSERT_OK(Flush(1));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
|
||||
int non_level0_num_files = 0;
|
||||
for (int i = 1; i < options.num_levels; i++) {
|
||||
non_level0_num_files += NumTableFilesAtLevel(i, 1);
|
||||
}
|
||||
ASSERT_EQ(non_level0_num_files, 0);
|
||||
|
||||
// Stage 2: reopen with universal compaction, num_levels=4
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.num_levels = 4;
|
||||
@@ -1228,8 +1003,6 @@ TEST_P(DBTestUniversalCompaction, IncreaseUniversalCompactionNumLevels) {
|
||||
for (int i = max_key1 + 1; i <= max_key2; i++) {
|
||||
// each value is 10K
|
||||
ASSERT_OK(Put(1, Key(i), RandomString(&rnd, 10000)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
}
|
||||
ASSERT_OK(Flush(1));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
@@ -1246,7 +1019,6 @@ TEST_P(DBTestUniversalCompaction, IncreaseUniversalCompactionNumLevels) {
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 0;
|
||||
compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
|
||||
dbfull()->CompactRange(compact_options, handles_[1], nullptr, nullptr);
|
||||
// Need to restart it once to remove higher level records in manifest.
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, options);
|
||||
@@ -1260,8 +1032,6 @@ TEST_P(DBTestUniversalCompaction, IncreaseUniversalCompactionNumLevels) {
|
||||
for (int i = max_key2 + 1; i <= max_key3; i++) {
|
||||
// each value is 10K
|
||||
ASSERT_OK(Put(1, Key(i), RandomString(&rnd, 10000)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
}
|
||||
ASSERT_OK(Flush(1));
|
||||
dbfull()->TEST_WaitForCompact();
|
||||
@@ -1273,17 +1043,16 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionSecondPathRatio) {
|
||||
if (!Snappy_Supported()) {
|
||||
return;
|
||||
}
|
||||
Options options = CurrentOptions();
|
||||
Options options;
|
||||
options.db_paths.emplace_back(dbname_, 500 * 1024);
|
||||
options.db_paths.emplace_back(dbname_ + "_2", 1024 * 1024 * 1024);
|
||||
options.compaction_style = kCompactionStyleUniversal;
|
||||
options.compaction_options_universal.size_ratio = 5;
|
||||
options.write_buffer_size = 111 << 10; // 114KB
|
||||
options.write_buffer_size = 110 << 10; // 105KB
|
||||
options.arena_block_size = 4 * 1024;
|
||||
options.arena_block_size = 4 << 10;
|
||||
options.level0_file_num_compaction_trigger = 2;
|
||||
options.num_levels = 1;
|
||||
options.memtable_factory.reset(
|
||||
new SpecialSkipListFactory(KNumKeysByGenerateNewFile - 1));
|
||||
options = CurrentOptions(options);
|
||||
|
||||
std::vector<std::string> filenames;
|
||||
env_->GetChildren(options.db_paths[1].path, &filenames);
|
||||
@@ -1317,12 +1086,12 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionSecondPathRatio) {
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 2, 4) -> (3, 4)
|
||||
// (1, 2, 4)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(2, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(2, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 3, 4) -> (8)
|
||||
// (1, 1, 2, 4) -> (8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
@@ -1337,20 +1106,20 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionSecondPathRatio) {
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 2, 8) -> (3, 8)
|
||||
// (1, 2, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(2, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 1, 2, 8) -> (4, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(2, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 3, 8) -> (4, 8)
|
||||
// (1, 4, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(2, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
|
||||
// (1, 4, 8) -> (5, 8)
|
||||
GenerateNewFile(&rnd, &key_idx);
|
||||
ASSERT_EQ(2, GetSstFileCount(options.db_paths[1].path));
|
||||
ASSERT_EQ(0, GetSstFileCount(dbname_));
|
||||
ASSERT_EQ(1, GetSstFileCount(dbname_));
|
||||
|
||||
for (int i = 0; i < key_idx; i++) {
|
||||
auto v = Get(Key(i));
|
||||
@@ -1370,8 +1139,7 @@ TEST_P(DBTestUniversalCompaction, UniversalCompactionSecondPathRatio) {
|
||||
}
|
||||
|
||||
INSTANTIATE_TEST_CASE_P(UniversalCompactionNumLevels, DBTestUniversalCompaction,
|
||||
::testing::Combine(::testing::Values(1, 3, 5),
|
||||
::testing::Bool()));
|
||||
::testing::Values(1, 3, 5));
|
||||
|
||||
class DBTestUniversalManualCompactionOutputPathId
|
||||
: public DBTestUniversalCompactionBase {
|
||||
@@ -1403,7 +1171,6 @@ TEST_P(DBTestUniversalManualCompactionOutputPathId,
|
||||
// Full compaction to DB path 0
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.target_path_id = 1;
|
||||
compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
|
||||
db_->CompactRange(compact_options, handles_[1], nullptr, nullptr);
|
||||
ASSERT_EQ(1, TotalLiveFiles(1));
|
||||
ASSERT_EQ(0, GetSstFileCount(options.db_paths[0].path));
|
||||
@@ -1426,7 +1193,6 @@ TEST_P(DBTestUniversalManualCompactionOutputPathId,
|
||||
|
||||
// Full compaction to DB path 0
|
||||
compact_options.target_path_id = 0;
|
||||
compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
|
||||
db_->CompactRange(compact_options, handles_[1], nullptr, nullptr);
|
||||
ASSERT_EQ(1, TotalLiveFiles(1));
|
||||
ASSERT_EQ(1, GetSstFileCount(options.db_paths[0].path));
|
||||
@@ -1434,22 +1200,20 @@ TEST_P(DBTestUniversalManualCompactionOutputPathId,
|
||||
|
||||
// Fail when compacting to an invalid path ID
|
||||
compact_options.target_path_id = 2;
|
||||
compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
|
||||
ASSERT_TRUE(db_->CompactRange(compact_options, handles_[1], nullptr, nullptr)
|
||||
.IsInvalidArgument());
|
||||
}
|
||||
|
||||
INSTANTIATE_TEST_CASE_P(DBTestUniversalManualCompactionOutputPathId,
|
||||
DBTestUniversalManualCompactionOutputPathId,
|
||||
::testing::Combine(::testing::Values(1, 8),
|
||||
::testing::Bool()));
|
||||
::testing::Values(1, 8));
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
#endif // !defined(ROCKSDB_LITE)
|
||||
#endif // !(defined NDEBUG) || !defined(OS_WIN)
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !defined(ROCKSDB_LITE)
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
|
||||
+9
-915
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -7,11 +7,11 @@
|
||||
// Use of this source code is governed by a BSD-style license that can be
|
||||
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
||||
|
||||
#include "db/db_test_util.h"
|
||||
#include "port/stack_trace.h"
|
||||
#include "util/fault_injection_test_env.h"
|
||||
#include "util/options_helper.h"
|
||||
#include "util/db_test_util.h"
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
#include "util/sync_point.h"
|
||||
#endif
|
||||
|
||||
namespace rocksdb {
|
||||
class DBWALTest : public DBTestBase {
|
||||
@@ -70,6 +70,7 @@ TEST_F(DBWALTest, RollLog) {
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
TEST_F(DBWALTest, SyncWALNotBlockWrite) {
|
||||
Options options = CurrentOptions();
|
||||
options.max_write_buffer_number = 4;
|
||||
@@ -129,922 +130,15 @@ TEST_F(DBWALTest, SyncWALNotWaitWrite) {
|
||||
ASSERT_EQ(Get("foo2"), "bar2");
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, Recover) {
|
||||
do {
|
||||
CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
|
||||
ASSERT_OK(Put(1, "foo", "v1"));
|
||||
ASSERT_OK(Put(1, "baz", "v5"));
|
||||
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
ASSERT_EQ("v1", Get(1, "foo"));
|
||||
|
||||
ASSERT_EQ("v1", Get(1, "foo"));
|
||||
ASSERT_EQ("v5", Get(1, "baz"));
|
||||
ASSERT_OK(Put(1, "bar", "v2"));
|
||||
ASSERT_OK(Put(1, "foo", "v3"));
|
||||
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
ASSERT_EQ("v3", Get(1, "foo"));
|
||||
ASSERT_OK(Put(1, "foo", "v4"));
|
||||
ASSERT_EQ("v4", Get(1, "foo"));
|
||||
ASSERT_EQ("v2", Get(1, "bar"));
|
||||
ASSERT_EQ("v5", Get(1, "baz"));
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoverWithTableHandle) {
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.disable_auto_compactions = true;
|
||||
options.avoid_flush_during_recovery = false;
|
||||
DestroyAndReopen(options);
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
|
||||
ASSERT_OK(Put(1, "foo", "v1"));
|
||||
ASSERT_OK(Put(1, "bar", "v2"));
|
||||
ASSERT_OK(Flush(1));
|
||||
ASSERT_OK(Put(1, "foo", "v3"));
|
||||
ASSERT_OK(Put(1, "bar", "v4"));
|
||||
ASSERT_OK(Flush(1));
|
||||
ASSERT_OK(Put(1, "big", std::string(100, 'a')));
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
|
||||
std::vector<std::vector<FileMetaData>> files;
|
||||
dbfull()->TEST_GetFilesMetaData(handles_[1], &files);
|
||||
size_t total_files = 0;
|
||||
for (const auto& level : files) {
|
||||
total_files += level.size();
|
||||
}
|
||||
ASSERT_EQ(total_files, 3);
|
||||
for (const auto& level : files) {
|
||||
for (const auto& file : level) {
|
||||
if (kInfiniteMaxOpenFiles == option_config_) {
|
||||
ASSERT_TRUE(file.table_reader_handle != nullptr);
|
||||
} else {
|
||||
ASSERT_TRUE(file.table_reader_handle == nullptr);
|
||||
}
|
||||
}
|
||||
}
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, IgnoreRecoveredLog) {
|
||||
std::string backup_logs = dbname_ + "/backup_logs";
|
||||
|
||||
// delete old files in backup_logs directory
|
||||
env_->CreateDirIfMissing(backup_logs);
|
||||
std::vector<std::string> old_files;
|
||||
env_->GetChildren(backup_logs, &old_files);
|
||||
for (auto& file : old_files) {
|
||||
if (file != "." && file != "..") {
|
||||
env_->DeleteFile(backup_logs + "/" + file);
|
||||
}
|
||||
}
|
||||
|
||||
do {
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.merge_operator = MergeOperators::CreateUInt64AddOperator();
|
||||
options.wal_dir = dbname_ + "/logs";
|
||||
DestroyAndReopen(options);
|
||||
|
||||
// fill up the DB
|
||||
std::string one, two;
|
||||
PutFixed64(&one, 1);
|
||||
PutFixed64(&two, 2);
|
||||
ASSERT_OK(db_->Merge(WriteOptions(), Slice("foo"), Slice(one)));
|
||||
ASSERT_OK(db_->Merge(WriteOptions(), Slice("foo"), Slice(one)));
|
||||
ASSERT_OK(db_->Merge(WriteOptions(), Slice("bar"), Slice(one)));
|
||||
|
||||
// copy the logs to backup
|
||||
std::vector<std::string> logs;
|
||||
env_->GetChildren(options.wal_dir, &logs);
|
||||
for (auto& log : logs) {
|
||||
if (log != ".." && log != ".") {
|
||||
CopyFile(options.wal_dir + "/" + log, backup_logs + "/" + log);
|
||||
}
|
||||
}
|
||||
|
||||
// recover the DB
|
||||
Reopen(options);
|
||||
ASSERT_EQ(two, Get("foo"));
|
||||
ASSERT_EQ(one, Get("bar"));
|
||||
Close();
|
||||
|
||||
// copy the logs from backup back to wal dir
|
||||
for (auto& log : logs) {
|
||||
if (log != ".." && log != ".") {
|
||||
CopyFile(backup_logs + "/" + log, options.wal_dir + "/" + log);
|
||||
}
|
||||
}
|
||||
// this should ignore the log files, recovery should not happen again
|
||||
// if the recovery happens, the same merge operator would be called twice,
|
||||
// leading to incorrect results
|
||||
Reopen(options);
|
||||
ASSERT_EQ(two, Get("foo"));
|
||||
ASSERT_EQ(one, Get("bar"));
|
||||
Close();
|
||||
Destroy(options);
|
||||
Reopen(options);
|
||||
Close();
|
||||
|
||||
// copy the logs from backup back to wal dir
|
||||
env_->CreateDirIfMissing(options.wal_dir);
|
||||
for (auto& log : logs) {
|
||||
if (log != ".." && log != ".") {
|
||||
CopyFile(backup_logs + "/" + log, options.wal_dir + "/" + log);
|
||||
}
|
||||
}
|
||||
// assert that we successfully recovered only from logs, even though we
|
||||
// destroyed the DB
|
||||
Reopen(options);
|
||||
ASSERT_EQ(two, Get("foo"));
|
||||
ASSERT_EQ(one, Get("bar"));
|
||||
|
||||
// Recovery will fail if DB directory doesn't exist.
|
||||
Destroy(options);
|
||||
// copy the logs from backup back to wal dir
|
||||
env_->CreateDirIfMissing(options.wal_dir);
|
||||
for (auto& log : logs) {
|
||||
if (log != ".." && log != ".") {
|
||||
CopyFile(backup_logs + "/" + log, options.wal_dir + "/" + log);
|
||||
// we won't be needing this file no more
|
||||
env_->DeleteFile(backup_logs + "/" + log);
|
||||
}
|
||||
}
|
||||
Status s = TryReopen(options);
|
||||
ASSERT_TRUE(!s.ok());
|
||||
} while (ChangeOptions(kSkipHashCuckoo));
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoveryWithEmptyLog) {
|
||||
do {
|
||||
CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
|
||||
ASSERT_OK(Put(1, "foo", "v1"));
|
||||
ASSERT_OK(Put(1, "foo", "v2"));
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
ASSERT_OK(Put(1, "foo", "v3"));
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
ASSERT_EQ("v3", Get(1, "foo"));
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
TEST_F(DBWALTest, GetSortedWalFiles) {
|
||||
do {
|
||||
CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
|
||||
VectorLogPtr log_files;
|
||||
ASSERT_OK(dbfull()->GetSortedWalFiles(log_files));
|
||||
ASSERT_EQ(0, log_files.size());
|
||||
|
||||
ASSERT_OK(Put(1, "foo", "v1"));
|
||||
ASSERT_OK(dbfull()->GetSortedWalFiles(log_files));
|
||||
ASSERT_EQ(1, log_files.size());
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoveryWithLogDataForSomeCFs) {
|
||||
// Test for regression of WAL cleanup missing files that don't contain data
|
||||
// for every column family.
|
||||
do {
|
||||
CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
|
||||
ASSERT_OK(Put(1, "foo", "v1"));
|
||||
ASSERT_OK(Put(1, "foo", "v2"));
|
||||
uint64_t earliest_log_nums[2];
|
||||
for (int i = 0; i < 2; ++i) {
|
||||
if (i > 0) {
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
|
||||
}
|
||||
VectorLogPtr log_files;
|
||||
ASSERT_OK(dbfull()->GetSortedWalFiles(log_files));
|
||||
if (log_files.size() > 0) {
|
||||
earliest_log_nums[i] = log_files[0]->LogNumber();
|
||||
} else {
|
||||
earliest_log_nums[i] = port::kMaxUint64;
|
||||
}
|
||||
}
|
||||
// Check at least the first WAL was cleaned up during the recovery.
|
||||
ASSERT_LT(earliest_log_nums[0], earliest_log_nums[1]);
|
||||
} while (ChangeOptions());
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoverWithLargeLog) {
|
||||
do {
|
||||
{
|
||||
Options options = CurrentOptions();
|
||||
CreateAndReopenWithCF({"pikachu"}, options);
|
||||
ASSERT_OK(Put(1, "big1", std::string(200000, '1')));
|
||||
ASSERT_OK(Put(1, "big2", std::string(200000, '2')));
|
||||
ASSERT_OK(Put(1, "small3", std::string(10, '3')));
|
||||
ASSERT_OK(Put(1, "small4", std::string(10, '4')));
|
||||
ASSERT_EQ(NumTableFilesAtLevel(0, 1), 0);
|
||||
}
|
||||
|
||||
// Make sure that if we re-open with a small write buffer size that
|
||||
// we flush table files in the middle of a large log file.
|
||||
Options options;
|
||||
options.write_buffer_size = 100000;
|
||||
options = CurrentOptions(options);
|
||||
ReopenWithColumnFamilies({"default", "pikachu"}, options);
|
||||
ASSERT_EQ(NumTableFilesAtLevel(0, 1), 3);
|
||||
ASSERT_EQ(std::string(200000, '1'), Get(1, "big1"));
|
||||
ASSERT_EQ(std::string(200000, '2'), Get(1, "big2"));
|
||||
ASSERT_EQ(std::string(10, '3'), Get(1, "small3"));
|
||||
ASSERT_EQ(std::string(10, '4'), Get(1, "small4"));
|
||||
ASSERT_GT(NumTableFilesAtLevel(0, 1), 1);
|
||||
} while (ChangeCompactOptions());
|
||||
}
|
||||
|
||||
// In https://reviews.facebook.net/D20661 we change
|
||||
// recovery behavior: previously for each log file each column family
|
||||
// memtable was flushed, even it was empty. Now it's changed:
|
||||
// we try to create the smallest number of table files by merging
|
||||
// updates from multiple logs
|
||||
TEST_F(DBWALTest, RecoverCheckFileAmountWithSmallWriteBuffer) {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 5000000;
|
||||
CreateAndReopenWithCF({"pikachu", "dobrynia", "nikitich"}, options);
|
||||
|
||||
// Since we will reopen DB with smaller write_buffer_size,
|
||||
// each key will go to new SST file
|
||||
ASSERT_OK(Put(1, Key(10), DummyString(1000000)));
|
||||
ASSERT_OK(Put(1, Key(10), DummyString(1000000)));
|
||||
ASSERT_OK(Put(1, Key(10), DummyString(1000000)));
|
||||
ASSERT_OK(Put(1, Key(10), DummyString(1000000)));
|
||||
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1)));
|
||||
// Make 'dobrynia' to be flushed and new WAL file to be created
|
||||
ASSERT_OK(Put(2, Key(10), DummyString(7500000)));
|
||||
ASSERT_OK(Put(2, Key(1), DummyString(1)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[2]);
|
||||
{
|
||||
auto tables = ListTableFiles(env_, dbname_);
|
||||
ASSERT_EQ(tables.size(), static_cast<size_t>(1));
|
||||
// Make sure 'dobrynia' was flushed: check sst files amount
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
|
||||
static_cast<uint64_t>(1));
|
||||
}
|
||||
// New WAL file
|
||||
ASSERT_OK(Put(1, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(1, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1)));
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1)));
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1)));
|
||||
|
||||
options.write_buffer_size = 4096;
|
||||
options.arena_block_size = 4096;
|
||||
ReopenWithColumnFamilies({"default", "pikachu", "dobrynia", "nikitich"},
|
||||
options);
|
||||
{
|
||||
// No inserts => default is empty
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
|
||||
static_cast<uint64_t>(0));
|
||||
// First 4 keys goes to separate SSTs + 1 more SST for 2 smaller keys
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
|
||||
static_cast<uint64_t>(5));
|
||||
// 1 SST for big key + 1 SST for small one
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
|
||||
static_cast<uint64_t>(2));
|
||||
// 1 SST for all keys
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
|
||||
static_cast<uint64_t>(1));
|
||||
}
|
||||
}
|
||||
|
||||
// In https://reviews.facebook.net/D20661 we change
|
||||
// recovery behavior: previously for each log file each column family
|
||||
// memtable was flushed, even it wasn't empty. Now it's changed:
|
||||
// we try to create the smallest number of table files by merging
|
||||
// updates from multiple logs
|
||||
TEST_F(DBWALTest, RecoverCheckFileAmount) {
|
||||
Options options = CurrentOptions();
|
||||
options.write_buffer_size = 100000;
|
||||
options.arena_block_size = 4 * 1024;
|
||||
options.avoid_flush_during_recovery = false;
|
||||
CreateAndReopenWithCF({"pikachu", "dobrynia", "nikitich"}, options);
|
||||
|
||||
ASSERT_OK(Put(0, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(1, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(2, Key(1), DummyString(1)));
|
||||
|
||||
// Make 'nikitich' memtable to be flushed
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1002400)));
|
||||
ASSERT_OK(Put(3, Key(1), DummyString(1)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[3]);
|
||||
// 4 memtable are not flushed, 1 sst file
|
||||
{
|
||||
auto tables = ListTableFiles(env_, dbname_);
|
||||
ASSERT_EQ(tables.size(), static_cast<size_t>(1));
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
|
||||
static_cast<uint64_t>(1));
|
||||
}
|
||||
// Memtable for 'nikitich' has flushed, new WAL file has opened
|
||||
// 4 memtable still not flushed
|
||||
|
||||
// Write to new WAL file
|
||||
ASSERT_OK(Put(0, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(1, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(2, Key(1), DummyString(1)));
|
||||
|
||||
// Fill up 'nikitich' one more time
|
||||
ASSERT_OK(Put(3, Key(10), DummyString(1002400)));
|
||||
// make it flush
|
||||
ASSERT_OK(Put(3, Key(1), DummyString(1)));
|
||||
dbfull()->TEST_WaitForFlushMemTable(handles_[3]);
|
||||
// There are still 4 memtable not flushed, and 2 sst tables
|
||||
ASSERT_OK(Put(0, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(1, Key(1), DummyString(1)));
|
||||
ASSERT_OK(Put(2, Key(1), DummyString(1)));
|
||||
|
||||
{
|
||||
auto tables = ListTableFiles(env_, dbname_);
|
||||
ASSERT_EQ(tables.size(), static_cast<size_t>(2));
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
|
||||
static_cast<uint64_t>(2));
|
||||
}
|
||||
|
||||
ReopenWithColumnFamilies({"default", "pikachu", "dobrynia", "nikitich"},
|
||||
options);
|
||||
{
|
||||
std::vector<uint64_t> table_files = ListTableFiles(env_, dbname_);
|
||||
// Check, that records for 'default', 'dobrynia' and 'pikachu' from
|
||||
// first, second and third WALs went to the same SST.
|
||||
// So, there is 6 SSTs: three for 'nikitich', one for 'default', one for
|
||||
// 'dobrynia', one for 'pikachu'
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
|
||||
static_cast<uint64_t>(1));
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
|
||||
static_cast<uint64_t>(3));
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
|
||||
static_cast<uint64_t>(1));
|
||||
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
|
||||
static_cast<uint64_t>(1));
|
||||
}
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, SyncMultipleLogs) {
|
||||
const uint64_t kNumBatches = 2;
|
||||
const int kBatchSize = 1000;
|
||||
|
||||
Options options = CurrentOptions();
|
||||
options.create_if_missing = true;
|
||||
options.write_buffer_size = 4096;
|
||||
Reopen(options);
|
||||
|
||||
WriteBatch batch;
|
||||
WriteOptions wo;
|
||||
wo.sync = true;
|
||||
|
||||
for (uint64_t b = 0; b < kNumBatches; b++) {
|
||||
batch.Clear();
|
||||
for (int i = 0; i < kBatchSize; i++) {
|
||||
batch.Put(Key(i), DummyString(128));
|
||||
}
|
||||
|
||||
dbfull()->Write(wo, &batch);
|
||||
}
|
||||
|
||||
ASSERT_OK(dbfull()->SyncWAL());
|
||||
}
|
||||
|
||||
// Github issue 1339. Prior the fix we read sequence id from the first log to
|
||||
// a local variable, then keep increase the variable as we replay logs,
|
||||
// ignoring actual sequence id of the records. This is incorrect if some writes
|
||||
// come with WAL disabled.
|
||||
TEST_F(DBWALTest, PartOfWritesWithWALDisabled) {
|
||||
std::unique_ptr<FaultInjectionTestEnv> fault_env(
|
||||
new FaultInjectionTestEnv(env_));
|
||||
Options options = CurrentOptions();
|
||||
options.env = fault_env.get();
|
||||
options.disable_auto_compactions = true;
|
||||
// TODO(yiwu): fix for 2PC.
|
||||
options.allow_2pc = false;
|
||||
WriteOptions wal_on, wal_off;
|
||||
wal_on.sync = true;
|
||||
wal_on.disableWAL = false;
|
||||
wal_off.disableWAL = true;
|
||||
CreateAndReopenWithCF({"dummy"}, options);
|
||||
ASSERT_OK(Put(1, "dummy", "d1", wal_on)); // seq id 1
|
||||
ASSERT_OK(Put(1, "dummy", "d2", wal_off));
|
||||
ASSERT_OK(Put(1, "dummy", "d3", wal_off));
|
||||
ASSERT_OK(Put(0, "key", "v4", wal_on)); // seq id 4
|
||||
ASSERT_OK(Flush(0));
|
||||
ASSERT_OK(Put(0, "key", "v5", wal_on)); // seq id 5
|
||||
ASSERT_EQ("v5", Get(0, "key"));
|
||||
// Simulate a crash.
|
||||
fault_env->SetFilesystemActive(false);
|
||||
Close();
|
||||
fault_env->ResetState();
|
||||
ReopenWithColumnFamilies({"default", "dummy"}, options);
|
||||
// Prior to the fix, we may incorrectly recover "v5" with sequence id = 3.
|
||||
ASSERT_EQ("v5", Get(0, "key"));
|
||||
// Destroy DB before destruct fault_env.
|
||||
Destroy(options);
|
||||
}
|
||||
|
||||
//
|
||||
// Test WAL recovery for the various modes available
|
||||
//
|
||||
class RecoveryTestHelper {
|
||||
public:
|
||||
// Number of WAL files to generate
|
||||
static const int kWALFilesCount = 10;
|
||||
// Starting number for the WAL file name like 00010.log
|
||||
static const int kWALFileOffset = 10;
|
||||
// Keys to be written per WAL file
|
||||
static const int kKeysPerWALFile = 1024;
|
||||
// Size of the value
|
||||
static const int kValueSize = 10;
|
||||
|
||||
// Create WAL files with values filled in
|
||||
static void FillData(DBWALTest* test, const Options& options,
|
||||
const size_t wal_count, size_t* count) {
|
||||
const DBOptions& db_options = options;
|
||||
|
||||
*count = 0;
|
||||
|
||||
shared_ptr<Cache> table_cache = NewLRUCache(50000, 16);
|
||||
EnvOptions env_options;
|
||||
WriteBufferManager write_buffer_manager(db_options.db_write_buffer_size);
|
||||
|
||||
unique_ptr<VersionSet> versions;
|
||||
unique_ptr<WalManager> wal_manager;
|
||||
WriteController write_controller;
|
||||
|
||||
versions.reset(new VersionSet(test->dbname_, &db_options, env_options,
|
||||
table_cache.get(), &write_buffer_manager,
|
||||
&write_controller));
|
||||
|
||||
wal_manager.reset(new WalManager(db_options, env_options));
|
||||
|
||||
std::unique_ptr<log::Writer> current_log_writer;
|
||||
|
||||
for (size_t j = kWALFileOffset; j < wal_count + kWALFileOffset; j++) {
|
||||
uint64_t current_log_number = j;
|
||||
std::string fname = LogFileName(test->dbname_, current_log_number);
|
||||
unique_ptr<WritableFile> file;
|
||||
ASSERT_OK(db_options.env->NewWritableFile(fname, &file, env_options));
|
||||
unique_ptr<WritableFileWriter> file_writer(
|
||||
new WritableFileWriter(std::move(file), env_options));
|
||||
current_log_writer.reset(
|
||||
new log::Writer(std::move(file_writer), current_log_number,
|
||||
db_options.recycle_log_file_num > 0));
|
||||
|
||||
for (int i = 0; i < kKeysPerWALFile; i++) {
|
||||
std::string key = "key" + ToString((*count)++);
|
||||
std::string value = test->DummyString(kValueSize);
|
||||
assert(current_log_writer.get() != nullptr);
|
||||
uint64_t seq = versions->LastSequence() + 1;
|
||||
WriteBatch batch;
|
||||
batch.Put(key, value);
|
||||
WriteBatchInternal::SetSequence(&batch, seq);
|
||||
current_log_writer->AddRecord(WriteBatchInternal::Contents(&batch));
|
||||
versions->SetLastSequence(seq);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Recreate and fill the store with some data
|
||||
static size_t FillData(DBWALTest* test, Options* options) {
|
||||
options->create_if_missing = true;
|
||||
test->DestroyAndReopen(*options);
|
||||
test->Close();
|
||||
|
||||
size_t count = 0;
|
||||
FillData(test, *options, kWALFilesCount, &count);
|
||||
return count;
|
||||
}
|
||||
|
||||
// Read back all the keys we wrote and return the number of keys found
|
||||
static size_t GetData(DBWALTest* test) {
|
||||
size_t count = 0;
|
||||
for (size_t i = 0; i < kWALFilesCount * kKeysPerWALFile; i++) {
|
||||
if (test->Get("key" + ToString(i)) != "NOT_FOUND") {
|
||||
++count;
|
||||
}
|
||||
}
|
||||
return count;
|
||||
}
|
||||
|
||||
// Manuall corrupt the specified WAL
|
||||
static void CorruptWAL(DBWALTest* test, const Options& options,
|
||||
const double off, const double len,
|
||||
const int wal_file_id, const bool trunc = false) {
|
||||
Env* env = options.env;
|
||||
std::string fname = LogFileName(test->dbname_, wal_file_id);
|
||||
uint64_t size;
|
||||
ASSERT_OK(env->GetFileSize(fname, &size));
|
||||
ASSERT_GT(size, 0);
|
||||
#ifdef OS_WIN
|
||||
// Windows disk cache behaves differently. When we truncate
|
||||
// the original content is still in the cache due to the original
|
||||
// handle is still open. Generally, in Windows, one prohibits
|
||||
// shared access to files and it is not needed for WAL but we allow
|
||||
// it to induce corruption at various tests.
|
||||
test->Close();
|
||||
#endif
|
||||
if (trunc) {
|
||||
ASSERT_EQ(0, truncate(fname.c_str(), static_cast<int64_t>(size * off)));
|
||||
} else {
|
||||
InduceCorruption(fname, static_cast<size_t>(size * off),
|
||||
static_cast<size_t>(size * len));
|
||||
}
|
||||
}
|
||||
|
||||
// Overwrite data with 'a' from offset for length len
|
||||
static void InduceCorruption(const std::string& filename, size_t offset,
|
||||
size_t len) {
|
||||
ASSERT_GT(len, 0U);
|
||||
|
||||
int fd = open(filename.c_str(), O_RDWR);
|
||||
|
||||
// On windows long is 32-bit
|
||||
ASSERT_LE(offset, std::numeric_limits<long>::max());
|
||||
|
||||
ASSERT_GT(fd, 0);
|
||||
ASSERT_EQ(offset, lseek(fd, static_cast<long>(offset), SEEK_SET));
|
||||
|
||||
void* buf = alloca(len);
|
||||
memset(buf, 'a', len);
|
||||
ASSERT_EQ(len, write(fd, buf, static_cast<unsigned int>(len)));
|
||||
|
||||
close(fd);
|
||||
}
|
||||
};
|
||||
|
||||
// Test scope:
|
||||
// - We expect to open the data store when there is incomplete trailing writes
|
||||
// at the end of any of the logs
|
||||
// - We do not expect to open the data store for corruption
|
||||
TEST_F(DBWALTest, kTolerateCorruptedTailRecords) {
|
||||
const int jstart = RecoveryTestHelper::kWALFileOffset;
|
||||
const int jend = jstart + RecoveryTestHelper::kWALFilesCount;
|
||||
|
||||
for (auto trunc : {true, false}) { /* Corruption style */
|
||||
for (int i = 0; i < 4; i++) { /* Corruption offset position */
|
||||
for (int j = jstart; j < jend; j++) { /* WAL file */
|
||||
// Fill data for testing
|
||||
Options options = CurrentOptions();
|
||||
const size_t row_count = RecoveryTestHelper::FillData(this, &options);
|
||||
// test checksum failure or parsing
|
||||
RecoveryTestHelper::CorruptWAL(this, options, /*off=*/i * .3,
|
||||
/*len%=*/.1, /*wal=*/j, trunc);
|
||||
|
||||
if (trunc) {
|
||||
options.wal_recovery_mode =
|
||||
WALRecoveryMode::kTolerateCorruptedTailRecords;
|
||||
options.create_if_missing = false;
|
||||
ASSERT_OK(TryReopen(options));
|
||||
const size_t recovered_row_count = RecoveryTestHelper::GetData(this);
|
||||
ASSERT_TRUE(i == 0 || recovered_row_count > 0);
|
||||
ASSERT_LT(recovered_row_count, row_count);
|
||||
} else {
|
||||
options.wal_recovery_mode =
|
||||
WALRecoveryMode::kTolerateCorruptedTailRecords;
|
||||
ASSERT_NOK(TryReopen(options));
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Test scope:
|
||||
// We don't expect the data store to be opened if there is any corruption
|
||||
// (leading, middle or trailing -- incomplete writes or corruption)
|
||||
TEST_F(DBWALTest, kAbsoluteConsistency) {
|
||||
const int jstart = RecoveryTestHelper::kWALFileOffset;
|
||||
const int jend = jstart + RecoveryTestHelper::kWALFilesCount;
|
||||
|
||||
// Verify clean slate behavior
|
||||
Options options = CurrentOptions();
|
||||
const size_t row_count = RecoveryTestHelper::FillData(this, &options);
|
||||
options.wal_recovery_mode = WALRecoveryMode::kAbsoluteConsistency;
|
||||
options.create_if_missing = false;
|
||||
ASSERT_OK(TryReopen(options));
|
||||
ASSERT_EQ(RecoveryTestHelper::GetData(this), row_count);
|
||||
|
||||
for (auto trunc : {true, false}) { /* Corruption style */
|
||||
for (int i = 0; i < 4; i++) { /* Corruption offset position */
|
||||
if (trunc && i == 0) {
|
||||
continue;
|
||||
}
|
||||
|
||||
for (int j = jstart; j < jend; j++) { /* wal files */
|
||||
// fill with new date
|
||||
RecoveryTestHelper::FillData(this, &options);
|
||||
// corrupt the wal
|
||||
RecoveryTestHelper::CorruptWAL(this, options, /*off=*/i * .3,
|
||||
/*len%=*/.1, j, trunc);
|
||||
// verify
|
||||
options.wal_recovery_mode = WALRecoveryMode::kAbsoluteConsistency;
|
||||
options.create_if_missing = false;
|
||||
ASSERT_NOK(TryReopen(options));
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Test scope:
|
||||
// - We expect to open data store under all circumstances
|
||||
// - We expect only data upto the point where the first error was encountered
|
||||
TEST_F(DBWALTest, kPointInTimeRecovery) {
|
||||
const int jstart = RecoveryTestHelper::kWALFileOffset;
|
||||
const int jend = jstart + RecoveryTestHelper::kWALFilesCount;
|
||||
const int maxkeys =
|
||||
RecoveryTestHelper::kWALFilesCount * RecoveryTestHelper::kKeysPerWALFile;
|
||||
|
||||
for (auto trunc : {true, false}) { /* Corruption style */
|
||||
for (int i = 0; i < 4; i++) { /* Offset of corruption */
|
||||
for (int j = jstart; j < jend; j++) { /* WAL file */
|
||||
// Fill data for testing
|
||||
Options options = CurrentOptions();
|
||||
const size_t row_count = RecoveryTestHelper::FillData(this, &options);
|
||||
|
||||
// Corrupt the wal
|
||||
RecoveryTestHelper::CorruptWAL(this, options, /*off=*/i * .3,
|
||||
/*len%=*/.1, j, trunc);
|
||||
|
||||
// Verify
|
||||
options.wal_recovery_mode = WALRecoveryMode::kPointInTimeRecovery;
|
||||
options.create_if_missing = false;
|
||||
ASSERT_OK(TryReopen(options));
|
||||
|
||||
// Probe data for invariants
|
||||
size_t recovered_row_count = RecoveryTestHelper::GetData(this);
|
||||
ASSERT_LT(recovered_row_count, row_count);
|
||||
|
||||
bool expect_data = true;
|
||||
for (size_t k = 0; k < maxkeys; ++k) {
|
||||
bool found = Get("key" + ToString(i)) != "NOT_FOUND";
|
||||
if (expect_data && !found) {
|
||||
expect_data = false;
|
||||
}
|
||||
ASSERT_EQ(found, expect_data);
|
||||
}
|
||||
|
||||
const size_t min = RecoveryTestHelper::kKeysPerWALFile *
|
||||
(j - RecoveryTestHelper::kWALFileOffset);
|
||||
ASSERT_GE(recovered_row_count, min);
|
||||
if (!trunc && i != 0) {
|
||||
const size_t max = RecoveryTestHelper::kKeysPerWALFile *
|
||||
(j - RecoveryTestHelper::kWALFileOffset + 1);
|
||||
ASSERT_LE(recovered_row_count, max);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Test scope:
|
||||
// - We expect to open the data store under all scenarios
|
||||
// - We expect to have recovered records past the corruption zone
|
||||
TEST_F(DBWALTest, kSkipAnyCorruptedRecords) {
|
||||
const int jstart = RecoveryTestHelper::kWALFileOffset;
|
||||
const int jend = jstart + RecoveryTestHelper::kWALFilesCount;
|
||||
|
||||
for (auto trunc : {true, false}) { /* Corruption style */
|
||||
for (int i = 0; i < 4; i++) { /* Corruption offset */
|
||||
for (int j = jstart; j < jend; j++) { /* wal files */
|
||||
// Fill data for testing
|
||||
Options options = CurrentOptions();
|
||||
const size_t row_count = RecoveryTestHelper::FillData(this, &options);
|
||||
|
||||
// Corrupt the WAL
|
||||
RecoveryTestHelper::CorruptWAL(this, options, /*off=*/i * .3,
|
||||
/*len%=*/.1, j, trunc);
|
||||
|
||||
// Verify behavior
|
||||
options.wal_recovery_mode = WALRecoveryMode::kSkipAnyCorruptedRecords;
|
||||
options.create_if_missing = false;
|
||||
ASSERT_OK(TryReopen(options));
|
||||
|
||||
// Probe data for invariants
|
||||
size_t recovered_row_count = RecoveryTestHelper::GetData(this);
|
||||
ASSERT_LT(recovered_row_count, row_count);
|
||||
|
||||
if (!trunc) {
|
||||
ASSERT_TRUE(i != 0 || recovered_row_count > 0);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, AvoidFlushDuringRecovery) {
|
||||
Options options = CurrentOptions();
|
||||
options.disable_auto_compactions = true;
|
||||
options.avoid_flush_during_recovery = false;
|
||||
|
||||
// Test with flush after recovery.
|
||||
Reopen(options);
|
||||
ASSERT_OK(Put("foo", "v1"));
|
||||
ASSERT_OK(Put("bar", "v2"));
|
||||
ASSERT_OK(Flush());
|
||||
ASSERT_OK(Put("foo", "v3"));
|
||||
ASSERT_OK(Put("bar", "v4"));
|
||||
ASSERT_EQ(1, TotalTableFiles());
|
||||
// Reopen DB. Check if WAL logs flushed.
|
||||
Reopen(options);
|
||||
ASSERT_EQ("v3", Get("foo"));
|
||||
ASSERT_EQ("v4", Get("bar"));
|
||||
ASSERT_EQ(2, TotalTableFiles());
|
||||
|
||||
// Test without flush after recovery.
|
||||
options.avoid_flush_during_recovery = true;
|
||||
DestroyAndReopen(options);
|
||||
ASSERT_OK(Put("foo", "v5"));
|
||||
ASSERT_OK(Put("bar", "v6"));
|
||||
ASSERT_OK(Flush());
|
||||
ASSERT_OK(Put("foo", "v7"));
|
||||
ASSERT_OK(Put("bar", "v8"));
|
||||
ASSERT_EQ(1, TotalTableFiles());
|
||||
// Reopen DB. WAL logs should not be flushed this time.
|
||||
Reopen(options);
|
||||
ASSERT_EQ("v7", Get("foo"));
|
||||
ASSERT_EQ("v8", Get("bar"));
|
||||
ASSERT_EQ(1, TotalTableFiles());
|
||||
|
||||
// Force flush with allow_2pc.
|
||||
options.avoid_flush_during_recovery = true;
|
||||
options.allow_2pc = true;
|
||||
ASSERT_OK(Put("foo", "v9"));
|
||||
ASSERT_OK(Put("bar", "v10"));
|
||||
ASSERT_OK(Flush());
|
||||
ASSERT_OK(Put("foo", "v11"));
|
||||
ASSERT_OK(Put("bar", "v12"));
|
||||
Reopen(options);
|
||||
ASSERT_EQ("v11", Get("foo"));
|
||||
ASSERT_EQ("v12", Get("bar"));
|
||||
ASSERT_EQ(2, TotalTableFiles());
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoverWithoutFlush) {
|
||||
Options options = CurrentOptions();
|
||||
options.avoid_flush_during_recovery = true;
|
||||
options.create_if_missing = false;
|
||||
options.disable_auto_compactions = true;
|
||||
options.write_buffer_size = 64 * 1024 * 1024;
|
||||
|
||||
size_t count = RecoveryTestHelper::FillData(this, &options);
|
||||
auto validateData = [this, count]() {
|
||||
for (size_t i = 0; i < count; i++) {
|
||||
ASSERT_NE(Get("key" + ToString(i)), "NOT_FOUND");
|
||||
}
|
||||
};
|
||||
Reopen(options);
|
||||
validateData();
|
||||
// Insert some data without flush
|
||||
ASSERT_OK(Put("foo", "foo_v1"));
|
||||
ASSERT_OK(Put("bar", "bar_v1"));
|
||||
Reopen(options);
|
||||
validateData();
|
||||
ASSERT_EQ(Get("foo"), "foo_v1");
|
||||
ASSERT_EQ(Get("bar"), "bar_v1");
|
||||
// Insert again and reopen
|
||||
ASSERT_OK(Put("foo", "foo_v2"));
|
||||
ASSERT_OK(Put("bar", "bar_v2"));
|
||||
Reopen(options);
|
||||
validateData();
|
||||
ASSERT_EQ(Get("foo"), "foo_v2");
|
||||
ASSERT_EQ(Get("bar"), "bar_v2");
|
||||
// manual flush and insert again
|
||||
Flush();
|
||||
ASSERT_EQ(Get("foo"), "foo_v2");
|
||||
ASSERT_EQ(Get("bar"), "bar_v2");
|
||||
ASSERT_OK(Put("foo", "foo_v3"));
|
||||
ASSERT_OK(Put("bar", "bar_v3"));
|
||||
Reopen(options);
|
||||
validateData();
|
||||
ASSERT_EQ(Get("foo"), "foo_v3");
|
||||
ASSERT_EQ(Get("bar"), "bar_v3");
|
||||
}
|
||||
|
||||
TEST_F(DBWALTest, RecoverWithoutFlushMultipleCF) {
|
||||
const std::string kSmallValue = "v";
|
||||
const std::string kLargeValue = DummyString(1024);
|
||||
Options options = CurrentOptions();
|
||||
options.avoid_flush_during_recovery = true;
|
||||
options.create_if_missing = false;
|
||||
options.disable_auto_compactions = true;
|
||||
|
||||
auto countWalFiles = [this]() {
|
||||
VectorLogPtr log_files;
|
||||
dbfull()->GetSortedWalFiles(log_files);
|
||||
return log_files.size();
|
||||
};
|
||||
|
||||
// Create DB with multiple column families and multiple log files.
|
||||
CreateAndReopenWithCF({"one", "two"}, options);
|
||||
ASSERT_OK(Put(0, "key1", kSmallValue));
|
||||
ASSERT_OK(Put(1, "key2", kLargeValue));
|
||||
Flush(1);
|
||||
ASSERT_EQ(1, countWalFiles());
|
||||
ASSERT_OK(Put(0, "key3", kSmallValue));
|
||||
ASSERT_OK(Put(2, "key4", kLargeValue));
|
||||
Flush(2);
|
||||
ASSERT_EQ(2, countWalFiles());
|
||||
|
||||
// Reopen, insert and flush.
|
||||
options.db_write_buffer_size = 64 * 1024 * 1024;
|
||||
ReopenWithColumnFamilies({"default", "one", "two"}, options);
|
||||
ASSERT_EQ(Get(0, "key1"), kSmallValue);
|
||||
ASSERT_EQ(Get(1, "key2"), kLargeValue);
|
||||
ASSERT_EQ(Get(0, "key3"), kSmallValue);
|
||||
ASSERT_EQ(Get(2, "key4"), kLargeValue);
|
||||
// Insert more data.
|
||||
ASSERT_OK(Put(0, "key5", kLargeValue));
|
||||
ASSERT_OK(Put(1, "key6", kLargeValue));
|
||||
ASSERT_EQ(3, countWalFiles());
|
||||
Flush(1);
|
||||
ASSERT_OK(Put(2, "key7", kLargeValue));
|
||||
ASSERT_EQ(4, countWalFiles());
|
||||
|
||||
// Reopen twice and validate.
|
||||
for (int i = 0; i < 2; i++) {
|
||||
ReopenWithColumnFamilies({"default", "one", "two"}, options);
|
||||
ASSERT_EQ(Get(0, "key1"), kSmallValue);
|
||||
ASSERT_EQ(Get(1, "key2"), kLargeValue);
|
||||
ASSERT_EQ(Get(0, "key3"), kSmallValue);
|
||||
ASSERT_EQ(Get(2, "key4"), kLargeValue);
|
||||
ASSERT_EQ(Get(0, "key5"), kLargeValue);
|
||||
ASSERT_EQ(Get(1, "key6"), kLargeValue);
|
||||
ASSERT_EQ(Get(2, "key7"), kLargeValue);
|
||||
ASSERT_EQ(4, countWalFiles());
|
||||
}
|
||||
}
|
||||
|
||||
// In this test we are trying to do the following:
|
||||
// 1. Create a DB with corrupted WAL log;
|
||||
// 2. Open with avoid_flush_during_recovery = true;
|
||||
// 3. Append more data without flushing, which creates new WAL log.
|
||||
// 4. Open again. See if it can correctly handle previous corruption.
|
||||
TEST_F(DBWALTest, RecoverFromCorruptedWALWithoutFlush) {
|
||||
const int jstart = RecoveryTestHelper::kWALFileOffset;
|
||||
const int jend = jstart + RecoveryTestHelper::kWALFilesCount;
|
||||
const int kAppendKeys = 100;
|
||||
Options options = CurrentOptions();
|
||||
options.avoid_flush_during_recovery = true;
|
||||
options.create_if_missing = false;
|
||||
options.disable_auto_compactions = true;
|
||||
options.write_buffer_size = 64 * 1024 * 1024;
|
||||
|
||||
auto getAll = [this]() {
|
||||
std::vector<std::pair<std::string, std::string>> data;
|
||||
ReadOptions ropt;
|
||||
Iterator* iter = dbfull()->NewIterator(ropt);
|
||||
for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
|
||||
data.push_back(
|
||||
std::make_pair(iter->key().ToString(), iter->value().ToString()));
|
||||
}
|
||||
delete iter;
|
||||
return data;
|
||||
};
|
||||
for (auto& mode : wal_recovery_mode_string_map) {
|
||||
options.wal_recovery_mode = mode.second;
|
||||
for (auto trunc : {true, false}) {
|
||||
for (int i = 0; i < 4; i++) {
|
||||
for (int j = jstart; j < jend; j++) {
|
||||
// Create corrupted WAL
|
||||
RecoveryTestHelper::FillData(this, &options);
|
||||
RecoveryTestHelper::CorruptWAL(this, options, /*off=*/i * .3,
|
||||
/*len%=*/.1, /*wal=*/j, trunc);
|
||||
// Skip the test if DB won't open.
|
||||
if (!TryReopen(options).ok()) {
|
||||
ASSERT_TRUE(options.wal_recovery_mode ==
|
||||
WALRecoveryMode::kAbsoluteConsistency ||
|
||||
(!trunc &&
|
||||
options.wal_recovery_mode ==
|
||||
WALRecoveryMode::kTolerateCorruptedTailRecords));
|
||||
continue;
|
||||
}
|
||||
ASSERT_OK(TryReopen(options));
|
||||
// Append some more data.
|
||||
for (int k = 0; k < kAppendKeys; k++) {
|
||||
std::string key = "extra_key" + ToString(k);
|
||||
std::string value = DummyString(RecoveryTestHelper::kValueSize);
|
||||
ASSERT_OK(Put(key, value));
|
||||
}
|
||||
// Save data for comparison.
|
||||
auto data = getAll();
|
||||
// Reopen. Verify data.
|
||||
ASSERT_OK(TryReopen(options));
|
||||
auto actual_data = getAll();
|
||||
ASSERT_EQ(data, actual_data);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#endif // ROCKSDB_LITE
|
||||
|
||||
} // namespace rocksdb
|
||||
|
||||
int main(int argc, char** argv) {
|
||||
#if !(defined NDEBUG) || !defined(OS_WIN)
|
||||
rocksdb::port::InstallStackTraceHandler();
|
||||
::testing::InitGoogleTest(&argc, argv);
|
||||
return RUN_ALL_TESTS();
|
||||
#else
|
||||
return 0;
|
||||
#endif
|
||||
}
|
||||
+3
-3
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -22,7 +22,7 @@ namespace rocksdb {
|
||||
|
||||
uint64_t PackSequenceAndType(uint64_t seq, ValueType t) {
|
||||
assert(seq <= kMaxSequenceNumber);
|
||||
assert(IsExtendedValueType(t));
|
||||
assert(IsValueType(t));
|
||||
return (seq << 8) | t;
|
||||
}
|
||||
|
||||
@@ -31,7 +31,7 @@ void UnPackSequenceAndType(uint64_t packed, uint64_t* seq, ValueType* t) {
|
||||
*t = static_cast<ValueType>(packed & 0xff);
|
||||
|
||||
assert(*seq <= kMaxSequenceNumber);
|
||||
assert(IsExtendedValueType(*t));
|
||||
assert(IsValueType(*t));
|
||||
}
|
||||
|
||||
void AppendInternalKey(std::string* result, const ParsedInternalKey& key) {
|
||||
|
||||
+28
-103
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -10,7 +10,6 @@
|
||||
#pragma once
|
||||
#include <stdio.h>
|
||||
#include <string>
|
||||
#include <utility>
|
||||
#include "rocksdb/comparator.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/filter_policy.h"
|
||||
@@ -40,13 +39,6 @@ enum ValueType : unsigned char {
|
||||
kTypeColumnFamilyMerge = 0x6, // WAL only.
|
||||
kTypeSingleDeletion = 0x7,
|
||||
kTypeColumnFamilySingleDeletion = 0x8, // WAL only.
|
||||
kTypeBeginPrepareXID = 0x9, // WAL only.
|
||||
kTypeEndPrepareXID = 0xA, // WAL only.
|
||||
kTypeCommitXID = 0xB, // WAL only.
|
||||
kTypeRollbackXID = 0xC, // WAL only.
|
||||
kTypeNoop = 0xD, // WAL only.
|
||||
kTypeColumnFamilyRangeDeletion = 0xE, // WAL only.
|
||||
kTypeRangeDeletion = 0xF, // meta block
|
||||
kMaxValue = 0x7F // Not used for storing records.
|
||||
};
|
||||
|
||||
@@ -58,18 +50,12 @@ enum ValueType : unsigned char {
|
||||
// ValueType, not the lowest).
|
||||
static const ValueType kValueTypeForSeek = kTypeSingleDeletion;
|
||||
|
||||
// Checks whether a type is an inline value type
|
||||
// (i.e. a type used in memtable skiplist and sst file datablock).
|
||||
// Checks whether a type is a value type (i.e. a type used in memtables and sst
|
||||
// files).
|
||||
inline bool IsValueType(ValueType t) {
|
||||
return t <= kTypeMerge || t == kTypeSingleDeletion;
|
||||
}
|
||||
|
||||
// Checks whether a type is from user operation
|
||||
// kTypeRangeDeletion is in meta block so this API is separated from above
|
||||
inline bool IsExtendedValueType(ValueType t) {
|
||||
return IsValueType(t) || t == kTypeRangeDeletion;
|
||||
}
|
||||
|
||||
// We leave eight bits empty at the bottom so a type and sequence#
|
||||
// can be packed together into 64-bits.
|
||||
static const SequenceNumber kMaxSequenceNumber =
|
||||
@@ -188,10 +174,6 @@ class InternalKey {
|
||||
Slice user_key() const { return ExtractUserKey(rep_); }
|
||||
size_t size() { return rep_.size(); }
|
||||
|
||||
void Set(const Slice& _user_key, SequenceNumber s, ValueType t) {
|
||||
SetFrom(ParsedInternalKey(_user_key, s, t));
|
||||
}
|
||||
|
||||
void SetFrom(const ParsedInternalKey& p) {
|
||||
rep_.clear();
|
||||
AppendInternalKey(&rep_, p);
|
||||
@@ -217,7 +199,7 @@ inline bool ParseInternalKey(const Slice& internal_key,
|
||||
result->type = static_cast<ValueType>(c);
|
||||
assert(result->type <= ValueType::kMaxValue);
|
||||
result->user_key = Slice(internal_key.data(), n - 8);
|
||||
return IsExtendedValueType(result->type);
|
||||
return IsValueType(result->type);
|
||||
}
|
||||
|
||||
// Update the sequence number in the internal key.
|
||||
@@ -289,8 +271,7 @@ inline LookupKey::~LookupKey() {
|
||||
|
||||
class IterKey {
|
||||
public:
|
||||
IterKey()
|
||||
: buf_(space_), buf_size_(sizeof(space_)), key_(buf_), key_size_(0) {}
|
||||
IterKey() : key_(space_), buf_size_(sizeof(space_)), key_size_(0) {}
|
||||
|
||||
~IterKey() { ResetBuffer(); }
|
||||
|
||||
@@ -312,41 +293,31 @@ class IterKey {
|
||||
void TrimAppend(const size_t shared_len, const char* non_shared_data,
|
||||
const size_t non_shared_len) {
|
||||
assert(shared_len <= key_size_);
|
||||
size_t total_size = shared_len + non_shared_len;
|
||||
|
||||
if (IsKeyPinned() /* key is not in buf_ */) {
|
||||
// Copy the key from external memory to buf_ (copy shared_len bytes)
|
||||
EnlargeBufferIfNeeded(total_size);
|
||||
memcpy(buf_, key_, shared_len);
|
||||
} else if (total_size > buf_size_) {
|
||||
size_t total_size = shared_len + non_shared_len;
|
||||
if (total_size <= buf_size_) {
|
||||
key_size_ = total_size;
|
||||
} else {
|
||||
// Need to allocate space, delete previous space
|
||||
char* p = new char[total_size];
|
||||
memcpy(p, key_, shared_len);
|
||||
|
||||
if (buf_ != space_) {
|
||||
delete[] buf_;
|
||||
if (key_ != space_) {
|
||||
delete[] key_;
|
||||
}
|
||||
|
||||
buf_ = p;
|
||||
key_ = p;
|
||||
key_size_ = total_size;
|
||||
buf_size_ = total_size;
|
||||
}
|
||||
|
||||
memcpy(buf_ + shared_len, non_shared_data, non_shared_len);
|
||||
key_ = buf_;
|
||||
key_size_ = total_size;
|
||||
memcpy(key_ + shared_len, non_shared_data, non_shared_len);
|
||||
}
|
||||
|
||||
Slice SetKey(const Slice& key, bool copy = true) {
|
||||
Slice SetKey(const Slice& key) {
|
||||
size_t size = key.size();
|
||||
if (copy) {
|
||||
// Copy key to buf_
|
||||
EnlargeBufferIfNeeded(size);
|
||||
memcpy(buf_, key.data(), size);
|
||||
key_ = buf_;
|
||||
} else {
|
||||
// Update key_ to point to external memory
|
||||
key_ = key.data();
|
||||
}
|
||||
EnlargeBufferIfNeeded(size);
|
||||
memcpy(key_, key.data(), size);
|
||||
key_size_ = size;
|
||||
return Slice(key_, key_size_);
|
||||
}
|
||||
@@ -364,14 +335,11 @@ class IterKey {
|
||||
// Update the sequence number in the internal key. Guarantees not to
|
||||
// invalidate slices to the key (and the user key).
|
||||
void UpdateInternalKey(uint64_t seq, ValueType t) {
|
||||
assert(!IsKeyPinned());
|
||||
assert(key_size_ >= 8);
|
||||
uint64_t newval = (seq << 8) | t;
|
||||
EncodeFixed64(&buf_[key_size_ - 8], newval);
|
||||
EncodeFixed64(&key_[key_size_ - 8], newval);
|
||||
}
|
||||
|
||||
bool IsKeyPinned() const { return (key_ != buf_); }
|
||||
|
||||
void SetInternalKey(const Slice& key_prefix, const Slice& user_key,
|
||||
SequenceNumber s,
|
||||
ValueType value_type = kValueTypeForSeek) {
|
||||
@@ -379,12 +347,10 @@ class IterKey {
|
||||
size_t usize = user_key.size();
|
||||
EnlargeBufferIfNeeded(psize + usize + sizeof(uint64_t));
|
||||
if (psize > 0) {
|
||||
memcpy(buf_, key_prefix.data(), psize);
|
||||
memcpy(key_, key_prefix.data(), psize);
|
||||
}
|
||||
memcpy(buf_ + psize, user_key.data(), usize);
|
||||
EncodeFixed64(buf_ + usize + psize, PackSequenceAndType(s, value_type));
|
||||
|
||||
key_ = buf_;
|
||||
memcpy(key_ + psize, user_key.data(), usize);
|
||||
EncodeFixed64(key_ + usize + psize, PackSequenceAndType(s, value_type));
|
||||
key_size_ = psize + usize + sizeof(uint64_t);
|
||||
}
|
||||
|
||||
@@ -411,22 +377,20 @@ class IterKey {
|
||||
void EncodeLengthPrefixedKey(const Slice& key) {
|
||||
auto size = key.size();
|
||||
EnlargeBufferIfNeeded(size + static_cast<size_t>(VarintLength(size)));
|
||||
char* ptr = EncodeVarint32(buf_, static_cast<uint32_t>(size));
|
||||
char* ptr = EncodeVarint32(key_, static_cast<uint32_t>(size));
|
||||
memcpy(ptr, key.data(), size);
|
||||
key_ = buf_;
|
||||
}
|
||||
|
||||
private:
|
||||
char* buf_;
|
||||
char* key_;
|
||||
size_t buf_size_;
|
||||
const char* key_;
|
||||
size_t key_size_;
|
||||
char space_[32]; // Avoid allocation for short keys
|
||||
|
||||
void ResetBuffer() {
|
||||
if (buf_ != space_) {
|
||||
delete[] buf_;
|
||||
buf_ = space_;
|
||||
if (key_ != space_) {
|
||||
delete[] key_;
|
||||
key_ = space_;
|
||||
}
|
||||
buf_size_ = sizeof(space_);
|
||||
key_size_ = 0;
|
||||
@@ -443,7 +407,7 @@ class IterKey {
|
||||
if (key_size > buf_size_) {
|
||||
// Need to enlarge the buffer.
|
||||
ResetBuffer();
|
||||
buf_ = new char[key_size];
|
||||
key_ = new char[key_size];
|
||||
buf_size_ = key_size;
|
||||
}
|
||||
}
|
||||
@@ -483,12 +447,6 @@ class InternalKeySliceTransform : public SliceTransform {
|
||||
const SliceTransform* const transform_;
|
||||
};
|
||||
|
||||
// Read the key of a record from a write batch.
|
||||
// if this record represent the default column family then cf_record
|
||||
// must be passed as false, otherwise it must be passed as true.
|
||||
extern bool ReadKeyFromWriteBatchEntry(Slice* input, Slice* key,
|
||||
bool cf_record);
|
||||
|
||||
// Read record from a write batch piece from input.
|
||||
// tag, column_family, key, value and blob are return values. Callers own the
|
||||
// Slice they point to.
|
||||
@@ -496,38 +454,5 @@ extern bool ReadKeyFromWriteBatchEntry(Slice* input, Slice* key,
|
||||
// input will be advanced to after the record.
|
||||
extern Status ReadRecordFromWriteBatch(Slice* input, char* tag,
|
||||
uint32_t* column_family, Slice* key,
|
||||
Slice* value, Slice* blob, Slice* xid);
|
||||
|
||||
// When user call DeleteRange() to delete a range of keys,
|
||||
// we will store a serialized RangeTombstone in MemTable and SST.
|
||||
// the struct here is a easy-understood form
|
||||
// start/end_key_ is the start/end user key of the range to be deleted
|
||||
struct RangeTombstone {
|
||||
Slice start_key_;
|
||||
Slice end_key_;
|
||||
SequenceNumber seq_;
|
||||
explicit RangeTombstone(Slice sk, Slice ek, SequenceNumber sn)
|
||||
: start_key_(sk), end_key_(ek), seq_(sn) {}
|
||||
|
||||
explicit RangeTombstone(Slice internal_key, Slice value) {
|
||||
ParsedInternalKey parsed_key;
|
||||
if (ParseInternalKey(internal_key, &parsed_key)) {
|
||||
start_key_ = parsed_key.user_key;
|
||||
seq_ = parsed_key.sequence;
|
||||
end_key_ = value;
|
||||
}
|
||||
}
|
||||
|
||||
// be careful to use Serialize(); InternalKey() allocates new memory
|
||||
std::pair<InternalKey, Slice> Serialize() {
|
||||
auto key = InternalKey(start_key_, seq_, kTypeRangeDeletion);
|
||||
Slice value = end_key_;
|
||||
return std::make_pair(std::move(key), std::move(value));
|
||||
}
|
||||
|
||||
InternalKey SerializeKey() {
|
||||
return InternalKey(start_key_, seq_, kTypeRangeDeletion);
|
||||
}
|
||||
};
|
||||
|
||||
Slice* value, Slice* blob);
|
||||
} // namespace rocksdb
|
||||
+1
-25
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -92,30 +92,6 @@ TEST_F(FormatTest, InternalKeyShortSeparator) {
|
||||
Shorten(IKey("foo", 100, kTypeValue),
|
||||
IKey("hello", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(IKey("ABC2", kMaxSequenceNumber, kValueTypeForSeek),
|
||||
Shorten(IKey("ABC1AAAAA", 100, kTypeValue),
|
||||
IKey("ABC2ABB", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(IKey("AAA2", kMaxSequenceNumber, kValueTypeForSeek),
|
||||
Shorten(IKey("AAA1AAA", 100, kTypeValue),
|
||||
IKey("AAA2AA", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(
|
||||
IKey("AAA2", kMaxSequenceNumber, kValueTypeForSeek),
|
||||
Shorten(IKey("AAA1AAA", 100, kTypeValue), IKey("AAA4", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(
|
||||
IKey("AAA1B", kMaxSequenceNumber, kValueTypeForSeek),
|
||||
Shorten(IKey("AAA1AAA", 100, kTypeValue), IKey("AAA2", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(IKey("AAA2", kMaxSequenceNumber, kValueTypeForSeek),
|
||||
Shorten(IKey("AAA1AAA", 100, kTypeValue),
|
||||
IKey("AAA2A", 200, kTypeValue)));
|
||||
|
||||
ASSERT_EQ(
|
||||
IKey("AAA1", 100, kTypeValue),
|
||||
Shorten(IKey("AAA1", 100, kTypeValue), IKey("AAA2", 200, kTypeValue)));
|
||||
|
||||
// When start user key is prefix of limit user key
|
||||
ASSERT_EQ(IKey("foo", 100, kTypeValue),
|
||||
Shorten(IKey("foo", 100, kTypeValue),
|
||||
|
||||
+8
-132
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -9,21 +9,20 @@
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
|
||||
#include <stdlib.h>
|
||||
#include <map>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
#include "rocksdb/db.h"
|
||||
#include "db/db_impl.h"
|
||||
#include "db/filename.h"
|
||||
#include "db/version_set.h"
|
||||
#include "db/write_batch_internal.h"
|
||||
#include "rocksdb/db.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/transaction_log.h"
|
||||
#include "util/string_util.h"
|
||||
#include "util/sync_point.h"
|
||||
#include "util/testharness.h"
|
||||
#include "util/testutil.h"
|
||||
#include "rocksdb/env.h"
|
||||
#include "rocksdb/transaction_log.h"
|
||||
#include <vector>
|
||||
#include <stdlib.h>
|
||||
#include <map>
|
||||
#include <string>
|
||||
|
||||
namespace rocksdb {
|
||||
|
||||
@@ -38,7 +37,6 @@ class DeleteFileTest : public testing::Test {
|
||||
DeleteFileTest() {
|
||||
db_ = nullptr;
|
||||
env_ = Env::Default();
|
||||
options_.delete_obsolete_files_period_micros = 0; // always do full purge
|
||||
options_.enable_thread_tracking = true;
|
||||
options_.write_buffer_size = 1024*1024*1000;
|
||||
options_.target_file_size_base = 1024*1024*1000;
|
||||
@@ -76,7 +74,6 @@ class DeleteFileTest : public testing::Test {
|
||||
|
||||
void CloseDB() {
|
||||
delete db_;
|
||||
db_ = nullptr;
|
||||
}
|
||||
|
||||
void AddKeys(int numkeys, int startkey = 0) {
|
||||
@@ -153,15 +150,6 @@ class DeleteFileTest : public testing::Test {
|
||||
ASSERT_EQ(required_manifest, manifest_cnt);
|
||||
}
|
||||
|
||||
static void DoSleep(void* arg) {
|
||||
auto test = reinterpret_cast<DeleteFileTest*>(arg);
|
||||
test->env_->SleepForMicroseconds(2 * 1000 * 1000);
|
||||
}
|
||||
|
||||
// An empty job to guard all jobs are processed
|
||||
static void GuardFinish(void* arg) {
|
||||
TEST_SYNC_POINT("DeleteFileTest::GuardFinish");
|
||||
}
|
||||
};
|
||||
|
||||
TEST_F(DeleteFileTest, AddKeysAndQueryLevels) {
|
||||
@@ -241,118 +229,6 @@ TEST_F(DeleteFileTest, PurgeObsoleteFilesTest) {
|
||||
CloseDB();
|
||||
}
|
||||
|
||||
TEST_F(DeleteFileTest, BackgroundPurgeTest) {
|
||||
std::string first("0"), last("999999");
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 2;
|
||||
Slice first_slice(first), last_slice(last);
|
||||
|
||||
// We keep an iterator alive
|
||||
Iterator* itr = 0;
|
||||
CreateTwoLevels();
|
||||
ReadOptions options;
|
||||
options.background_purge_on_iterator_cleanup = true;
|
||||
itr = db_->NewIterator(options);
|
||||
db_->CompactRange(compact_options, &first_slice, &last_slice);
|
||||
// 3 sst after compaction with live iterator
|
||||
CheckFileTypeCounts(dbname_, 0, 3, 1);
|
||||
test::SleepingBackgroundTask sleeping_task_before;
|
||||
env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask,
|
||||
&sleeping_task_before, Env::Priority::HIGH);
|
||||
delete itr;
|
||||
test::SleepingBackgroundTask sleeping_task_after;
|
||||
env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask,
|
||||
&sleeping_task_after, Env::Priority::HIGH);
|
||||
|
||||
// Make sure no purges are executed foreground
|
||||
CheckFileTypeCounts(dbname_, 0, 3, 1);
|
||||
sleeping_task_before.WakeUp();
|
||||
sleeping_task_before.WaitUntilDone();
|
||||
|
||||
// Make sure all background purges are executed
|
||||
sleeping_task_after.WakeUp();
|
||||
sleeping_task_after.WaitUntilDone();
|
||||
// 1 sst after iterator deletion
|
||||
CheckFileTypeCounts(dbname_, 0, 1, 1);
|
||||
|
||||
CloseDB();
|
||||
}
|
||||
|
||||
// This test is to reproduce a bug that read invalid ReadOption in iterator
|
||||
// cleanup function
|
||||
TEST_F(DeleteFileTest, BackgroundPurgeCopyOptions) {
|
||||
std::string first("0"), last("999999");
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 2;
|
||||
Slice first_slice(first), last_slice(last);
|
||||
|
||||
// We keep an iterator alive
|
||||
Iterator* itr = 0;
|
||||
CreateTwoLevels();
|
||||
ReadOptions* options = new ReadOptions();
|
||||
options->background_purge_on_iterator_cleanup = true;
|
||||
itr = db_->NewIterator(*options);
|
||||
// ReadOptions is deleted, but iterator cleanup function should not be
|
||||
// affected
|
||||
delete options;
|
||||
|
||||
db_->CompactRange(compact_options, &first_slice, &last_slice);
|
||||
// 3 sst after compaction with live iterator
|
||||
CheckFileTypeCounts(dbname_, 0, 3, 1);
|
||||
delete itr;
|
||||
|
||||
test::SleepingBackgroundTask sleeping_task_after;
|
||||
env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask,
|
||||
&sleeping_task_after, Env::Priority::HIGH);
|
||||
|
||||
// Make sure all background purges are executed
|
||||
sleeping_task_after.WakeUp();
|
||||
sleeping_task_after.WaitUntilDone();
|
||||
// 1 sst after iterator deletion
|
||||
CheckFileTypeCounts(dbname_, 0, 1, 1);
|
||||
|
||||
CloseDB();
|
||||
}
|
||||
|
||||
TEST_F(DeleteFileTest, BackgroundPurgeTestMultipleJobs) {
|
||||
std::string first("0"), last("999999");
|
||||
CompactRangeOptions compact_options;
|
||||
compact_options.change_level = true;
|
||||
compact_options.target_level = 2;
|
||||
Slice first_slice(first), last_slice(last);
|
||||
|
||||
// We keep an iterator alive
|
||||
CreateTwoLevels();
|
||||
ReadOptions options;
|
||||
options.background_purge_on_iterator_cleanup = true;
|
||||
Iterator* itr1 = db_->NewIterator(options);
|
||||
CreateTwoLevels();
|
||||
Iterator* itr2 = db_->NewIterator(options);
|
||||
db_->CompactRange(compact_options, &first_slice, &last_slice);
|
||||
// 5 sst files after 2 compactions with 2 live iterators
|
||||
CheckFileTypeCounts(dbname_, 0, 5, 1);
|
||||
|
||||
// ~DBImpl should wait until all BGWorkPurge are finished
|
||||
rocksdb::SyncPoint::GetInstance()->LoadDependency(
|
||||
{{"DBImpl::~DBImpl:WaitJob", "DBImpl::BGWorkPurge"},
|
||||
{"DeleteFileTest::GuardFinish",
|
||||
"DeleteFileTest::BackgroundPurgeTestMultipleJobs:DBClose"}});
|
||||
rocksdb::SyncPoint::GetInstance()->EnableProcessing();
|
||||
|
||||
delete itr1;
|
||||
env_->Schedule(&DeleteFileTest::DoSleep, this, Env::Priority::HIGH);
|
||||
delete itr2;
|
||||
env_->Schedule(&DeleteFileTest::GuardFinish, nullptr, Env::Priority::HIGH);
|
||||
CloseDB();
|
||||
|
||||
TEST_SYNC_POINT("DeleteFileTest::BackgroundPurgeTestMultipleJobs:DBClose");
|
||||
// 1 sst after iterator deletion
|
||||
CheckFileTypeCounts(dbname_, 0, 1, 1);
|
||||
rocksdb::SyncPoint::GetInstance()->DisableProcessing();
|
||||
}
|
||||
|
||||
TEST_F(DeleteFileTest, DeleteFileWithIterator) {
|
||||
CreateTwoLevels();
|
||||
ReadOptions options;
|
||||
|
||||
+40
-69
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -8,10 +8,7 @@
|
||||
namespace rocksdb {
|
||||
|
||||
namespace {
|
||||
template<class T>
|
||||
inline T SafeDivide(T a, T b) {
|
||||
return b == 0 ? 0 : a / b;
|
||||
}
|
||||
inline double SafeDivide(double a, double b) { return b == 0.0 ? 0 : a / b; }
|
||||
} // namespace
|
||||
|
||||
void EventHelpers::AppendCurrentTime(JSONWriter* jwriter) {
|
||||
@@ -20,83 +17,57 @@ void EventHelpers::AppendCurrentTime(JSONWriter* jwriter) {
|
||||
std::chrono::system_clock::now().time_since_epoch()).count();
|
||||
}
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
void EventHelpers::NotifyTableFileCreationStarted(
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners,
|
||||
const std::string& db_name, const std::string& cf_name,
|
||||
const std::string& file_path, int job_id, TableFileCreationReason reason) {
|
||||
TableFileCreationBriefInfo info;
|
||||
info.db_name = db_name;
|
||||
info.cf_name = cf_name;
|
||||
info.file_path = file_path;
|
||||
info.job_id = job_id;
|
||||
info.reason = reason;
|
||||
for (auto& listener : listeners) {
|
||||
listener->OnTableFileCreationStarted(info);
|
||||
}
|
||||
}
|
||||
#endif // !ROCKSDB_LITE
|
||||
|
||||
void EventHelpers::LogAndNotifyTableFileCreationFinished(
|
||||
void EventHelpers::LogAndNotifyTableFileCreation(
|
||||
EventLogger* event_logger,
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners,
|
||||
const std::string& db_name, const std::string& cf_name,
|
||||
const std::string& file_path, int job_id, const FileDescriptor& fd,
|
||||
const TableProperties& table_properties, TableFileCreationReason reason,
|
||||
const Status& s) {
|
||||
if (s.ok() && event_logger) {
|
||||
JSONWriter jwriter;
|
||||
AppendCurrentTime(&jwriter);
|
||||
jwriter << "cf_name" << cf_name << "job" << job_id << "event"
|
||||
<< "table_file_creation"
|
||||
<< "file_number" << fd.GetNumber() << "file_size"
|
||||
<< fd.GetFileSize();
|
||||
const FileDescriptor& fd, const TableFileCreationInfo& info) {
|
||||
assert(event_logger);
|
||||
JSONWriter jwriter;
|
||||
AppendCurrentTime(&jwriter);
|
||||
jwriter << "cf_name" << info.cf_name
|
||||
<< "job" << info.job_id
|
||||
<< "event" << "table_file_creation"
|
||||
<< "file_number" << fd.GetNumber()
|
||||
<< "file_size" << fd.GetFileSize();
|
||||
|
||||
// table_properties
|
||||
{
|
||||
jwriter << "table_properties";
|
||||
jwriter.StartObject();
|
||||
// table_properties
|
||||
{
|
||||
jwriter << "table_properties";
|
||||
jwriter.StartObject();
|
||||
|
||||
// basic properties:
|
||||
jwriter << "data_size" << table_properties.data_size << "index_size"
|
||||
<< table_properties.index_size << "filter_size"
|
||||
<< table_properties.filter_size << "raw_key_size"
|
||||
<< table_properties.raw_key_size << "raw_average_key_size"
|
||||
<< SafeDivide(table_properties.raw_key_size,
|
||||
table_properties.num_entries)
|
||||
<< "raw_value_size" << table_properties.raw_value_size
|
||||
<< "raw_average_value_size"
|
||||
<< SafeDivide(table_properties.raw_value_size,
|
||||
table_properties.num_entries)
|
||||
<< "num_data_blocks" << table_properties.num_data_blocks
|
||||
<< "num_entries" << table_properties.num_entries
|
||||
<< "filter_policy_name" << table_properties.filter_policy_name;
|
||||
// basic properties:
|
||||
jwriter << "data_size" << info.table_properties.data_size
|
||||
<< "index_size" << info.table_properties.index_size
|
||||
<< "filter_size" << info.table_properties.filter_size
|
||||
<< "raw_key_size" << info.table_properties.raw_key_size
|
||||
<< "raw_average_key_size" << SafeDivide(
|
||||
info.table_properties.raw_key_size,
|
||||
info.table_properties.num_entries)
|
||||
<< "raw_value_size" << info.table_properties.raw_value_size
|
||||
<< "raw_average_value_size" << SafeDivide(
|
||||
info.table_properties.raw_value_size,
|
||||
info.table_properties.num_entries)
|
||||
<< "num_data_blocks" << info.table_properties.num_data_blocks
|
||||
<< "num_entries" << info.table_properties.num_entries
|
||||
<< "filter_policy_name" <<
|
||||
info.table_properties.filter_policy_name;
|
||||
|
||||
// user collected properties
|
||||
for (const auto& prop : table_properties.readable_properties) {
|
||||
jwriter << prop.first << prop.second;
|
||||
}
|
||||
jwriter.EndObject();
|
||||
// user collected properties
|
||||
for (const auto& prop : info.table_properties.user_collected_properties) {
|
||||
jwriter << prop.first << prop.second;
|
||||
}
|
||||
jwriter.EndObject();
|
||||
|
||||
event_logger->Log(jwriter);
|
||||
}
|
||||
jwriter.EndObject();
|
||||
|
||||
event_logger->Log(jwriter);
|
||||
|
||||
#ifndef ROCKSDB_LITE
|
||||
if (listeners.size() == 0) {
|
||||
return;
|
||||
}
|
||||
TableFileCreationInfo info;
|
||||
info.db_name = db_name;
|
||||
info.cf_name = cf_name;
|
||||
info.file_path = file_path;
|
||||
info.file_size = fd.file_size;
|
||||
info.job_id = job_id;
|
||||
info.table_properties = table_properties;
|
||||
info.reason = reason;
|
||||
info.status = s;
|
||||
for (auto& listener : listeners) {
|
||||
|
||||
for (auto listener : listeners) {
|
||||
listener->OnTableFileCreated(info);
|
||||
}
|
||||
#endif // !ROCKSDB_LITE
|
||||
|
||||
+3
-18
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2013, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
@@ -19,30 +19,15 @@ namespace rocksdb {
|
||||
class EventHelpers {
|
||||
public:
|
||||
static void AppendCurrentTime(JSONWriter* json_writer);
|
||||
#ifndef ROCKSDB_LITE
|
||||
static void NotifyTableFileCreationStarted(
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners,
|
||||
const std::string& db_name, const std::string& cf_name,
|
||||
const std::string& file_path, int job_id, TableFileCreationReason reason);
|
||||
#endif // !ROCKSDB_LITE
|
||||
static void LogAndNotifyTableFileCreationFinished(
|
||||
static void LogAndNotifyTableFileCreation(
|
||||
EventLogger* event_logger,
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners,
|
||||
const std::string& db_name, const std::string& cf_name,
|
||||
const std::string& file_path, int job_id, const FileDescriptor& fd,
|
||||
const TableProperties& table_properties, TableFileCreationReason reason,
|
||||
const Status& s);
|
||||
const FileDescriptor& fd, const TableFileCreationInfo& info);
|
||||
static void LogAndNotifyTableFileDeletion(
|
||||
EventLogger* event_logger, int job_id,
|
||||
uint64_t file_number, const std::string& file_path,
|
||||
const Status& status, const std::string& db_name,
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners);
|
||||
|
||||
private:
|
||||
static void LogAndNotifyTableFileCreation(
|
||||
EventLogger* event_logger,
|
||||
const std::vector<std::shared_ptr<EventListener>>& listeners,
|
||||
const FileDescriptor& fd, const TableFileCreationInfo& info);
|
||||
};
|
||||
|
||||
} // namespace rocksdb
|
||||
+1
-1
@@ -1,4 +1,4 @@
|
||||
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
||||
// Copyright (c) 2014, Facebook, Inc. All rights reserved.
|
||||
// This source code is licensed under the BSD-style license found in the
|
||||
// LICENSE file in the root directory of this source tree. An additional grant
|
||||
// of patent rights can be found in the PATENTS file in the same directory.
|
||||
|
||||
Loaded 100 of 994 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user