Bumping version to 5.8

Summary: Closes https://github.com/facebook/rocksdb/pull/2738

Differential Revision: D5736261

Pulled By: maysamyabandeh

fbshipit-source-id: 49d27e9ccd786c4056a3d586a060fe460ea883ac

.arcconfig

@@ -1,17 +0,0 @@
-{
-  "project_id" : "rocksdb",
-  "conduit_uri" : "https://phabricator.fb.com/api/",
-  "copyright_holder" : "Facebook",
-  "load" : [
-    "arcanist_util"
-  ],
-  "lint.engine" : "FacebookFbcodeLintEngine",
-  "lint.engine.single.linter" : "FbcodeCppLinter",
-  "unit.engine" : "FacebookFbcodeUnitTestEngine",
-  "arcanist_configuration" : "FacebookArcanistConfiguration",
-  "base" : "git:HEAD^, hg:.^",
-  "git.default-relative-commit" : "HEAD^",
-  "git:arc.feature.start.default" : "origin/master",
-  "arc.feature.start.default" : "master",
-  "history.immutable" : false
-}

.gitignore

@@ -1,4 +1,3 @@
-TARGETS
 make_config.mk
 
 *.a
@@ -32,6 +31,7 @@ build/
 ldb
 manifest_dump
 sst_dump
+blob_dump
 column_aware_encoding_exp
 util/build_version.cc
 build_tools/VALGRIND_LOGS/
@@ -39,9 +39,9 @@ coverage/COVERAGE_REPORT
 .gdbhistory
 .gdb_history
 package/
-.phutil_module_cache
 unity.a
 tags
+etags
 rocksdb_dump
 rocksdb_undump
 db_test2
@@ -60,7 +60,7 @@ rocksdb.h
 unity.cc
 java/crossbuild/.vagrant
 .vagrant/
-java/**.asc
+java/**/*.asc
 java/javadoc
 
 scan_build_report/
@@ -71,3 +71,4 @@ db_logs/
 tp2/
 fbcode/
 fbcode
+buckifier/*.pyc

.travis.yml

@@ -7,6 +7,7 @@ os:
 compiler:
   - clang
   - gcc
+osx_image: xcode8.3
 jdk:
   - oraclejdk7
 cache:
@@ -15,44 +16,55 @@ cache:
 
 addons:
    apt:
-      sources: ['ubuntu-toolchain-r-test', 'llvm-toolchain-precise-3.6']
-      packages: ['clang-3.6' , 'g++-6', 'zlib1g-dev', 'libbz2-dev', 'libsnappy-dev', 'curl', 'libgflags-dev']
+      packages: ['zlib1g-dev', 'libbz2-dev', 'libsnappy-dev', 'curl', 'libgflags-dev', 'mingw-w64']
 env:
-  - TEST_GROUP=platform_dependent1
-  - TEST_GROUP=platform_dependent2
-  - TEST_GROUP=platform_dependent3
-  - TEST_GROUP=1
-  - TEST_GROUP=2
+  - TEST_GROUP=platform_dependent # 16-18 minutes
+  - TEST_GROUP=1 # 33-35 minutes
+  - TEST_GROUP=2 # 30-32 minutes
   # Run java tests
-  - JOB_NAME=java_test
+  - JOB_NAME=java_test # 4-11 minutes
   # Build ROCKSDB_LITE
-  - JOB_NAME=lite_build
+  - JOB_NAME=lite_build # 3-4 minutes
+  # Build examples
+  - JOB_NAME=examples # 5-7 minutes
+  - JOB_NAME=cmake # 3-5 minutes
+  - JOB_NAME=cmake-mingw # 3 minutes
 
 matrix:
   exclude:
-  - os: osx
-    compiler: gcc
   - os: osx
     env: TEST_GROUP=1
   - os: osx
     env: TEST_GROUP=2
+  - os : osx
+    env: JOB_NAME=cmake-mingw
+  - os : linux
+    compiler: clang
+  - os : osx
+    compiler: gcc
+
+# https://docs.travis-ci.com/user/caching/#ccache-cache
+install:
+  - if [ "${TRAVIS_OS_NAME}" == osx ]; then
+      brew install ccache;
+      PATH=$PATH:/usr/local/opt/ccache/libexec;
+    fi
 
 before_script:
-  - if [[ "${TRAVIS_OS_NAME}" == 'linux' && "${CXX}" == 'clang++' ]]; then CXX=clang++-3.6; fi
-  # test one linux g++ build with g++-6
-  - if [[ "${TRAVIS_OS_NAME}" == 'linux' && "${CXX}" == 'g++' && "${JOB_NAME}" == 'unittests' ]]; then CXX=g++-6; fi
-  # Limit the maximum number of open file descriptors to 8192
-  - ulimit -n 8192 || true
+  # Increase the maximum number of open file descriptors, since some tests use
+  # more FDs than the default limit.
+  - ulimit -n 8192
 
 script:
   - ${CXX} --version
-  - if [[ "${TEST_GROUP}" == 'platform_dependent1' ]]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_END=external_sst_file_basic_test make -j4 check_some; fi
-  - if [[ "${TEST_GROUP}" == 'platform_dependent2' ]]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=external_sst_file_basic_test ROCKSDBTESTS_END=checkpoint_test make -j4 check_some; fi
-  - if [[ "${TEST_GROUP}" == 'platform_dependent3' ]]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=checkpoint_test ROCKSDBTESTS_END=db_block_cache_test make -j4 check_some; fi
-  - if [[ "${TEST_GROUP}" == '1' ]]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=db_block_cache_test ROCKSDBTESTS_END=comparator_db_test make -j4 check_some; fi
-  - if [[ "${TEST_GROUP}" == '2' ]]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=comparator_db_test 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 [ "${TEST_GROUP}" == 'platform_dependent' ]; then ccache -C && OPT=-DTRAVIS V=1 ROCKSDBTESTS_END=db_block_cache_test make -j4 all_but_some_tests check_some; fi
+  - if [ "${TEST_GROUP}" == '1' ]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=db_block_cache_test ROCKSDBTESTS_END=comparator_db_test make -j4 check_some; fi
+  - if [ "${TEST_GROUP}" == '2' ]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=comparator_db_test make -j4 check_some; fi
+  - if [ "${JOB_NAME}" == 'java_test' ]; then OPT=-DTRAVIS V=1 make clean jclean && make rocksdbjava jtest; fi
+  - if [ "${JOB_NAME}" == 'lite_build' ]; then OPT="-DTRAVIS -DROCKSDB_LITE" V=1 make -j4 static_lib; fi
+  - if [ "${JOB_NAME}" == 'examples' ]; then OPT=-DTRAVIS V=1 make -j4 static_lib; cd examples; make -j4; fi
+  - if [ "${JOB_NAME}" == 'cmake' ]; then mkdir build && cd build && cmake .. && make -j4 rocksdb; fi
+  - if [ "${JOB_NAME}" == 'cmake-mingw' ]; then mkdir build && cd build && cmake .. -DCMAKE_C_COMPILER=x86_64-w64-mingw32-gcc -DCMAKE_CXX_COMPILER=x86_64-w64-mingw32-g++ -DCMAKE_SYSTEM_NAME=Windows && make -j4 rocksdb; fi
 notifications:
     email:
       - leveldb@fb.com

CMakeLists.txt

@@ -41,20 +41,22 @@ endif()
 
 list(APPEND CMAKE_MODULE_PATH "${CMAKE_SOURCE_DIR}/cmake/modules/")
 
-if(WIN32)
+if(MSVC)
   include(${CMAKE_CURRENT_SOURCE_DIR}/thirdparty.inc)
 else()
   option(WITH_JEMALLOC "build with JeMalloc" OFF)
-  if(WITH_JEMALLOC)
-    find_package(JeMalloc REQUIRED)
-    add_definitions(-DROCKSDB_JEMALLOC)
-    include_directories(${JEMALLOC_INCLUDE_DIR})
-  endif()
-  if(CMAKE_SYSTEM_NAME STREQUAL "FreeBSD")
+  if(CMAKE_SYSTEM_NAME MATCHES "FreeBSD")
     # FreeBSD has jemaloc as default malloc
-    add_definitions(-DROCKSDB_JEMALLOC)
+    # but it does not have all the jemalloc files in include/...
     set(WITH_JEMALLOC ON)
+  else()
+    if(WITH_JEMALLOC)
+      find_package(JeMalloc REQUIRED)
+      add_definitions(-DROCKSDB_JEMALLOC -DJEMALLOC_NO_DEMANGLE)
+      include_directories(${JEMALLOC_INCLUDE_DIR})
+    endif()
   endif()
+
   option(WITH_SNAPPY "build with SNAPPY" OFF)
   if(WITH_SNAPPY)
     find_package(snappy REQUIRED)
@@ -62,11 +64,43 @@ else()
     include_directories(${SNAPPY_INCLUDE_DIR})
     list(APPEND THIRDPARTY_LIBS ${SNAPPY_LIBRARIES})
   endif()
+
+  option(WITH_ZLIB "build with zlib" OFF)
+  if(WITH_ZLIB)
+    find_package(zlib REQUIRED)
+    add_definitions(-DZLIB)
+    include_directories(${ZLIB_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${ZLIB_LIBRARIES})
+  endif()
+
+  option(WITH_BZ2 "build with bzip2" OFF)
+  if(WITH_BZ2)
+    find_package(bzip2 REQUIRED)
+    add_definitions(-DBZIP2)
+    include_directories(${BZIP2_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${BZIP2_LIBRARIES})
+  endif()
+
+  option(WITH_LZ4 "build with lz4" OFF)
+  if(WITH_LZ4)
+    find_package(lz4 REQUIRED)
+    add_definitions(-DLZ4)
+    include_directories(${LZ4_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${LZ4_LIBRARIES})
+  endif()
+
+  option(WITH_ZSTD "build with zstd" OFF)
+  if(WITH_ZSTD)
+    find_package(zstd REQUIRED)
+    add_definitions(-DZSTD)
+    include_directories(${ZSTD_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${ZSTD_LIBRARIES})
+  endif()
 endif()
 
 if(WIN32)
-  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)
+  execute_process(COMMAND powershell -noprofile -Command "Get-Date -format MM_dd_yyyy" OUTPUT_VARIABLE DATE)
+  execute_process(COMMAND powershell -noprofile -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}")
   set(GIT_DATE_TIME "${DATE} ${TIME}")
@@ -78,7 +112,7 @@ endif()
 
 find_package(Git)
 
-if (GIT_FOUND AND EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
+if(GIT_FOUND AND EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
   if(WIN32)
     execute_process(COMMAND $ENV{COMSPEC} /C ${GIT_EXECUTABLE} -C ${CMAKE_CURRENT_SOURCE_DIR} rev-parse HEAD OUTPUT_VARIABLE GIT_SHA)
   else()
@@ -90,7 +124,7 @@ endif()
 
 string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA "${GIT_SHA}")
 
-if (NOT WIN32)
+if(NOT WIN32)
   execute_process(COMMAND
       "./build_tools/version.sh" "full"
       WORKING_DIRECTORY ${PROJECT_SOURCE_DIR}
@@ -105,21 +139,32 @@ if (NOT WIN32)
   string(STRIP "${ROCKSDB_VERSION_MAJOR}" ROCKSDB_VERSION_MAJOR)
 endif()
 
+option(WITH_MD_LIBRARY "build with MD" ON)
+if(WIN32 AND MSVC)
+  if(WITH_MD_LIBRARY)
+    set(RUNTIME_LIBRARY "MD")
+  else()
+    set(RUNTIME_LIBRARY "MT")
+  endif()
+endif()
+
 set(BUILD_VERSION_CC ${CMAKE_BINARY_DIR}/build_version.cc)
 configure_file(util/build_version.cc.in ${BUILD_VERSION_CC} @ONLY)
 add_library(build_version OBJECT ${BUILD_VERSION_CC})
 target_include_directories(build_version PRIVATE
   ${CMAKE_CURRENT_SOURCE_DIR}/util)
-if(WIN32)
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo  /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
-
+if(MSVC)
+  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 /wd4127 /wd4800 /wd4996 /wd4351")
 else()
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -W -Wextra -Wall")
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wsign-compare -Wshadow -Wno-unused-parameter -Wno-unused-variable -Woverloaded-virtual -Wnon-virtual-dtor -Wno-missing-field-initializers")
+  if(MINGW)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-format")
+  endif()
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
   if(NOT CMAKE_BUILD_TYPE STREQUAL "Debug")
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-omit-frame-pointer")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O2 -fno-omit-frame-pointer")
     include(CheckCXXCompilerFlag)
     CHECK_CXX_COMPILER_FLAG("-momit-leaf-frame-pointer" HAVE_OMIT_LEAF_FRAME_POINTER)
     if(HAVE_OMIT_LEAF_FRAME_POINTER)
@@ -128,6 +173,49 @@ else()
   endif()
 endif()
 
+option(PORTABLE "build a portable binary" OFF)
+option(FORCE_SSE42 "force building with SSE4.2, even when PORTABLE=ON" OFF)
+if(PORTABLE)
+  # MSVC does not need a separate compiler flag to enable SSE4.2; if nmmintrin.h
+  # is available, it is available by default.
+  if(FORCE_SSE42 AND NOT MSVC)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse4.2")
+  endif()
+else()
+  if(MSVC)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
+  else()
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native")
+  endif()
+endif()
+
+set(CMAKE_REQUIRED_FLAGS ${CMAKE_CXX_FLAGS})
+include(CheckCXXSourceCompiles)
+CHECK_CXX_SOURCE_COMPILES("
+#include <cstdint>
+#include <nmmintrin.h>
+int main() {
+  volatile uint32_t x = _mm_crc32_u32(0, 0);
+}
+" HAVE_SSE42)
+if(HAVE_SSE42)
+  add_definitions(-DHAVE_SSE42)
+elseif(FORCE_SSE42)
+  message(FATAL_ERROR "FORCE_SSE42=ON but unable to compile with SSE4.2 enabled")
+endif()
+
+CHECK_CXX_SOURCE_COMPILES("
+#if defined(_MSC_VER) && !defined(__thread)
+#define __thread __declspec(thread)
+#endif
+int main() {
+  static __thread int tls;
+}
+" HAVE_THREAD_LOCAL)
+if(HAVE_THREAD_LOCAL)
+  add_definitions(-DROCKSDB_SUPPORT_THREAD_LOCAL)
+endif()
+
 option(FAIL_ON_WARNINGS "Treat compile warnings as errors" ON)
 if(FAIL_ON_WARNINGS)
   if(MSVC)
@@ -139,7 +227,6 @@ endif()
 
 option(WITH_ASAN "build with ASAN" OFF)
 if(WITH_ASAN)
-  add_definitions(-DROCKSDB_TSAN_RUN)
   set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -fsanitize=address")
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsanitize=address")
   set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fsanitize=address")
@@ -151,7 +238,6 @@ endif()
 option(WITH_TSAN "build with TSAN" OFF)
 if(WITH_TSAN)
   set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -fsanitize=thread -pie")
-  add_definitions(-DROCKSDB_TSAN_RUN)
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsanitize=thread -fPIC")
   set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fsanitize=thread -fPIC")
   if(WITH_JEMALLOC)
@@ -161,6 +247,7 @@ endif()
 
 option(WITH_UBSAN "build with UBSAN" OFF)
 if(WITH_UBSAN)
+  add_definitions(-DROCKSDB_UBSAN_RUN)
   set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -fsanitize=undefined")
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsanitize=undefined")
   set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fsanitize=undefined")
@@ -178,14 +265,14 @@ else()
    set(OPTIMIZE_DEBUG ${OPTIMIZE_DEBUG_DEFAULT})
 endif()
 
-if(WIN32)
+if(MSVC)
   if((${OPTIMIZE_DEBUG} EQUAL 1))
     message(STATUS "Debug optimization is enabled")
-    set(CMAKE_CXX_FLAGS_DEBUG "/Oxt /MDd")
+    set(CMAKE_CXX_FLAGS_DEBUG "/Oxt /${RUNTIME_LIBRARY}d")
   else()
-    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /RTC1 /Gm /MDd")
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /RTC1 /Gm /${RUNTIME_LIBRARY}d")
   endif()
-  set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oxt /Zp8 /Gm- /Gy /MD")
+  set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oxt /Zp8 /Gm- /Gy /${RUNTIME_LIBRARY}")
 
   set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} /DEBUG")
   set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /DEBUG")
@@ -226,6 +313,9 @@ elseif(CMAKE_SYSTEM_NAME MATCHES "Android")
   add_definitions(-DOS_ANDROID)
 elseif(CMAKE_SYSTEM_NAME MATCHES "Windows")
   add_definitions(-DWIN32 -DOS_WIN -D_MBCS -DWIN64 -DNOMINMAX)
+  if(MINGW)
+    add_definitions(-D_WIN32_WINNT=_WIN32_WINNT_VISTA)
+  endif()
 endif()
 
 if(NOT WIN32)
@@ -235,6 +325,7 @@ endif()
 option(WITH_FALLOCATE "build with fallocate" ON)
 
 if(WITH_FALLOCATE)
+  set(CMAKE_REQUIRED_FLAGS ${CMAKE_C_FLAGS})
   include(CheckCSourceCompiles)
   CHECK_C_SOURCE_COMPILES("
 #include <fcntl.h>
@@ -265,7 +356,9 @@ add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
 # Main library source code
 
 set(SOURCES
-        db/auto_roll_logger.cc
+        cache/clock_cache.cc
+        cache/lru_cache.cc
+        cache/sharded_cache.cc
         db/builder.cc
         db/c.cc
         db/column_family.cc
@@ -274,19 +367,23 @@ set(SOURCES
         db/compaction_iterator.cc
         db/compaction_job.cc
         db/compaction_picker.cc
+        db/compaction_picker_universal.cc
         db/convenience.cc
-        db/dbformat.cc
         db/db_filesnapshot.cc
         db/db_impl.cc
+        db/db_impl_write.cc
+        db/db_impl_compaction_flush.cc
+        db/db_impl_files.cc
+        db/db_impl_open.cc
         db/db_impl_debug.cc
         db/db_impl_experimental.cc
         db/db_impl_readonly.cc
         db/db_info_dumper.cc
         db/db_iter.cc
+        db/dbformat.cc
         db/event_helpers.cc
-        db/external_sst_file_ingestion_job.cc
         db/experimental.cc
-        db/filename.cc
+        db/external_sst_file_ingestion_job.cc
         db/file_indexer.cc
         db/flush_job.cc
         db/flush_scheduler.cc
@@ -294,9 +391,9 @@ set(SOURCES
         db/internal_stats.cc
         db/log_reader.cc
         db/log_writer.cc
+        db/malloc_stats.cc
         db/managed_iterator.cc
         db/memtable.cc
-        db/memtable_allocator.cc
         db/memtable_list.cc
         db/merge_helper.cc
         db/merge_operator.cc
@@ -314,11 +411,35 @@ set(SOURCES
         db/write_batch_base.cc
         db/write_controller.cc
         db/write_thread.cc
+        env/env.cc
+        env/env_chroot.cc
+        env/env_encryption.cc
+        env/env_hdfs.cc
+        env/mock_env.cc
+        memtable/alloc_tracker.cc
         memtable/hash_cuckoo_rep.cc
         memtable/hash_linklist_rep.cc
         memtable/hash_skiplist_rep.cc
         memtable/skiplistrep.cc
         memtable/vectorrep.cc
+        memtable/write_buffer_manager.cc
+        monitoring/histogram.cc
+        monitoring/histogram_windowing.cc
+        monitoring/instrumented_mutex.cc
+        monitoring/iostats_context.cc
+        monitoring/perf_context.cc
+        monitoring/perf_level.cc
+        monitoring/statistics.cc
+        monitoring/thread_status_impl.cc
+        monitoring/thread_status_updater.cc
+        monitoring/thread_status_util.cc
+        monitoring/thread_status_util_debug.cc
+        options/cf_options.cc
+        options/db_options.cc
+        options/options.cc
+        options/options_helper.cc
+        options/options_parser.cc
+        options/options_sanity_check.cc
         port/stack_trace.cc
         table/adaptive_table_factory.cc
         table/block.cc
@@ -339,64 +460,44 @@ set(SOURCES
         table/index_builder.cc
         table/iterator.cc
         table/merging_iterator.cc
-        table/sst_file_writer.cc
         table/meta_blocks.cc
         table/partitioned_filter_block.cc
+        table/persistent_cache_helper.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/sst_file_writer.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
+        tools/ldb_cmd.cc
+        tools/ldb_tool.cc
+        tools/sst_dump_tool.cc
         util/arena.cc
+        util/auto_roll_logger.cc
         util/bloom.cc
-        util/cf_options.cc
-        util/clock_cache.cc
         util/coding.cc
         util/compaction_job_stats_impl.cc
         util/comparator.cc
         util/concurrent_arena.cc
         util/crc32c.cc
-        util/db_options.cc
         util/delete_scheduler.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/file_util.cc
+        util/filename.cc
         util/filter_policy.cc
         util/hash.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/logging.cc
         util/log_buffer.cc
-        util/memenv.cc
         util/murmurhash.cc
-        util/options.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/slice.cc
-        util/statistics.cc
+        util/sst_file_manager_impl.cc
         util/status.cc
         util/status_message.cc
         util/string_util.cc
@@ -404,29 +505,40 @@ set(SOURCES
         util/testutil.cc
         util/thread_local.cc
         util/threadpool_imp.cc
-        util/thread_status_impl.cc
-        util/thread_status_updater.cc
-        util/thread_status_util.cc
-        util/thread_status_util_debug.cc
         util/transaction_test_util.cc
         util/xxhash.cc
         utilities/backupable/backupable_db.cc
         utilities/blob_db/blob_db.cc
-        utilities/checkpoint/checkpoint.cc
+        utilities/blob_db/blob_db_impl.cc
+        utilities/blob_db/blob_dump_tool.cc
+        utilities/blob_db/blob_file.cc
+        utilities/blob_db/blob_log_reader.cc
+        utilities/blob_db/blob_log_writer.cc
+        utilities/blob_db/blob_log_format.cc
+        utilities/blob_db/ttl_extractor.cc
+        utilities/cassandra/cassandra_compaction_filter.cc
+        utilities/cassandra/format.cc
+        utilities/cassandra/merge_operator.cc
+        utilities/checkpoint/checkpoint_impl.cc
+        utilities/col_buf_decoder.cc
+        utilities/col_buf_encoder.cc
+        utilities/column_aware_encoding_util.cc
         utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc
         utilities/date_tiered/date_tiered_db_impl.cc
+        utilities/debug.cc
         utilities/document/document_db.cc
         utilities/document/json_document.cc
         utilities/document/json_document_builder.cc
         utilities/env_mirror.cc
+        utilities/env_timed.cc
         utilities/geodb/geodb_impl.cc
         utilities/leveldb_options/leveldb_options.cc
         utilities/lua/rocks_lua_compaction_filter.cc
         utilities/memory/memory_util.cc
+        utilities/merge_operators/max.cc
+        utilities/merge_operators/put.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
@@ -439,20 +551,18 @@ set(SOURCES
         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
         utilities/transactions/optimistic_transaction_db_impl.cc
+        utilities/transactions/optimistic_transaction.cc
         utilities/transactions/transaction_base.cc
-        utilities/transactions/transaction_impl.cc
-        utilities/transactions/transaction_db_impl.cc
+        utilities/transactions/pessimistic_transaction_db.cc
         utilities/transactions/transaction_db_mutex_impl.cc
+        utilities/transactions/pessimistic_transaction.cc
         utilities/transactions/transaction_lock_mgr.cc
         utilities/transactions/transaction_util.cc
+        utilities/transactions/write_prepared_txn.cc
         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
         $<TARGET_OBJECTS:build_version>)
 
 if(WIN32)
@@ -467,20 +577,18 @@ if(WIN32)
 else()
   list(APPEND SOURCES
     port/port_posix.cc
-    util/env_posix.cc
-    util/io_posix.cc)
+    env/env_posix.cc
+    env/io_posix.cc)
 endif()
 
+set(ROCKSDB_STATIC_LIB rocksdb${ARTIFACT_SUFFIX})
+set(ROCKSDB_SHARED_LIB rocksdb-shared${ARTIFACT_SUFFIX})
+set(ROCKSDB_IMPORT_LIB ${ROCKSDB_SHARED_LIB})
 if(WIN32)
   set(SYSTEM_LIBS ${SYSTEM_LIBS} Shlwapi.lib Rpcrt4.lib)
-  set(ROCKSDB_STATIC_LIB rocksdblib${ARTIFACT_SUFFIX})
-  set(ROCKSDB_IMPORT_LIB rocksdb${ARTIFACT_SUFFIX})
   set(LIBS ${ROCKSDB_STATIC_LIB} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
 else()
-  set(SYSTEM_LIBS ${CMAKE_THREAD_LIBS_INIT} rt)
-  set(ROCKSDB_STATIC_LIB rocksdb${ARTIFACT_SUFFIX})
-  set(ROCKSDB_SHARED_LIB rocksdb-shared)
-  set(ROCKSDB_IMPORT_LIB ${ROCKSDB_SHARED_LIB})
+  set(SYSTEM_LIBS ${CMAKE_THREAD_LIBS_INIT})
   set(LIBS ${ROCKSDB_SHARED_LIB} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
 
   add_library(${ROCKSDB_SHARED_LIB} SHARED ${SOURCES})
@@ -505,18 +613,18 @@ add_library(${ROCKSDB_STATIC_LIB} STATIC ${SOURCES})
 target_link_libraries(${ROCKSDB_STATIC_LIB}
   ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
 
-if(WIN32)
-  set_target_properties(${ROCKSDB_STATIC_LIB} PROPERTIES
-    COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/${ROCKSDB_STATIC_LIB}.pdb")
-endif()
-
 if(WIN32)
   add_library(${ROCKSDB_IMPORT_LIB} SHARED ${SOURCES})
   target_link_libraries(${ROCKSDB_IMPORT_LIB}
     ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
   set_target_properties(${ROCKSDB_IMPORT_LIB} PROPERTIES
-    COMPILE_FLAGS "-DROCKSDB_DLL -DROCKSDB_LIBRARY_EXPORTS /Fd${CMAKE_CFG_INTDIR}/${ROCKSDB_IMPORT_LIB}.pdb")
-else()
+    COMPILE_DEFINITIONS "ROCKSDB_DLL;ROCKSDB_LIBRARY_EXPORTS")
+  if(MSVC)
+    set_target_properties(${ROCKSDB_STATIC_LIB} PROPERTIES
+      COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/${ROCKSDB_STATIC_LIB}.pdb")
+    set_target_properties(${ROCKSDB_IMPORT_LIB} PROPERTIES
+      COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/${ROCKSDB_IMPORT_LIB}.pdb")
+  endif()
 endif()
 
 option(WITH_JNI "build with JNI" OFF)
@@ -527,61 +635,128 @@ else()
   message(STATUS "JNI library is disabled")
 endif()
 
-set(TESTS
-        db/auto_roll_logger_test.cc
+# Installation and packaging
+if(WIN32)
+  option(ROCKSDB_INSTALL_ON_WINDOWS "Enable install target on Windows" OFF)
+endif()
+if(NOT WIN32 OR ROCKSDB_INSTALL_ON_WINDOWS)
+  if(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT)
+    if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+      # Change default installation prefix on Linux to /usr
+      set(CMAKE_INSTALL_PREFIX /usr CACHE PATH "Install path prefix, prepended onto install directories." FORCE)
+    endif()
+  endif()
+
+  include(GNUInstallDirs)
+  include(CMakePackageConfigHelpers)
+
+  set(package_config_destination ${CMAKE_INSTALL_LIBDIR}/cmake/rocksdb)
+
+  configure_package_config_file(
+    ${CMAKE_SOURCE_DIR}/cmake/RocksDBConfig.cmake.in RocksDBConfig.cmake
+    INSTALL_DESTINATION ${package_config_destination}
+  )
+
+  write_basic_package_version_file(
+    RocksDBConfigVersion.cmake
+    VERSION ${ROCKSDB_VERSION}
+    COMPATIBILITY SameMajorVersion
+  )
+
+  install(DIRECTORY include/rocksdb COMPONENT devel DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}")
+
+  install(
+    TARGETS ${ROCKSDB_STATIC_LIB}
+    EXPORT RocksDBTargets
+    COMPONENT devel
+    ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
+  )
+
+  install(
+    TARGETS ${ROCKSDB_SHARED_LIB}
+    EXPORT RocksDBTargets
+    COMPONENT runtime
+    RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}"
+    LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
+  )
+
+  install(
+    EXPORT RocksDBTargets
+    COMPONENT devel
+    DESTINATION ${package_config_destination}
+    NAMESPACE RocksDB::
+  )
+
+  install(
+    FILES
+    ${CMAKE_CURRENT_BINARY_DIR}/RocksDBConfig.cmake
+    ${CMAKE_CURRENT_BINARY_DIR}/RocksDBConfigVersion.cmake
+    COMPONENT devel
+    DESTINATION ${package_config_destination}
+  )
+endif()
+
+option(WITH_TESTS "build with tests" ON)
+if(WITH_TESTS)
+  set(TESTS
+        cache/cache_test.cc
+        cache/lru_cache_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_job_test.cc
         db/compaction_picker_test.cc
         db/comparator_db_test.cc
         db/corruption_test.cc
         db/cuckoo_table_db_test.cc
         db/db_basic_test.cc
+        db/db_block_cache_test.cc
+        db/db_bloom_filter_test.cc
         db/db_compaction_filter_test.cc
         db/db_compaction_test.cc
         db/db_dynamic_level_test.cc
+        db/db_flush_test.cc
         db/db_inplace_update_test.cc
+        db/db_io_failure_test.cc
         db/db_iter_test.cc
+        db/db_iterator_test.cc
         db/db_log_iter_test.cc
         db/db_memtable_test.cc
         db/db_merge_operator_test.cc
         db/db_options_test.cc
         db/db_properties_test.cc
+        db/db_range_del_test.cc
+        db/db_sst_test.cc
+        db/db_statistics_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_range_del_test.cc
-        db/db_sst_test.cc
-        db/external_sst_file_basic_test.cc
-        db/external_sst_file_test.cc
         db/db_universal_compaction_test.cc
         db/db_wal_test.cc
+        db/db_write_test.cc
         db/dbformat_test.cc
         db/deletefile_test.cc
+        db/external_sst_file_basic_test.cc
+        db/external_sst_file_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/merge_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
         db/version_edit_test.cc
@@ -590,7 +765,17 @@ set(TESTS
         db/write_batch_test.cc
         db/write_callback_test.cc
         db/write_controller_test.cc
-        db/db_io_failure_test.cc
+        env/env_basic_test.cc
+        env/env_test.cc
+        env/mock_env_test.cc
+        memtable/inlineskiplist_test.cc
+        memtable/skiplist_test.cc
+        memtable/write_buffer_manager_test.cc
+        monitoring/histogram_test.cc
+        monitoring/iostats_context_test.cc
+        monitoring/statistics_test.cc
+        options/options_settable_test.cc
+        options/options_test.cc
         table/block_based_filter_block_test.cc
         table/block_test.cc
         table/cuckoo_table_builder_test.cc
@@ -602,140 +787,133 @@ set(TESTS
         tools/reduce_levels_test.cc
         tools/sst_dump_test.cc
         util/arena_test.cc
+        util/auto_roll_logger_test.cc
         util/autovector_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/filelock_test.cc
+        util/hash_test.cc
         util/heap_test.cc
-        util/histogram_test.cc
-        util/iostats_context_test.cc
-        util/lru_cache_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/timer_queue_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/cassandra/cassandra_functional_test.cc
+        utilities/cassandra/cassandra_format_test.cc
+        utilities/cassandra/cassandra_row_merge_test.cc
+        utilities/cassandra/cassandra_serialize_test.cc
         utilities/checkpoint/checkpoint_test.cc
+        utilities/column_aware_encoding_test.cc
         utilities/date_tiered/date_tiered_test.cc
         utilities/document/document_db_test.cc
         utilities/document/json_document_test.cc
-        utilities/object_registry_test.cc
         utilities/geodb/geodb_test.cc
+        utilities/lua/rocks_lua_test.cc
         utilities/memory/memory_test.cc
         utilities/merge_operators/string_append/stringappend_test.cc
+        utilities/object_registry_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/simulator_cache/sim_cache_test.cc
         utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
         utilities/transactions/optimistic_transaction_test.cc
         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
-        utilities/lua/rocks_lua_test.cc
-)
-if(WITH_LIBRADOS)
-  list(APPEND TESTS utilities/env_librados_test.cc)
-endif()
+  )
+  if(WITH_LIBRADOS)
+    list(APPEND TESTS utilities/env_librados_test.cc)
+  endif()
 
-set(BENCHMARKS
-  tools/db_bench.cc
-  table/table_reader_bench.cc
-  util/cache_bench.cc
-  db/memtablerep_bench.cc
-  utilities/column_aware_encoding_exp.cc
-  utilities/persistent_cache/hash_table_bench.cc)
-add_library(testharness OBJECT util/testharness.cc)
-foreach(sourcefile ${BENCHMARKS})
-  get_filename_component(exename ${sourcefile} NAME_WE)
-  add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile}
-    $<TARGET_OBJECTS:testharness>)
-  target_link_libraries(${exename}${ARTIFACT_SUFFIX} gtest ${LIBS})
-endforeach(sourcefile ${BENCHMARKS})
-
-# 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
-)
-# test utilities are only build in debug
-enable_testing()
-add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND})
-set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
-add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
-if(WIN32)
-  set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
-endif()
-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})
+  set(BENCHMARKS
+    cache/cache_bench.cc
+    memtable/memtablerep_bench.cc
+    tools/db_bench.cc
+    table/table_reader_bench.cc
+    utilities/column_aware_encoding_exp.cc
+    utilities/persistent_cache/hash_table_bench.cc)
+  add_library(testharness OBJECT util/testharness.cc)
+  foreach(sourcefile ${BENCHMARKS})
     get_filename_component(exename ${sourcefile} NAME_WE)
     add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile}
       $<TARGET_OBJECTS:testharness>)
-    set_target_properties(${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}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} gtest ${LIBS})
-    if(NOT "${exename}" MATCHES "db_sanity_test")
+    target_link_libraries(${exename}${ARTIFACT_SUFFIX} gtest ${LIBS})
+  endforeach(sourcefile ${BENCHMARKS})
+
+  # 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
+      monitoring/thread_status_updater_debug.cc
+      table/mock_table.cc
+      util/fault_injection_test_env.cc
+      utilities/cassandra/test_utils.cc
+  )
+  # test utilities are only build in debug
+  enable_testing()
+  add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND})
+  set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
+  add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
+  if(MSVC)
+    set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
+  endif()
+  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})
+      get_filename_component(exename ${sourcefile} NAME_WE)
+      add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile}
+        $<TARGET_OBJECTS:testharness>)
+      set_target_properties(${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}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} gtest ${LIBS})
+      if(NOT "${exename}" MATCHES "db_sanity_test")
+        add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
+        add_dependencies(check ${exename}${ARTIFACT_SUFFIX})
+      endif()
+  endforeach(sourcefile ${TEST_EXES})
+
+  # C executables must link to a shared object
+  set(C_TESTS db/c_test.c)
+  set(C_TEST_EXES ${C_TESTS})
+
+  foreach(sourcefile ${C_TEST_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} ${ROCKSDB_IMPORT_LIB} testutillib${ARTIFACT_SUFFIX})
       add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
       add_dependencies(check ${exename}${ARTIFACT_SUFFIX})
-    endif()
-endforeach(sourcefile ${TEST_EXES})
-
-# C executables must link to a shared object
-set(C_TESTS db/c_test.c)
-set(C_TEST_EXES ${C_TESTS})
-
-foreach(sourcefile ${C_TEST_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} ${ROCKSDB_IMPORT_LIB} testutillib${ARTIFACT_SUFFIX})
-    add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
-    add_dependencies(check ${exename}${ARTIFACT_SUFFIX})
-endforeach(sourcefile ${C_TEST_EXES})
-add_subdirectory(tools)
-
-# Installation and packaging for Linux
-if (NOT WIN32)
-install(TARGETS ${ROCKSDB_STATIC_LIB} COMPONENT devel ARCHIVE DESTINATION lib64)
-install(TARGETS ${ROCKSDB_SHARED_LIB} COMPONENT runtime DESTINATION lib64)
-install(DIRECTORY "${PROJECT_SOURCE_DIR}/include/rocksdb/"
-        COMPONENT devel
-        DESTINATION include/rocksdb)
-set(CMAKE_INSTALL_PREFIX /usr)
+  endforeach(sourcefile ${C_TEST_EXES})
+endif()
+
+option(WITH_TOOLS "build with tools" ON)
+if(WITH_TOOLS)
+  add_subdirectory(tools)
 endif()

DEFAULT_OPTIONS_HISTORY.md

@@ -1,5 +1,8 @@
 # RocksDB default options change log
 ## Unreleased
+* delayed_write_rate takes the rate given by rate_limiter if not specified.
+
+## 5.2
 * Change the default of delayed slowdown value to 16MB/s and further increase the L0 stop condition to 36 files.
 
 ## 5.0 (11/17/2016)

HISTORY.md

@@ -1,6 +1,94 @@
 # Rocksdb Change Log
-## 5.3.7
-* Change license to Apache + GPLv2
+## Unreleased
+### Public API Change
+### New Features
+### Bug Fixes
+
+## 5.8.0 (08/30/2017)
+### Public API Change
+* Users of `Statistics::getHistogramString()` will see fewer histogram buckets and different bucket endpoints.
+* `Slice::compare` and BytewiseComparator `Compare` no longer accept `Slice`s containing nullptr.
+* `Transaction::Get` and `Transaction::GetForUpdate` variants with `PinnableSlice` added.
+
+### New Features
+* Add Iterator::Refresh(), which allows users to update the iterator state so that they can avoid some initialization costs of recreating iterators.
+* Replace dynamic_cast<> (except unit test) so people can choose to build with RTTI off. With make, release mode is by default built with -fno-rtti and debug mode is built without it. Users can override it by setting USE_RTTI=0 or 1.
+* Universal compactions including the bottom level can be executed in a dedicated thread pool. This alleviates head-of-line blocking in the compaction queue, which cause write stalling, particularly in multi-instance use cases. Users can enable this feature via `Env::SetBackgroundThreads(N, Env::Priority::BOTTOM)`, where `N > 0`.
+* Allow merge operator to be called even with a single merge operand during compactions, by appropriately overriding `MergeOperator::AllowSingleOperand`.
+* Add `DB::VerifyChecksum()`, which verifies the checksums in all SST files in a running DB.
+* Block-based table support for disabling checksums by setting `BlockBasedTableOptions::checksum = kNoChecksum`.
+
+### Bug Fixes
+* Fix wrong latencies in `rocksdb.db.get.micros`, `rocksdb.db.write.micros`, and `rocksdb.sst.read.micros`.
+* Fix incorrect dropping of deletions during intra-L0 compaction.
+* Fix transient reappearance of keys covered by range deletions when memtable prefix bloom filter is enabled.
+* Fix potentially wrong file smallest key when range deletions separated by snapshot are written together.
+
+## 5.7.0 (07/13/2017)
+### Public API Change
+* DB property "rocksdb.sstables" now prints keys in hex form.
+
+### New Features
+* Measure estimated number of reads per file. The information can be accessed through DB::GetColumnFamilyMetaData or "rocksdb.sstables" DB property.
+* RateLimiter support for throttling background reads, or throttling the sum of background reads and writes. This can give more predictable I/O usage when compaction reads more data than it writes, e.g., due to lots of deletions.
+* [Experimental] FIFO compaction with TTL support. It can be enabled by setting CompactionOptionsFIFO.ttl > 0.
+* Introduce `EventListener::OnBackgroundError()` callback. Users can implement it to be notified of errors causing the DB to enter read-only mode, and optionally override them.
+* Partitioned Index/Filters exiting the experimental mode. To enable partitioned indexes set index_type to kTwoLevelIndexSearch and to further enable partitioned filters set partition_filters to true. To configure the partition size set metadata_block_size.
+
+
+### Bug Fixes
+* Fix discarding empty compaction output files when `DeleteRange()` is used together with subcompactions.
+
+## 5.6.0 (06/06/2017)
+### Public API Change
+* Scheduling flushes and compactions in the same thread pool is no longer supported by setting `max_background_flushes=0`. Instead, users can achieve this by configuring their high-pri thread pool to have zero threads.
+* Replace `Options::max_background_flushes`, `Options::max_background_compactions`, and `Options::base_background_compactions` all with `Options::max_background_jobs`, which automatically decides how many threads to allocate towards flush/compaction.
+* options.delayed_write_rate by default take the value of options.rate_limiter rate.
+* Replace global variable `IOStatsContext iostats_context` with `IOStatsContext* get_iostats_context()`; replace global variable `PerfContext perf_context` with `PerfContext* get_perf_context()`.
+
+### New Features
+* Change ticker/histogram statistics implementations to use core-local storage. This improves aggregation speed compared to our previous thread-local approach, particularly for applications with many threads.
+* Users can pass a cache object to write buffer manager, so that they can cap memory usage for memtable and block cache using one single limit.
+* Flush will be triggered when 7/8 of the limit introduced by write_buffer_manager or db_write_buffer_size is triggered, so that the hard threshold is hard to hit.
+* Introduce WriteOptions.low_pri. If it is true, low priority writes will be throttled if the compaction is behind.
+* `DB::IngestExternalFile()` now supports ingesting files into a database containing range deletions.
+
+### Bug Fixes
+* Shouldn't ignore return value of fsync() in flush.
+
+## 5.5.0 (05/17/2017)
+### New Features
+* FIFO compaction to support Intra L0 compaction too with CompactionOptionsFIFO.allow_compaction=true.
+* DB::ResetStats() to reset internal stats.
+* Statistics::Reset() to reset user stats.
+* ldb add option --try_load_options, which will open DB with its own option file.
+* Introduce WriteBatch::PopSavePoint to pop the most recent save point explicitly.
+* Support dynamically change `max_open_files` option via SetDBOptions()
+* Added DB::CreateColumnFamilie() and DB::DropColumnFamilies() to bulk create/drop column families.
+* Add debugging function `GetAllKeyVersions` to see internal versions of a range of keys.
+* Support file ingestion with universal compaction style
+* Support file ingestion behind with option `allow_ingest_behind`
+* New option enable_pipelined_write which may improve write throughput in case writing from multiple threads and WAL enabled.
+
+### Bug Fixes
+* Fix the bug that Direct I/O uses direct reads for non-SST file
+
+## 5.4.0 (04/11/2017)
+### Public API Change
+* random_access_max_buffer_size no longer has any effect
+* Removed Env::EnableReadAhead(), Env::ShouldForwardRawRequest()
+* Support dynamically change `stats_dump_period_sec` option via SetDBOptions().
+* Added ReadOptions::max_skippable_internal_keys to set a threshold to fail a request as incomplete when too many keys are being skipped when using iterators.
+* DB::Get in place of std::string accepts PinnableSlice, which avoids the extra memcpy of value to std::string in most of cases.
+    * PinnableSlice releases the pinned resources that contain the value when it is destructed or when ::Reset() is called on it.
+    * The old API that accepts std::string, although discouraged, is still supported.
+* Replace Options::use_direct_writes with Options::use_direct_io_for_flush_and_compaction. Read Direct IO wiki for details.
+* Added CompactionEventListener and EventListener::OnFlushBegin interfaces.
+
+### New Features
+* Memtable flush can be avoided during checkpoint creation if total log file size is smaller than a threshold specified by the user.
+* Introduce level-based L0->L0 compactions to reduce file count, so write delays are incurred less often.
+* (Experimental) Partitioning filters which creates an index on the partitions. The feature can be enabled by setting partition_filters when using kFullFilter. Currently the feature also requires two-level indexing to be enabled. Number of partitions is the same as the number of partitions for indexes, which is controlled by metadata_block_size.
 
 ## 5.3.0 (03/08/2017)
 ### Public API Change
@@ -9,9 +97,6 @@
 * Remove option min_partial_merge_operands. Partial merge operands will always be merged in flush or compaction if there are more than one.
 * Remove option verify_checksums_in_compaction. Compaction will always verify checksum.
 
-### New Features
-* Memtable flush can be avoided during checkpoint creation if total log file size is smaller than a threshold specified by the user.
-
 ### Bug Fixes
 * Fix the bug that iterator may skip keys
 
@@ -230,7 +315,7 @@
 * Added a new way to report QPS from db_bench (check out --report_file and --report_interval_seconds)
 * Added a cache for individual rows. See DBOptions::row_cache for more info.
 * Several new features on EventListener (see include/rocksdb/listener.h):
- - OnCompationCompleted() now returns per-compaciton job statistics, defined in include/rocksdb/compaction_job_stats.h.
+ - OnCompationCompleted() now returns per-compaction job statistics, defined in include/rocksdb/compaction_job_stats.h.
  - Added OnTableFileCreated() and OnTableFileDeleted().
 * Add compaction_options_universal.enable_trivial_move to true, to allow trivial move while performing universal compaction. Trivial move will happen only when all the input files are non overlapping.
 

INSTALL.md

@@ -1,6 +1,6 @@
 ## Compilation
 
-**Important**: If you plan to run RocksDB in production, don't compile using default 
+**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.
 
@@ -21,14 +21,18 @@ depend on gflags. You will need to have gflags installed to run `make all`. This
 use binaries compiled by `make all` in production.
 
 * 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
-your make commands, like this: `PORTABLE=1 make static_lib`
+(`-march=native` or the equivalent). SSE4.2 will thus be enabled automatically if your
+CPU supports it. To print a warning if your CPU does not support SSE4.2, build with
+`USE_SSE=1 make static_lib` or, if using CMake, `cmake -DFORCE_SSE42=ON`. If you want
+to build a portable binary, add `PORTABLE=1` before your make commands, like this:
+`PORTABLE=1 make static_lib`.
 
 ## Dependencies
 
 * You can link RocksDB with following compression libraries:
   - [zlib](http://www.zlib.net/) - a library for data compression.
   - [bzip2](http://www.bzip.org/) - a library for data compression.
+  - [lz4](https://github.com/lz4/lz4) - a library for extremely fast data compression.
   - [snappy](http://google.github.io/snappy/) - a library for fast
       data compression.
   - [zstandard](http://www.zstd.net) - Fast real-time compression
@@ -50,6 +54,7 @@ your make commands, like this: `PORTABLE=1 make static_lib`
       `sudo apt-get install libsnappy-dev`.
     * Install zlib. Try: `sudo apt-get install zlib1g-dev`.
     * Install bzip2: `sudo apt-get install libbz2-dev`.
+    * Install lz4: `sudo apt-get install liblz4-dev`.
     * Install zstandard: `sudo apt-get install libzstd-dev`.
 
 * **Linux - CentOS / RHEL**
@@ -58,31 +63,38 @@ your make commands, like this: `PORTABLE=1 make static_lib`
     * Install gflags:
 
               git clone https://github.com/gflags/gflags.git
-              git checkout v2.0
               cd gflags
+              git checkout v2.0
               ./configure && make && sudo make install
 
+      **Notice**: Once installed, please add the include path for gflags to your `CPATH` environment variable and the
+      lib path to `LIBRARY_PATH`. If installed with default settings, the include path will be `/usr/local/include`
+      and the lib path will be `/usr/local/lib`.
+
     * Install snappy:
 
-              wget https://github.com/google/snappy/releases/download/1.1.4/snappy-1.1.4.tar.gz
-              tar -xzvf snappy-1.1.4.tar.gz
-              cd snappy-1.1.4
-              ./configure && make && sudo make install
+              sudo yum install snappy snappy-devel
 
     * Install zlib:
 
-              sudo yum install zlib
-              sudo yum install zlib-devel
+              sudo yum install zlib zlib-devel
 
     * Install bzip2:
 
-              sudo yum install bzip2
-              sudo yum install bzip2-devel
+              sudo yum install bzip2 bzip2-devel
+
+    * Install lz4:
+
+              sudo yum install lz4-devel
+
+    * Install ASAN (optional for debugging):
+
+              sudo yum install libasan
 
     * Install zstandard:
 
              wget https://github.com/facebook/zstd/archive/v1.1.3.tar.gz
-             mv v1.1.3.tar.gz zstandard-1.1.3.tar.gz
+             mv v1.1.3.tar.gz zstd-1.1.3.tar.gz
              tar zxvf zstd-1.1.3.tar.gz
              cd zstd-1.1.3
              make && sudo make install
@@ -101,3 +113,32 @@ your make commands, like this: `PORTABLE=1 make static_lib`
 * **Windows**:
   * For building with MS Visual Studio 13 you will need Update 4 installed.
   * Read and follow the instructions at CMakeLists.txt
+  * Or install via [vcpkg](https://github.com/microsoft/vcpkg) 
+       * run `vcpkg install rocksdb`
+
+* **AIX 6.1**
+    * Install AIX Toolbox rpms with gcc
+    * Use these environment variables:
+  
+             export PORTABLE=1
+             export CC=gcc
+             export AR="ar -X64"
+             export EXTRA_ARFLAGS=-X64
+             export EXTRA_CFLAGS=-maix64
+             export EXTRA_CXXFLAGS=-maix64
+             export PLATFORM_LDFLAGS="-static-libstdc++ -static-libgcc"
+             export LIBPATH=/opt/freeware/lib
+             export JAVA_HOME=/usr/java8_64
+             export PATH=/opt/freeware/bin:$PATH
+  
+* **Solaris Sparc**
+    * Install GCC 4.8.2 and higher.
+    * Use these environment variables:
+
+             export CC=gcc
+             export EXTRA_CFLAGS=-m64
+             export EXTRA_CXXFLAGS=-m64
+             export EXTRA_LDFLAGS=-m64
+             export PORTABLE=1
+             export PLATFORM_LDFLAGS="-static-libstdc++ -static-libgcc"
+

LANGUAGE-BINDINGS.md

@@ -9,5 +9,8 @@ This is the list of all known third-party language bindings for RocksDB. If some
 * 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
+* Rust
+    * https://github.com/spacejam/rust-rocksdb
+    * https://github.com/bh1xuw/rust-rocks
 * D programming language - https://github.com/b1naryth1ef/rocksdb
+* Erlang - https://gitlab.com/barrel-db/erlang-rocksdb

Makefile

@@ -14,7 +14,8 @@ CFLAGS += ${EXTRA_CFLAGS}
 CXXFLAGS += ${EXTRA_CXXFLAGS}
 LDFLAGS += $(EXTRA_LDFLAGS)
 MACHINE ?= $(shell uname -m)
-ARFLAGS = rs
+ARFLAGS = ${EXTRA_ARFLAGS} rs
+STRIPFLAGS = -S -x
 
 # Transform parallel LOG output into something more readable.
 perl_command = perl -n \
@@ -100,7 +101,19 @@ endif
 ifeq ($(DEBUG_LEVEL),0)
 OPT += -DNDEBUG
 DISABLE_WARNING_AS_ERROR=1
+
+ifneq ($(USE_RTTI), 1)
+	CXXFLAGS += -fno-rtti
 else
+	CXXFLAGS += -DROCKSDB_USE_RTTI
+endif
+else
+ifneq ($(USE_RTTI), 0)
+	CXXFLAGS += -DROCKSDB_USE_RTTI
+else
+	CXXFLAGS += -fno-rtti
+endif
+
 $(warning Warning: Compiling in debug mode. Don't use the resulting binary in production)
 endif
 
@@ -154,7 +167,9 @@ missing_make_config_paths := $(shell				\
 $(foreach path, $(missing_make_config_paths), \
 	$(warning Warning: $(path) dont exist))
 
-ifneq ($(PLATFORM), IOS)
+ifeq ($(PLATFORM), OS_AIX)
+# no debug info
+else ifneq ($(PLATFORM), IOS)
 CFLAGS += -g
 CXXFLAGS += -g
 else
@@ -162,6 +177,11 @@ else
 OPT += -DNDEBUG
 endif
 
+ifeq ($(PLATFORM), OS_AIX)
+ARFLAGS = -X64 rs
+STRIPFLAGS = -X64 -x
+endif
+
 ifeq ($(PLATFORM), OS_SOLARIS)
 	PLATFORM_CXXFLAGS += -D _GLIBCXX_USE_C99
 endif
@@ -184,9 +204,9 @@ endif
 # TSAN doesn't work well with jemalloc. If we're compiling with TSAN, we should use regular malloc.
 ifdef COMPILE_WITH_TSAN
 	DISABLE_JEMALLOC=1
-	EXEC_LDFLAGS += -fsanitize=thread -pie
-	PLATFORM_CCFLAGS += -fsanitize=thread -fPIC -DROCKSDB_TSAN_RUN
-	PLATFORM_CXXFLAGS += -fsanitize=thread -fPIC -DROCKSDB_TSAN_RUN
+	EXEC_LDFLAGS += -fsanitize=thread
+	PLATFORM_CCFLAGS += -fsanitize=thread -fPIC
+	PLATFORM_CXXFLAGS += -fsanitize=thread -fPIC
         # Turn off -pg when enabling TSAN testing, because that induces
         # a link failure.  TODO: find the root cause
 	PROFILING_FLAGS =
@@ -194,28 +214,44 @@ ifdef COMPILE_WITH_TSAN
 	LUA_PATH =
 endif
 
+# AIX doesn't work with -pg
+ifeq ($(PLATFORM), OS_AIX)
+	PROFILING_FLAGS =
+endif
+
 # USAN doesn't work well with jemalloc. If we're compiling with USAN, we should use regular malloc.
 ifdef COMPILE_WITH_UBSAN
 	DISABLE_JEMALLOC=1
 	EXEC_LDFLAGS += -fsanitize=undefined
-	PLATFORM_CCFLAGS += -fsanitize=undefined
-	PLATFORM_CXXFLAGS += -fsanitize=undefined
+	PLATFORM_CCFLAGS += -fsanitize=undefined -DROCKSDB_UBSAN_RUN
+	PLATFORM_CXXFLAGS += -fsanitize=undefined -DROCKSDB_UBSAN_RUN
 endif
 
 ifndef DISABLE_JEMALLOC
 	ifdef JEMALLOC
-		PLATFORM_CXXFLAGS += "-DROCKSDB_JEMALLOC"
-		PLATFORM_CCFLAGS +=  "-DROCKSDB_JEMALLOC"
+		PLATFORM_CXXFLAGS += -DROCKSDB_JEMALLOC -DJEMALLOC_NO_DEMANGLE
+		PLATFORM_CCFLAGS  += -DROCKSDB_JEMALLOC -DJEMALLOC_NO_DEMANGLE
+	endif
+	ifdef WITH_JEMALLOC_FLAG
+		PLATFORM_LDFLAGS += -ljemalloc
+		JAVA_LDFLAGS += -ljemalloc
 	endif
 	EXEC_LDFLAGS := $(JEMALLOC_LIB) $(EXEC_LDFLAGS)
 	PLATFORM_CXXFLAGS += $(JEMALLOC_INCLUDE)
 	PLATFORM_CCFLAGS += $(JEMALLOC_INCLUDE)
 endif
 
-export GTEST_THROW_ON_FAILURE=1 GTEST_HAS_EXCEPTIONS=1
+export GTEST_THROW_ON_FAILURE=1
+export GTEST_HAS_EXCEPTIONS=1
 GTEST_DIR = ./third-party/gtest-1.7.0/fused-src
-PLATFORM_CCFLAGS += -isystem $(GTEST_DIR)
-PLATFORM_CXXFLAGS += -isystem $(GTEST_DIR)
+# AIX: pre-defined system headers are surrounded by an extern "C" block
+ifeq ($(PLATFORM), OS_AIX)
+	PLATFORM_CCFLAGS += -I$(GTEST_DIR)
+	PLATFORM_CXXFLAGS += -I$(GTEST_DIR)
+else
+	PLATFORM_CCFLAGS += -isystem $(GTEST_DIR)
+	PLATFORM_CXXFLAGS += -isystem $(GTEST_DIR)
+endif
 
 # This (the first rule) must depend on "all".
 default: all
@@ -223,6 +259,11 @@ default: all
 WARNING_FLAGS = -W -Wextra -Wall -Wsign-compare -Wshadow \
   -Wno-unused-parameter
 
+ifndef DISABLE_WARNING_AS_ERROR
+	WARNING_FLAGS += -Werror
+endif
+
+
 ifdef LUA_PATH
 
 ifndef LUA_INCLUDE
@@ -255,6 +296,9 @@ CXXFLAGS += $(WARNING_FLAGS) -I. -I./include $(PLATFORM_CXXFLAGS) $(OPT) -Woverl
 
 LDFLAGS += $(PLATFORM_LDFLAGS)
 
+# If NO_UPDATE_BUILD_VERSION is set we don't update util/build_version.cc, but
+# the file needs to already exist or else the build will fail
+ifndef NO_UPDATE_BUILD_VERSION
 date := $(shell date +%F)
 ifdef FORCE_GIT_SHA
 	git_sha := $(FORCE_GIT_SHA)
@@ -268,7 +312,6 @@ gen_build_version = sed -e s/@@GIT_SHA@@/$(git_sha)/ -e s/@@GIT_DATE_TIME@@/$(da
 # as a regular source file as part of the compilation process.
 # One can run "strings executable_filename | grep _build_" to find
 # the version of the source that we used to build the executable file.
-CLEAN_FILES += util/build_version.cc:
 FORCE:
 util/build_version.cc: FORCE
 	$(AM_V_GEN)rm -f $@-t
@@ -276,6 +319,7 @@ util/build_version.cc: FORCE
 	$(AM_V_at)if test -f $@; then					\
 	  cmp -s $@-t $@ && rm -f $@-t || mv -f $@-t $@;		\
 	else mv -f $@-t $@; fi
+endif
 
 LIBOBJECTS = $(LIB_SOURCES:.cc=.o)
 LIBOBJECTS += $(TOOL_LIB_SOURCES:.cc=.o)
@@ -295,6 +339,7 @@ EXPOBJECTS = $(EXP_LIB_SOURCES:.cc=.o) $(LIBOBJECTS) $(TESTUTIL)
 
 TESTS = \
 	db_basic_test \
+	db_encryption_test \
 	db_test2 \
 	external_sst_file_basic_test \
 	auto_roll_logger_test \
@@ -307,6 +352,7 @@ TESTS = \
 	inlineskiplist_test \
 	env_basic_test \
 	env_test \
+	hash_test \
 	thread_local_test \
 	rate_limiter_test \
 	perf_context_test \
@@ -333,7 +379,10 @@ TESTS = \
 	db_io_failure_test \
 	db_properties_test \
 	db_table_properties_test \
+	db_statistics_test \
+	db_write_test \
 	autovector_test \
+	blob_db_test \
 	cleanable_test \
 	column_family_test \
 	table_properties_collector_test \
@@ -349,6 +398,7 @@ TESTS = \
 	file_reader_writer_test \
 	block_based_filter_block_test \
 	full_filter_block_test \
+	partitioned_filter_block_test \
 	hash_table_test \
 	histogram_test \
 	log_test \
@@ -368,11 +418,15 @@ TESTS = \
 	external_sst_file_test \
 	prefix_test \
 	skiplist_test \
+	write_buffer_manager_test \
 	stringappend_test \
+	cassandra_format_test \
+	cassandra_functional_test \
+	cassandra_row_merge_test \
+	cassandra_serialize_test \
 	ttl_test \
 	date_tiered_test \
 	backupable_db_test \
-	blob_db_test \
 	document_db_test \
 	json_document_test \
 	sim_cache_test \
@@ -393,6 +447,7 @@ TESTS = \
 	options_settable_test \
 	options_util_test \
 	event_logger_test \
+	timer_queue_test \
 	cuckoo_table_builder_test \
 	cuckoo_table_reader_test \
 	cuckoo_table_db_test \
@@ -420,6 +475,7 @@ TESTS = \
 	lru_cache_test \
 	object_registry_test \
 	repair_test \
+	env_timed_test \
 
 PARALLEL_TEST = \
 	backupable_db_test \
@@ -454,7 +510,8 @@ TOOLS = \
 	ldb \
 	db_repl_stress \
 	rocksdb_dump \
-	rocksdb_undump
+	rocksdb_undump \
+	blob_dump \
 
 TEST_LIBS = \
 	librocksdb_env_basic_test.a
@@ -515,9 +572,14 @@ $(SHARED3): $(SHARED4)
 	ln -fs $(SHARED4) $(SHARED3)
 endif
 
-$(SHARED4):
-	$(CXX) $(PLATFORM_SHARED_LDFLAGS)$(SHARED3) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) $(LIB_SOURCES) $(TOOL_LIB_SOURCES) \
-		$(LDFLAGS) -o $@
+shared_libobjects = $(patsubst %,shared-objects/%,$(LIBOBJECTS))
+CLEAN_FILES += shared-objects
+
+$(shared_libobjects): shared-objects/%.o: %.cc
+	$(AM_V_CC)mkdir -p $(@D) && $(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) -c $< -o $@
+
+$(SHARED4): $(shared_libobjects)
+	$(CXX) $(PLATFORM_SHARED_LDFLAGS)$(SHARED3) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) $(shared_libobjects) $(LDFLAGS) -o $@
 
 endif  # PLATFORM_SHARED_EXT
 
@@ -529,6 +591,8 @@ endif  # PLATFORM_SHARED_EXT
 
 all: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(TESTS)
 
+all_but_some_tests: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(SUBSET)
+
 static_lib: $(LIBRARY)
 
 shared_lib: $(SHARED)
@@ -558,7 +622,8 @@ ifneq (,$(filter check parallel_check,$(MAKECMDGOALS)),)
 # and create a randomly-named rocksdb.XXXX directory therein.
 # We'll use that directory in the "make check" rules.
 ifeq ($(TMPD),)
-TMPD := $(shell f=/dev/shm; test -k $$f || f=/tmp;			\
+TMPDIR := $(shell echo $${TMPDIR:-/tmp})
+TMPD := $(shell f=/dev/shm; test -k $$f || f=$(TMPDIR);     \
   perl -le 'use File::Temp "tempdir";'					\
     -e 'print tempdir("'$$f'/rocksdb.XXXX", CLEANUP => 0)')
 endif
@@ -713,10 +778,12 @@ check: all
 	      echo "===== Running $$t"; ./$$t || exit 1; done;          \
 	fi
 	rm -rf $(TMPD)
+ifneq ($(PLATFORM), OS_AIX)
 ifeq ($(filter -DROCKSDB_LITE,$(OPT)),)
 	python tools/ldb_test.py
 	sh tools/rocksdb_dump_test.sh
 endif
+endif
 
 # TODO add ldb_tests
 check_some: $(SUBSET)
@@ -729,8 +796,8 @@ ldb_tests: ldb
 crash_test: whitebox_crash_test blackbox_crash_test
 
 blackbox_crash_test: db_stress
-	python -u tools/db_crashtest.py --simple blackbox
-	python -u tools/db_crashtest.py blackbox
+	python -u tools/db_crashtest.py --simple blackbox $(CRASH_TEST_EXT_ARGS)
+	python -u tools/db_crashtest.py blackbox $(CRASH_TEST_EXT_ARGS)
 
 ifeq ($(CRASH_TEST_KILL_ODD),)
   CRASH_TEST_KILL_ODD=888887
@@ -738,9 +805,9 @@ endif
 
 whitebox_crash_test: db_stress
 	python -u tools/db_crashtest.py --simple whitebox --random_kill_odd \
-      $(CRASH_TEST_KILL_ODD)
+      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)
 	python -u tools/db_crashtest.py whitebox  --random_kill_odd \
-      $(CRASH_TEST_KILL_ODD)
+      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)
 
 asan_check:
 	$(MAKE) clean
@@ -862,12 +929,13 @@ clean:
 	rm -rf $(CLEAN_FILES) ios-x86 ios-arm scan_build_report
 	find . -name "*.[oda]" -exec rm -f {} \;
 	find . -type f -regex ".*\.\(\(gcda\)\|\(gcno\)\)" -exec rm {} \;
-	rm -rf bzip2* snappy* zlib* lz4*
+	rm -rf bzip2* snappy* zlib* lz4* zstd*
 	cd java; $(MAKE) clean
 
 tags:
 	ctags * -R
 	cscope -b `find . -name '*.cc'` `find . -name '*.h'` `find . -name '*.c'`
+	ctags -e -R -o etags *
 
 format:
 	build_tools/format-diff.sh
@@ -886,20 +954,20 @@ $(TOOLS_LIBRARY): $(BENCH_LIB_SOURCES:.cc=.o) $(TOOL_LIB_SOURCES:.cc=.o) $(LIB_S
 	$(AM_V_AR)rm -f $@
 	$(AM_V_at)$(AR) $(ARFLAGS) $@ $^
 
-librocksdb_env_basic_test.a: util/env_basic_test.o $(LIBOBJECTS) $(TESTHARNESS)
+librocksdb_env_basic_test.a: env/env_basic_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_V_AR)rm -f $@
 	$(AM_V_at)$(AR) $(ARFLAGS) $@ $^
 
 db_bench: tools/db_bench.o $(BENCHTOOLOBJECTS)
 	$(AM_LINK)
 
-cache_bench: util/cache_bench.o $(LIBOBJECTS) $(TESTUTIL)
+cache_bench: cache/cache_bench.o $(LIBOBJECTS) $(TESTUTIL)
 	$(AM_LINK)
 
 persistent_cache_bench: utilities/persistent_cache/persistent_cache_bench.o $(LIBOBJECTS) $(TESTUTIL)
 	$(AM_LINK)
 
-memtablerep_bench: db/memtablerep_bench.o $(LIBOBJECTS) $(TESTUTIL)
+memtablerep_bench: memtable/memtablerep_bench.o $(LIBOBJECTS) $(TESTUTIL)
 	$(AM_LINK)
 
 db_stress: tools/db_stress.o $(LIBOBJECTS) $(TESTUTIL)
@@ -935,25 +1003,40 @@ dynamic_bloom_test: util/dynamic_bloom_test.o $(LIBOBJECTS) $(TESTHARNESS)
 c_test: db/c_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-cache_test: util/cache_test.o $(LIBOBJECTS) $(TESTHARNESS)
+cache_test: cache/cache_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 coding_test: util/coding_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+hash_test: util/hash_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 option_change_migration_test: utilities/option_change_migration/option_change_migration_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 stringappend_test: utilities/merge_operators/string_append/stringappend_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+cassandra_format_test: utilities/cassandra/cassandra_format_test.o utilities/cassandra/test_utils.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+cassandra_functional_test: utilities/cassandra/cassandra_functional_test.o utilities/cassandra/test_utils.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+cassandra_row_merge_test: utilities/cassandra/cassandra_row_merge_test.o utilities/cassandra/test_utils.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+cassandra_serialize_test: utilities/cassandra/cassandra_serialize_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 redis_test: utilities/redis/redis_lists_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 hash_table_test: utilities/persistent_cache/hash_table_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-histogram_test: util/histogram_test.o $(LIBOBJECTS) $(TESTHARNESS)
+histogram_test: monitoring/histogram_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 thread_local_test: util/thread_local_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -971,6 +1054,9 @@ slice_transform_test: util/slice_transform_test.o $(LIBOBJECTS) $(TESTHARNESS)
 db_basic_test: db/db_basic_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+db_encryption_test: db/db_encryption_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 db_test: db/db_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -1019,6 +1105,12 @@ db_range_del_test: db/db_range_del_test.o db/db_test_util.o $(LIBOBJECTS) $(TEST
 db_sst_test: db/db_sst_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+db_statistics_test: db/db_statistics_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+db_write_test: db/db_write_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 external_sst_file_basic_test: db/external_sst_file_basic_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -1067,9 +1159,6 @@ prefix_test: db/prefix_test.o $(LIBOBJECTS) $(TESTHARNESS)
 backupable_db_test: utilities/backupable/backupable_db_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-blob_db_test: utilities/blob_db/blob_db_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 checkpoint_test: utilities/checkpoint/checkpoint_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -1088,6 +1177,9 @@ spatial_db_test: utilities/spatialdb/spatial_db_test.o $(LIBOBJECTS) $(TESTHARNE
 env_mirror_test: utilities/env_mirror_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+env_timed_test: utilities/env_timed_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 ifdef ROCKSDB_USE_LIBRADOS
 env_librados_test: utilities/env_librados_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_V_CCLD)$(CXX) $^ $(EXEC_LDFLAGS) -o $@ $(LDFLAGS) $(COVERAGEFLAGS)
@@ -1126,10 +1218,10 @@ wal_manager_test: db/wal_manager_test.o $(LIBOBJECTS) $(TESTHARNESS)
 dbformat_test: db/dbformat_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-env_basic_test: util/env_basic_test.o $(LIBOBJECTS) $(TESTHARNESS)
+env_basic_test: env/env_basic_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-env_test: util/env_test.o $(LIBOBJECTS) $(TESTHARNESS)
+env_test: env/env_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 fault_injection_test: db/fault_injection_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1153,6 +1245,9 @@ block_based_filter_block_test: table/block_based_filter_block_test.o $(LIBOBJECT
 full_filter_block_test: table/full_filter_block_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+partitioned_filter_block_test: table/partitioned_filter_block_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 log_test: db/log_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -1165,10 +1260,13 @@ table_test: table/table_test.o $(LIBOBJECTS) $(TESTHARNESS)
 block_test: table/block_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-inlineskiplist_test: db/inlineskiplist_test.o $(LIBOBJECTS) $(TESTHARNESS)
+inlineskiplist_test: memtable/inlineskiplist_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-skiplist_test: db/skiplist_test.o $(LIBOBJECTS) $(TESTHARNESS)
+skiplist_test: memtable/skiplist_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+write_buffer_manager_test: memtable/write_buffer_manager_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 version_edit_test: db/version_edit_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1243,10 +1341,10 @@ thread_list_test: util/thread_list_test.o $(LIBOBJECTS) $(TESTHARNESS)
 compact_files_test: db/compact_files_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-options_test: util/options_test.o $(LIBOBJECTS) $(TESTHARNESS)
+options_test: options/options_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-options_settable_test: util/options_settable_test.o $(LIBOBJECTS) $(TESTHARNESS)
+options_settable_test: options/options_settable_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 options_util_test: utilities/options/options_util_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1258,6 +1356,9 @@ db_bench_tool_test: tools/db_bench_tool_test.o $(BENCHTOOLOBJECTS) $(TESTHARNESS
 event_logger_test: util/event_logger_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+timer_queue_test: util/timer_queue_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 sst_dump_test: tools/sst_dump_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -1267,7 +1368,7 @@ column_aware_encoding_test: utilities/column_aware_encoding_test.o $(TESTHARNESS
 optimistic_transaction_test: utilities/transactions/optimistic_transaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-mock_env_test : util/mock_env_test.o $(LIBOBJECTS) $(TESTHARNESS)
+mock_env_test : env/mock_env_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 manual_compaction_test: db/manual_compaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1276,7 +1377,7 @@ manual_compaction_test: db/manual_compaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
 filelock_test: util/filelock_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-auto_roll_logger_test: db/auto_roll_logger_test.o $(LIBOBJECTS) $(TESTHARNESS)
+auto_roll_logger_test: util/auto_roll_logger_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 memtable_list_test: db/memtable_list_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1294,6 +1395,9 @@ transaction_test: utilities/transactions/transaction_test.o $(LIBOBJECTS) $(TEST
 sst_dump: tools/sst_dump.o $(LIBOBJECTS)
 	$(AM_LINK)
 
+blob_dump: tools/blob_dump.o $(LIBOBJECTS)
+	$(AM_LINK)
+
 column_aware_encoding_exp: utilities/column_aware_encoding_exp.o $(EXPOBJECTS)
 	$(AM_LINK)
 
@@ -1306,16 +1410,16 @@ ldb_cmd_test: tools/ldb_cmd_test.o $(LIBOBJECTS) $(TESTHARNESS)
 ldb: tools/ldb.o $(LIBOBJECTS)
 	$(AM_LINK)
 
-iostats_context_test: util/iostats_context_test.o $(LIBOBJECTS) $(TESTHARNESS)
+iostats_context_test: monitoring/iostats_context_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_V_CCLD)$(CXX) $^ $(EXEC_LDFLAGS) -o $@ $(LDFLAGS)
 
 persistent_cache_test: utilities/persistent_cache/persistent_cache_test.o  db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-statistics_test: util/statistics_test.o $(LIBOBJECTS) $(TESTHARNESS)
+statistics_test: monitoring/statistics_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-lru_cache_test: util/lru_cache_test.o $(LIBOBJECTS) $(TESTHARNESS)
+lru_cache_test: cache/lru_cache_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 lua_test: utilities/lua/rocks_lua_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -1324,6 +1428,9 @@ lua_test: utilities/lua/rocks_lua_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTH
 range_del_aggregator_test: db/range_del_aggregator_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+blob_db_test: utilities/blob_db/blob_db_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 #-------------------------------------------------
 # make install related stuff
 INSTALL_PATH ?= /usr/local
@@ -1371,21 +1478,30 @@ ifeq ($(PLATFORM), OS_SOLARIS)
 else
 	ARCH := $(shell getconf LONG_BIT)
 endif
-ROCKSDBJNILIB = librocksdbjni-linux$(ARCH).so
+
+ifeq (,$(findstring ppc,$(MACHINE)))
+        ROCKSDBJNILIB = librocksdbjni-linux$(ARCH).so
+else
+        ROCKSDBJNILIB = librocksdbjni-linux-$(MACHINE).so
+endif
 ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux$(ARCH).jar
 ROCKSDB_JAR_ALL = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH).jar
 ROCKSDB_JAVADOCS_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-javadoc.jar
 ROCKSDB_SOURCES_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-sources.jar
 SHA256_CMD = sha256sum
 
-ZLIB_VER = 1.2.11
-ZLIB_SHA256 = c3e5e9fdd5004dcb542feda5ee4f0ff0744628baf8ed2dd5d66f8ca1197cb1a1
-BZIP2_VER = 1.0.6
-BZIP2_SHA256 = a2848f34fcd5d6cf47def00461fcb528a0484d8edef8208d6d2e2909dc61d9cd
-SNAPPY_VER = 1.1.4
-SNAPPY_SHA256 = 134bfe122fd25599bb807bb8130e7ba6d9bdb851e0b16efcb83ac4f5d0b70057
-LZ4_VER = 1.7.5
-LZ4_SHA256 = 0190cacd63022ccb86f44fa5041dc6c3804407ad61550ca21c382827319e7e7e
+ZLIB_VER ?= 1.2.11
+ZLIB_SHA256 ?= c3e5e9fdd5004dcb542feda5ee4f0ff0744628baf8ed2dd5d66f8ca1197cb1a1
+ZLIB_DOWNLOAD_BASE ?= http://zlib.net
+BZIP2_VER ?= 1.0.6
+BZIP2_SHA256 ?= a2848f34fcd5d6cf47def00461fcb528a0484d8edef8208d6d2e2909dc61d9cd
+BZIP2_DOWNLOAD_BASE ?= http://www.bzip.org
+SNAPPY_VER ?= 1.1.4
+SNAPPY_SHA256 ?= 134bfe122fd25599bb807bb8130e7ba6d9bdb851e0b16efcb83ac4f5d0b70057
+SNAPPY_DOWNLOAD_BASE ?= https://github.com/google/snappy/releases/download
+LZ4_VER ?= 1.7.5
+LZ4_SHA256 ?= 0190cacd63022ccb86f44fa5041dc6c3804407ad61550ca21c382827319e7e7e
+LZ4_DOWNLOAD_BASE ?= https://github.com/lz4/lz4/archive
 ZSTD_VER ?= 1.2.0
 ZSTD_SHA256 ?= 4a7e4593a3638276ca7f2a09dc4f38e674d8317bbea51626393ca73fc047cbfb
 ZSTD_DOWNLOAD_BASE ?= https://github.com/facebook/zstd/archive
@@ -1411,47 +1527,53 @@ ifeq ($(PLATFORM), OS_SOLARIS)
 	JAVA_INCLUDE = -I$(JAVA_HOME)/include/ -I$(JAVA_HOME)/include/solaris
 	SHA256_CMD = digest -a sha256
 endif
+ifeq ($(PLATFORM), OS_AIX)
+	JAVA_INCLUDE = -I$(JAVA_HOME)/include/ -I$(JAVA_HOME)/include/aix
+	ROCKSDBJNILIB = librocksdbjni-aix.so
+	EXTRACT_SOURCES = gunzip < TAR_GZ | tar xvf -
+	SNAPPY_MAKE_TARGET = libsnappy.la
+endif
 
 libz.a:
 	-rm -rf zlib-$(ZLIB_VER)
-	curl -O -L http://zlib.net/zlib-$(ZLIB_VER).tar.gz
+	curl -O -L ${ZLIB_DOWNLOAD_BASE}/zlib-$(ZLIB_VER).tar.gz
 	ZLIB_SHA256_ACTUAL=`$(SHA256_CMD) zlib-$(ZLIB_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(ZLIB_SHA256)" != "$$ZLIB_SHA256_ACTUAL" ]; then \
 		echo zlib-$(ZLIB_VER).tar.gz checksum mismatch, expected=\"$(ZLIB_SHA256)\" actual=\"$$ZLIB_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
 	tar xvzf zlib-$(ZLIB_VER).tar.gz
-	cd zlib-$(ZLIB_VER) && CFLAGS='-fPIC' ./configure --static && make
+	cd zlib-$(ZLIB_VER) && CFLAGS='-fPIC ${EXTRA_CFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --static && make
 	cp zlib-$(ZLIB_VER)/libz.a .
 
 libbz2.a:
 	-rm -rf bzip2-$(BZIP2_VER)
-	curl -O -L http://www.bzip.org/$(BZIP2_VER)/bzip2-$(BZIP2_VER).tar.gz
+	curl -O -L ${BZIP2_DOWNLOAD_BASE}/$(BZIP2_VER)/bzip2-$(BZIP2_VER).tar.gz
 	BZIP2_SHA256_ACTUAL=`$(SHA256_CMD) bzip2-$(BZIP2_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(BZIP2_SHA256)" != "$$BZIP2_SHA256_ACTUAL" ]; then \
 		echo bzip2-$(BZIP2_VER).tar.gz checksum mismatch, expected=\"$(BZIP2_SHA256)\" actual=\"$$BZIP2_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
 	tar xvzf bzip2-$(BZIP2_VER).tar.gz
-	cd bzip2-$(BZIP2_VER) && make CFLAGS='-fPIC -O2 -g -D_FILE_OFFSET_BITS=64'
+	cd bzip2-$(BZIP2_VER) && make CFLAGS='-fPIC -O2 -g -D_FILE_OFFSET_BITS=64 ${EXTRA_CFLAGS}' AR='ar ${EXTRA_ARFLAGS}'
 	cp bzip2-$(BZIP2_VER)/libbz2.a .
 
 libsnappy.a:
 	-rm -rf snappy-$(SNAPPY_VER)
-	curl -O -L https://github.com/google/snappy/releases/download/$(SNAPPY_VER)/snappy-$(SNAPPY_VER).tar.gz
+	curl -O -L ${SNAPPY_DOWNLOAD_BASE}/$(SNAPPY_VER)/snappy-$(SNAPPY_VER).tar.gz
 	SNAPPY_SHA256_ACTUAL=`$(SHA256_CMD) snappy-$(SNAPPY_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(SNAPPY_SHA256)" != "$$SNAPPY_SHA256_ACTUAL" ]; then \
 		echo snappy-$(SNAPPY_VER).tar.gz checksum mismatch, expected=\"$(SNAPPY_SHA256)\" actual=\"$$SNAPPY_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
 	tar xvzf snappy-$(SNAPPY_VER).tar.gz
-	cd snappy-$(SNAPPY_VER) && ./configure --with-pic --enable-static
-	cd snappy-$(SNAPPY_VER) && make
+	cd snappy-$(SNAPPY_VER) && CFLAGS='${EXTRA_CFLAGS}' CXXFLAGS='${EXTRA_CXXFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --with-pic --enable-static --disable-shared
+	cd snappy-$(SNAPPY_VER) && make ${SNAPPY_MAKE_TARGET}
 	cp snappy-$(SNAPPY_VER)/.libs/libsnappy.a .
 
 liblz4.a:
 	-rm -rf lz4-$(LZ4_VER)
-	curl -O -L https://github.com/lz4/lz4/archive/v$(LZ4_VER).tar.gz
+	curl -O -L ${LZ4_DOWNLOAD_BASE}/v$(LZ4_VER).tar.gz
 	mv v$(LZ4_VER).tar.gz lz4-$(LZ4_VER).tar.gz
 	LZ4_SHA256_ACTUAL=`$(SHA256_CMD) lz4-$(LZ4_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(LZ4_SHA256)" != "$$LZ4_SHA256_ACTUAL" ]; then \
@@ -1459,7 +1581,7 @@ liblz4.a:
 		exit 1; \
 	fi
 	tar xvzf lz4-$(LZ4_VER).tar.gz
-	cd lz4-$(LZ4_VER)/lib && make CFLAGS='-fPIC' all
+	cd lz4-$(LZ4_VER)/lib && make CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' all
 	cp lz4-$(LZ4_VER)/lib/liblz4.a .
 
 libzstd.a:
@@ -1496,7 +1618,7 @@ rocksdbjavastatic: $(java_static_libobjects)
 	  -o ./java/target/$(ROCKSDBJNILIB) $(JNI_NATIVE_SOURCES) \
 	  $(java_static_libobjects) $(COVERAGEFLAGS) \
 	  $(JAVA_COMPRESSIONS) $(JAVA_STATIC_LDFLAGS)
-	cd java/target;strip -S -x $(ROCKSDBJNILIB)
+	cd java/target;strip $(STRIPFLAGS) $(ROCKSDBJNILIB)
 	cd java;jar -cf target/$(ROCKSDB_JAR) HISTORY*.md
 	cd java/target;jar -uf $(ROCKSDB_JAR) $(ROCKSDBJNILIB)
 	cd java/target/classes;jar -uf ../$(ROCKSDB_JAR) org/rocksdb/*.class org/rocksdb/util/*.class
@@ -1512,12 +1634,12 @@ rocksdbjavastaticrelease: rocksdbjavastatic
 rocksdbjavastaticreleasedocker: rocksdbjavastatic
 	DOCKER_LINUX_X64_CONTAINER=`docker ps -aqf name=rocksdb_linux_x64-be`; \
 	if [ -z "$$DOCKER_LINUX_X64_CONTAINER" ]; then \
-		docker container create --attach stdin --attach stdout --attach stderr --volume `pwd`:/rocksdb-host --name rocksdb_linux_x64-be evolvedbinary/rocksjava:centos5_x64-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh; \
+		docker container create --attach stdin --attach stdout --attach stderr --volume `pwd`:/rocksdb-host --name rocksdb_linux_x64-be evolvedbinary/rocksjava:centos6_x64-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh; \
 	fi
 	docker start -a rocksdb_linux_x64-be
 	DOCKER_LINUX_X86_CONTAINER=`docker ps -aqf name=rocksdb_linux_x86-be`; \
 	if [ -z "$$DOCKER_LINUX_X86_CONTAINER" ]; then \
-		docker container create --attach stdin --attach stdout --attach stderr --volume `pwd`:/rocksdb-host --name rocksdb_linux_x86-be evolvedbinary/rocksjava:centos5_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh; \
+		docker container create --attach stdin --attach stdout --attach stderr --volume `pwd`:/rocksdb-host --name rocksdb_linux_x86-be evolvedbinary/rocksjava:centos6_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh; \
 	fi
 	docker start -a rocksdb_linux_x86-be
 	cd java;jar -cf target/$(ROCKSDB_JAR_ALL) HISTORY*.md

TARGETS

@@ -0,0 +1,533 @@
+
+import os
+
+TARGETS_PATH = os.path.dirname(__file__)
+REPO_PATH = TARGETS_PATH[(TARGETS_PATH.find('fbcode/') + len('fbcode/')):] + "/"
+BUCK_BINS = "buck-out/gen/" + REPO_PATH
+TEST_RUNNER = REPO_PATH + "buckifier/rocks_test_runner.sh"
+rocksdb_compiler_flags = [
+  "-fno-builtin-memcmp",
+  "-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",
+  "-DOS_LINUX",
+  "-DROCKSDB_UBSAN_RUN",
+  # Flags to enable libs we include
+  "-DSNAPPY",
+  "-DZLIB",
+  "-DBZIP2",
+  "-DLZ4",
+  "-DZSTD",
+  "-DGFLAGS=gflags",
+  "-DNUMA",
+  "-DTBB",
+  # Needed to compile in fbcode
+  "-Wno-expansion-to-defined",
+]
+
+rocksdb_external_deps = [
+  ('bzip2', None, 'bz2'),
+  ('snappy', None, "snappy"),
+  ('zlib', None, 'z'),
+  ('gflags', None, 'gflags'),
+  ('lz4', None, 'lz4'),
+  ('zstd', None),
+  ('tbb', None),
+  ("numa", None, "numa"),
+  ("googletest", None, "gtest"),
+]
+
+rocksdb_preprocessor_flags = [
+  # Directories with files for #include
+  "-I" + REPO_PATH + "include/",
+  "-I" + REPO_PATH,
+]
+
+rocksdb_arch_preprocessor_flags = {
+  "x86_64": ["-DHAVE_SSE42"],
+}
+
+cpp_library(
+    name = "rocksdb_lib",
+    headers = AutoHeaders.RECURSIVE_GLOB,
+    srcs = [
+      "cache/clock_cache.cc",
+      "cache/lru_cache.cc",
+      "cache/sharded_cache.cc",
+      "db/builder.cc",
+      "db/c.cc",
+      "db/column_family.cc",
+      "db/compacted_db_impl.cc",
+      "db/compaction.cc",
+      "db/compaction_iterator.cc",
+      "db/compaction_job.cc",
+      "db/compaction_picker.cc",
+      "db/compaction_picker_universal.cc",
+      "db/convenience.cc",
+      "db/db_filesnapshot.cc",
+      "db/db_impl.cc",
+      "db/db_impl_write.cc",
+      "db/db_impl_compaction_flush.cc",
+      "db/db_impl_files.cc",
+      "db/db_impl_open.cc",
+      "db/db_impl_debug.cc",
+      "db/db_impl_experimental.cc",
+      "db/db_impl_readonly.cc",
+      "db/db_info_dumper.cc",
+      "db/db_iter.cc",
+      "db/dbformat.cc",
+      "db/event_helpers.cc",
+      "db/experimental.cc",
+      "db/external_sst_file_ingestion_job.cc",
+      "db/file_indexer.cc",
+      "db/flush_job.cc",
+      "db/flush_scheduler.cc",
+      "db/forward_iterator.cc",
+      "db/internal_stats.cc",
+      "db/log_reader.cc",
+      "db/log_writer.cc",
+      "db/malloc_stats.cc",
+      "db/managed_iterator.cc",
+      "db/memtable.cc",
+      "db/memtable_list.cc",
+      "db/merge_helper.cc",
+      "db/merge_operator.cc",
+      "db/range_del_aggregator.cc",
+      "db/repair.cc",
+      "db/snapshot_impl.cc",
+      "db/table_cache.cc",
+      "db/table_properties_collector.cc",
+      "db/transaction_log_impl.cc",
+      "db/version_builder.cc",
+      "db/version_edit.cc",
+      "db/version_set.cc",
+      "db/wal_manager.cc",
+      "db/write_batch.cc",
+      "db/write_batch_base.cc",
+      "db/write_controller.cc",
+      "db/write_thread.cc",
+      "env/env.cc",
+      "env/env_chroot.cc",
+      "env/env_encryption.cc",
+      "env/env_hdfs.cc",
+      "env/env_posix.cc",
+      "env/io_posix.cc",
+      "env/mock_env.cc",
+      "memtable/alloc_tracker.cc",
+      "memtable/hash_cuckoo_rep.cc",
+      "memtable/hash_linklist_rep.cc",
+      "memtable/hash_skiplist_rep.cc",
+      "memtable/skiplistrep.cc",
+      "memtable/vectorrep.cc",
+      "memtable/write_buffer_manager.cc",
+      "monitoring/histogram.cc",
+      "monitoring/histogram_windowing.cc",
+      "monitoring/instrumented_mutex.cc",
+      "monitoring/iostats_context.cc",
+      "monitoring/perf_context.cc",
+      "monitoring/perf_level.cc",
+      "monitoring/statistics.cc",
+      "monitoring/thread_status_impl.cc",
+      "monitoring/thread_status_updater.cc",
+      "monitoring/thread_status_updater_debug.cc",
+      "monitoring/thread_status_util.cc",
+      "monitoring/thread_status_util_debug.cc",
+      "options/cf_options.cc",
+      "options/db_options.cc",
+      "options/options.cc",
+      "options/options_helper.cc",
+      "options/options_parser.cc",
+      "options/options_sanity_check.cc",
+      "port/port_posix.cc",
+      "port/stack_trace.cc",
+      "table/adaptive_table_factory.cc",
+      "table/block.cc",
+      "table/block_based_filter_block.cc",
+      "table/block_based_table_builder.cc",
+      "table/block_based_table_factory.cc",
+      "table/block_based_table_reader.cc",
+      "table/block_builder.cc",
+      "table/block_prefix_index.cc",
+      "table/bloom_block.cc",
+      "table/cuckoo_table_builder.cc",
+      "table/cuckoo_table_factory.cc",
+      "table/cuckoo_table_reader.cc",
+      "table/flush_block_policy.cc",
+      "table/format.cc",
+      "table/full_filter_block.cc",
+      "table/get_context.cc",
+      "table/index_builder.cc",
+      "table/iterator.cc",
+      "table/merging_iterator.cc",
+      "table/meta_blocks.cc",
+      "table/partitioned_filter_block.cc",
+      "table/persistent_cache_helper.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/sst_file_writer.cc",
+      "table/table_properties.cc",
+      "table/two_level_iterator.cc",
+      "tools/dump/db_dump_tool.cc",
+      "util/arena.cc",
+      "util/auto_roll_logger.cc",
+      "util/bloom.cc",
+      "util/build_version.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/dynamic_bloom.cc",
+      "util/event_logger.cc",
+      "util/file_reader_writer.cc",
+      "util/file_util.cc",
+      "util/filename.cc",
+      "util/filter_policy.cc",
+      "util/hash.cc",
+      "util/log_buffer.cc",
+      "util/murmurhash.cc",
+      "util/random.cc",
+      "util/rate_limiter.cc",
+      "util/slice.cc",
+      "util/sst_file_manager_impl.cc",
+      "util/status.cc",
+      "util/status_message.cc",
+      "util/string_util.cc",
+      "util/sync_point.cc",
+      "util/thread_local.cc",
+      "util/threadpool_imp.cc",
+      "util/transaction_test_util.cc",
+      "util/xxhash.cc",
+      "utilities/backupable/backupable_db.cc",
+      "utilities/blob_db/blob_db.cc",
+      "utilities/blob_db/blob_db_impl.cc",
+      "utilities/blob_db/blob_file.cc",
+      "utilities/blob_db/blob_log_reader.cc",
+      "utilities/blob_db/blob_log_writer.cc",
+      "utilities/blob_db/blob_log_format.cc",
+      "utilities/blob_db/ttl_extractor.cc",
+      "utilities/cassandra/cassandra_compaction_filter.cc",
+      "utilities/cassandra/format.cc",
+      "utilities/cassandra/merge_operator.cc",
+      "utilities/checkpoint/checkpoint_impl.cc",
+      "utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc",
+      "utilities/convenience/info_log_finder.cc",
+      "utilities/date_tiered/date_tiered_db_impl.cc",
+      "utilities/debug.cc",
+      "utilities/document/document_db.cc",
+      "utilities/document/json_document.cc",
+      "utilities/document/json_document_builder.cc",
+      "utilities/env_mirror.cc",
+      "utilities/env_timed.cc",
+      "utilities/geodb/geodb_impl.cc",
+      "utilities/leveldb_options/leveldb_options.cc",
+      "utilities/lua/rocks_lua_compaction_filter.cc",
+      "utilities/memory/memory_util.cc",
+      "utilities/merge_operators/max.cc",
+      "utilities/merge_operators/put.cc",
+      "utilities/merge_operators/string_append/stringappend.cc",
+      "utilities/merge_operators/string_append/stringappend2.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_db_impl.cc",
+      "utilities/transactions/optimistic_transaction.cc",
+      "utilities/transactions/transaction_base.cc",
+      "utilities/transactions/pessimistic_transaction_db.cc",
+      "utilities/transactions/transaction_db_mutex_impl.cc",
+      "utilities/transactions/pessimistic_transaction.cc",
+      "utilities/transactions/transaction_lock_mgr.cc",
+      "utilities/transactions/transaction_util.cc",
+      "utilities/transactions/write_prepared_txn.cc",
+      "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",
+      "tools/ldb_cmd.cc",
+      "tools/ldb_tool.cc",
+      "tools/sst_dump_tool.cc",
+      "utilities/blob_db/blob_dump_tool.cc",
+    ],
+    deps = [],
+    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+    compiler_flags = rocksdb_compiler_flags,
+    external_deps = rocksdb_external_deps,
+)
+
+cpp_library(
+    name = "rocksdb_test_lib",
+    headers = AutoHeaders.RECURSIVE_GLOB,
+    srcs = [
+      "table/mock_table.cc",
+      "util/fault_injection_test_env.cc",
+      "util/testharness.cc",
+      "util/testutil.cc",
+      "db/db_test_util.cc",
+      "utilities/cassandra/test_utils.cc",
+      "utilities/col_buf_encoder.cc",
+      "utilities/col_buf_decoder.cc",
+      "utilities/column_aware_encoding_util.cc",
+    ],
+    deps = [":rocksdb_lib"],
+    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+    compiler_flags = rocksdb_compiler_flags,
+    external_deps = rocksdb_external_deps,
+)
+
+cpp_library(
+    name = "rocksdb_tools_lib",
+    headers = AutoHeaders.RECURSIVE_GLOB,
+    srcs = [
+      "tools/db_bench_tool.cc",
+      "util/testutil.cc",
+    ],
+    deps = [":rocksdb_lib"],
+    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+    compiler_flags = rocksdb_compiler_flags,
+    external_deps = rocksdb_external_deps,
+)
+
+cpp_library(
+    name = "env_basic_test_lib",
+    headers = AutoHeaders.RECURSIVE_GLOB,
+    srcs = ["env/env_basic_test.cc"],
+    deps = [":rocksdb_test_lib"],
+    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+    compiler_flags = rocksdb_compiler_flags,
+    external_deps = rocksdb_external_deps,
+)
+
+# [test_name, test_src, test_type]
+ROCKS_TESTS = [['arena_test', 'util/arena_test.cc', 'serial'],
+ ['auto_roll_logger_test', 'util/auto_roll_logger_test.cc', 'serial'],
+ ['autovector_test', 'util/autovector_test.cc', 'serial'],
+ ['backupable_db_test',
+  'utilities/backupable/backupable_db_test.cc',
+  'parallel'],
+ ['blob_db_test', 'utilities/blob_db/blob_db_test.cc', 'serial'],
+ ['block_based_filter_block_test',
+  'table/block_based_filter_block_test.cc',
+  'serial'],
+ ['block_test', 'table/block_test.cc', 'serial'],
+ ['bloom_test', 'util/bloom_test.cc', 'serial'],
+ ['c_test', 'db/c_test.c', 'serial'],
+ ['cache_test', 'cache/cache_test.cc', 'serial'],
+ ['cassandra_format_test',
+  'utilities/cassandra/cassandra_format_test.cc',
+  'serial'],
+ ['cassandra_functional_test',
+  'utilities/cassandra/cassandra_functional_test.cc',
+  'serial'],
+ ['cassandra_row_merge_test',
+  'utilities/cassandra/cassandra_row_merge_test.cc',
+  'serial'],
+ ['cassandra_serialize_test',
+  'utilities/cassandra/cassandra_serialize_test.cc',
+  'serial'],
+ ['checkpoint_test', 'utilities/checkpoint/checkpoint_test.cc', 'serial'],
+ ['cleanable_test', 'table/cleanable_test.cc', 'serial'],
+ ['coding_test', 'util/coding_test.cc', 'serial'],
+ ['column_aware_encoding_test',
+  'utilities/column_aware_encoding_test.cc',
+  'serial'],
+ ['column_family_test', 'db/column_family_test.cc', 'serial'],
+ ['compact_files_test', 'db/compact_files_test.cc', 'serial'],
+ ['compact_on_deletion_collector_test',
+  'utilities/table_properties_collectors/compact_on_deletion_collector_test.cc',
+  'serial'],
+ ['compaction_iterator_test', 'db/compaction_iterator_test.cc', 'serial'],
+ ['compaction_job_stats_test', 'db/compaction_job_stats_test.cc', 'serial'],
+ ['compaction_job_test', 'db/compaction_job_test.cc', 'serial'],
+ ['compaction_picker_test', 'db/compaction_picker_test.cc', 'serial'],
+ ['comparator_db_test', 'db/comparator_db_test.cc', 'serial'],
+ ['corruption_test', 'db/corruption_test.cc', 'serial'],
+ ['crc32c_test', 'util/crc32c_test.cc', 'serial'],
+ ['cuckoo_table_builder_test', 'table/cuckoo_table_builder_test.cc', 'serial'],
+ ['cuckoo_table_db_test', 'db/cuckoo_table_db_test.cc', 'serial'],
+ ['cuckoo_table_reader_test', 'table/cuckoo_table_reader_test.cc', 'serial'],
+ ['date_tiered_test', 'utilities/date_tiered/date_tiered_test.cc', 'serial'],
+ ['db_basic_test', 'db/db_basic_test.cc', 'serial'],
+ ['db_block_cache_test', 'db/db_block_cache_test.cc', 'serial'],
+ ['db_bloom_filter_test', 'db/db_bloom_filter_test.cc', 'serial'],
+ ['db_compaction_filter_test', 'db/db_compaction_filter_test.cc', 'parallel'],
+ ['db_compaction_test', 'db/db_compaction_test.cc', 'parallel'],
+ ['db_dynamic_level_test', 'db/db_dynamic_level_test.cc', 'serial'],
+ ['db_encryption_test', 'db/db_encryption_test.cc', 'serial'],
+ ['db_flush_test', 'db/db_flush_test.cc', 'serial'],
+ ['db_inplace_update_test', 'db/db_inplace_update_test.cc', 'serial'],
+ ['db_io_failure_test', 'db/db_io_failure_test.cc', 'serial'],
+ ['db_iter_test', 'db/db_iter_test.cc', 'serial'],
+ ['db_iterator_test', 'db/db_iterator_test.cc', 'serial'],
+ ['db_log_iter_test', 'db/db_log_iter_test.cc', 'serial'],
+ ['db_memtable_test', 'db/db_memtable_test.cc', 'serial'],
+ ['db_merge_operator_test', 'db/db_merge_operator_test.cc', 'serial'],
+ ['db_options_test', 'db/db_options_test.cc', 'serial'],
+ ['db_properties_test', 'db/db_properties_test.cc', 'serial'],
+ ['db_range_del_test', 'db/db_range_del_test.cc', 'serial'],
+ ['db_sst_test', 'db/db_sst_test.cc', 'parallel'],
+ ['db_statistics_test', 'db/db_statistics_test.cc', 'serial'],
+ ['db_table_properties_test', 'db/db_table_properties_test.cc', 'serial'],
+ ['db_tailing_iter_test', 'db/db_tailing_iter_test.cc', 'serial'],
+ ['db_test', 'db/db_test.cc', 'parallel'],
+ ['db_test2', 'db/db_test2.cc', 'serial'],
+ ['db_universal_compaction_test',
+  'db/db_universal_compaction_test.cc',
+  'parallel'],
+ ['db_wal_test', 'db/db_wal_test.cc', 'parallel'],
+ ['db_write_test', 'db/db_write_test.cc', 'serial'],
+ ['dbformat_test', 'db/dbformat_test.cc', 'serial'],
+ ['delete_scheduler_test', 'util/delete_scheduler_test.cc', 'serial'],
+ ['deletefile_test', 'db/deletefile_test.cc', 'serial'],
+ ['document_db_test', 'utilities/document/document_db_test.cc', 'serial'],
+ ['dynamic_bloom_test', 'util/dynamic_bloom_test.cc', 'serial'],
+ ['env_basic_test', 'env/env_basic_test.cc', 'serial'],
+ ['env_test', 'env/env_test.cc', 'serial'],
+ ['env_timed_test', 'utilities/env_timed_test.cc', 'serial'],
+ ['event_logger_test', 'util/event_logger_test.cc', 'serial'],
+ ['external_sst_file_basic_test',
+  'db/external_sst_file_basic_test.cc',
+  'serial'],
+ ['external_sst_file_test', 'db/external_sst_file_test.cc', 'parallel'],
+ ['fault_injection_test', 'db/fault_injection_test.cc', 'parallel'],
+ ['file_indexer_test', 'db/file_indexer_test.cc', 'serial'],
+ ['file_reader_writer_test', 'util/file_reader_writer_test.cc', 'serial'],
+ ['filelock_test', 'util/filelock_test.cc', 'serial'],
+ ['filename_test', 'db/filename_test.cc', 'serial'],
+ ['flush_job_test', 'db/flush_job_test.cc', 'serial'],
+ ['full_filter_block_test', 'table/full_filter_block_test.cc', 'serial'],
+ ['geodb_test', 'utilities/geodb/geodb_test.cc', 'serial'],
+ ['hash_table_test',
+  'utilities/persistent_cache/hash_table_test.cc',
+  'serial'],
+ ['hash_test', 'util/hash_test.cc', 'serial'],
+ ['heap_test', 'util/heap_test.cc', 'serial'],
+ ['histogram_test', 'monitoring/histogram_test.cc', 'serial'],
+ ['inlineskiplist_test', 'memtable/inlineskiplist_test.cc', 'parallel'],
+ ['iostats_context_test', 'monitoring/iostats_context_test.cc', 'serial'],
+ ['json_document_test', 'utilities/document/json_document_test.cc', 'serial'],
+ ['ldb_cmd_test', 'tools/ldb_cmd_test.cc', 'serial'],
+ ['listener_test', 'db/listener_test.cc', 'serial'],
+ ['log_test', 'db/log_test.cc', 'serial'],
+ ['lru_cache_test', 'cache/lru_cache_test.cc', 'serial'],
+ ['manual_compaction_test', 'db/manual_compaction_test.cc', 'parallel'],
+ ['memory_test', 'utilities/memory/memory_test.cc', 'serial'],
+ ['memtable_list_test', 'db/memtable_list_test.cc', 'serial'],
+ ['merge_helper_test', 'db/merge_helper_test.cc', 'serial'],
+ ['merge_test', 'db/merge_test.cc', 'serial'],
+ ['merger_test', 'table/merger_test.cc', 'serial'],
+ ['mock_env_test', 'env/mock_env_test.cc', 'serial'],
+ ['object_registry_test', 'utilities/object_registry_test.cc', 'serial'],
+ ['optimistic_transaction_test',
+  'utilities/transactions/optimistic_transaction_test.cc',
+  'serial'],
+ ['option_change_migration_test',
+  'utilities/option_change_migration/option_change_migration_test.cc',
+  'serial'],
+ ['options_file_test', 'db/options_file_test.cc', 'serial'],
+ ['options_settable_test', 'options/options_settable_test.cc', 'serial'],
+ ['options_test', 'options/options_test.cc', 'serial'],
+ ['options_util_test', 'utilities/options/options_util_test.cc', 'serial'],
+ ['partitioned_filter_block_test',
+  'table/partitioned_filter_block_test.cc',
+  'serial'],
+ ['perf_context_test', 'db/perf_context_test.cc', 'serial'],
+ ['persistent_cache_test',
+  'utilities/persistent_cache/persistent_cache_test.cc',
+  'parallel'],
+ ['plain_table_db_test', 'db/plain_table_db_test.cc', 'serial'],
+ ['prefix_test', 'db/prefix_test.cc', 'serial'],
+ ['range_del_aggregator_test', 'db/range_del_aggregator_test.cc', 'serial'],
+ ['rate_limiter_test', 'util/rate_limiter_test.cc', 'serial'],
+ ['reduce_levels_test', 'tools/reduce_levels_test.cc', 'serial'],
+ ['repair_test', 'db/repair_test.cc', 'serial'],
+ ['sim_cache_test', 'utilities/simulator_cache/sim_cache_test.cc', 'serial'],
+ ['skiplist_test', 'memtable/skiplist_test.cc', 'serial'],
+ ['slice_transform_test', 'util/slice_transform_test.cc', 'serial'],
+ ['spatial_db_test', 'utilities/spatialdb/spatial_db_test.cc', 'serial'],
+ ['sst_dump_test', 'tools/sst_dump_test.cc', 'serial'],
+ ['statistics_test', 'monitoring/statistics_test.cc', 'serial'],
+ ['stringappend_test',
+  'utilities/merge_operators/string_append/stringappend_test.cc',
+  'serial'],
+ ['table_properties_collector_test',
+  'db/table_properties_collector_test.cc',
+  'serial'],
+ ['table_test', 'table/table_test.cc', 'parallel'],
+ ['thread_list_test', 'util/thread_list_test.cc', 'serial'],
+ ['thread_local_test', 'util/thread_local_test.cc', 'serial'],
+ ['timer_queue_test', 'util/timer_queue_test.cc', 'serial'],
+ ['transaction_test', 'utilities/transactions/transaction_test.cc', 'serial'],
+ ['ttl_test', 'utilities/ttl/ttl_test.cc', 'serial'],
+ ['util_merge_operators_test',
+  'utilities/util_merge_operators_test.cc',
+  'serial'],
+ ['version_builder_test', 'db/version_builder_test.cc', 'serial'],
+ ['version_edit_test', 'db/version_edit_test.cc', 'serial'],
+ ['version_set_test', 'db/version_set_test.cc', 'serial'],
+ ['wal_manager_test', 'db/wal_manager_test.cc', 'serial'],
+ ['write_batch_test', 'db/write_batch_test.cc', 'serial'],
+ ['write_batch_with_index_test',
+  'utilities/write_batch_with_index/write_batch_with_index_test.cc',
+  'serial'],
+ ['write_buffer_manager_test',
+  'memtable/write_buffer_manager_test.cc',
+  'serial'],
+ ['write_callback_test', 'db/write_callback_test.cc', 'serial'],
+ ['write_controller_test', 'db/write_controller_test.cc', 'serial']]
+
+
+# Generate a test rule for each entry in ROCKS_TESTS
+for test_cfg in ROCKS_TESTS:
+    test_name = test_cfg[0]
+    test_cc = test_cfg[1]
+    ttype = "gtest" if test_cfg[2] == "parallel" else "simple"
+    test_bin = test_name + "_bin"
+
+    cpp_binary (
+      name = test_bin,
+      srcs = [test_cc],
+      deps = [":rocksdb_test_lib"],
+      preprocessor_flags = rocksdb_preprocessor_flags,
+      arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+      compiler_flags = rocksdb_compiler_flags,
+      external_deps = rocksdb_external_deps,
+    )
+
+    custom_unittest(
+      name = test_name,
+      type = ttype,
+      deps = [":" + test_bin],
+      command = [TEST_RUNNER, BUCK_BINS + test_bin]
+    )
+
+custom_unittest(
+    name = "make_rocksdbjavastatic",
+    type = "simple",
+    command = ["internal_repo_rocksdb/make_rocksdbjavastatic.sh"],
+)
+
+custom_unittest(
+    name = "make_rocksdb_lite_release",
+    type = "simple",
+    command = ["internal_repo_rocksdb/make_rocksdb_lite_release.sh"],
+)

USERS.md

@@ -76,7 +76,10 @@ quasardb uses a heavily tuned RocksDB as its persistence layer.
 [Dgraph](https://github.com/dgraph-io/dgraph) is an open-source, scalable, distributed, low latency, high throughput Graph database .They use RocksDB to store state locally on a machine.
 
 ## Uber
-[Uber](http://eng.uber.com/cherami/) uses RocksDB as a durable and and scalable task queue.
+[Uber](http://eng.uber.com/cherami/) uses RocksDB as a durable and scalable task queue.
 
 ## 360 Pika
 [360](http://www.360.cn/) [Pika](https://github.com/Qihoo360/pika) is a nosql compatible with redis. With the huge amount of data stored, redis may suffer for a capacity bottleneck, and pika was born for solving it. It has widely been widely used in many company
+
+## LzLabs
+LzLabs is using RocksDB as a storage engine in their multi-database distributed framework to store application configuration and user data.

appveyor.yml

@@ -11,8 +11,5 @@ build:
   verbosity: minimal
 test:
 test_script:
-- ps: build_tools\run_ci_db_test.ps1 -Run db_basic_test -Concurrency 8
-- ps: build_tools\run_ci_db_test.ps1 -Run db_test2 -Concurrency 8
-- ps: build_tools\run_ci_db_test.ps1 -Run db_test -Concurrency 8
-- ps: build_tools\run_ci_db_test.ps1 -Run env_test,env_basic_test -Concurrency 2
+- ps: build_tools\run_ci_db_test.ps1 -SuiteRun db_basic_test,db_test2,db_test,env_basic_test,env_test -Concurrency 8
 

buckifier/buckify_rocksdb.py

@@ -0,0 +1,172 @@
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+from __future__ import unicode_literals
+from targets_builder import TARGETSBuilder
+from optparse import OptionParser
+import os
+import fnmatch
+import sys
+import tempfile
+
+from util import ColorString
+import util
+
+# tests to export as libraries for inclusion in other projects
+_EXPORTED_TEST_LIBS = ["env_basic_test"]
+
+# Parse src.mk files as a Dictionary of
+# VAR_NAME => list of files
+def parse_src_mk(repo_path):
+    src_mk = repo_path + "/src.mk"
+    src_files = {}
+    for line in open(src_mk):
+        line = line.strip()
+        if len(line) == 0 or line[0] == '#':
+            continue
+        if '=' in line:
+            current_src = line.split('=')[0].strip()
+            src_files[current_src] = []
+        elif '.cc' in line:
+            src_path = line.split('.cc')[0].strip() + '.cc'
+            src_files[current_src].append(src_path)
+    return src_files
+
+
+# get all .cc / .c files
+def get_cc_files(repo_path):
+    cc_files = []
+    for root, dirnames, filenames in os.walk(repo_path):
+        root = root[(len(repo_path) + 1):]
+        if "java" in root:
+            # Skip java
+            continue
+        for filename in fnmatch.filter(filenames, '*.cc'):
+            cc_files.append(os.path.join(root, filename))
+        for filename in fnmatch.filter(filenames, '*.c'):
+            cc_files.append(os.path.join(root, filename))
+    return cc_files
+
+
+# Get tests from Makefile
+def get_tests(repo_path):
+    Makefile = repo_path + "/Makefile"
+
+    # Dictionary TEST_NAME => IS_PARALLEL
+    tests = {}
+
+    found_tests = False
+    for line in open(Makefile):
+        line = line.strip()
+        if line.startswith("TESTS ="):
+            found_tests = True
+        elif found_tests:
+            if line.endswith("\\"):
+                # remove the trailing \
+                line = line[:-1]
+                line = line.strip()
+                tests[line] = False
+            else:
+                # we consumed all the tests
+                break
+
+    found_parallel_tests = False
+    for line in open(Makefile):
+        line = line.strip()
+        if line.startswith("PARALLEL_TEST ="):
+            found_parallel_tests = True
+        elif found_parallel_tests:
+            if line.endswith("\\"):
+                # remove the trailing \
+                line = line[:-1]
+                line = line.strip()
+                tests[line] = True
+            else:
+                # we consumed all the parallel tests
+                break
+    
+    return tests
+
+
+# Prepare TARGETS file for buck
+def generate_targets(repo_path):
+    print(ColorString.info("Generating TARGETS"))
+    # parsed src.mk file
+    src_mk = parse_src_mk(repo_path)
+    # get all .cc files
+    cc_files = get_cc_files(repo_path)
+    # get tests from Makefile
+    tests = get_tests(repo_path)
+
+    if src_mk is None or cc_files is None or tests is None:
+        return False
+
+    TARGETS = TARGETSBuilder("%s/TARGETS" % repo_path)
+    # rocksdb_lib
+    TARGETS.add_library(
+        "rocksdb_lib",
+        src_mk["LIB_SOURCES"] +
+        src_mk["TOOL_LIB_SOURCES"])
+    # rocksdb_test_lib
+    TARGETS.add_library(
+        "rocksdb_test_lib",
+        src_mk.get("MOCK_LIB_SOURCES", []) +
+        src_mk.get("TEST_LIB_SOURCES", []) +
+        src_mk.get("EXP_LIB_SOURCES", []),
+        [":rocksdb_lib"])
+    # rocksdb_tools_lib
+    TARGETS.add_library(
+        "rocksdb_tools_lib",
+        src_mk.get("BENCH_LIB_SOURCES", []) +
+        ["util/testutil.cc"],
+        [":rocksdb_lib"])
+
+    # test for every test we found in the Makefile
+    for test in sorted(tests):
+        match_src = [src for src in cc_files if ("/%s.c" % test) in src]
+        if len(match_src) == 0:
+            print(ColorString.warning("Cannot find .cc file for %s" % test))
+            continue
+        elif len(match_src) > 1:
+            print(ColorString.warning("Found more than one .cc for %s" % test))
+            print(match_src)
+            continue
+
+        assert(len(match_src) == 1)
+        is_parallel = tests[test]
+        TARGETS.register_test(test, match_src[0], is_parallel)
+
+        if test in _EXPORTED_TEST_LIBS:
+            test_library = "%s_lib" % test
+            TARGETS.add_library(test_library, match_src, [":rocksdb_test_lib"])
+    TARGETS.flush_tests()
+
+    print(ColorString.info("Generated TARGETS Summary:"))
+    print(ColorString.info("- %d libs" % TARGETS.total_lib))
+    print(ColorString.info("- %d binarys" % TARGETS.total_bin))
+    print(ColorString.info("- %d tests" % TARGETS.total_test))
+    return True
+
+
+def get_rocksdb_path():
+    # rocksdb = {script_dir}/..
+    script_dir = os.path.dirname(sys.argv[0])
+    script_dir = os.path.abspath(script_dir)
+    rocksdb_path = os.path.abspath(
+        os.path.join(script_dir, "../"))
+
+    return rocksdb_path
+
+def exit_with_error(msg):
+    print(ColorString.error(msg))
+    sys.exit(1)
+
+
+def main():
+    # Generate TARGETS file for buck
+    ok = generate_targets(get_rocksdb_path())
+    if not ok:
+        exit_with_error("Failed to generate TARGETS files")
+
+if __name__ == "__main__":
+    main()

buckifier/rocks_test_runner.sh

@@ -0,0 +1,4 @@
+#!/usr/bin/env bash
+# Create a tmp directory for the test to use
+TEST_DIR=$(mktemp -d /dev/shm/fbcode_rocksdb_XXXXXXX)
+TEST_TMPDIR="$TEST_DIR" $@ && rm -rf "$TEST_DIR"

buckifier/targets_builder.py

@@ -0,0 +1,65 @@
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+from __future__ import unicode_literals
+import targets_cfg
+import pprint
+
+# TODO(tec): replace this with PrettyPrinter
+def pretty_list(lst, indent=6):
+    if lst is None or len(lst) == 0:
+        return ""
+
+    if len(lst) == 1:
+        return "\"%s\"" % lst[0]
+
+    separator = "\",\n%s\"" % (" " * indent)
+    res = separator.join(lst)
+    res = "\n" + (" " * indent) + "\"" + res + "\",\n" + (" " * (indent - 2))
+    return res
+
+
+class TARGETSBuilder:
+    def __init__(self, path):
+        self.path = path
+        self.targets_file = open(path, 'w')
+        self.targets_file.write(targets_cfg.rocksdb_target_header)
+        self.total_lib = 0
+        self.total_bin = 0
+        self.total_test = 0
+        self.tests_cfg = []
+
+    def __del__(self):
+        self.targets_file.close()
+
+    def add_library(self, name, srcs, deps=None, headers=None):
+        if headers is None:
+            headers = "AutoHeaders.RECURSIVE_GLOB"
+        self.targets_file.write(targets_cfg.library_template % (
+            name,
+            headers,
+            pretty_list(srcs),
+            pretty_list(deps)))
+        self.total_lib = self.total_lib + 1
+
+    def add_binary(self, name, srcs, deps=None):
+        self.targets_file.write(targets_cfg.binary_template % (
+            name,
+            pretty_list(srcs),
+            pretty_list(deps)))
+        self.total_bin = self.total_bin + 1
+
+    def register_test(self, test_name, src, is_parallel):
+        exec_mode = "serial"
+        if is_parallel:
+            exec_mode = "parallel"
+        self.tests_cfg.append([test_name, str(src), str(exec_mode)])
+
+        self.total_test = self.total_test + 1
+
+    def flush_tests(self):
+        self.targets_file.write(targets_cfg.unittests_template % (
+            pprint.PrettyPrinter().pformat(self.tests_cfg)
+        ))
+
+        self.tests_cfg = []

buckifier/targets_cfg.py

@@ -0,0 +1,124 @@
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+from __future__ import unicode_literals
+rocksdb_target_header = """
+import os
+
+TARGETS_PATH = os.path.dirname(__file__)
+REPO_PATH = "rocksdb/src/"
+BUCK_BINS = "buck-out/gen/" + REPO_PATH
+TEST_RUNNER = REPO_PATH + "buckifier/rocks_test_runner.sh"
+rocksdb_compiler_flags = [
+  "-fno-builtin-memcmp",
+  "-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",
+  "-DOS_LINUX",
+  # Flags to enable libs we include
+  "-DSNAPPY",
+  "-DZLIB",
+  "-DBZIP2",
+  "-DLZ4",
+  "-DZSTD",
+  "-DGFLAGS=gflags",
+  "-DNUMA",
+  "-DTBB",
+  # Needed to compile in fbcode
+  "-Wno-expansion-to-defined",
+]
+
+rocksdb_external_deps = [
+  ('bzip2', None, 'bz2'),
+  ('snappy', None, "snappy"),
+  ('zlib', None, 'z'),
+  ('gflags', None, 'gflags'),
+  ('lz4', None, 'lz4'),
+  ('zstd', None),
+  ('tbb', None),
+  ("numa", None, "numa"),
+  ("googletest", None, "gtest"),
+]
+
+rocksdb_preprocessor_flags = [
+  # Directories with files for #include
+  "-I" + REPO_PATH + "include/",
+  "-I" + REPO_PATH,
+]
+
+rocksdb_arch_preprocessor_flags = {
+  "x86_64": ["-DHAVE_SSE42"],
+}
+"""
+
+
+library_template = """
+cpp_library(
+    name = "%s",
+    headers = %s,
+    srcs = [%s],
+    deps = [%s],
+    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+    compiler_flags = rocksdb_compiler_flags,
+    external_deps = rocksdb_external_deps,
+)
+"""
+
+binary_template = """
+cpp_binary(
+  name = "%s",
+  srcs = [%s],
+  deps = [%s],
+  preprocessor_flags = rocksdb_preprocessor_flags,
+  arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+  compiler_flags = rocksdb_compiler_flags,
+  external_deps = rocksdb_external_deps,
+)
+"""
+
+unittests_template = """
+# [test_name, test_src, test_type]
+ROCKS_TESTS = %s
+
+
+# Generate a test rule for each entry in ROCKS_TESTS
+for test_cfg in ROCKS_TESTS:
+    test_name = test_cfg[0]
+    test_cc = test_cfg[1]
+    ttype = "gtest" if test_cfg[2] == "parallel" else "simple"
+    test_bin = test_name + "_bin"
+
+    cpp_binary (
+      name = test_bin,
+      srcs = [test_cc],
+      deps = [":rocksdb_test_lib"],
+      preprocessor_flags = rocksdb_preprocessor_flags,
+      arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
+      compiler_flags = rocksdb_compiler_flags,
+      external_deps = rocksdb_external_deps,
+    )
+
+    custom_unittest(
+      name = test_name,
+      type = ttype,
+      deps = [":" + test_bin],
+      command = [TEST_RUNNER, BUCK_BINS + test_bin]
+    )
+
+custom_unittest(
+    name = "make_rocksdbjavastatic",
+    type = "simple",
+    command = ["internal_repo_rocksdb/make_rocksdbjavastatic.sh"],
+)
+
+custom_unittest(
+    name = "make_rocksdb_lite_release",
+    type = "simple",
+    command = ["internal_repo_rocksdb/make_rocksdb_lite_release.sh"],
+)
+"""

buckifier/util.py

@@ -0,0 +1,107 @@
+"""
+This module keeps commonly used components.
+"""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+from __future__ import unicode_literals
+import subprocess
+import os
+import time
+
+class ColorString:
+    """ Generate colorful strings on terminal """
+    HEADER = '\033[95m'
+    BLUE = '\033[94m'
+    GREEN = '\033[92m'
+    WARNING = '\033[93m'
+    FAIL = '\033[91m'
+    ENDC = '\033[0m'
+
+    @staticmethod
+    def _make_color_str(text, color):
+        return "".join([color, text.encode('utf-8'), ColorString.ENDC])
+
+    @staticmethod
+    def ok(text):
+        if ColorString.is_disabled:
+            return text
+        return ColorString._make_color_str(text, ColorString.GREEN)
+
+    @staticmethod
+    def info(text):
+        if ColorString.is_disabled:
+            return text
+        return ColorString._make_color_str(text, ColorString.BLUE)
+
+    @staticmethod
+    def header(text):
+        if ColorString.is_disabled:
+            return text
+        return ColorString._make_color_str(text, ColorString.HEADER)
+
+    @staticmethod
+    def error(text):
+        if ColorString.is_disabled:
+            return text
+        return ColorString._make_color_str(text, ColorString.FAIL)
+
+    @staticmethod
+    def warning(text):
+        if ColorString.is_disabled:
+            return text
+        return ColorString._make_color_str(text, ColorString.WARNING)
+
+    is_disabled = False
+
+
+def run_shell_command(shell_cmd, cmd_dir=None):
+    """ Run a single shell command.
+        @returns a tuple of shell command return code, stdout, stderr """
+
+    if cmd_dir is not None and not os.path.exists(cmd_dir):
+        run_shell_command("mkdir -p %s" % cmd_dir)
+
+    start = time.time()
+    print("\t>>> Running: " + shell_cmd)
+    p = subprocess.Popen(shell_cmd,
+                         shell=True,
+                         stdout=subprocess.PIPE,
+                         stderr=subprocess.PIPE,
+                         cwd=cmd_dir)
+    stdout, stderr = p.communicate()
+    end = time.time()
+
+    # Report time if we spent more than 5 minutes executing a command
+    execution_time = end - start
+    if execution_time > (60 * 5):
+        mins = (execution_time / 60)
+        secs = (execution_time % 60)
+        print("\t>time spent: %d minutes %d seconds" % (mins, secs))
+
+
+    return p.returncode, stdout, stderr
+
+
+def run_shell_commands(shell_cmds, cmd_dir=None, verbose=False):
+    """ Execute a sequence of shell commands, which is equivalent to
+        running `cmd1 && cmd2 && cmd3`
+        @returns boolean indication if all commands succeeds.
+    """
+
+    if cmd_dir:
+        print("\t=== Set current working directory => %s" % cmd_dir)
+
+    for shell_cmd in shell_cmds:
+        ret_code, stdout, stderr = run_shell_command(shell_cmd, cmd_dir)
+        if stdout:
+            if verbose or ret_code != 0:
+                print(ColorString.info("stdout: \n"), stdout)
+        if stderr:
+            # contents in stderr is not necessarily to be error messages.
+            if verbose or ret_code != 0:
+                print(ColorString.error("stderr: \n"), stderr)
+        if ret_code != 0:
+            return False
+
+    return True

build_tools/RocksDBCommonHelper.php

@@ -0,0 +1,377 @@
+<?php
+// Copyright 2004-present Facebook. All Rights Reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root 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("CONT_RUN_ALIAS", "leveldb");
+
+//////////////////////////////////////////////////////////////////////
+/*  Run tests in sandcastle */
+function postURL($diffID, $url) {
+  assert(strlen($diffID) > 0);
+  assert(is_numeric($diffID));
+  assert(strlen($url) > 0);
+
+  $cmd_args = array(
+    'diff_id' => (int)$diffID,
+    'name' => sprintf(
+      'click here for sandcastle tests for D%d',
+      (int)$diffID
+    ),
+    'link' => $url
+  );
+  $cmd = 'echo ' . escapeshellarg(json_encode($cmd_args))
+         . ' | 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);
+
+  $cmd_args = array(
+    'diff_id' => (int)$diffID,
+    'name' => $test,
+    'result' => $status
+  );
+
+  $cmd = 'echo ' . escapeshellarg(json_encode($cmd_args))
+         . ' | 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.
+    $setup = array(
+      "name" => "Setup arcrc",
+      "shell" => "echo " . escapeshellarg($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"
+  );
+
+  // This fixes "FATAL: ThreadSanitizer can not mmap the shadow memory"
+  // Source:
+  // https://github.com/google/sanitizers/wiki/ThreadSanitizerCppManual#FAQ
+  $fix_kernel_issue = array(
+    "name" => "Fix kernel issue with tsan",
+    "shell" => "echo 2 >/proc/sys/kernel/randomize_va_space",
+    "user" => "root"
+  );
+
+  $steps[] = $fix_git_ignore;
+  $steps[] = $fix_kernel_issue;
+
+  // 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 " . $diffID,
+      "shell" => "arc --arcrc-file ~/.arcrc "
+                  . "patch --nocommit --diff " . escapeshellarg($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 " .
+           escapeshellarg($test) . "; exit_code=$?; ";
+
+  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" => $cmd,
+    "user" => "root",
+    "parser" => "python build_tools/error_filter.py " . escapeshellarg($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();
+
+  $cwd = getcwd();
+  $cwd_token_file = "{$cwd}/.sandcastle";
+  // 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);
+    $sandcastle_config[1] = getenv(ENV_HTTPS_TOKEN_VALUE);
+  } else {
+    // This is a typical `[p]arc diff` case. Fetch the values from the specific
+    // configuration files.
+    for ($i = 0; $i < 50; $i++) {
+      if (file_exists(PRIMARY_TOKEN_FILE) ||
+          file_exists($cwd_token_file)) {
+        break;
+      }
+      // If we failed to fetch the tokens, sleep for 0.2 second and try again
+      usleep(200000);
+    }
+    assert(file_exists(PRIMARY_TOKEN_FILE) ||
+           file_exists($cwd_token_file));
+
+    // Try the primary location first, followed by a secondary.
+    if (file_exists(PRIMARY_TOKEN_FILE)) {
+      $cmd = 'cat ' . PRIMARY_TOKEN_FILE;
+    } else {
+      $cmd = 'cat ' . escapeshellarg($cwd_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", "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/create';
+
+  $job = array(
+    'command' => 'SandcastleUniversalCommand',
+    'args' => $command,
+    'capabilities' => array(
+      'vcs' => 'rocksdb-int-git',
+      'type' => 'lego',
+    ),
+    'hash' => 'origin/master',
+    'user' => $username,
+    'alias' => 'rocksdb-precommit',
+    'tags' => array('rocksdb'),
+    'description' => 'Rocksdb precommit job',
+  );
+
+  // Fetch the configuration necessary to submit a successful HTTPS request.
+  $sandcastle_config = getSandcastleConfig();
+
+  $app = $sandcastle_config[0];
+  $token = $sandcastle_config[1];
+
+  $cmd = 'curl -s -k '
+          . ' -F app=' . escapeshellarg($app)
+          . ' -F token=' . escapeshellarg($token)
+          . ' -F job=' . escapeshellarg(json_encode($job))
+          .' ' . escapeshellarg($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 */);
+}

build_tools/build_detect_platform

@@ -51,7 +51,12 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
     FBCODE_BUILD="true"
     # If we're compiling with TSAN we need pic build
     PIC_BUILD=$COMPILE_WITH_TSAN
-    source "$PWD/build_tools/fbcode_config.sh"
+    if [ -z "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
+      source "$PWD/build_tools/fbcode_config.sh"
+    else
+      # we need this to build with MySQL. Don't use for other purposes.
+      source "$PWD/build_tools/fbcode_config4.8.1.sh"
+    fi
 fi
 
 # Delete existing output, if it exists
@@ -123,10 +128,17 @@ case "$TARGET_OS" in
         ;;
     SunOS)
         PLATFORM=OS_SOLARIS
-        COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_SOLARIS"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt"
+        COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_SOLARIS -m64"
+        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt -static-libstdc++ -static-libgcc -m64"
         # PORT_FILES=port/sunos/sunos_specific.cc
         ;;
+    AIX)
+        PLATFORM=OS_AIX
+        CC=gcc
+        COMMON_FLAGS="$COMMON_FLAGS -maix64 -pthread -fno-builtin-memcmp -D_REENTRANT -DOS_AIX -D__STDC_FORMAT_MACROS"
+        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -pthread -lpthread -lrt -maix64 -static-libstdc++ -static-libgcc"
+        # PORT_FILES=port/aix/aix_specific.cc
+        ;;
     FreeBSD)
         PLATFORM=OS_FREEBSD
         COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_FREEBSD"
@@ -305,9 +317,11 @@ EOF
     # Test whether jemalloc is available
     if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null -ljemalloc \
       2>/dev/null; then
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ljemalloc"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -ljemalloc"
+        # This will enable some preprocessor identifiers in the Makefile
         JEMALLOC=1
+        # JEMALLOC can be enabled either using the flag (like here) or by
+        # providing direct link to the jemalloc library
+        WITH_JEMALLOC_FLAG=1
     else
         # jemalloc is not available. Let's try tcmalloc
         if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null \
@@ -377,6 +391,29 @@ EOF
     if [ "$?" = 0 ]; then
         PROFILING_FLAGS=-pg
     fi
+
+    # Test whether sync_file_range is supported for compatibility with an old glibc
+    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+      #include <fcntl.h>
+      int main() {
+        int fd = open("/dev/null", 0);
+        sync_file_range(fd, 0, 1024, SYNC_FILE_RANGE_WRITE);
+      }
+EOF
+    if [ "$?" = 0 ]; then
+        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_RANGESYNC_PRESENT"
+    fi
+
+    # Test whether sched_getcpu is supported
+    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+      #include <sched.h>
+      int main() {
+        int cpuid = sched_getcpu();
+      }
+EOF
+    if [ "$?" = 0 ]; then
+        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_SCHED_GETCPU_PRESENT"
+    fi
 fi
 
 # TODO(tec): Fix -Wshorten-64-to-32 errors on FreeBSD and enable the warning.
@@ -407,14 +444,8 @@ if test "$USE_HDFS"; then
   JAVA_LDFLAGS="$JAVA_LDFLAGS $HDFS_LDFLAGS"
 fi
 
-if [ "$TARGET_OS" = FreeBSD -a "$TARGET_ARCHITECTURE" = i386 ]; then
-  # Intel SSE instructions breaks compilation on FreeBSD i386
-  unset USE_SSE
-fi
-
 if test "$USE_SSE"; then
-  # if Intel SSE instruction set is supported, set USE_SSE=1
-  COMMON_FLAGS="$COMMON_FLAGS -msse -msse4.2 "
+  COMMON_FLAGS="$COMMON_FLAGS -msse4.2"
 elif test -z "$PORTABLE"; then
   if test -n "`echo $TARGET_ARCHITECTURE | grep ^ppc64`"; then
     # Tune for this POWER processor, treating '+' models as base models
@@ -422,11 +453,41 @@ elif test -z "$PORTABLE"; then
     COMMON_FLAGS="$COMMON_FLAGS -mcpu=$POWER -mtune=$POWER "
   elif test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
     COMMON_FLAGS="$COMMON_FLAGS -march=z10 "
-  else
+  elif [ "$TARGET_OS" != AIX ] && [ "$TARGET_OS" != SunOS ]; then
     COMMON_FLAGS="$COMMON_FLAGS -march=native "
   fi
 fi
 
+$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
+  #include <cstdint>
+  #include <nmmintrin.h>
+  int main() {
+    volatile uint32_t x = _mm_crc32_u32(0, 0);
+  }
+EOF
+if [ "$?" = 0 ]; then
+  COMMON_FLAGS="$COMMON_FLAGS -DHAVE_SSE42"
+elif test "$USE_SSE"; then
+  echo "warning: USE_SSE specified but compiler could not use SSE intrinsics, disabling"
+fi
+
+# iOS doesn't support thread-local storage, but this check would erroneously
+# succeed because the cross-compiler flags are added by the Makefile, not this
+# script.
+if [ "$PLATFORM" != IOS ]; then
+  $CXX $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
+  #if defined(_MSC_VER) && !defined(__thread)
+  #define __thread __declspec(thread)
+  #endif
+  int main() {
+    static __thread int tls;
+  }
+EOF
+  if [ "$?" = 0 ]; then
+    COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_SUPPORT_THREAD_LOCAL"
+  fi
+fi
+
 PLATFORM_CCFLAGS="$PLATFORM_CCFLAGS $COMMON_FLAGS"
 PLATFORM_CXXFLAGS="$PLATFORM_CXXFLAGS $COMMON_FLAGS"
 
@@ -458,7 +519,14 @@ 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"
+# This will enable some related identifiers for the preprocessor
 if test -n "$JEMALLOC"; then
   echo "JEMALLOC=1" >> "$OUTPUT"
 fi
+# Indicates that jemalloc should be enabled using -ljemalloc flag
+# The alternative is to porvide a direct link to the library via JEMALLOC_LIB
+# and JEMALLOC_INCLUDE
+if test -n "$WITH_JEMALLOC_FLAG"; then
+  echo "WITH_JEMALLOC_FLAG=$WITH_JEMALLOC_FLAG" >> "$OUTPUT"
+fi
 echo "LUA_PATH=$LUA_PATH" >> "$OUTPUT"

build_tools/cont_integration.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 #
 # Copyright (c) 2016, Facebook. All rights reserved.
 #
@@ -67,7 +67,7 @@ function update_repo_status {
 #
 
 # Path to the determinator from the root of the RocksDB repo.
-CONTRUN_DETERMINATOR=./arcanist_util/config/RocksDBCommonHelper.php
+CONTRUN_DETERMINATOR=./build_tools/RocksDBCommonHelper.php
 
 # Value of the previous commit.
 PREV_COMMIT=

build_tools/dependencies.sh

@@ -1,19 +1,18 @@
-# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
-GCC_BASE=/mnt/gvfs/third-party2/gcc/7331085db891a2ef4a88a48a751d834e8d68f4cb/7.x/centos7-native/b2ef2b6
-CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/963d9aeda70cc4779885b1277484fe7544a04e3e/9.0.0/platform007/9e92d53/
-LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/6ace84e956873d53638c738b6f65f3f469cca74c/7.x/platform007/5620abc
-GLIBC_BASE=/mnt/gvfs/third-party2/glibc/192b0f42d63dcf6210d6ceae387b49af049e6e0c/2.26/platform007/f259413
-SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/7f9bdaada18f59bc27ec2b0871eb8a6144343aef/1.1.3/platform007/ca4da3d
-ZLIB_BASE=/mnt/gvfs/third-party2/zlib/2d9f0b9a4274cc21f61272a9e89bdb859bce8f1f/1.2.8/platform007/ca4da3d
-BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/dc49a21c5fceec6456a7a28a94dcd16690af1337/1.0.6/platform007/ca4da3d
-LZ4_BASE=/mnt/gvfs/third-party2/lz4/0f607f8fc442ea7d6b876931b1898bb573d5e5da/1.9.1/platform007/ca4da3d
-ZSTD_BASE=/mnt/gvfs/third-party2/zstd/ca22bc441a4eb709e9e0b1f9fec9750fed7b31c5/1.4.x/platform007/15a3614
-GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/0b9929d2588991c65a57168bf88aff2db87c5d48/2.2.0/platform007/ca4da3d
-JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/c26f08f47ac35fc31da2633b7da92d6b863246eb/master/platform007/c26c002
-NUMA_BASE=/mnt/gvfs/third-party2/numa/3f3fb57a5ccc5fd21c66416c0b83e0aa76a05376/2.0.11/platform007/ca4da3d
-LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/40c73d874898b386a71847f1b99115d93822d11f/1.4/platform007/6f3e0a9
-TBB_BASE=/mnt/gvfs/third-party2/tbb/4ce8e8dba77cdbd81b75d6f0c32fd7a1b76a11ec/2018_U5/platform007/ca4da3d
-KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/fb251ecd2f5ae16f8671f7014c246e52a748fe0b/fb/platform007/da39a3e
-BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/ab9f09bba370e7066cafd4eb59752db93f2e8312/2.29.1/platform007/15a3614
-VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/d42d152a15636529b0861ec493927200ebebca8e/3.15.0/platform007/ca4da3d
-LUA_BASE=/mnt/gvfs/third-party2/lua/f0cd714433206d5139df61659eb7b28b1dea6683/5.3.4/platform007/5007832
+GCC_BASE=/mnt/gvfs/third-party2/gcc/2928bb3ed95bf64f5b388ee88c30dc74710c3b35/5.x/centos6-native/f4950a1
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/a5fea028cb7ba43498976e1f8054b0b2e790c295/stable/centos6-native/6aaf4de
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/7a9099f6587ee4378c0b1fa32bb8934019d30ca4/5.x/gcc-5-glibc-2.23/339d858
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/3b7c6469854dfc7832a1c3cc5b86919a84e5f865/2.23/gcc-5-glibc-2.23/ca1d1c0
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/8c38a4c1e52b4c2cc8a9cdc31b9c947ed7dbfcb4/1.1.3/gcc-5-glibc-2.23/9bc6787
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/d7861abe6f0e27ab98c9303b95a662f0e4cdedb5/1.2.8/gcc-5-glibc-2.23/9bc6787
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/740325875f6729f42d28deaa2147b0854f3a347e/1.0.6/gcc-5-glibc-2.23/9bc6787
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/0815d59804160c96caac5f27ca004f51af893dc6/r131/gcc-5-glibc-2.23/9bc6787
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/c15a4f5f619a2930478d01e2e34dc1e0652b0873/1.1.4/gcc-5-glibc-2.23/03859b5
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/f905a5e1032fb30c05db3d3752319857388c0c49/2.2.0/gcc-5-glibc-2.23/9bc6787
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/8d60633d822a2a55849c73db24e74a25e52b71db/master/gcc-5-glibc-2.23/1c32b4b
+NUMA_BASE=/mnt/gvfs/third-party2/numa/17c514c4d102a25ca15f4558be564eeed76f4b6a/2.0.8/gcc-5-glibc-2.23/9bc6787
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/8db74270cd6d0212ac92d69e7fc7beefe617d772/trunk/gcc-5-glibc-2.23/b1847cb
+TBB_BASE=/mnt/gvfs/third-party2/tbb/9d9a554877d0c5bef330fe818ab7178806dd316a/4.0_update2/gcc-5-glibc-2.23/9bc6787
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/90c9734afc5579c9d1db529fa788d09f97763b85/4.0.9-36_fbk5_2933_gd092e3f/gcc-5-glibc-2.23/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/9e829389ef61b92c62de8748c80169aaf25ce1f0/2.26.1/centos6-native/da39a3e
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/d7f4d4d86674a57668e3a96f76f0e17dd0eb8765/3.11.0/gcc-5-glibc-2.23/9bc6787
+LUA_BASE=/mnt/gvfs/third-party2/lua/61e4abf5813bbc39bc4f548757ccfcadde175a48/5.2.3/gcc-5-glibc-2.23/65372bd

build_tools/dockerbuild.sh

@@ -1,2 +1,2 @@
-#!/bin/bash
+#!/usr/bin/env bash
 docker run -v $PWD:/rocks -w /rocks buildpack-deps make

build_tools/fbcode_config.sh

@@ -13,7 +13,7 @@ source "$BASEDIR/dependencies.sh"
 CFLAGS=""
 
 # libgcc
-LIBGCC_INCLUDE="$LIBGCC_BASE/include/c++/7.3.0"
+LIBGCC_INCLUDE="$LIBGCC_BASE/include"
 LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
 
 # glibc
@@ -43,15 +43,11 @@ if test -z $PIC_BUILD; then
   LZ4_INCLUDE=" -I $LZ4_BASE/include/"
   LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
   CFLAGS+=" -DLZ4"
-fi
 
-ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
-if test -z $PIC_BUILD; then
+  ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
   ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
-else
-  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
+  CFLAGS+=" -DZSTD"
 fi
-CFLAGS+=" -DZSTD"
 
 # location of gflags headers and libraries
 GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
@@ -108,8 +104,8 @@ if [ -z "$USE_CLANG" ]; then
   CXX="$GCC_BASE/bin/g++"
 
   CFLAGS+=" -B$BINUTILS/gold"
-  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
   CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
   JEMALLOC=1
 else
   # clang
@@ -120,8 +116,8 @@ else
   KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
 
   CFLAGS+=" -B$BINUTILS/gold -nostdinc -nostdlib"
-  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x "
-  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x/x86_64-facebook-linux "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/5.x "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/5.x/x86_64-facebook-linux "
   CFLAGS+=" -isystem $GLIBC_INCLUDE"
   CFLAGS+=" -isystem $LIBGCC_INCLUDE"
   CFLAGS+=" -isystem $CLANG_INCLUDE"
@@ -137,9 +133,9 @@ CXXFLAGS+=" $CFLAGS"
 
 EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_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+=" -Wl,--dynamic-linker,/usr/local/fbcode/gcc-5-glibc-2.23/lib/ld.so"
 EXEC_LDFLAGS+=" $LIBUNWIND"
-EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/platform007/lib"
+EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/gcc-5-glibc-2.23/lib"
 # required by libtbb
 EXEC_LDFLAGS+=" -ldl"
 
@@ -149,4 +145,12 @@ EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GF
 
 VALGRIND_VER="$VALGRIND_BASE/bin/"
 
-export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD 
+LUA_PATH="$LUA_BASE"
+
+if test -z $PIC_BUILD; then
+  LUA_LIB=" $LUA_PATH/lib/liblua.a"
+else
+  LUA_LIB=" $LUA_PATH/lib/liblua_pic.a"
+fi
+
+export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD LUA_PATH LUA_LIB

build_tools/fbcode_config4.8.1.sh

@@ -93,7 +93,7 @@ else
 fi
 
 CFLAGS+=" $DEPS_INCLUDE"
-CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE"
+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"
 CFLAGS+=" -DSNAPPY -DGFLAGS=google -DZLIB -DBZIP2 -DLZ4 -DZSTD -DNUMA -DTBB"
 CXXFLAGS+=" $CFLAGS"
 

build_tools/format-diff.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # If clang_format_diff.py command is not specfied, we assume we are able to
 # access directly without any path.
 if [ -z $CLANG_FORMAT_DIFF ]
@@ -9,10 +9,12 @@ fi
 # Check clang-format-diff.py
 if ! which $CLANG_FORMAT_DIFF &> /dev/null
 then
-  echo "You didn't have clang-format-diff.py available in your computer!"
-  echo "You can download it by running: "
+  echo "You didn't have clang-format-diff.py and/or clang-format available in your computer!"
+  echo "You can download clang-format-diff.py by running: "
   echo "    curl --location http://goo.gl/iUW1u2 -o ${CLANG_FORMAT_DIFF}"
-  echo "and move ${CLANG_FORMAT_DIFF} to some directory within PATH=${PATH}"
+  echo "You can download clang-format by running: "
+  echo "    brew install clang-format"
+  echo "Then, move both files (i.e. ${CLANG_FORMAT_DIFF} and clang-format) to some directory within PATH=${PATH}"
   exit 128
 fi
 

build_tools/make_package.sh

@@ -99,6 +99,18 @@ function main() {
 
   make static_lib
   make install INSTALL_PATH=package
+
+  cd package
+
+  LIB_DIR=lib
+  if [[ -z "$ARCH" ]]; then
+      ARCH=$(getconf LONG_BIT)
+  fi
+  if [[ ("$FPM_OUTPUT" = "rpm") && ($ARCH -eq 64) ]]; then
+      mv lib lib64
+      LIB_DIR=lib64
+  fi
+
   fpm \
     -s dir \
     -t $FPM_OUTPUT \
@@ -110,7 +122,7 @@ function main() {
     --license BSD \
     --vendor Facebook \
     --description "RocksDB is an embeddable persistent key-value store for fast storage." \
-    package
+    include $LIB_DIR
 }
 
 main $@

build_tools/regression_build_test.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 set -e
 

build_tools/rocksdb-lego-determinator

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # This script is executed by Sandcastle
 # to determine next steps to run
 
@@ -81,6 +81,7 @@ PARALLELISM="$PARALLEL_J $PARALLEL_j"
 
 DEBUG="OPT=-g"
 SHM="TEST_TMPDIR=/dev/shm/rocksdb"
+NON_SHM="TMPD=/tmp/rocksdb_test_tmp"
 GCC_481="ROCKSDB_FBCODE_BUILD_WITH_481=1"
 ASAN="COMPILE_WITH_ASAN=1"
 CLANG="USE_CLANG=1"
@@ -129,35 +130,6 @@ DISABLE_COMMANDS="[
     }
 ]"
 
-#
-# RocksDB unit test in parallel
-# Currently we always have noise in our parallel runs. This job is to help
-# manage the noise
-#
-PARALLEL_UNIT_TEST_COMMANDS="[
-    {
-        'name':'Rocksdb Parallel Unit Test',
-        'oncall':'$ONCALL',
-        'steps': [
-            $CLEANUP_ENV,
-            {
-                'name':'Build and test RocksDB debug version',
-                'shell':'$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
-            },
-            $CLEANUP_ENV,
-            {
-                'name':'Build and test RocksDB debug version under gcc-4.8.1',
-                'shell':'$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
-            },
-        ],
-        $REPORT
-    }
-]"
-
 #
 # RocksDB unit test
 #
@@ -191,7 +163,7 @@ UNIT_TEST_NON_SHM_COMMANDS="[
             {
                 'name':'Build and test RocksDB debug version',
                 'timeout': 86400,
-                'shell':'$DEBUG make $PARALLELISM check || $CONTRUN_NAME=non_shm_check $TASK_CREATION_TOOL',
+                'shell':'$NON_SHM $DEBUG make $PARALLELISM check || $CONTRUN_NAME=non_shm_check $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -592,7 +564,7 @@ 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',
+                'shell':'set -o pipefail && $SHM $DEBUG $TSAN CRASH_TEST_KILL_ODD=1887 CRASH_TEST_EXT_ARGS=--log2_keys_per_lock=22  make J=1 crash_test || $CONTRUN_NAME=tsan_crash_test $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -611,39 +583,7 @@ run_format_compatible()
   rm -rf /dev/shm/rocksdb
   mkdir /dev/shm/rocksdb
 
-  echo '
-  if [ -e "build_tools/build_detect_platform" ]
-  then
-    sed "s/tcmalloc/nothingnothingnothing/g" build_tools/build_detect_platform > $TEST_TMPDIR/temp_build_file
-    rm -rf build_tools/build_detect_platform
-    cp $TEST_TMPDIR/temp_build_file build_tools/build_detect_platform
-    chmod +x build_tools/build_detect_platform
-  fi
-
-  if [ -e "build_detect_platform" ]
-  then
-    sed "s/tcmalloc/nothingnothingnothing/g" build_detect_platform > $TEST_TMPDIR/temp_build_file
-    rm -rf build_detect_platform
-    cp $TEST_TMPDIR/temp_build_file build_detect_platform
-    chmod +x build_detect_platform
-  fi
-
-  make ldb -j32
-
-  if [ -e "build_detect_platform" ]
-  then
-    git checkout -- build_detect_platform
-  fi
-
-  if [ -e "build_tools/build_detect_platform" ]
-  then
-    git checkout -- build_tools/build_detect_platform
-  fi
-  ' > temp_build_ldb.sh
-
-  sed "s/make ldb -j32/source temp_build_ldb.sh/g" tools/check_format_compatible.sh > tools/temp_check_format_compatible.sh
-  chmod +x tools/temp_check_format_compatible.sh
-  tools/temp_check_format_compatible.sh
+  tools/check_format_compatible.sh
 }
 
 FORMAT_COMPATIBLE_COMMANDS="[
@@ -775,9 +715,6 @@ JAVA_BUILD_TEST_COMMANDS="[
 
 
 case $1 in
-  punit)
-    echo $PARALLEL_UNIT_TEST_COMMANDS
-    ;;
   unit)
     echo $UNIT_TEST_COMMANDS
     ;;

build_tools/run_ci_db_test.ps1

@@ -1,24 +1,50 @@
 # This script enables you running RocksDB tests by running
-# All the tests in parallel 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
+# All the tests concurrently and utilizing all the cores
 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]$Run = "db_test",  # Run db_test|db_test2|tests|testname1,testname2...
+  [switch]$EnableJE = $false,  # Look for and use _je executable, append _je to listed exclusions
+  [switch]$RunAll = $false,    # Will attempt discover all *_test[_je].exe binaries and run all
+                               # of them as Google suites. I.e. It will run test cases concurrently
+                               # except those mentioned as $Run, those will run as individual test cases
+                               # And any execlued with $ExcludeExes or $ExcludeCases
+                               # It will also not run any individual test cases
+                               # excluded but $ExcludeCasese
+  [string]$SuiteRun = "",      # Split test suites in test cases and run in parallel, not compatible with $RunAll
+  [string]$Run = "",           # Run specified executables in parallel but do not split to test cases
+  [string]$ExcludeCases = "",  # Exclude test cases, expects a comma separated list, no spaces
+                               # Takes effect when $RunAll or $SuiteRun is specified. Must have full
+                               # Test cases name including a group and a parameter if any
+  [string]$ExcludeExes = "",   # Exclude exes from consideration, expects a comma separated list,
+                               # no spaces. Takes effect only when $RunAll is specified
+  [string]$WorkFolder = "",    # Direct tests to use that folder. SSD or Ram drive are better options.
    # 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]$Concurrency = 16
+  [int]$Concurrency = 8,
+  [int]$Limit = -1 # -1 means do not limit for test purposes
 )
 
 # Folders and commands must be fullpath to run assuming
 # the current folder is at the root of the git enlistment
-Get-Date
+$StartDate = (Get-Date)
+$StartDate
+
+
+$DebugPreference = "Continue"
+
+# These tests are not google test suites and we should guard
+# Against running them as suites
+$RunOnly = New-Object System.Collections.Generic.HashSet[string]
+$RunOnly.Add("c_test") | Out-Null
+$RunOnly.Add("compact_on_deletion_collector_test") | Out-Null
+$RunOnly.Add("merge_test") | Out-Null
+$RunOnly.Add("stringappend_test") | Out-Null # Apparently incorrectly written
+$RunOnly.Add("backupable_db_test") | Out-Null # Disabled
+
+
+if($RunAll -and $SuiteRun -ne "") {
+    Write-Error "$RunAll and $SuiteRun are not compatible"
+    exit 1
+}
 
 # If running under Appveyor assume that root
 [string]$Appveyor = $Env:APPVEYOR_BUILD_FOLDER
@@ -47,40 +73,32 @@ if($WorkFolder -eq "") {
 }
 
 Write-Output "Root: $RootFolder, WorkFolder: $WorkFolder"
-
-# Use JEMALLOC executables
-if($Run -ceq "db_test" -or
-   $Run -ceq "db_test2" -or
-   $Run -ceq "db_basic_test") {
-
-   $file_name = $Run
-
-   if($EnableJE) {
-     $file_name += "_je"
-   }
-
-   $file_name += ".exe"
-
-   $db_test = -Join ($BinariesFolder, $file_name)
-
-   Write-Output "Binaries: $BinariesFolder db_test: $db_test"
-}
-
-
-#Exclusions that we do not want to run
-$ExcludeTests = New-Object System.Collections.Generic.HashSet[string]
-
-
-if($Exclude -ne "") {
-    Write-Host "Exclude: $Exclude"
-    $l = $Exclude -split ' '
-    ForEach($t in $l) { $ExcludeTests.Add($t) | Out-Null }
-}
+Write-Output "BinariesFolder: $BinariesFolder, LogFolder: $LogFolder"
 
 # Create test directories in the current folder
 md -Path $WorkFolder -ErrorAction Ignore | Out-Null
 md -Path $LogFolder -ErrorAction Ignore | Out-Null
 
+
+$ExcludeCasesSet = New-Object System.Collections.Generic.HashSet[string]
+if($ExcludeCases -ne "") {
+    Write-Host "ExcludeCases: $ExcludeCases"
+    $l = $ExcludeCases -split ' '
+    ForEach($t in $l) { 
+      $ExcludeCasesSet.Add($t) | Out-Null
+    }
+}
+
+$ExcludeExesSet = New-Object System.Collections.Generic.HashSet[string]
+if($ExcludeExes -ne "") {
+    Write-Host "ExcludeExe: $ExcludeExes"
+    $l = $ExcludeExes -split ' '
+    ForEach($t in $l) { 
+      $ExcludeExesSet.Add($t) | Out-Null
+    }
+}
+
+
 # Extract the names of its tests by running db_test with --gtest_list_tests.
 # This filter removes the "#"-introduced comments, and expands to
 # fully-qualified names by changing input like this:
@@ -98,28 +116,37 @@ md -Path $LogFolder -ErrorAction Ignore | Out-Null
 #   DBTest.WriteEmptyBatch
 #   MultiThreaded/MultiThreadedDBTest.MultiThreaded/0
 #   MultiThreaded/MultiThreadedDBTest.MultiThreaded/1
+#
 # Output into the parameter in a form TestName -> Log File Name
-function Normalize-DbTests($HashTable) {
+function ExtractTestCases([string]$GTestExe, $HashTable) {
 
     $Tests = @()
 # Run db_test to get a list of tests and store it into $a array
-    &$db_test --gtest_list_tests | tee -Variable Tests | Out-Null
+    &$GTestExe --gtest_list_tests | tee -Variable Tests | Out-Null
 
     # Current group
     $Group=""
 
     ForEach( $l in $Tests) {
-      # Trailing dot is a test group
-      if( $l -match "\.$") {
+
+      # Leading whitespace is fine
+      $l = $l -replace '^\s+',''
+      # but no whitespace any other place
+      if($l -match "\s+") {
+        continue
+      }
+      # Trailing dot is a test group but no whitespace
+      elseif ( $l -match "\.$" ) {
         $Group = $l
       }  else {
         # Otherwise it is a test name, remove leading space
-        $test = $l -replace '^\s+',''
+        $test = $l
         # remove trailing comment if any and create a log name
         $test = $test -replace '\s+\#.*',''
         $test = "$Group$test"
 
-        if($ExcludeTests.Contains($test)) {
+        if($ExcludeCasesSet.Contains($test)) {
+            Write-Warning "$test case is excluded"
             continue
         }
 
@@ -135,59 +162,130 @@ function Normalize-DbTests($HashTable) {
 
 # The function removes trailing .exe siffix if any,
 # creates a name for the log file
+# Then adds the test name if it was not excluded into
+# a HashTable in a form of test_name -> log_path
 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"
-    }
+    $HashTable.Add($test_name, $log_path)
 }
 
-# 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, $HashTable) {
+# This function takes a list of Suites to run
+# Lists all the test cases in each of the suite
+# and populates HashOfHashes
+# Ordered by suite(exe) @{ Exe = @{ TestCase = LogName }}
+function ProcessSuites($ListOfSuites, $HashOfHashes) {
 
-    $Exclusions = @("db_test*", "db_sanity_test*", "db_basic_test*")
+  $suite_list = $ListOfSuites
+  # Problem: if you run --gtest_list_tests on
+  # a non Google Test executable then it will start executing
+  # and we will get nowhere
+  ForEach($suite in $suite_list) {
 
-    $p = -join ($BinariesFolder, $pattern)
-
-    Write-Host "Path: $p"
-
-    dir -Path $p -Exclude $Exclusions | ForEach-Object {
-       MakeAndAdd -token ($_.Name) -HashTable $HashTable
+    if($RunOnly.Contains($suite)) {
+      Write-Warning "$suite is excluded from running as Google test suite"
+      continue
     }
-}
 
-$TestsToRun = [ordered]@{}
-
-if($Run -ceq "db_test" -or
-   $Run -ceq "db_test2" -or
-   $Run -ceq "db_basic_test") {
-    Normalize-DbTests -HashTable $TestsToRun
-} elseif($Run -ceq "tests") {
     if($EnableJE) {
-        $pattern = "*_test_je.exe"
+      $suite += "_je"
+    }
+
+    $Cases = [ordered]@{}
+    $Cases.Clear()
+    $suite_exe = -Join ($BinariesFolder, $suite)
+    ExtractTestCases -GTestExe $suite_exe -HashTable $Cases
+    if($Cases.Count -gt 0) {
+      $HashOfHashes.Add($suite, $Cases);
+    }
+  }
+
+  # Make logs and run
+  if($CasesToRun.Count -lt 1) {
+     Write-Error "Failed to extract tests from $SuiteRun"
+     exit 1
+  }
+
+}
+
+# This will contain all test executables to run
+
+# Hash table that contains all non suite
+# Test executable to run
+$TestExes = [ordered]@{}
+
+# Check for test exe that are not
+# Google Test Suites
+# Since this is explicitely mentioned it is not subject
+# for exclusions
+if($Run -ne "") {
+
+  $test_list = $Run -split ' '
+
+  ForEach($t in $test_list) {
+
+    if($EnableJE) {
+      $t += "_je"
+    }
+
+    MakeAndAdd -token $t -HashTable $TestExes
+  }
+
+  if($TestExes.Count -lt 1) {
+     Write-Error "Failed to extract tests from $Run"
+     exit 1
+  }
+}
+
+# Ordered by exe @{ Exe = @{ TestCase = LogName }}
+$CasesToRun = [ordered]@{}
+
+if($SuiteRun -ne "") {
+  $suite_list = $SuiteRun -split ' '
+  ProcessSuites -ListOfSuites $suite_list -HashOfHashes $CasesToRun
+}
+
+if($RunAll) {
+# Discover all the test binaries
+  if($EnableJE) {
+    $pattern = "*_test_je.exe"
+  } else {
+    $pattern = "*_test.exe"
+  }
+
+
+  $search_path = -join ($BinariesFolder, $pattern)
+  Write-Host "Binaries Search Path: $search_path"
+
+  $ListOfExe = @()
+  dir -Path $search_path | ForEach-Object {
+     $ListOfExe += ($_.Name)     
+  }
+
+  # Exclude those in RunOnly from running as suites
+  $ListOfSuites = @()
+  ForEach($e in $ListOfExe) {
+
+    $e = $e -replace '.exe$', ''
+    $bare_name = $e -replace '_je$', ''
+
+    if($ExcludeExesSet.Contains($bare_name)) {
+      Write-Warning "Test $e is excluded"
+      continue
+    }
+
+    if($RunOnly.Contains($bare_name)) {
+      MakeAndAdd -token $e -HashTable $TestExes
     } else {
-        $pattern = "*_test.exe"
+      $ListOfSuites += $bare_name
     }
-    Discover-TestBinaries -Pattern $pattern -HashTable $TestsToRun
-} else {
+  }
 
-    $test_list = $Run -split ' '
-
-    ForEach($t in $test_list) {
-       MakeAndAdd -token $t -HashTable $TestsToRun
-    }
+  ProcessSuites -ListOfSuites $ListOfSuites -HashOfHashes $CasesToRun
 }
 
-$NumTestsToStart = $TestsToRun.Count
-if($Limit -ge 0 -and $NumTestsToStart -gt $Limit) {
-    $NumTestsToStart = $Limit
-}
 
 Write-Host "Attempting to start: $NumTestsToStart tests"
 
@@ -205,13 +303,13 @@ $InvokeTestAsync = {
 
 # Hash that contains tests to rerun if any failed
 # Those tests will be rerun sequentially
-$Rerun = [ordered]@{}
+# $Rerun = [ordered]@{}
 # Test limiting factor here
-$count = 0
+[int]$count = 0
 # Overall status
 [bool]$success = $true;
 
-function RunJobs($TestToLog, [int]$ConcurrencyVal, [bool]$AddForRerun)
+function RunJobs($Suites, $TestCmds, [int]$ConcurrencyVal)
 {
     # Array to wait for any of the running jobs
     $jobs = @()
@@ -220,38 +318,66 @@ function RunJobs($TestToLog, [int]$ConcurrencyVal, [bool]$AddForRerun)
 
     # Wait for all to finish and get the results
     while(($JobToLog.Count -gt 0) -or
-          ($TestToLog.Count -gt 0)) {
+          ($TestCmds.Count -gt 0) -or 
+           ($Suites.Count -gt 0)) {
 
         # 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
+              ((($TestCmds.Count -gt 0) -or ($Suites.Count -gt 0)) -and
               (($Limit -lt 0) -or ($count -lt $Limit)))) {
 
+            # We always favore suites to run if available
+            [string]$exe_name = ""
+            [string]$log_path = ""
+            $Cases = @{}
 
-            # We only need the first key
-            foreach($key in $TestToLog.keys) {
-                $k = $key
+            if($Suites.Count -gt 0) {
+              # Will the first one
+              ForEach($e in $Suites.Keys) {
+                $exe_name = $e
+                $Cases = $Suites[$e]
                 break
-            }
+              }
+              [string]$test_case = ""
+              [string]$log_path = ""
+              ForEach($c in $Cases.Keys) {
+                 $test_case = $c
+                 $log_path = $Cases[$c]
+                 break
+              }
 
-            Write-Host "Starting $k"
-            $log_path = ($TestToLog.$k)
+              Write-Host "Starting $exe_name::$test_case"
+              [string]$Exe =  -Join ($BinariesFolder, $exe_name)
+              $job = Start-Job -Name "$exe_name::$test_case" -ArgumentList @($Exe,$test_case,$log_path) -ScriptBlock $InvokeTestCase
+              $JobToLog.Add($job, $log_path)
+
+              $Cases.Remove($test_case)
+              if($Cases.Count -lt 1) {
+                $Suites.Remove($exe_name)
+              }
+
+            } elseif ($TestCmds.Count -gt 0) {
+
+               ForEach($e in $TestCmds.Keys) {
+                 $exe_name = $e
+                 $log_path = $TestCmds[$e]
+                 break
+               }
+
+              [string]$Exe =  -Join ($BinariesFolder, $exe_name)
+              $job = Start-Job -Name $exe_name -ScriptBlock $InvokeTestAsync -ArgumentList @($Exe,$log_path)
+              $JobToLog.Add($job, $log_path)
+
+              $TestCmds.Remove($exe_name)
 
-            if($Run -ceq "db_test" -or
-               $Run -ceq "db_test2" -or
-               $Run -ceq "db_basic_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)
+                Write-Error "In the job loop but nothing to run"
+                exit 1
             }
 
-            $JobToLog.Add($job, $log_path)
-            $TestToLog.Remove($k)
-
             ++$count
-        }
+        } # End of Job starting loop
 
         if($JobToLog.Count -lt 1) {
           break
@@ -296,9 +422,6 @@ function RunJobs($TestToLog, [int]$ConcurrencyVal, [bool]$AddForRerun)
                 $success = $false;
                 Write-Warning $message
                 $log_content | Write-Warning
-                if($AddForRerun) {
-                    $Rerun.Add($completed.Name, $log)
-                }
             } else {
                 Write-Host $message
             }
@@ -310,16 +433,14 @@ function RunJobs($TestToLog, [int]$ConcurrencyVal, [bool]$AddForRerun)
     }
 }
 
-RunJobs -TestToLog $TestsToRun -ConcurrencyVal $Concurrency -AddForRerun $EnableRerun
+RunJobs -Suites $CasesToRun -TestCmds $TestExes -ConcurrencyVal $Concurrency
 
-if($Rerun.Count -gt 0) {
-    Write-Host "Rerunning " ($Rerun.Count) " tests sequentially"
-    $success = $true
-    $count = 0
-    RunJobs -TestToLog $Rerun -ConcurrencyVal 1 -AddForRerun $false
-}
+$EndDate = (Get-Date)
 
-Get-Date
+New-TimeSpan -Start $StartDate -End $EndDate | 
+  ForEach-Object { 
+    "Elapsed time: {0:g}" -f $_
+  }
 
 
 if(!$success) {
@@ -327,7 +448,9 @@ if(!$success) {
 # So we simply exit
 #    Remove-Job -Job $jobs -Force
 # indicate failure using this exit code
-    exit 12345
+    exit 1
  }
 
+ exit 0
+
  

build_tools/update_dependencies.sh

@@ -53,7 +53,7 @@ function get_lib_base()
 }
 
 ###########################################################
-#                   4.9.x dependencies                    #
+#                   5.x dependencies                      #
 ###########################################################
 
 OUTPUT="$BASEDIR/dependencies.sh"
@@ -64,16 +64,16 @@ touch "$OUTPUT"
 echo "Writing dependencies to $OUTPUT"
 
 # Compilers locations
-GCC_BASE=`readlink -f $TP2_LATEST/gcc/4.9.x/centos6-native/*/`
+GCC_BASE=`readlink -f $TP2_LATEST/gcc/5.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 libgcc     5.x
+get_lib_base glibc      2.23
+get_lib_base snappy     LATEST gcc-5-glibc-2.23
 get_lib_base zlib       LATEST
 get_lib_base bzip2      LATEST
 get_lib_base lz4        LATEST
@@ -82,12 +82,12 @@ 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 tbb        4.0_update2 gcc-5-glibc-2.23
 
 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
-get_lib_base lua        5.2.3 gcc-4.9-glibc-2.20
+get_lib_base valgrind   3.10.0 gcc-5-glibc-2.23
+get_lib_base lua        5.2.3 gcc-5-glibc-2.23
 
 git diff $OUTPUT
 

cache/cache_test.cc

@@ -14,9 +14,9 @@
 #include <iostream>
 #include <string>
 #include <vector>
-#include "util/clock_cache.h"
+#include "cache/clock_cache.h"
+#include "cache/lru_cache.h"
 #include "util/coding.h"
-#include "util/lru_cache.h"
 #include "util/string_util.h"
 #include "util/testharness.h"
 
@@ -475,6 +475,37 @@ void deleter(const Slice& key, void* value) {
 }
 }  // namespace
 
+TEST_P(CacheTest, ReleaseAndErase) {
+  std::shared_ptr<Cache> cache = NewCache(5, 0, false);
+  Cache::Handle* handle;
+  Status s = cache->Insert(EncodeKey(100), EncodeValue(100), 1,
+                           &CacheTest::Deleter, &handle);
+  ASSERT_TRUE(s.ok());
+  ASSERT_EQ(5U, cache->GetCapacity());
+  ASSERT_EQ(1U, cache->GetUsage());
+  ASSERT_EQ(0U, deleted_keys_.size());
+  auto erased = cache->Release(handle, true);
+  ASSERT_TRUE(erased);
+  // This tests that deleter has been called
+  ASSERT_EQ(1U, deleted_keys_.size());
+}
+
+TEST_P(CacheTest, ReleaseWithoutErase) {
+  std::shared_ptr<Cache> cache = NewCache(5, 0, false);
+  Cache::Handle* handle;
+  Status s = cache->Insert(EncodeKey(100), EncodeValue(100), 1,
+                           &CacheTest::Deleter, &handle);
+  ASSERT_TRUE(s.ok());
+  ASSERT_EQ(5U, cache->GetCapacity());
+  ASSERT_EQ(1U, cache->GetUsage());
+  ASSERT_EQ(0U, deleted_keys_.size());
+  auto erased = cache->Release(handle);
+  ASSERT_FALSE(erased);
+  // This tests that deleter is not called. When cache has free capacity it is
+  // not expected to immediately erase the released items.
+  ASSERT_EQ(0U, deleted_keys_.size());
+}
+
 TEST_P(CacheTest, SetCapacity) {
   // test1: increase capacity
   // lets create a cache with capacity 5,

cache/clock_cache.cc

@@ -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.
 
-#include "util/clock_cache.h"
+#include "cache/clock_cache.h"
 
 #ifndef SUPPORT_CLOCK_CACHE
 
@@ -26,14 +26,18 @@ std::shared_ptr<Cache> NewClockCache(size_t capacity, int num_shard_bits,
 #include <assert.h>
 #include <atomic>
 #include <deque>
-#include <limits>
 
+// "tbb/concurrent_hash_map.h" requires RTTI if exception is enabled.
+// Disable it so users can chooose to disable RTTI.
+#ifndef ROCKSDB_USE_RTTI
+#define TBB_USE_EXCEPTIONS 0
+#endif
 #include "tbb/concurrent_hash_map.h"
 
+#include "cache/sharded_cache.h"
 #include "port/port.h"
 #include "util/autovector.h"
 #include "util/mutexlock.h"
-#include "util/sharded_cache.h"
 
 namespace rocksdb {
 
@@ -252,8 +256,11 @@ class ClockCacheShard : public CacheShard {
   //
   // Not necessary to hold mutex_ before being called.
   virtual bool Ref(Cache::Handle* handle) override;
-  virtual void Release(Cache::Handle* handle) override;
+  virtual bool Release(Cache::Handle* handle,
+                       bool force_erase = false) override;
   virtual void Erase(const Slice& key, uint32_t hash) override;
+  bool EraseAndConfirm(const Slice& key, uint32_t hash,
+                       CleanupContext* context);
   virtual size_t GetUsage() const override;
   virtual size_t GetPinnedUsage() const override;
   virtual void EraseUnRefEntries() override;
@@ -274,13 +281,17 @@ class ClockCacheShard : public CacheShard {
   // Decrease reference count of the entry. If this decreases the count to 0,
   // recycle the entry. If set_usage is true, also set the usage bit.
   //
+  // returns true if a value is erased.
+  //
   // Not necessary to hold mutex_ before being called.
-  void Unref(CacheHandle* handle, bool set_usage, CleanupContext* context);
+  bool Unref(CacheHandle* handle, bool set_usage, CleanupContext* context);
 
   // Unset in-cache bit of the entry. Recycle the handle if necessary.
   //
+  // returns true if a value is erased.
+  //
   // Has to hold mutex_ before being called.
-  void UnsetInCache(CacheHandle* handle, CleanupContext* context);
+  bool UnsetInCache(CacheHandle* handle, CleanupContext* context);
 
   // Put the handle back to recycle_ list, and put the value associated with
   // it into to-be-deleted list. It doesn't cleanup the key as it might be
@@ -432,7 +443,7 @@ bool ClockCacheShard::Ref(Cache::Handle* h) {
   return false;
 }
 
-void ClockCacheShard::Unref(CacheHandle* handle, bool set_usage,
+bool ClockCacheShard::Unref(CacheHandle* handle, bool set_usage,
                             CleanupContext* context) {
   if (set_usage) {
     handle->flags.fetch_or(kUsageBit, std::memory_order_relaxed);
@@ -451,9 +462,10 @@ void ClockCacheShard::Unref(CacheHandle* handle, bool set_usage,
       RecycleHandle(handle, context);
     }
   }
+  return context->to_delete_value.size();
 }
 
-void ClockCacheShard::UnsetInCache(CacheHandle* handle,
+bool ClockCacheShard::UnsetInCache(CacheHandle* handle,
                                    CleanupContext* context) {
   mutex_.AssertHeld();
   // Use acquire-release semantics as previous operations on the cache entry
@@ -465,6 +477,7 @@ void ClockCacheShard::UnsetInCache(CacheHandle* handle,
   if (InCache(flags) && CountRefs(flags) == 0) {
     RecycleHandle(handle, context);
   }
+  return context->to_delete_value.size();
 }
 
 bool ClockCacheShard::TryEvict(CacheHandle* handle, CleanupContext* context) {
@@ -618,27 +631,36 @@ Cache::Handle* ClockCacheShard::Lookup(const Slice& key, uint32_t hash) {
   return reinterpret_cast<Cache::Handle*>(handle);
 }
 
-void ClockCacheShard::Release(Cache::Handle* h) {
+bool ClockCacheShard::Release(Cache::Handle* h, bool force_erase) {
   CleanupContext context;
   CacheHandle* handle = reinterpret_cast<CacheHandle*>(h);
-  Unref(handle, true, &context);
+  bool erased = Unref(handle, true, &context);
+  if (force_erase && !erased) {
+    erased = EraseAndConfirm(handle->key, handle->hash, &context);
+  }
   Cleanup(context);
+  return erased;
 }
 
 void ClockCacheShard::Erase(const Slice& key, uint32_t hash) {
   CleanupContext context;
-  {
-    MutexLock l(&mutex_);
-    HashTable::accessor accessor;
-    if (table_.find(accessor, CacheKey(key, hash))) {
-      CacheHandle* handle = accessor->second;
-      table_.erase(accessor);
-      UnsetInCache(handle, &context);
-    }
-  }
+  EraseAndConfirm(key, hash, &context);
   Cleanup(context);
 }
 
+bool ClockCacheShard::EraseAndConfirm(const Slice& key, uint32_t hash,
+                                      CleanupContext* context) {
+  MutexLock l(&mutex_);
+  HashTable::accessor accessor;
+  bool erased = false;
+  if (table_.find(accessor, CacheKey(key, hash))) {
+    CacheHandle* handle = accessor->second;
+    table_.erase(accessor);
+    erased = UnsetInCache(handle, context);
+  }
+  return erased;
+}
+
 void ClockCacheShard::EraseUnRefEntries() {
   CleanupContext context;
   {

cache/lru_cache.cc

@@ -11,7 +11,7 @@
 #define __STDC_FORMAT_MACROS
 #endif
 
-#include "util/lru_cache.h"
+#include "cache/lru_cache.h"
 
 #include <assert.h>
 #include <stdio.h>
@@ -22,7 +22,7 @@
 
 namespace rocksdb {
 
-LRUHandleTable::LRUHandleTable() : length_(0), elems_(0), list_(nullptr) {
+LRUHandleTable::LRUHandleTable() : list_(nullptr), length_(0), elems_(0) {
   Resize();
 }
 
@@ -100,7 +100,7 @@ void LRUHandleTable::Resize() {
 }
 
 LRUCacheShard::LRUCacheShard()
-    : usage_(0), lru_usage_(0), high_pri_pool_usage_(0) {
+    : high_pri_pool_usage_(0), usage_(0), lru_usage_(0) {
   // Make empty circular linked list
   lru_.next = &lru_;
   lru_.prev = &lru_;
@@ -157,6 +157,16 @@ void LRUCacheShard::TEST_GetLRUList(LRUHandle** lru, LRUHandle** lru_low_pri) {
   *lru_low_pri = lru_low_pri_;
 }
 
+size_t LRUCacheShard::TEST_GetLRUSize() {
+  LRUHandle* lru_handle = lru_.next;
+  size_t lru_size = 0;
+  while (lru_handle != &lru_) {
+    lru_size++;
+    lru_handle = lru_handle->next;
+  }
+  return lru_size;
+}
+
 void LRUCacheShard::LRU_Remove(LRUHandle* e) {
   assert(e->next != nullptr);
   assert(e->prev != nullptr);
@@ -223,6 +233,22 @@ void LRUCacheShard::EvictFromLRU(size_t charge,
   }
 }
 
+void* LRUCacheShard::operator new(size_t size) {
+  return port::cacheline_aligned_alloc(size);
+}
+
+void* LRUCacheShard::operator new[](size_t size) {
+  return port::cacheline_aligned_alloc(size);
+}
+
+void LRUCacheShard::operator delete(void *memblock) {
+  port::cacheline_aligned_free(memblock);
+}
+
+void LRUCacheShard::operator delete[](void* memblock) {
+  port::cacheline_aligned_free(memblock);
+}
+
 void LRUCacheShard::SetCapacity(size_t capacity) {
   autovector<LRUHandle*> last_reference_list;
   {
@@ -273,9 +299,9 @@ void LRUCacheShard::SetHighPriorityPoolRatio(double high_pri_pool_ratio) {
   MaintainPoolSize();
 }
 
-void LRUCacheShard::Release(Cache::Handle* handle) {
+bool LRUCacheShard::Release(Cache::Handle* handle, bool force_erase) {
   if (handle == nullptr) {
-    return;
+    return false;
   }
   LRUHandle* e = reinterpret_cast<LRUHandle*>(handle);
   bool last_reference = false;
@@ -287,10 +313,10 @@ void LRUCacheShard::Release(Cache::Handle* handle) {
     }
     if (e->refs == 1 && e->InCache()) {
       // The item is still in cache, and nobody else holds a reference to it
-      if (usage_ > capacity_) {
+      if (usage_ > capacity_ || force_erase) {
         // the cache is full
         // The LRU list must be empty since the cache is full
-        assert(lru_.next == &lru_);
+        assert(!(usage_ > capacity_) || lru_.next == &lru_);
         // take this opportunity and remove the item
         table_.Remove(e->key(), e->hash);
         e->SetInCache(false);
@@ -308,6 +334,7 @@ void LRUCacheShard::Release(Cache::Handle* handle) {
   if (last_reference) {
     e->Free();
   }
+  return last_reference;
 }
 
 Status LRUCacheShard::Insert(const Slice& key, uint32_t hash, void* value,
@@ -437,11 +464,18 @@ std::string LRUCacheShard::GetPrintableOptions() const {
 LRUCache::LRUCache(size_t capacity, int num_shard_bits,
                    bool strict_capacity_limit, double high_pri_pool_ratio)
     : ShardedCache(capacity, num_shard_bits, strict_capacity_limit) {
-  int num_shards = 1 << num_shard_bits;
-  shards_ = new LRUCacheShard[num_shards];
+  num_shards_ = 1 << num_shard_bits;
+#if defined(_MSC_VER)
+#pragma warning(push)
+#pragma warning(disable: 4316) // We've validated the alignment with the new operators
+#endif
+  shards_ = new LRUCacheShard[num_shards_];
+#if defined(_MSC_VER)
+#pragma warning(pop)
+#endif
   SetCapacity(capacity);
   SetStrictCapacityLimit(strict_capacity_limit);
-  for (int i = 0; i < num_shards; i++) {
+  for (int i = 0; i < num_shards_; i++) {
     shards_[i].SetHighPriorityPoolRatio(high_pri_pool_ratio);
   }
 }
@@ -468,7 +502,20 @@ uint32_t LRUCache::GetHash(Handle* handle) const {
   return reinterpret_cast<const LRUHandle*>(handle)->hash;
 }
 
-void LRUCache::DisownData() { shards_ = nullptr; }
+void LRUCache::DisownData() {
+// Do not drop data if compile with ASAN to suppress leak warning.
+#ifndef __SANITIZE_ADDRESS__
+  shards_ = nullptr;
+#endif  // !__SANITIZE_ADDRESS__
+}
+
+size_t LRUCache::TEST_GetLRUSize() {
+  size_t lru_size_of_all_shards = 0;
+  for (int i = 0; i < num_shards_; i++) {
+    lru_size_of_all_shards += shards_[i].TEST_GetLRUSize();
+  }
+  return lru_size_of_all_shards;
+}
 
 std::shared_ptr<Cache> NewLRUCache(size_t capacity, int num_shard_bits,
                                    bool strict_capacity_limit,

cache/lru_cache.h

@@ -10,7 +10,7 @@
 
 #include <string>
 
-#include "util/sharded_cache.h"
+#include "cache/sharded_cache.h"
 
 #include "port/port.h"
 #include "util/autovector.h"
@@ -148,13 +148,13 @@ class LRUHandleTable {
 
   // The table consists of an array of buckets where each bucket is
   // a linked list of cache entries that hash into the bucket.
+  LRUHandle** list_;
   uint32_t length_;
   uint32_t elems_;
-  LRUHandle** list_;
 };
 
 // A single shard of sharded cache.
-class LRUCacheShard : public CacheShard {
+class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard : public CacheShard {
  public:
   LRUCacheShard();
   virtual ~LRUCacheShard();
@@ -178,7 +178,8 @@ class LRUCacheShard : public CacheShard {
                         Cache::Priority priority) override;
   virtual Cache::Handle* Lookup(const Slice& key, uint32_t hash) override;
   virtual bool Ref(Cache::Handle* handle) override;
-  virtual void Release(Cache::Handle* handle) override;
+  virtual bool Release(Cache::Handle* handle,
+                       bool force_erase = false) override;
   virtual void Erase(const Slice& key, uint32_t hash) override;
 
   // Although in some platforms the update of size_t is atomic, to make sure
@@ -197,6 +198,19 @@ class LRUCacheShard : public CacheShard {
 
   void TEST_GetLRUList(LRUHandle** lru, LRUHandle** lru_low_pri);
 
+  //  Retrieves number of elements in LRU, for unit test purpose only
+  //  not threadsafe
+  size_t TEST_GetLRUSize();
+
+  // Overloading to aligned it to cache line size
+  void* operator new(size_t);
+
+  void* operator new[](size_t);
+
+  void operator delete(void *);
+
+  void operator delete[](void*);
+
  private:
   void LRU_Remove(LRUHandle* e);
   void LRU_Insert(LRUHandle* e);
@@ -218,12 +232,6 @@ class LRUCacheShard : public CacheShard {
   // Initialized before use.
   size_t capacity_;
 
-  // Memory size for entries residing in the cache
-  size_t usage_;
-
-  // Memory size for entries residing only in the LRU list
-  size_t lru_usage_;
-
   // Memory size for entries in high-pri pool.
   size_t high_pri_pool_usage_;
 
@@ -237,11 +245,6 @@ class LRUCacheShard : public CacheShard {
   // Remember the value to avoid recomputing each time.
   double high_pri_pool_capacity_;
 
-  // mutex_ protects the following state.
-  // We don't count mutex_ as the cache's internal state so semantically we
-  // don't mind mutex_ invoking the non-const actions.
-  mutable port::Mutex mutex_;
-
   // Dummy head of LRU list.
   // lru.prev is newest entry, lru.next is oldest entry.
   // LRU contains items which can be evicted, ie reference only by cache
@@ -250,7 +253,29 @@ class LRUCacheShard : public CacheShard {
   // Pointer to head of low-pri pool in LRU list.
   LRUHandle* lru_low_pri_;
 
+  // ------------^^^^^^^^^^^^^-----------
+  // Not frequently modified data members
+  // ------------------------------------
+  //
+  // We separate data members that are updated frequently from the ones that
+  // are not frequently updated so that they don't share the same cache line
+  // which will lead into false cache sharing
+  //
+  // ------------------------------------
+  // Frequently modified data members
+  // ------------vvvvvvvvvvvvv-----------
   LRUHandleTable table_;
+
+  // Memory size for entries residing in the cache
+  size_t usage_;
+
+  // Memory size for entries residing only in the LRU list
+  size_t lru_usage_;
+
+  // mutex_ protects the following state.
+  // We don't count mutex_ as the cache's internal state so semantically we
+  // don't mind mutex_ invoking the non-const actions.
+  mutable port::Mutex mutex_;
 };
 
 class LRUCache : public ShardedCache {
@@ -266,8 +291,12 @@ class LRUCache : public ShardedCache {
   virtual uint32_t GetHash(Handle* handle) const override;
   virtual void DisownData() override;
 
+  //  Retrieves number of elements in LRU, for unit test purpose only
+  size_t TEST_GetLRUSize();
+
  private:
   LRUCacheShard* shards_;
+  int num_shards_ = 0;
 };
 
 }  // namespace rocksdb

cache/lru_cache_test.cc

@@ -3,7 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 
-#include "util/lru_cache.h"
+#include "cache/lru_cache.h"
 
 #include <string>
 #include <vector>
@@ -17,7 +17,16 @@ class LRUCacheTest : public testing::Test {
   ~LRUCacheTest() {}
 
   void NewCache(size_t capacity, double high_pri_pool_ratio = 0.0) {
-    cache_.reset(new LRUCacheShard());
+    cache_.reset(
+#if defined(_MSC_VER)
+#pragma warning(push)
+#pragma warning(disable: 4316) // We've validated the alignment with the new operators
+#endif
+      new LRUCacheShard()
+#if defined(_MSC_VER)
+#pragma warning(pop)
+#endif
+    );
     cache_->SetCapacity(capacity);
     cache_->SetStrictCapacityLimit(false);
     cache_->SetHighPriorityPoolRatio(high_pri_pool_ratio);

cache/sharded_cache.cc

@@ -11,7 +11,7 @@
 #define __STDC_FORMAT_MACROS
 #endif
 
-#include "util/sharded_cache.h"
+#include "cache/sharded_cache.h"
 
 #include <string>
 
@@ -63,9 +63,9 @@ bool ShardedCache::Ref(Handle* handle) {
   return GetShard(Shard(hash))->Ref(handle);
 }
 
-void ShardedCache::Release(Handle* handle) {
+bool ShardedCache::Release(Handle* handle, bool force_erase) {
   uint32_t hash = GetHash(handle);
-  GetShard(Shard(hash))->Release(handle);
+  return GetShard(Shard(hash))->Release(handle, force_erase);
 }
 
 void ShardedCache::Erase(const Slice& key) {

cache/sharded_cache.h

@@ -30,7 +30,7 @@ class CacheShard {
                         Cache::Handle** handle, Cache::Priority priority) = 0;
   virtual Cache::Handle* Lookup(const Slice& key, uint32_t hash) = 0;
   virtual bool Ref(Cache::Handle* handle) = 0;
-  virtual void Release(Cache::Handle* handle) = 0;
+  virtual bool Release(Cache::Handle* handle, bool force_erase = false) = 0;
   virtual void Erase(const Slice& key, uint32_t hash) = 0;
   virtual void SetCapacity(size_t capacity) = 0;
   virtual void SetStrictCapacityLimit(bool strict_capacity_limit) = 0;
@@ -65,7 +65,7 @@ class ShardedCache : public Cache {
                         Handle** handle, Priority priority) override;
   virtual Handle* Lookup(const Slice& key, Statistics* stats) override;
   virtual bool Ref(Handle* handle) override;
-  virtual void Release(Handle* handle) override;
+  virtual bool Release(Handle* handle, bool force_erase = false) override;
   virtual void Erase(const Slice& key) override;
   virtual uint64_t NewId() override;
   virtual size_t GetCapacity() const override;

cmake/RocksDBConfig.cmake.in

@@ -0,0 +1,3 @@
+@PACKAGE_INIT@
+include("${CMAKE_CURRENT_LIST_DIR}/RocksDBTargets.cmake")
+check_required_components(RocksDB)

cmake/modules/FindJeMalloc.cmake

@@ -1,22 +1,21 @@
 # - Find JeMalloc library
 # Find the native JeMalloc includes and library
-# This module defines
-#  JEMALLOC_INCLUDE_DIRS, where to find jemalloc.h, Set when
-#                        JEMALLOC_INCLUDE_DIR is found.
-#  JEMALLOC_LIBRARIES, libraries to link against to use JeMalloc.
-#  JEMALLOC_FOUND, If false, do not try to use JeMalloc.
 #
+# JEMALLOC_INCLUDE_DIR - where to find jemalloc.h, etc.
+# JEMALLOC_LIBRARIES - List of libraries when using jemalloc.
+# JEMALLOC_FOUND - True if jemalloc found.
+
 find_path(JEMALLOC_INCLUDE_DIR
-  jemalloc/jemalloc.h)
+  NAMES jemalloc/jemalloc.h
+  HINTS ${JEMALLOC_ROOT_DIR}/include)
 
 find_library(JEMALLOC_LIBRARIES
-  jemalloc)
+  NAMES jemalloc
+  HINTS ${JEMALLOC_ROOT_DIR}/lib)
 
 include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(JeMalloc DEFAULT_MSG
-    JEMALLOC_LIBRARIES JEMALLOC_INCLUDE_DIR)
+find_package_handle_standard_args(jemalloc DEFAULT_MSG JEMALLOC_LIBRARIES JEMALLOC_INCLUDE_DIR)
 
-MARK_AS_ADVANCED(
-  JEMALLOC_INCLUDE_DIR
+mark_as_advanced(
   JEMALLOC_LIBRARIES
-)
+  JEMALLOC_INCLUDE_DIR)

cmake/modules/Findbzip2.cmake

@@ -0,0 +1,21 @@
+# - Find Bzip2
+# Find the bzip2 compression library and includes
+#
+# BZIP2_INCLUDE_DIR - where to find bzlib.h, etc.
+# BZIP2_LIBRARIES - List of libraries when using bzip2.
+# BZIP2_FOUND - True if bzip2 found.
+
+find_path(BZIP2_INCLUDE_DIR
+  NAMES bzlib.h
+  HINTS ${BZIP2_ROOT_DIR}/include)
+
+find_library(BZIP2_LIBRARIES
+  NAMES bz2
+  HINTS ${BZIP2_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(bzip2 DEFAULT_MSG BZIP2_LIBRARIES BZIP2_INCLUDE_DIR)
+
+mark_as_advanced(
+  BZIP2_LIBRARIES
+  BZIP2_INCLUDE_DIR)

cmake/modules/Findlz4.cmake

@@ -0,0 +1,21 @@
+# - Find Lz4
+# Find the lz4 compression library and includes
+#
+# LZ4_INCLUDE_DIR - where to find lz4.h, etc.
+# LZ4_LIBRARIES - List of libraries when using lz4.
+# LZ4_FOUND - True if lz4 found.
+
+find_path(LZ4_INCLUDE_DIR
+  NAMES lz4.h
+  HINTS ${LZ4_ROOT_DIR}/include)
+
+find_library(LZ4_LIBRARIES
+  NAMES lz4
+  HINTS ${LZ4_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(lz4 DEFAULT_MSG LZ4_LIBRARIES LZ4_INCLUDE_DIR)
+
+mark_as_advanced(
+  LZ4_LIBRARIES
+  LZ4_INCLUDE_DIR)

cmake/modules/Findsnappy.cmake

@@ -13,8 +13,6 @@ find_library(SNAPPY_LIBRARIES
   NAMES snappy
   HINTS ${SNAPPY_ROOT_DIR}/lib)
 
-# handle the QUIETLY and REQUIRED arguments and set UUID_FOUND to TRUE if
-# all listed variables are TRUE
 include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(snappy DEFAULT_MSG SNAPPY_LIBRARIES SNAPPY_INCLUDE_DIR)
 

cmake/modules/Findzlib.cmake

@@ -0,0 +1,21 @@
+# - Find zlib
+# Find the zlib compression library and includes
+#
+# ZLIB_INCLUDE_DIR - where to find zlib.h, etc.
+# ZLIB_LIBRARIES - List of libraries when using zlib.
+# ZLIB_FOUND - True if zlib found.
+
+find_path(ZLIB_INCLUDE_DIR
+  NAMES zlib.h
+  HINTS ${ZLIB_ROOT_DIR}/include)
+
+find_library(ZLIB_LIBRARIES
+  NAMES z
+  HINTS ${ZLIB_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(zlib DEFAULT_MSG ZLIB_LIBRARIES ZLIB_INCLUDE_DIR)
+
+mark_as_advanced(
+  ZLIB_LIBRARIES
+  ZLIB_INCLUDE_DIR)

cmake/modules/Findzstd.cmake

@@ -0,0 +1,21 @@
+# - Find zstd
+# Find the zstd compression library and includes
+#
+# ZSTD_INCLUDE_DIR - where to find zstd.h, etc.
+# ZSTD_LIBRARIES - List of libraries when using zstd.
+# ZSTD_FOUND - True if zstd found.
+
+find_path(ZSTD_INCLUDE_DIR
+  NAMES zstd.h
+  HINTS ${ZSTD_ROOT_DIR}/include)
+
+find_library(ZSTD_LIBRARIES
+  NAMES zstd
+  HINTS ${ZSTD_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(zstd DEFAULT_MSG ZSTD_LIBRARIES ZSTD_INCLUDE_DIR)
+
+mark_as_advanced(
+  ZSTD_LIBRARIES
+  ZSTD_INCLUDE_DIR)

coverage/coverage_test.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Exit on error.
 set -e

db/builder.cc

@@ -16,11 +16,12 @@
 #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"
 #include "db/table_cache.h"
 #include "db/version_edit.h"
+#include "monitoring/iostats_context_imp.h"
+#include "monitoring/thread_status_util.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
 #include "rocksdb/iterator.h"
@@ -29,9 +30,9 @@
 #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/filename.h"
 #include "util/stop_watch.h"
-#include "util/thread_status_util.h"
+#include "util/sync_point.h"
 
 namespace rocksdb {
 
@@ -44,9 +45,9 @@ TableBuilder* NewTableBuilder(
         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,
-    int level,
-    const std::string* compression_dict, const bool skip_filters) {
+    const CompressionOptions& compression_opts, int level,
+    const std::string* compression_dict, const bool skip_filters,
+    const uint64_t creation_time) {
   assert((column_family_id ==
           TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
          column_family_name.empty());
@@ -54,7 +55,7 @@ TableBuilder* NewTableBuilder(
       TableBuilderOptions(ioptions, internal_comparator,
                           int_tbl_prop_collector_factories, compression_type,
                           compression_opts, compression_dict, skip_filters,
-                          column_family_name, level),
+                          column_family_name, level, creation_time),
       column_family_id, file);
 }
 
@@ -73,7 +74,8 @@ Status BuildTable(
     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) {
+    TableProperties* table_properties, int level,
+    const uint64_t creation_time) {
   assert((column_family_id ==
           TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
          column_family_name.empty());
@@ -103,6 +105,10 @@ Status BuildTable(
     unique_ptr<WritableFileWriter> file_writer;
     {
       unique_ptr<WritableFile> file;
+#ifndef NDEBUG
+      bool use_direct_writes = env_options.use_direct_writes;
+      TEST_SYNC_POINT_CALLBACK("BuildTable:create_file", &use_direct_writes);
+#endif  // !NDEBUG
       s = NewWritableFile(env, fname, &file, env_options);
       if (!s.ok()) {
         EventHelpers::LogAndNotifyTableFileCreationFinished(
@@ -118,7 +124,8 @@ Status BuildTable(
       builder = NewTableBuilder(
           ioptions, internal_comparator, int_tbl_prop_collector_factories,
           column_family_id, column_family_name, file_writer.get(), compression,
-          compression_opts, level);
+          compression_opts, level, nullptr /* compression_dict */,
+          false /* skip_filters */, creation_time);
     }
 
     MergeHelper merge(env, internal_comparator.user_comparator(),
@@ -172,7 +179,7 @@ Status BuildTable(
     // Finish and check for file errors
     if (s.ok() && !empty) {
       StopWatch sw(env, ioptions.statistics, TABLE_SYNC_MICROS);
-      file_writer->Sync(ioptions.use_fsync);
+      s = file_writer->Sync(ioptions.use_fsync);
     }
     if (s.ok() && !empty) {
       s = file_writer->Close();
@@ -180,6 +187,11 @@ Status BuildTable(
 
     if (s.ok() && !empty) {
       // Verify that the table is usable
+      // We set for_compaction to false and don't OptimizeForCompactionTableRead
+      // here because this is a special case after we finish the table building
+      // No matter whether use_direct_io_for_flush_and_compaction is true,
+      // we will regrad this verification as user reads since the goal is
+      // to cache it here for further user reads
       std::unique_ptr<InternalIterator> it(table_cache->NewIterator(
           ReadOptions(), env_options, internal_comparator, meta->fd,
           nullptr /* range_del_agg */, nullptr,

db/builder.h

@@ -10,6 +10,7 @@
 #include <utility>
 #include <vector>
 #include "db/table_properties_collector.h"
+#include "options/cf_options.h"
 #include "rocksdb/comparator.h"
 #include "rocksdb/env.h"
 #include "rocksdb/listener.h"
@@ -18,7 +19,6 @@
 #include "rocksdb/table_properties.h"
 #include "rocksdb/types.h"
 #include "table/scoped_arena_iterator.h"
-#include "util/cf_options.h"
 #include "util/event_logger.h"
 
 namespace rocksdb {
@@ -48,10 +48,9 @@ TableBuilder* NewTableBuilder(
         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,
-    int level,
+    const CompressionOptions& compression_opts, int level,
     const std::string* compression_dict = nullptr,
-    const bool skip_filters = false);
+    const bool skip_filters = false, const uint64_t creation_time = 0);
 
 // Build a Table file from the contents of *iter.  The generated file
 // will be named according to number specified in meta. On success, the rest of
@@ -77,6 +76,7 @@ extern Status BuildTable(
     InternalStats* internal_stats, TableFileCreationReason reason,
     EventLogger* event_logger = nullptr, int job_id = 0,
     const Env::IOPriority io_priority = Env::IO_HIGH,
-    TableProperties* table_properties = nullptr, int level = -1);
+    TableProperties* table_properties = nullptr, int level = -1,
+    const uint64_t creation_time = 0);
 
 }  // namespace rocksdb

db/c_test.c

@@ -20,7 +20,7 @@
 // Can not use port/port.h macros as this is a c file
 #ifdef OS_WIN
 
-#include <Windows.h>
+#include <windows.h>
 
 #define snprintf _snprintf
 
@@ -40,6 +40,8 @@ const char* phase = "";
 static char dbname[200];
 static char sstfilename[200];
 static char dbbackupname[200];
+static char dbcheckpointname[200];
+static char dbpathname[200];
 
 static void StartPhase(const char* name) {
   fprintf(stderr, "=== Test %s\n", name);
@@ -87,6 +89,16 @@ static void Free(char** ptr) {
   }
 }
 
+static void CheckValue(
+    char* err,
+    const char* expected,
+    char** actual,
+    size_t actual_length) {
+  CheckNoError(err);
+  CheckEqual(expected, *actual, actual_length);
+  Free(actual);
+}
+
 static void CheckGet(
     rocksdb_t* db,
     const rocksdb_readoptions_t* options,
@@ -116,6 +128,32 @@ static void CheckGetCF(
   Free(&val);
 }
 
+static void CheckPinGet(rocksdb_t* db, const rocksdb_readoptions_t* options,
+                        const char* key, const char* expected) {
+  char* err = NULL;
+  size_t val_len;
+  const char* val;
+  rocksdb_pinnableslice_t* p;
+  p = rocksdb_get_pinned(db, options, key, strlen(key), &err);
+  CheckNoError(err);
+  val = rocksdb_pinnableslice_value(p, &val_len);
+  CheckEqual(expected, val, val_len);
+  rocksdb_pinnableslice_destroy(p);
+}
+
+static void CheckPinGetCF(rocksdb_t* db, const rocksdb_readoptions_t* options,
+                          rocksdb_column_family_handle_t* handle,
+                          const char* key, const char* expected) {
+  char* err = NULL;
+  size_t val_len;
+  const char* val;
+  rocksdb_pinnableslice_t* p;
+  p = rocksdb_get_pinned_cf(db, options, handle, key, strlen(key), &err);
+  CheckNoError(err);
+  val = rocksdb_pinnableslice_value(p, &val_len);
+  CheckEqual(expected, val, val_len);
+  rocksdb_pinnableslice_destroy(p);
+}
 
 static void CheckIter(rocksdb_iterator_t* iter,
                       const char* key, const char* val) {
@@ -282,10 +320,53 @@ static char* MergeOperatorPartialMerge(
   return result;
 }
 
+static void CheckTxnGet(
+        rocksdb_transaction_t* txn,
+        const rocksdb_readoptions_t* options,
+        const char* key,
+        const char* expected) {
+        char* err = NULL;
+        size_t val_len;
+        char* val;
+        val = rocksdb_transaction_get(txn, options, key, strlen(key), &val_len, &err);
+        CheckNoError(err);
+        CheckEqual(expected, val, val_len);
+        Free(&val);
+}
+
+static void CheckTxnDBGet(
+        rocksdb_transactiondb_t* txn_db,
+        const rocksdb_readoptions_t* options,
+        const char* key,
+        const char* expected) {
+        char* err = NULL;
+        size_t val_len;
+        char* val;
+        val = rocksdb_transactiondb_get(txn_db, options, key, strlen(key), &val_len, &err);
+        CheckNoError(err);
+        CheckEqual(expected, val, val_len);
+        Free(&val);
+}
+
+static void CheckTxnDBGetCF(rocksdb_transactiondb_t* txn_db,
+                            const rocksdb_readoptions_t* options,
+                            rocksdb_column_family_handle_t* column_family,
+                            const char* key, const char* expected) {
+  char* err = NULL;
+  size_t val_len;
+  char* val;
+  val = rocksdb_transactiondb_get_cf(txn_db, options, column_family, key,
+                                     strlen(key), &val_len, &err);
+  CheckNoError(err);
+  CheckEqual(expected, val, val_len);
+  Free(&val);
+}
+
 int main(int argc, char** argv) {
   rocksdb_t* db;
   rocksdb_comparator_t* cmp;
   rocksdb_cache_t* cache;
+  rocksdb_dbpath_t *dbpath;
   rocksdb_env_t* env;
   rocksdb_options_t* options;
   rocksdb_compactoptions_t* coptions;
@@ -293,6 +374,10 @@ int main(int argc, char** argv) {
   rocksdb_readoptions_t* roptions;
   rocksdb_writeoptions_t* woptions;
   rocksdb_ratelimiter_t* rate_limiter;
+  rocksdb_transactiondb_t* txn_db;
+  rocksdb_transactiondb_options_t* txn_db_options;
+  rocksdb_transaction_t* txn;
+  rocksdb_transaction_options_t* txn_options;
   char* err = NULL;
   int run = -1;
 
@@ -306,13 +391,24 @@ int main(int argc, char** argv) {
            GetTempDir(),
            ((int) geteuid()));
 
+  snprintf(dbcheckpointname, sizeof(dbcheckpointname),
+           "%s/rocksdb_c_test-%d-checkpoint",
+           GetTempDir(),
+           ((int) geteuid()));
+
   snprintf(sstfilename, sizeof(sstfilename),
            "%s/rocksdb_c_test-%d-sst",
            GetTempDir(),
            ((int)geteuid()));
 
+  snprintf(dbpathname, sizeof(dbpathname),
+           "%s/rocksdb_c_test-%d-dbpath",
+           GetTempDir(),
+           ((int) geteuid()));
+
   StartPhase("create_objects");
   cmp = rocksdb_comparator_create(NULL, CmpDestroy, CmpCompare, CmpName);
+  dbpath = rocksdb_dbpath_create(dbpathname, 1024 * 1024);
   env = rocksdb_create_default_env();
   cache = rocksdb_cache_create_lru(100000);
 
@@ -421,6 +517,36 @@ int main(int argc, char** argv) {
     rocksdb_backup_engine_close(be);
   }
 
+  StartPhase("checkpoint");
+  {
+    rocksdb_destroy_db(options, dbcheckpointname, &err);
+    CheckNoError(err);
+
+    rocksdb_checkpoint_t* checkpoint = rocksdb_checkpoint_object_create(db, &err);
+    CheckNoError(err);
+
+    rocksdb_checkpoint_create(checkpoint, dbcheckpointname, 0, &err);
+    CheckNoError(err);
+
+    // start a new database from the checkpoint
+    rocksdb_close(db);
+    rocksdb_options_set_error_if_exists(options, 0);
+    db = rocksdb_open(options, dbcheckpointname, &err);
+    CheckNoError(err);
+
+    CheckGet(db, roptions, "foo", "hello");
+
+    rocksdb_checkpoint_object_destroy(checkpoint);
+
+    rocksdb_close(db);
+    rocksdb_destroy_db(options, dbcheckpointname, &err);
+    CheckNoError(err);
+
+    db = rocksdb_open(options, dbname, &err);
+    CheckNoError(err);
+    rocksdb_options_set_error_if_exists(options, 1);
+  }
+
   StartPhase("compactall");
   rocksdb_compact_range(db, NULL, 0, NULL, 0);
   CheckGet(db, roptions, "foo", "hello");
@@ -465,11 +591,11 @@ int main(int argc, char** argv) {
     unlink(sstfilename);
     rocksdb_sstfilewriter_open(writer, sstfilename, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk1", 5, "v1", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk1", 5, "v1", 2, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk2", 5, "v2", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk2", 5, "v2", 2, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk3", 5, "v3", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk3", 5, "v3", 2, &err);
     CheckNoError(err);
     rocksdb_sstfilewriter_finish(writer, &err);
     CheckNoError(err);
@@ -486,11 +612,11 @@ int main(int argc, char** argv) {
     unlink(sstfilename);
     rocksdb_sstfilewriter_open(writer, sstfilename, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk2", 5, "v4", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk2", 5, "v4", 2, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk22", 6, "v5", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk22", 6, "v5", 2, &err);
     CheckNoError(err);
-    rocksdb_sstfilewriter_add(writer, "sstk3", 5, "v6", 2, &err);
+    rocksdb_sstfilewriter_put(writer, "sstk3", 5, "v6", 2, &err);
     CheckNoError(err);
     rocksdb_sstfilewriter_finish(writer, &err);
     CheckNoError(err);
@@ -577,10 +703,13 @@ int main(int argc, char** argv) {
   {
     rocksdb_writebatch_t* wb = rocksdb_writebatch_create();
     rocksdb_writebatch_set_save_point(wb);
+    rocksdb_writebatch_set_save_point(wb);
     const char* k_list[2] = {"z", "ap"};
     const size_t k_sizes[2] = {1, 2};
     const char* v_list[3] = {"x", "y", "z"};
     const size_t v_sizes[3] = {1, 1, 1};
+    rocksdb_writebatch_pop_save_point(wb, &err);
+    CheckNoError(err);
     rocksdb_writebatch_putv(wb, 2, k_list, k_sizes, 3, v_list, v_sizes);
     rocksdb_writebatch_rollback_to_save_point(wb, &err);
     CheckNoError(err);
@@ -608,6 +737,90 @@ int main(int argc, char** argv) {
     rocksdb_writebatch_destroy(wb2);
   }
 
+  StartPhase("writebatch_wi");
+  {
+    rocksdb_writebatch_wi_t* wbi = rocksdb_writebatch_wi_create(0, 1);
+    rocksdb_writebatch_wi_put(wbi, "foo", 3, "a", 1);
+    rocksdb_writebatch_wi_clear(wbi);
+    rocksdb_writebatch_wi_put(wbi, "bar", 3, "b", 1);
+    rocksdb_writebatch_wi_put(wbi, "box", 3, "c", 1);
+    rocksdb_writebatch_wi_delete(wbi, "bar", 3);
+    int count = rocksdb_writebatch_wi_count(wbi);
+    CheckCondition(count == 3);
+    size_t size;
+    char* value;
+    value = rocksdb_writebatch_wi_get_from_batch(wbi, options, "box", 3, &size, &err);
+    CheckValue(err, "c", &value, size);
+    value = rocksdb_writebatch_wi_get_from_batch(wbi, options, "bar", 3, &size, &err);
+    CheckValue(err, NULL, &value, size);
+    value = rocksdb_writebatch_wi_get_from_batch_and_db(wbi, db, roptions, "foo", 3, &size, &err);
+    CheckValue(err, "hello", &value, size);
+    value = rocksdb_writebatch_wi_get_from_batch_and_db(wbi, db, roptions, "box", 3, &size, &err);
+    CheckValue(err, "c", &value, size);
+    rocksdb_write_writebatch_wi(db, woptions, wbi, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "foo", "hello");
+    CheckGet(db, roptions, "bar", NULL);
+    CheckGet(db, roptions, "box", "c");
+    int pos = 0;
+    rocksdb_writebatch_wi_iterate(wbi, &pos, CheckPut, CheckDel);
+    CheckCondition(pos == 3);
+    rocksdb_writebatch_wi_clear(wbi);
+    rocksdb_writebatch_wi_put(wbi, "bar", 3, "b", 1);
+    rocksdb_writebatch_wi_put(wbi, "bay", 3, "d", 1);
+    rocksdb_writebatch_wi_delete_range(wbi, "bar", 3, "bay", 3);
+    rocksdb_write_writebatch_wi(db, woptions, wbi, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "bar", NULL);
+    CheckGet(db, roptions, "bay", "d");
+    rocksdb_writebatch_wi_clear(wbi);
+    const char* start_list[1] = {"bay"};
+    const size_t start_sizes[1] = {3};
+    const char* end_list[1] = {"baz"};
+    const size_t end_sizes[1] = {3};
+    rocksdb_writebatch_wi_delete_rangev(wbi, 1, start_list, start_sizes, end_list,
+                                     end_sizes);
+    rocksdb_write_writebatch_wi(db, woptions, wbi, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "bay", NULL);
+    rocksdb_writebatch_wi_destroy(wbi);
+  }
+
+  StartPhase("writebatch_wi_vectors");
+  {
+    rocksdb_writebatch_wi_t* wb = rocksdb_writebatch_wi_create(0, 1);
+    const char* k_list[2] = { "z", "ap" };
+    const size_t k_sizes[2] = { 1, 2 };
+    const char* v_list[3] = { "x", "y", "z" };
+    const size_t v_sizes[3] = { 1, 1, 1 };
+    rocksdb_writebatch_wi_putv(wb, 2, k_list, k_sizes, 3, v_list, v_sizes);
+    rocksdb_write_writebatch_wi(db, woptions, wb, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "zap", "xyz");
+    rocksdb_writebatch_wi_delete(wb, "zap", 3);
+    rocksdb_write_writebatch_wi(db, woptions, wb, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "zap", NULL);
+    rocksdb_writebatch_wi_destroy(wb);
+  }
+
+  StartPhase("writebatch_wi_savepoint");
+  {
+    rocksdb_writebatch_wi_t* wb = rocksdb_writebatch_wi_create(0, 1);
+    rocksdb_writebatch_wi_set_save_point(wb);
+    const char* k_list[2] = {"z", "ap"};
+    const size_t k_sizes[2] = {1, 2};
+    const char* v_list[3] = {"x", "y", "z"};
+    const size_t v_sizes[3] = {1, 1, 1};
+    rocksdb_writebatch_wi_putv(wb, 2, k_list, k_sizes, 3, v_list, v_sizes);
+    rocksdb_writebatch_wi_rollback_to_save_point(wb, &err);
+    CheckNoError(err);
+    rocksdb_write_writebatch_wi(db, woptions, wb, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "zap", NULL);
+    rocksdb_writebatch_wi_destroy(wb);
+  }
+
   StartPhase("iter");
   {
     rocksdb_iterator_t* iter = rocksdb_create_iterator(db, roptions);
@@ -634,6 +847,37 @@ int main(int argc, char** argv) {
     rocksdb_iter_destroy(iter);
   }
 
+  StartPhase("wbwi_iter");
+  {
+    rocksdb_iterator_t* base_iter = rocksdb_create_iterator(db, roptions);
+    rocksdb_writebatch_wi_t* wbi = rocksdb_writebatch_wi_create(0, 1);
+    rocksdb_writebatch_wi_put(wbi, "bar", 3, "b", 1);
+    rocksdb_writebatch_wi_delete(wbi, "foo", 3);
+    rocksdb_iterator_t* iter = rocksdb_writebatch_wi_create_iterator_with_base(wbi, base_iter);
+    CheckCondition(!rocksdb_iter_valid(iter));
+    rocksdb_iter_seek_to_first(iter);
+    CheckCondition(rocksdb_iter_valid(iter));
+    CheckIter(iter, "bar", "b");
+    rocksdb_iter_next(iter);
+    CheckIter(iter, "box", "c");
+    rocksdb_iter_prev(iter);
+    CheckIter(iter, "bar", "b");
+    rocksdb_iter_prev(iter);
+    CheckCondition(!rocksdb_iter_valid(iter));
+    rocksdb_iter_seek_to_last(iter);
+    CheckIter(iter, "box", "c");
+    rocksdb_iter_seek(iter, "b", 1);
+    CheckIter(iter, "bar", "b");
+    rocksdb_iter_seek_for_prev(iter, "c", 1);
+    CheckIter(iter, "box", "c");
+    rocksdb_iter_seek_for_prev(iter, "box", 3);
+    CheckIter(iter, "box", "c");
+    rocksdb_iter_get_error(iter, &err);
+    CheckNoError(err);
+    rocksdb_iter_destroy(iter);
+    rocksdb_writebatch_wi_destroy(wbi);
+  }
+
   StartPhase("multiget");
   {
     const char* keys[3] = { "box", "foo", "notfound" };
@@ -661,6 +905,13 @@ int main(int argc, char** argv) {
     }
   }
 
+  StartPhase("pin_get");
+  {
+    CheckPinGet(db, roptions, "box", "c");
+    CheckPinGet(db, roptions, "foo", "hello");
+    CheckPinGet(db, roptions, "notfound", NULL);
+  }
+
   StartPhase("approximate_sizes");
   {
     int i;
@@ -872,11 +1123,13 @@ int main(int argc, char** argv) {
     CheckNoError(err);
 
     CheckGetCF(db, roptions, handles[1], "foo", "hello");
+    CheckPinGetCF(db, roptions, handles[1], "foo", "hello");
 
     rocksdb_delete_cf(db, woptions, handles[1], "foo", 3, &err);
     CheckNoError(err);
 
     CheckGetCF(db, roptions, handles[1], "foo", NULL);
+    CheckPinGetCF(db, roptions, handles[1], "foo", NULL);
 
     rocksdb_writebatch_t* wb = rocksdb_writebatch_create();
     rocksdb_writebatch_put_cf(wb, handles[1], "baz", 3, "a", 1);
@@ -889,6 +1142,9 @@ int main(int argc, char** argv) {
     CheckGetCF(db, roptions, handles[1], "baz", NULL);
     CheckGetCF(db, roptions, handles[1], "bar", NULL);
     CheckGetCF(db, roptions, handles[1], "box", "c");
+    CheckPinGetCF(db, roptions, handles[1], "baz", NULL);
+    CheckPinGetCF(db, roptions, handles[1], "bar", NULL);
+    CheckPinGetCF(db, roptions, handles[1], "box", "c");
     rocksdb_writebatch_destroy(wb);
 
     const char* keys[3] = { "box", "box", "barfooxx" };
@@ -1100,14 +1356,127 @@ int main(int argc, char** argv) {
     }
   }
 
-  // Simple sanity check that setting memtable rep works.
-  StartPhase("memtable_reps");
+  StartPhase("transactions");
   {
-    // Create database with vector memtable.
     rocksdb_close(db);
     rocksdb_destroy_db(options, dbname, &err);
     CheckNoError(err);
 
+    // open a TransactionDB
+    txn_db_options = rocksdb_transactiondb_options_create();
+    txn_options = rocksdb_transaction_options_create();
+    rocksdb_options_set_create_if_missing(options, 1);
+    txn_db = rocksdb_transactiondb_open(options, txn_db_options, dbname, &err);
+    CheckNoError(err);
+
+    // put outside a transaction
+    rocksdb_transactiondb_put(txn_db, woptions, "foo", 3, "hello", 5, &err);
+    CheckNoError(err);
+    CheckTxnDBGet(txn_db, roptions, "foo", "hello");
+
+    // delete from outside transaction
+    rocksdb_transactiondb_delete(txn_db, woptions, "foo", 3, &err);
+    CheckNoError(err);
+    CheckTxnDBGet(txn_db, roptions, "foo", NULL);
+
+    // write batch into TransactionDB
+    rocksdb_writebatch_t* wb = rocksdb_writebatch_create();
+    rocksdb_writebatch_put(wb, "foo", 3, "a", 1);
+    rocksdb_writebatch_clear(wb);
+    rocksdb_writebatch_put(wb, "bar", 3, "b", 1);
+    rocksdb_writebatch_put(wb, "box", 3, "c", 1);
+    rocksdb_writebatch_delete(wb, "bar", 3);
+    rocksdb_transactiondb_write(txn_db, woptions, wb, &err);
+    rocksdb_writebatch_destroy(wb);
+    CheckTxnDBGet(txn_db, roptions, "box", "c");
+    CheckNoError(err);
+
+    // begin a transaction
+    txn = rocksdb_transaction_begin(txn_db, woptions, txn_options, NULL);
+    // put
+    rocksdb_transaction_put(txn, "foo", 3, "hello", 5, &err);
+    CheckNoError(err);
+    CheckTxnGet(txn, roptions, "foo", "hello");
+    // delete
+    rocksdb_transaction_delete(txn, "foo", 3, &err);
+    CheckNoError(err);
+    CheckTxnGet(txn, roptions, "foo", NULL);
+
+    rocksdb_transaction_put(txn, "foo", 3, "hello", 5, &err);
+    CheckNoError(err);
+
+    // read from outside transaction, before commit
+    CheckTxnDBGet(txn_db, roptions, "foo", NULL);
+
+    // commit
+    rocksdb_transaction_commit(txn, &err);
+    CheckNoError(err);
+
+    // read from outside transaction, after commit
+    CheckTxnDBGet(txn_db, roptions, "foo", "hello");
+
+    // reuse old transaction
+    txn = rocksdb_transaction_begin(txn_db, woptions, txn_options, txn);
+
+    // snapshot
+    const rocksdb_snapshot_t* snapshot;
+    snapshot = rocksdb_transactiondb_create_snapshot(txn_db);
+    rocksdb_readoptions_set_snapshot(roptions, snapshot);
+  
+    rocksdb_transactiondb_put(txn_db, woptions, "foo", 3, "hey", 3,  &err);
+    CheckNoError(err);
+
+    CheckTxnDBGet(txn_db, roptions, "foo", "hello");
+    rocksdb_readoptions_set_snapshot(roptions, NULL);
+    rocksdb_transactiondb_release_snapshot(txn_db, snapshot);
+    CheckTxnDBGet(txn_db, roptions, "foo", "hey");
+
+    // iterate
+    rocksdb_transaction_put(txn, "bar", 3, "hi", 2, &err);
+    rocksdb_iterator_t* iter = rocksdb_transaction_create_iterator(txn, roptions);
+    CheckCondition(!rocksdb_iter_valid(iter));
+    rocksdb_iter_seek_to_first(iter);
+    CheckCondition(rocksdb_iter_valid(iter));
+    CheckIter(iter, "bar", "hi");
+    rocksdb_iter_get_error(iter, &err);
+    CheckNoError(err);
+    rocksdb_iter_destroy(iter);
+
+    // rollback
+    rocksdb_transaction_rollback(txn, &err);
+    CheckNoError(err);
+    CheckTxnDBGet(txn_db, roptions, "bar", NULL);
+
+    // Column families.
+    rocksdb_column_family_handle_t* cfh;
+    cfh = rocksdb_transactiondb_create_column_family(txn_db, options,
+                                                     "txn_db_cf", &err);
+    CheckNoError(err);
+
+    rocksdb_transactiondb_put_cf(txn_db, woptions, cfh, "cf_foo", 6, "cf_hello",
+                                 8, &err);
+    CheckNoError(err);
+    CheckTxnDBGetCF(txn_db, roptions, cfh, "cf_foo", "cf_hello");
+
+    rocksdb_transactiondb_delete_cf(txn_db, woptions, cfh, "cf_foo", 6, &err);
+    CheckNoError(err);
+    CheckTxnDBGetCF(txn_db, roptions, cfh, "cf_foo", NULL);
+
+    rocksdb_column_family_handle_destroy(cfh);
+
+    // close and destroy
+    rocksdb_transaction_destroy(txn);
+    rocksdb_transactiondb_close(txn_db);
+    rocksdb_destroy_db(options, dbname, &err);
+    CheckNoError(err);
+    rocksdb_transaction_options_destroy(txn_options);
+    rocksdb_transactiondb_options_destroy(txn_db_options);
+  }
+
+  // Simple sanity check that setting memtable rep works.
+  StartPhase("memtable_reps");
+  {
+    // Create database with vector memtable.
     rocksdb_options_set_memtable_vector_rep(options);
     db = rocksdb_open(options, dbname, &err);
     CheckNoError(err);
@@ -1122,6 +1491,18 @@ int main(int argc, char** argv) {
     CheckNoError(err);
   }
 
+  // Simple sanity check that options setting db_paths work.
+  StartPhase("open_db_paths");
+  {
+    rocksdb_close(db);
+    rocksdb_destroy_db(options, dbname, &err);
+
+    const rocksdb_dbpath_t* paths[1] = {dbpath};
+    rocksdb_options_set_db_paths(options, paths, 1);
+    db = rocksdb_open(options, dbname, &err);
+    CheckNoError(err);
+  }
+  
   StartPhase("cleanup");
   rocksdb_close(db);
   rocksdb_options_destroy(options);
@@ -1131,6 +1512,7 @@ int main(int argc, char** argv) {
   rocksdb_compactoptions_destroy(coptions);
   rocksdb_cache_destroy(cache);
   rocksdb_comparator_destroy(cmp);
+  rocksdb_dbpath_destroy(dbpath);
   rocksdb_env_destroy(env);
 
   fprintf(stderr, "PASS\n");

db/column_family.cc

@@ -20,6 +20,7 @@
 #include <limits>
 
 #include "db/compaction_picker.h"
+#include "db/compaction_picker_universal.h"
 #include "db/db_impl.h"
 #include "db/internal_stats.h"
 #include "db/job_context.h"
@@ -27,11 +28,11 @@
 #include "db/version_set.h"
 #include "db/write_controller.h"
 #include "memtable/hash_skiplist_rep.h"
+#include "monitoring/thread_status_util.h"
+#include "options/options_helper.h"
 #include "table/block_based_table_factory.h"
 #include "util/autovector.h"
 #include "util/compression.h"
-#include "util/options_helper.h"
-#include "util/thread_status_util.h"
 
 namespace rocksdb {
 
@@ -169,6 +170,12 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
       result.num_levels < 2) {
     result.num_levels = 2;
   }
+
+  if (result.compaction_style == kCompactionStyleUniversal &&
+      db_options.allow_ingest_behind && result.num_levels < 3) {
+    result.num_levels = 3;
+  }
+
   if (result.max_write_buffer_number < 2) {
     result.max_write_buffer_number = 2;
   }
@@ -195,7 +202,6 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
     result.num_levels = 1;
     // since we delete level0 files in FIFO compaction when there are too many
     // of them, these options don't really mean anything
-    result.level0_file_num_compaction_trigger = std::numeric_limits<int>::max();
     result.level0_slowdown_writes_trigger = std::numeric_limits<int>::max();
     result.level0_stop_writes_trigger = std::numeric_limits<int>::max();
   }
@@ -205,8 +211,8 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
   }
 
   if (result.level0_file_num_compaction_trigger == 0) {
-    Warn(db_options.info_log.get(),
-         "level0_file_num_compaction_trigger cannot be 0");
+    ROCKS_LOG_WARN(db_options.info_log.get(),
+                   "level0_file_num_compaction_trigger cannot be 0");
     result.level0_file_num_compaction_trigger = 1;
   }
 
@@ -214,14 +220,14 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
           result.level0_slowdown_writes_trigger ||
       result.level0_slowdown_writes_trigger <
           result.level0_file_num_compaction_trigger) {
-    Warn(db_options.info_log.get(),
-         "This condition must be satisfied: "
-         "level0_stop_writes_trigger(%d) >= "
-         "level0_slowdown_writes_trigger(%d) >= "
-         "level0_file_num_compaction_trigger(%d)",
-         result.level0_stop_writes_trigger,
-         result.level0_slowdown_writes_trigger,
-         result.level0_file_num_compaction_trigger);
+    ROCKS_LOG_WARN(db_options.info_log.get(),
+                   "This condition must be satisfied: "
+                   "level0_stop_writes_trigger(%d) >= "
+                   "level0_slowdown_writes_trigger(%d) >= "
+                   "level0_file_num_compaction_trigger(%d)",
+                   result.level0_stop_writes_trigger,
+                   result.level0_slowdown_writes_trigger,
+                   result.level0_file_num_compaction_trigger);
     if (result.level0_slowdown_writes_trigger <
         result.level0_file_num_compaction_trigger) {
       result.level0_slowdown_writes_trigger =
@@ -231,14 +237,14 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
         result.level0_slowdown_writes_trigger) {
       result.level0_stop_writes_trigger = result.level0_slowdown_writes_trigger;
     }
-    Warn(db_options.info_log.get(),
-         "Adjust the value to "
-         "level0_stop_writes_trigger(%d)"
-         "level0_slowdown_writes_trigger(%d)"
-         "level0_file_num_compaction_trigger(%d)",
-         result.level0_stop_writes_trigger,
-         result.level0_slowdown_writes_trigger,
-         result.level0_file_num_compaction_trigger);
+    ROCKS_LOG_WARN(db_options.info_log.get(),
+                   "Adjust the value to "
+                   "level0_stop_writes_trigger(%d)"
+                   "level0_slowdown_writes_trigger(%d)"
+                   "level0_file_num_compaction_trigger(%d)",
+                   result.level0_stop_writes_trigger,
+                   result.level0_slowdown_writes_trigger,
+                   result.level0_file_num_compaction_trigger);
   }
 
   if (result.soft_pending_compaction_bytes_limit == 0) {
@@ -341,13 +347,14 @@ ColumnFamilyData::ColumnFamilyData(
       dummy_versions_(_dummy_versions),
       current_(nullptr),
       refs_(0),
+      initialized_(false),
       dropped_(false),
       internal_comparator_(cf_options.comparator),
       initial_cf_options_(SanitizeOptions(db_options, cf_options)),
       ioptions_(db_options, initial_cf_options_),
       mutable_cf_options_(initial_cf_options_),
-      is_delete_range_supported_(strcmp(cf_options.table_factory->Name(),
-                                        BlockBasedTableFactory().Name()) == 0),
+      is_delete_range_supported_(
+          cf_options.table_factory->IsDeleteRangeSupported()),
       write_buffer_manager_(write_buffer_manager),
       mem_(nullptr),
       imm_(ioptions_.min_write_buffer_number_to_merge,
@@ -386,27 +393,27 @@ ColumnFamilyData::ColumnFamilyData(
     } else if (ioptions_.compaction_style == kCompactionStyleNone) {
       compaction_picker_.reset(new NullCompactionPicker(
           ioptions_, &internal_comparator_));
-      Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
-          "Column family %s does not use any background compaction. "
-          "Compactions can only be done via CompactFiles\n",
-          GetName().c_str());
+      ROCKS_LOG_WARN(ioptions_.info_log,
+                     "Column family %s does not use any background compaction. "
+                     "Compactions can only be done via CompactFiles\n",
+                     GetName().c_str());
 #endif  // !ROCKSDB_LITE
     } else {
-      Log(InfoLogLevel::ERROR_LEVEL, ioptions_.info_log,
-          "Unable to recognize the specified compaction style %d. "
-          "Column family %s will use kCompactionStyleLevel.\n",
-          ioptions_.compaction_style, GetName().c_str());
+      ROCKS_LOG_ERROR(ioptions_.info_log,
+                      "Unable to recognize the specified compaction style %d. "
+                      "Column family %s will use kCompactionStyleLevel.\n",
+                      ioptions_.compaction_style, GetName().c_str());
       compaction_picker_.reset(
           new LevelCompactionPicker(ioptions_, &internal_comparator_));
     }
 
     if (column_family_set_->NumberOfColumnFamilies() < 10) {
-      Log(InfoLogLevel::INFO_LEVEL, ioptions_.info_log,
-          "--------------- Options for column family [%s]:\n", name.c_str());
+      ROCKS_LOG_INFO(ioptions_.info_log,
+                     "--------------- Options for column family [%s]:\n",
+                     name.c_str());
       initial_cf_options_.Dump(ioptions_.info_log);
     } else {
-      Log(InfoLogLevel::INFO_LEVEL, ioptions_.info_log,
-          "\t(skipping printing options)\n");
+      ROCKS_LOG_INFO(ioptions_.info_log, "\t(skipping printing options)\n");
     }
   }
 
@@ -543,7 +550,7 @@ std::unique_ptr<WriteControllerToken> SetupDelay(
     // If DB just falled into the stop condition, we need to further reduce
     // the write rate to avoid the stop condition.
     if (penalize_stop) {
-      // Penalize the near stop or stop condition by more agressive slowdown.
+      // Penalize the near stop or stop condition by more aggressive slowdown.
       // This is to provide the long term slowdown increase signal.
       // The penalty is more than the reward of recovering to the normal
       // condition.
@@ -620,7 +627,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
     if (imm()->NumNotFlushed() >= mutable_cf_options.max_write_buffer_number) {
       write_controller_token_ = write_controller->GetStopToken();
       internal_stats_->AddCFStats(InternalStats::MEMTABLE_COMPACTION, 1);
-      Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+      ROCKS_LOG_WARN(
+          ioptions_.info_log,
           "[%s] Stopping writes because we have %d immutable memtables "
           "(waiting for flush), max_write_buffer_number is set to %d",
           name_.c_str(), imm()->NumNotFlushed(),
@@ -634,9 +642,9 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
         internal_stats_->AddCFStats(
             InternalStats::LEVEL0_NUM_FILES_WITH_COMPACTION, 1);
       }
-      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());
+      ROCKS_LOG_WARN(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 >=
@@ -644,7 +652,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
       write_controller_token_ = write_controller->GetStopToken();
       internal_stats_->AddCFStats(
           InternalStats::HARD_PENDING_COMPACTION_BYTES_LIMIT, 1);
-      Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+      ROCKS_LOG_WARN(
+          ioptions_.info_log,
           "[%s] Stopping writes because of estimated pending compaction "
           "bytes %" PRIu64,
           name_.c_str(), compaction_needed_bytes);
@@ -656,7 +665,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
                      prev_compaction_needed_bytes_, was_stopped,
                      mutable_cf_options.disable_auto_compactions);
       internal_stats_->AddCFStats(InternalStats::MEMTABLE_SLOWDOWN, 1);
-      Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+      ROCKS_LOG_WARN(
+          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,
@@ -679,11 +689,11 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
         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());
+      ROCKS_LOG_WARN(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() >=
@@ -705,7 +715,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
                      mutable_cf_options.disable_auto_compactions);
       internal_stats_->AddCFStats(
           InternalStats::SOFT_PENDING_COMPACTION_BYTES_LIMIT, 1);
-      Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+      ROCKS_LOG_WARN(
+          ioptions_.info_log,
           "[%s] Stalling writes because of estimated pending compaction "
           "bytes %" PRIu64 " rate %" PRIu64,
           name_.c_str(), vstorage->estimated_compaction_needed_bytes(),
@@ -717,7 +728,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
               mutable_cf_options.level0_slowdown_writes_trigger)) {
         write_controller_token_ =
             write_controller->GetCompactionPressureToken();
-        Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+        ROCKS_LOG_INFO(
+            ioptions_.info_log,
             "[%s] Increasing compaction threads because we have %d level-0 "
             "files ",
             name_.c_str(), vstorage->l0_delay_trigger_count());
@@ -730,7 +742,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
         write_controller_token_ =
             write_controller->GetCompactionPressureToken();
         if (mutable_cf_options.soft_pending_compaction_bytes_limit > 0) {
-          Log(InfoLogLevel::WARN_LEVEL, ioptions_.info_log,
+          ROCKS_LOG_INFO(
+              ioptions_.info_log,
               "[%s] Increasing compaction threads because of estimated pending "
               "compaction "
               "bytes %" PRIu64,
@@ -746,6 +759,12 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
         uint64_t write_rate = write_controller->delayed_write_rate();
         write_controller->set_delayed_write_rate(static_cast<uint64_t>(
             static_cast<double>(write_rate) * kDelayRecoverSlowdownRatio));
+        // Set the low pri limit to be 1/4 the delayed write rate.
+        // Note we don't reset this value even after delay condition is relased.
+        // Low-pri rate will continue to apply if there is a compaction
+        // pressure.
+        write_controller->low_pri_rate_limiter()->SetBytesPerSecond(write_rate /
+                                                                    4);
       }
     }
     prev_compaction_needed_bytes_ = compaction_needed_bytes;
@@ -770,9 +789,8 @@ uint64_t ColumnFamilyData::GetTotalSstFilesSize() const {
 
 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_manager_, earliest_seq, id_);
 }
 
 void ColumnFamilyData::CreateNewMemtable(

db/column_family.h

@@ -19,11 +19,11 @@
 #include "db/table_properties_collector.h"
 #include "db/write_batch_internal.h"
 #include "db/write_controller.h"
+#include "options/cf_options.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/thread_local.h"
 
 namespace rocksdb {
@@ -333,6 +333,10 @@ class ColumnFamilyData {
   void RecalculateWriteStallConditions(
       const MutableCFOptions& mutable_cf_options);
 
+  void set_initialized() { initialized_.store(true); }
+
+  bool initialized() const { return initialized_.load(); }
+
  private:
   friend class ColumnFamilySet;
   ColumnFamilyData(uint32_t id, const std::string& name,
@@ -349,6 +353,7 @@ class ColumnFamilyData {
   Version* current_;         // == dummy_versions->prev_
 
   std::atomic<int> refs_;      // outstanding references to ColumnFamilyData
+  std::atomic<bool> initialized_;
   bool dropped_;               // true if client dropped it
 
   const InternalKeyComparator internal_comparator_;

db/column_family_test.cc

@@ -14,13 +14,13 @@
 
 #include "db/db_impl.h"
 #include "db/db_test_util.h"
+#include "options/options_parser.h"
 #include "port/port.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
 #include "rocksdb/iterator.h"
 #include "util/coding.h"
 #include "util/fault_injection_test_env.h"
-#include "util/options_parser.h"
 #include "util/string_util.h"
 #include "util/sync_point.h"
 #include "util/testharness.h"
@@ -54,7 +54,7 @@ class EnvCounter : public EnvWrapper {
   }
 
  private:
-  int num_new_writable_file_;
+  std::atomic<int> num_new_writable_file_;
 };
 
 class ColumnFamilyTest : public testing::Test {
@@ -302,6 +302,7 @@ class ColumnFamilyTest : public testing::Test {
         ASSERT_OK(Put(cf, key, RandomString(&rnd_, key_value_size - 10)));
       }
     }
+    db_->FlushWAL(false);
   }
 
 #ifndef ROCKSDB_LITE  // TEST functions in DB are not supported in lite
@@ -520,6 +521,30 @@ TEST_F(ColumnFamilyTest, DontReuseColumnFamilyID) {
   }
 }
 
+#ifndef ROCKSDB_LITE
+TEST_F(ColumnFamilyTest, CreateCFRaceWithGetAggProperty) {
+  Open();
+
+  rocksdb::SyncPoint::GetInstance()->LoadDependency(
+      {{"DBImpl::WriteOptionsFile:1",
+        "ColumnFamilyTest.CreateCFRaceWithGetAggProperty:1"},
+       {"ColumnFamilyTest.CreateCFRaceWithGetAggProperty:2",
+        "DBImpl::WriteOptionsFile:2"}});
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  rocksdb::port::Thread thread([&] { CreateColumnFamilies({"one"}); });
+
+  TEST_SYNC_POINT("ColumnFamilyTest.CreateCFRaceWithGetAggProperty:1");
+  uint64_t pv;
+  db_->GetAggregatedIntProperty(DB::Properties::kEstimateTableReadersMem, &pv);
+  TEST_SYNC_POINT("ColumnFamilyTest.CreateCFRaceWithGetAggProperty:2");
+
+  thread.join();
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+#endif  // !ROCKSDB_LITE
+
 class FlushEmptyCFTestWithParam : public ColumnFamilyTest,
                                   public testing::WithParamInterface<bool> {
  public:
@@ -557,6 +582,7 @@ TEST_P(FlushEmptyCFTestWithParam, FlushEmptyCFTest) {
   Flush(0);
   ASSERT_OK(Put(1, "bar", "v3"));  // seqID 4
   ASSERT_OK(Put(1, "foo", "v4"));  // seqID 5
+  db_->FlushWAL(false);
 
   // Preserve file system state up to here to simulate a crash condition.
   fault_env->SetFilesystemActive(false);
@@ -619,6 +645,7 @@ TEST_P(FlushEmptyCFTestWithParam, FlushEmptyCFTest2) {
   // Write to log file D
   ASSERT_OK(Put(1, "bar", "v4"));  // seqID 7
   ASSERT_OK(Put(1, "bar", "v5"));  // seqID 8
+  db_->FlushWAL(false);
   // Preserve file system state up to here to simulate a crash condition.
   fault_env->SetFilesystemActive(false);
   std::vector<std::string> names;
@@ -675,6 +702,44 @@ TEST_F(ColumnFamilyTest, AddDrop) {
               std::vector<std::string>({"default", "four", "three"}));
 }
 
+TEST_F(ColumnFamilyTest, BulkAddDrop) {
+  constexpr int kNumCF = 1000;
+  ColumnFamilyOptions cf_options;
+  WriteOptions write_options;
+  Open();
+  std::vector<std::string> cf_names;
+  std::vector<ColumnFamilyHandle*> cf_handles;
+  for (int i = 1; i <= kNumCF; i++) {
+    cf_names.push_back("cf1-" + ToString(i));
+  }
+  ASSERT_OK(db_->CreateColumnFamilies(cf_options, cf_names, &cf_handles));
+  for (int i = 1; i <= kNumCF; i++) {
+    ASSERT_OK(db_->Put(write_options, cf_handles[i - 1], "foo", "bar"));
+  }
+  ASSERT_OK(db_->DropColumnFamilies(cf_handles));
+  std::vector<ColumnFamilyDescriptor> cf_descriptors;
+  for (auto* handle : cf_handles) {
+    delete handle;
+  }
+  cf_handles.clear();
+  for (int i = 1; i <= kNumCF; i++) {
+    cf_descriptors.emplace_back("cf2-" + ToString(i), ColumnFamilyOptions());
+  }
+  ASSERT_OK(db_->CreateColumnFamilies(cf_descriptors, &cf_handles));
+  for (int i = 1; i <= kNumCF; i++) {
+    ASSERT_OK(db_->Put(write_options, cf_handles[i - 1], "foo", "bar"));
+  }
+  ASSERT_OK(db_->DropColumnFamilies(cf_handles));
+  for (auto* handle : cf_handles) {
+    delete handle;
+  }
+  Close();
+  std::vector<std::string> families;
+  ASSERT_OK(DB::ListColumnFamilies(db_options_, dbname_, &families));
+  std::sort(families.begin(), families.end());
+  ASSERT_TRUE(families == std::vector<std::string>({"default"}));
+}
+
 TEST_F(ColumnFamilyTest, DropTest) {
   // first iteration - dont reopen DB before dropping
   // second iteration - reopen DB before dropping
@@ -1341,7 +1406,6 @@ TEST_F(ColumnFamilyTest, AutomaticAndManualCompactions) {
   ColumnFamilyOptions default_cf, one, two;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1365,6 +1429,9 @@ TEST_F(ColumnFamilyTest, AutomaticAndManualCompactions) {
   two.write_buffer_size = 100000;
 
   Reopen({default_cf, one, two});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   bool cf_1_1 = true;
   rocksdb::SyncPoint::GetInstance()->LoadDependency(
@@ -1424,7 +1491,6 @@ TEST_F(ColumnFamilyTest, AutomaticAndManualCompactions) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-  Close();
 }
 
 TEST_F(ColumnFamilyTest, ManualAndAutomaticCompactions) {
@@ -1433,7 +1499,6 @@ TEST_F(ColumnFamilyTest, ManualAndAutomaticCompactions) {
   ColumnFamilyOptions default_cf, one, two;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1457,6 +1522,9 @@ TEST_F(ColumnFamilyTest, ManualAndAutomaticCompactions) {
   two.write_buffer_size = 100000;
 
   Reopen({default_cf, one, two});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   // SETUP column family "one" -- universal style
   for (int i = 0; i < one.level0_file_num_compaction_trigger - 2; ++i) {
@@ -1519,7 +1587,6 @@ TEST_F(ColumnFamilyTest, ManualAndAutomaticCompactions) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-  Close();
 }
 
 TEST_F(ColumnFamilyTest, SameCFManualManualCompactions) {
@@ -1528,7 +1595,6 @@ TEST_F(ColumnFamilyTest, SameCFManualManualCompactions) {
   ColumnFamilyOptions default_cf, one;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1547,6 +1613,9 @@ TEST_F(ColumnFamilyTest, SameCFManualManualCompactions) {
   one.write_buffer_size = 120000;
 
   Reopen({default_cf, one});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   // SETUP column family "one" -- universal style
   for (int i = 0; i < one.level0_file_num_compaction_trigger - 2; ++i) {
@@ -1617,7 +1686,6 @@ TEST_F(ColumnFamilyTest, SameCFManualManualCompactions) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-  Close();
 }
 
 TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactions) {
@@ -1626,7 +1694,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactions) {
   ColumnFamilyOptions default_cf, one;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1645,6 +1712,9 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactions) {
   one.write_buffer_size = 120000;
 
   Reopen({default_cf, one});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   // SETUP column family "one" -- universal style
   for (int i = 0; i < one.level0_file_num_compaction_trigger - 2; ++i) {
@@ -1706,7 +1776,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactions) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-  Close();
 }
 
 TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactionsLevel) {
@@ -1715,7 +1784,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactionsLevel) {
   ColumnFamilyOptions default_cf, one;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1730,10 +1798,13 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactionsLevel) {
 
   one.num_levels = 1;
   // trigger compaction if there are >= 4 files
-  one.level0_file_num_compaction_trigger = 4;
+  one.level0_file_num_compaction_trigger = 3;
   one.write_buffer_size = 120000;
 
   Reopen({default_cf, one});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   // SETUP column family "one" -- level style
   for (int i = 0; i < one.level0_file_num_compaction_trigger - 2; ++i) {
@@ -1795,7 +1866,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticCompactionsLevel) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-  Close();
 }
 
 // This test checks for automatic getting a conflict if there is a
@@ -1811,7 +1881,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticConflict) {
   ColumnFamilyOptions default_cf, one;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1830,6 +1899,9 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticConflict) {
   one.write_buffer_size = 120000;
 
   Reopen({default_cf, one});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   // SETUP column family "one" -- universal style
   for (int i = 0; i < one.level0_file_num_compaction_trigger - 2; ++i) {
@@ -1914,8 +1986,6 @@ TEST_F(ColumnFamilyTest, SameCFManualAutomaticConflict) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-
-  Close();
 }
 
 // In this test, we generate enough files to trigger automatic compactions.
@@ -1930,7 +2000,6 @@ TEST_F(ColumnFamilyTest, SameCFAutomaticManualCompactions) {
   ColumnFamilyOptions default_cf, one;
   db_options_.max_open_files = 20;  // only 10 files in file cache
   db_options_.max_background_compactions = 3;
-  db_options_.base_background_compactions = 3;
 
   default_cf.compaction_style = kCompactionStyleLevel;
   default_cf.num_levels = 3;
@@ -1949,6 +2018,9 @@ TEST_F(ColumnFamilyTest, SameCFAutomaticManualCompactions) {
   one.write_buffer_size = 120000;
 
   Reopen({default_cf, one});
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
 
   bool cf_1_1 = true;
   bool cf_1_2 = true;
@@ -2003,8 +2075,6 @@ TEST_F(ColumnFamilyTest, SameCFAutomaticManualCompactions) {
     ASSERT_NE("NOT_FOUND", Get(1, *key_iter));
     key_iter++;
   }
-
-  Close();
 }
 #endif  // !ROCKSDB_LITE
 
@@ -2457,7 +2527,6 @@ TEST_F(ColumnFamilyTest, CreateAndDropRace) {
 TEST_F(ColumnFamilyTest, WriteStallSingleColumnFamily) {
   const uint64_t kBaseRate = 800000u;
   db_options_.delayed_write_rate = kBaseRate;
-  db_options_.base_background_compactions = 2;
   db_options_.max_background_compactions = 6;
 
   Open({"default"});
@@ -2638,7 +2707,6 @@ TEST_F(ColumnFamilyTest, WriteStallSingleColumnFamily) {
 }
 
 TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
-  db_options_.base_background_compactions = 2;
   db_options_.max_background_compactions = 6;
   Open({"default"});
   ColumnFamilyData* cfd =
@@ -2658,7 +2726,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
 
   vstorage->TEST_set_estimated_compaction_needed_bytes(40);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->TEST_set_estimated_compaction_needed_bytes(50);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
@@ -2670,11 +2738,11 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
 
   vstorage->TEST_set_estimated_compaction_needed_bytes(45);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->set_l0_delay_trigger_count(7);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->set_l0_delay_trigger_count(9);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
@@ -2682,7 +2750,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
 
   vstorage->set_l0_delay_trigger_count(6);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   // Speed up threshold = min(4 * 2, 4 + (12 - 4)/4) = 6
   mutable_cf_options.level0_file_num_compaction_trigger = 4;
@@ -2691,7 +2759,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
 
   vstorage->set_l0_delay_trigger_count(5);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->set_l0_delay_trigger_count(7);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
@@ -2699,7 +2767,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupSingleColumnFamily) {
 
   vstorage->set_l0_delay_trigger_count(3);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 }
 
 TEST_F(ColumnFamilyTest, WriteStallTwoColumnFamilies) {
@@ -2778,7 +2846,6 @@ TEST_F(ColumnFamilyTest, WriteStallTwoColumnFamilies) {
 }
 
 TEST_F(ColumnFamilyTest, CompactionSpeedupTwoColumnFamilies) {
-  db_options_.base_background_compactions = 2;
   db_options_.max_background_compactions = 6;
   column_family_options_.soft_pending_compaction_bytes_limit = 200;
   column_family_options_.hard_pending_compaction_bytes_limit = 2000;
@@ -2806,11 +2873,11 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupTwoColumnFamilies) {
 
   vstorage->TEST_set_estimated_compaction_needed_bytes(40);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->TEST_set_estimated_compaction_needed_bytes(60);
   cfd1->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
   ASSERT_EQ(6, dbfull()->TEST_BGCompactionsAllowed());
 
@@ -2828,7 +2895,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupTwoColumnFamilies) {
 
   vstorage1->TEST_set_estimated_compaction_needed_bytes(3);
   cfd1->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 
   vstorage->set_l0_delay_trigger_count(9);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
@@ -2840,7 +2907,7 @@ TEST_F(ColumnFamilyTest, CompactionSpeedupTwoColumnFamilies) {
 
   vstorage->set_l0_delay_trigger_count(0);
   cfd->RecalculateWriteStallConditions(mutable_cf_options);
-  ASSERT_EQ(2, dbfull()->TEST_BGCompactionsAllowed());
+  ASSERT_EQ(1, dbfull()->TEST_BGCompactionsAllowed());
 }
 
 #ifndef ROCKSDB_LITE

db/compacted_db_impl.cc

@@ -42,8 +42,8 @@ size_t CompactedDBImpl::FindFile(const Slice& key) {
   return right;
 }
 
-Status CompactedDBImpl::Get(const ReadOptions& options,
-     ColumnFamilyHandle*, const Slice& key, std::string* value) {
+Status CompactedDBImpl::Get(const ReadOptions& options, ColumnFamilyHandle*,
+                            const Slice& key, PinnableSlice* value) {
   GetContext get_context(user_comparator_, nullptr, nullptr, nullptr,
                          GetContext::kNotFound, key, value, nullptr, nullptr,
                          nullptr, nullptr);
@@ -75,11 +75,14 @@ std::vector<Status> CompactedDBImpl::MultiGet(const ReadOptions& options,
   int idx = 0;
   for (auto* r : reader_list) {
     if (r != nullptr) {
+      PinnableSlice pinnable_val;
+      std::string& value = (*values)[idx];
       GetContext get_context(user_comparator_, nullptr, nullptr, nullptr,
-                             GetContext::kNotFound, keys[idx], &(*values)[idx],
+                             GetContext::kNotFound, keys[idx], &pinnable_val,
                              nullptr, nullptr, nullptr, nullptr);
       LookupKey lkey(keys[idx], kMaxSequenceNumber);
       r->Get(options, lkey.internal_key(), &get_context);
+      value.assign(pinnable_val.data(), pinnable_val.size());
       if (get_context.State() == GetContext::kFound) {
         statuses[idx] = Status::OK();
       }
@@ -151,8 +154,8 @@ Status CompactedDBImpl::Open(const Options& options,
   std::unique_ptr<CompactedDBImpl> db(new CompactedDBImpl(db_options, dbname));
   Status s = db->Init(options);
   if (s.ok()) {
-    Log(INFO_LEVEL, db->immutable_db_options_.info_log,
-        "Opened the db as fully compacted mode");
+    ROCKS_LOG_INFO(db->immutable_db_options_.info_log,
+                   "Opened the db as fully compacted mode");
     LogFlush(db->immutable_db_options_.info_log);
     *dbptr = db.release();
   }

db/compacted_db_impl.h

@@ -23,7 +23,7 @@ class CompactedDBImpl : public DBImpl {
   using DB::Get;
   virtual Status Get(const ReadOptions& options,
                      ColumnFamilyHandle* column_family, const Slice& key,
-                     std::string* value) override;
+                     PinnableSlice* value) override;
   using DB::MultiGet;
   virtual std::vector<Status> MultiGet(
       const ReadOptions& options,

db/compaction.cc

@@ -18,7 +18,7 @@
 
 #include "db/column_family.h"
 #include "rocksdb/compaction_filter.h"
-#include "util/logging.h"
+#include "util/string_util.h"
 #include "util/sync_point.h"
 
 namespace rocksdb {
@@ -169,6 +169,7 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
       bottommost_level_(IsBottommostLevel(output_level_, vstorage, inputs_)),
       is_full_compaction_(IsFullCompaction(vstorage, inputs_)),
       is_manual_compaction_(_manual_compaction),
+      is_trivial_move_(false),
       compaction_reason_(_compaction_reason) {
   MarkFilesBeingCompacted(true);
   if (is_manual_compaction_) {
@@ -283,28 +284,32 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
   assert(input_version_ != nullptr);
   assert(level_ptrs != nullptr);
   assert(level_ptrs->size() == static_cast<size_t>(number_levels_));
-  assert(cfd_->ioptions()->compaction_style != kCompactionStyleFIFO);
-  if (cfd_->ioptions()->compaction_style == kCompactionStyleUniversal) {
-    return bottommost_level_;
-  }
-  // Maybe use binary search to find right entry instead of linear search?
-  const Comparator* user_cmp = cfd_->user_comparator();
-  for (int lvl = output_level_ + 1; lvl < number_levels_; lvl++) {
-    const std::vector<FileMetaData*>& files = input_vstorage_->LevelFiles(lvl);
-    for (; level_ptrs->at(lvl) < files.size(); level_ptrs->at(lvl)++) {
-      auto* f = files[level_ptrs->at(lvl)];
-      if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
-        // We've advanced far enough
-        if (user_cmp->Compare(user_key, f->smallest.user_key()) >= 0) {
-          // Key falls in this file's range, so definitely
-          // exists beyond output level
-          return false;
+  if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel) {
+    if (output_level_ == 0) {
+      return false;
+    }
+    // Maybe use binary search to find right entry instead of linear search?
+    const Comparator* user_cmp = cfd_->user_comparator();
+    for (int lvl = output_level_ + 1; lvl < number_levels_; lvl++) {
+      const std::vector<FileMetaData*>& files =
+          input_vstorage_->LevelFiles(lvl);
+      for (; level_ptrs->at(lvl) < files.size(); level_ptrs->at(lvl)++) {
+        auto* f = files[level_ptrs->at(lvl)];
+        if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
+          // We've advanced far enough
+          if (user_cmp->Compare(user_key, f->smallest.user_key()) >= 0) {
+            // Key falls in this file's range, so definitely
+            // exists beyond output level
+            return false;
+          }
+          break;
         }
-        break;
       }
     }
+    return true;
+  } else {
+    return bottommost_level_;
   }
-  return true;
 }
 
 // Mark (or clear) each file that is being compacted
@@ -413,14 +418,15 @@ void Compaction::Summary(char* output, int len) {
 uint64_t Compaction::OutputFilePreallocationSize() const {
   uint64_t preallocation_size = 0;
 
-  if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel ||
-      output_level() > 0) {
+  if (max_output_file_size_ != port::kMaxUint64 &&
+      (cfd_->ioptions()->compaction_style == kCompactionStyleLevel ||
+       output_level() > 0)) {
     preallocation_size = max_output_file_size_;
   } else {
-    // output_level() == 0
-    assert(num_input_levels() > 0);
-    for (const auto& f : inputs_[0].files) {
-      preallocation_size += f->fd.GetFileSize();
+    for (const auto& level_files : inputs_) {
+      for (const auto& file : level_files.files) {
+        preallocation_size += file->fd.GetFileSize();
+      }
     }
   }
   // Over-estimate slightly so we don't end up just barely crossing
@@ -450,7 +456,7 @@ bool Compaction::ShouldFormSubcompactions() const {
     return false;
   }
   if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel) {
-    return start_level_ == 0 && !IsOutputLevelEmpty();
+    return start_level_ == 0 && output_level_ > 0 && !IsOutputLevelEmpty();
   } else if (cfd_->ioptions()->compaction_style == kCompactionStyleUniversal) {
     return number_levels_ > 1 && output_level_ > 0;
   } else {
@@ -458,4 +464,17 @@ bool Compaction::ShouldFormSubcompactions() const {
   }
 }
 
+uint64_t Compaction::MaxInputFileCreationTime() const {
+  uint64_t max_creation_time = 0;
+  for (const auto& file : inputs_[0].files) {
+    if (file->fd.table_reader != nullptr &&
+        file->fd.table_reader->GetTableProperties() != nullptr) {
+      uint64_t creation_time =
+          file->fd.table_reader->GetTableProperties()->creation_time;
+      max_creation_time = std::max(max_creation_time, creation_time);
+    }
+  }
+  return max_creation_time;
+}
+
 }  // namespace rocksdb

db/compaction.h

@@ -9,9 +9,9 @@
 
 #pragma once
 #include "db/version_set.h"
+#include "options/cf_options.h"
 #include "util/arena.h"
 #include "util/autovector.h"
-#include "util/cf_options.h"
 
 namespace rocksdb {
 
@@ -99,7 +99,8 @@ class Compaction {
   // input level.
   // REQUIREMENT: "compaction_input_level" must be >= 0 and
   //              < "input_levels()"
-  const std::vector<FileMetaData*>* inputs(size_t compaction_input_level) {
+  const std::vector<FileMetaData*>* inputs(
+      size_t compaction_input_level) const {
     assert(compaction_input_level < inputs_.size());
     return &inputs_[compaction_input_level].files;
   }
@@ -240,6 +241,8 @@ class Compaction {
 
   uint64_t max_compaction_bytes() const { return max_compaction_bytes_; }
 
+  uint64_t MaxInputFileCreationTime() const;
+
  private:
   // mark (or clear) all files that are being compacted
   void MarkFilesBeingCompacted(bool mark_as_compacted);

db/compaction_iteration_stats.h

@@ -16,6 +16,8 @@ struct CompactionIterationStats {
   int64_t num_record_drop_obsolete = 0;
   int64_t num_record_drop_range_del = 0;
   int64_t num_range_del_drop_obsolete = 0;
+  // Deletions obsoleted before bottom level due to file gap optimization.
+  int64_t num_optimized_del_drop_obsolete = 0;
   uint64_t total_filter_time = 0;
 
   // Input statistics

db/compaction_iterator.cc

@@ -4,16 +4,41 @@
 //  (found in the LICENSE.Apache file in the root directory).
 
 #include "db/compaction_iterator.h"
+#include "rocksdb/listener.h"
 #include "table/internal_iterator.h"
 
 namespace rocksdb {
 
+#ifndef ROCKSDB_LITE
+CompactionEventListener::CompactionListenerValueType fromInternalValueType(
+    ValueType vt) {
+  switch (vt) {
+    case kTypeDeletion:
+      return CompactionEventListener::CompactionListenerValueType::kDelete;
+    case kTypeValue:
+      return CompactionEventListener::CompactionListenerValueType::kValue;
+    case kTypeMerge:
+      return CompactionEventListener::CompactionListenerValueType::
+          kMergeOperand;
+    case kTypeSingleDeletion:
+      return CompactionEventListener::CompactionListenerValueType::
+          kSingleDelete;
+    case kTypeRangeDeletion:
+      return CompactionEventListener::CompactionListenerValueType::kRangeDelete;
+    default:
+      assert(false);
+      return CompactionEventListener::CompactionListenerValueType::kInvalid;
+  }
+}
+#endif  // ROCKSDB_LITE
+
 CompactionIterator::CompactionIterator(
     InternalIterator* 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, RangeDelAggregator* range_del_agg,
     const Compaction* compaction, const CompactionFilter* compaction_filter,
+    CompactionEventListener* compaction_listener,
     const std::atomic<bool>* shutting_down)
     : CompactionIterator(
           input, cmp, merge_helper, last_sequence, snapshots,
@@ -21,7 +46,7 @@ CompactionIterator::CompactionIterator(
           range_del_agg,
           std::unique_ptr<CompactionProxy>(
               compaction ? new CompactionProxy(compaction) : nullptr),
-          compaction_filter, shutting_down) {}
+          compaction_filter, compaction_listener, shutting_down) {}
 
 CompactionIterator::CompactionIterator(
     InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
@@ -30,6 +55,7 @@ CompactionIterator::CompactionIterator(
     bool expect_valid_internal_key, RangeDelAggregator* range_del_agg,
     std::unique_ptr<CompactionProxy> compaction,
     const CompactionFilter* compaction_filter,
+    CompactionEventListener* compaction_listener,
     const std::atomic<bool>* shutting_down)
     : input_(input),
       cmp_(cmp),
@@ -41,7 +67,11 @@ CompactionIterator::CompactionIterator(
       range_del_agg_(range_del_agg),
       compaction_(std::move(compaction)),
       compaction_filter_(compaction_filter),
+#ifndef ROCKSDB_LITE
+      compaction_listener_(compaction_listener),
+#endif  // ROCKSDB_LITE
       shutting_down_(shutting_down),
+      ignore_snapshots_(false),
       merge_out_iter_(merge_helper_) {
   assert(compaction_filter_ == nullptr || compaction_ != nullptr);
   bottommost_level_ =
@@ -53,15 +83,17 @@ CompactionIterator::CompactionIterator(
   if (snapshots_->size() == 0) {
     // optimize for fast path if there are no snapshots
     visible_at_tip_ = true;
-    earliest_snapshot_ = last_sequence;
+    earliest_snapshot_ = kMaxSequenceNumber;
     latest_snapshot_ = 0;
   } else {
     visible_at_tip_ = false;
     earliest_snapshot_ = snapshots_->at(0);
     latest_snapshot_ = snapshots_->back();
   }
-  if (compaction_filter_ != nullptr && compaction_filter_->IgnoreSnapshots()) {
-    ignore_snapshots_ = true;
+  if (compaction_filter_ != nullptr) {
+    if (compaction_filter_->IgnoreSnapshots()) {
+      ignore_snapshots_ = true;
+    }
   } else {
     ignore_snapshots_ = false;
   }
@@ -79,6 +111,7 @@ void CompactionIterator::ResetRecordCounts() {
   iter_stats_.num_record_drop_obsolete = 0;
   iter_stats_.num_record_drop_range_del = 0;
   iter_stats_.num_range_del_drop_obsolete = 0;
+  iter_stats_.num_optimized_del_drop_obsolete = 0;
 }
 
 void CompactionIterator::SeekToFirst() {
@@ -103,7 +136,7 @@ void CompactionIterator::Next() {
       assert(valid_key);
       // Keep current_key_ in sync.
       current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
-      key_ = current_key_.GetKey();
+      key_ = current_key_.GetInternalKey();
       ikey_.user_key = current_key_.GetUserKey();
       valid_ = true;
     } else {
@@ -149,7 +182,7 @@ void CompactionIterator::NextFromInput() {
         status_ = Status::Corruption("Corrupted internal key not expected.");
         break;
       }
-      key_ = current_key_.SetKey(key_);
+      key_ = current_key_.SetInternalKey(key_);
       has_current_user_key_ = false;
       current_user_key_sequence_ = kMaxSequenceNumber;
       current_user_key_snapshot_ = 0;
@@ -179,13 +212,21 @@ void CompactionIterator::NextFromInput() {
         !cmp_->Equal(ikey_.user_key, current_user_key_)) {
       // First occurrence of this user key
       // Copy key for output
-      key_ = current_key_.SetKey(key_, &ikey_);
+      key_ = current_key_.SetInternalKey(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;
 
+#ifndef ROCKSDB_LITE
+      if (compaction_listener_) {
+        compaction_listener_->OnCompaction(compaction_->level(), ikey_.user_key,
+                                           fromInternalValueType(ikey_.type),
+                                           value_, ikey_.sequence, true);
+      }
+#endif  // ROCKSDB_LITE
+
       // 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_ ||
@@ -233,13 +274,21 @@ void CompactionIterator::NextFromInput() {
         }
       }
     } else {
+#ifndef ROCKSDB_LITE
+      if (compaction_listener_) {
+        compaction_listener_->OnCompaction(compaction_->level(), ikey_.user_key,
+                                           fromInternalValueType(ikey_.type),
+                                           value_, ikey_.sequence, false);
+      }
+#endif  // ROCKSDB_LITE
+
       // 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();
+      key_ = current_key_.GetInternalKey();
       ikey_.user_key = current_key_.GetUserKey();
     }
 
@@ -378,6 +427,9 @@ void CompactionIterator::NextFromInput() {
           // Can compact out this SingleDelete.
           ++iter_stats_.num_record_drop_obsolete;
           ++iter_stats_.num_single_del_fallthru;
+          if (!bottommost_level_) {
+            ++iter_stats_.num_optimized_del_drop_obsolete;
+          }
         } else {
           // Output SingleDelete
           valid_ = true;
@@ -392,7 +444,7 @@ void CompactionIterator::NextFromInput() {
       // is the same as the visibility of a previous instance of the
       // same key, then this kv is not visible in any snapshot.
       // Hidden by an newer entry for same user key
-      // TODO: why not > ?
+      // TODO(noetzli): why not > ?
       //
       // Note: Dropping this key will not affect TransactionDB write-conflict
       // checking since there has already been a record returned for this key
@@ -419,6 +471,9 @@ void CompactionIterator::NextFromInput() {
       // Note:  Dropping this Delete will not affect TransactionDB
       // write-conflict checking since it is earlier than any snapshot.
       ++iter_stats_.num_record_drop_obsolete;
+      if (!bottommost_level_) {
+        ++iter_stats_.num_optimized_del_drop_obsolete;
+      }
       input_->Next();
     } else if (ikey_.type == kTypeMerge) {
       if (!merge_helper_->HasOperator()) {
@@ -451,7 +506,7 @@ void CompactionIterator::NextFromInput() {
         assert(valid_key);
         // Keep current_key_ in sync.
         current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
-        key_ = current_key_.GetKey();
+        key_ = current_key_.GetInternalKey();
         ikey_.user_key = current_key_.GetUserKey();
         valid_ = true;
       } else {
@@ -498,7 +553,8 @@ void CompactionIterator::PrepareOutput() {
 
   // 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_ &&
+  if ((compaction_ != nullptr && !compaction_->allow_ingest_behind()) &&
+      bottommost_level_ && valid_ && ikey_.sequence <= earliest_snapshot_ &&
       ikey_.type != kTypeMerge &&
       !cmp_->Equal(compaction_->GetLargestUserKey(), ikey_.user_key)) {
     assert(ikey_.type != kTypeDeletion && ikey_.type != kTypeSingleDeletion);

db/compaction_iterator.h

@@ -14,10 +14,13 @@
 #include "db/merge_helper.h"
 #include "db/pinned_iterators_manager.h"
 #include "db/range_del_aggregator.h"
+#include "options/cf_options.h"
 #include "rocksdb/compaction_filter.h"
 
 namespace rocksdb {
 
+class CompactionEventListener;
+
 class CompactionIterator {
  public:
   // A wrapper around Compaction. Has a much smaller interface, only what
@@ -42,6 +45,9 @@ class CompactionIterator {
     virtual Slice GetLargestUserKey() const {
       return compaction_->GetLargestUserKey();
     }
+    virtual bool allow_ingest_behind() const {
+      return compaction_->immutable_cf_options()->allow_ingest_behind;
+    }
 
    protected:
     CompactionProxy() = default;
@@ -58,6 +64,7 @@ class CompactionIterator {
                      RangeDelAggregator* range_del_agg,
                      const Compaction* compaction = nullptr,
                      const CompactionFilter* compaction_filter = nullptr,
+                     CompactionEventListener* compaction_listener = nullptr,
                      const std::atomic<bool>* shutting_down = nullptr);
 
   // Constructor with custom CompactionProxy, used for tests.
@@ -69,6 +76,7 @@ class CompactionIterator {
                      RangeDelAggregator* range_del_agg,
                      std::unique_ptr<CompactionProxy> compaction,
                      const CompactionFilter* compaction_filter = nullptr,
+                     CompactionEventListener* compaction_listener = nullptr,
                      const std::atomic<bool>* shutting_down = nullptr);
 
   ~CompactionIterator();
@@ -122,6 +130,9 @@ class CompactionIterator {
   RangeDelAggregator* range_del_agg_;
   std::unique_ptr<CompactionProxy> compaction_;
   const CompactionFilter* compaction_filter_;
+#ifndef ROCKSDB_LITE
+  CompactionEventListener* compaction_listener_;
+#endif  // ROCKSDB_LITE
   const std::atomic<bool>* shutting_down_;
   bool bottommost_level_;
   bool valid_ = false;

db/compaction_iterator_test.cc

@@ -154,6 +154,7 @@ class FakeCompaction : public CompactionIterator::CompactionProxy {
   virtual Slice GetLargestUserKey() const {
     return "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
   }
+  virtual bool allow_ingest_behind() const { return false; }
 
   bool key_not_exists_beyond_output_level = false;
 };
@@ -189,7 +190,7 @@ class CompactionIteratorTest : public testing::Test {
     c_iter_.reset(new CompactionIterator(
         iter_.get(), cmp_, merge_helper_.get(), last_sequence, &snapshots_,
         kMaxSequenceNumber, Env::Default(), false, range_del_agg_.get(),
-        std::move(compaction), filter, &shutting_down_));
+        std::move(compaction), filter, nullptr, &shutting_down_));
   }
 
   void AddSnapshot(SequenceNumber snapshot) { snapshots_.push_back(snapshot); }
@@ -454,6 +455,111 @@ TEST_F(CompactionIteratorTest, ShuttingDownInMerge) {
   EXPECT_EQ(2, filter.last_seen.load());
 }
 
+TEST_F(CompactionIteratorTest, SingleMergeOperand) {
+  class Filter : public CompactionFilter {
+    virtual Decision FilterV2(int level, const Slice& key, ValueType t,
+                              const Slice& existing_value,
+                              std::string* new_value,
+                              std::string* skip_until) const override {
+      std::string k = key.ToString();
+      std::string v = existing_value.ToString();
+
+      // See InitIterators() call below for the sequence of keys and their
+      // filtering decisions. Here we closely assert that compaction filter is
+      // called with the expected keys and only them, and with the right values.
+      if (k == "a") {
+        EXPECT_EQ(ValueType::kMergeOperand, t);
+        EXPECT_EQ("av1", v);
+        return Decision::kKeep;
+      } else if (k == "b") {
+        EXPECT_EQ(ValueType::kMergeOperand, t);
+        return Decision::kKeep;
+      } else if (k == "c") {
+        return Decision::kKeep;
+      }
+
+      ADD_FAILURE();
+      return Decision::kKeep;
+    }
+
+    const char* Name() const override {
+      return "CompactionIteratorTest.SingleMergeOperand::Filter";
+    }
+  };
+
+  class SingleMergeOp : public MergeOperator {
+   public:
+    bool FullMergeV2(const MergeOperationInput& merge_in,
+                     MergeOperationOutput* merge_out) const override {
+      // See InitIterators() call below for why "c" is the only key for which
+      // FullMergeV2 should be called.
+      EXPECT_EQ("c", merge_in.key.ToString());
+
+      std::string temp_value;
+      if (merge_in.existing_value != nullptr) {
+        temp_value = merge_in.existing_value->ToString();
+      }
+
+      for (auto& operand : merge_in.operand_list) {
+        temp_value.append(operand.ToString());
+      }
+      merge_out->new_value = temp_value;
+
+      return true;
+    }
+
+    bool PartialMergeMulti(const Slice& key,
+                           const std::deque<Slice>& operand_list,
+                           std::string* new_value,
+                           Logger* logger) const override {
+      std::string string_key = key.ToString();
+      EXPECT_TRUE(string_key == "a" || string_key == "b");
+
+      if (string_key == "a") {
+        EXPECT_EQ(1, operand_list.size());
+      } else if (string_key == "b") {
+        EXPECT_EQ(2, operand_list.size());
+      }
+
+      std::string temp_value;
+      for (auto& operand : operand_list) {
+        temp_value.append(operand.ToString());
+      }
+      swap(temp_value, *new_value);
+
+      return true;
+    }
+
+    const char* Name() const override {
+      return "CompactionIteratorTest SingleMergeOp";
+    }
+
+    bool AllowSingleOperand() const override { return true; }
+  };
+
+  SingleMergeOp merge_op;
+  Filter filter;
+  InitIterators(
+      // a should invoke PartialMergeMulti with a single merge operand.
+      {test::KeyStr("a", 50, kTypeMerge),
+       // b should invoke PartialMergeMulti with two operands.
+       test::KeyStr("b", 70, kTypeMerge), test::KeyStr("b", 60, kTypeMerge),
+       // c should invoke FullMerge due to kTypeValue at the beginning.
+       test::KeyStr("c", 90, kTypeMerge), test::KeyStr("c", 80, kTypeValue)},
+      {"av1", "bv2", "bv1", "cv2", "cv1"}, {}, {}, kMaxSequenceNumber,
+      &merge_op, &filter);
+
+  c_iter_->SeekToFirst();
+  ASSERT_TRUE(c_iter_->Valid());
+  ASSERT_EQ(test::KeyStr("a", 50, kTypeMerge), c_iter_->key().ToString());
+  ASSERT_EQ("av1", c_iter_->value().ToString());
+  c_iter_->Next();
+  ASSERT_TRUE(c_iter_->Valid());
+  ASSERT_EQ("bv1bv2", c_iter_->value().ToString());
+  c_iter_->Next();
+  ASSERT_EQ("cv1cv2", c_iter_->value().ToString());
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/compaction_job.cc

@@ -28,7 +28,6 @@
 #include "db/db_iter.h"
 #include "db/dbformat.h"
 #include "db/event_helpers.h"
-#include "db/filename.h"
 #include "db/log_reader.h"
 #include "db/log_writer.h"
 #include "db/memtable.h"
@@ -36,6 +35,9 @@
 #include "db/merge_context.h"
 #include "db/merge_helper.h"
 #include "db/version_set.h"
+#include "monitoring/iostats_context_imp.h"
+#include "monitoring/perf_context_imp.h"
+#include "monitoring/thread_status_util.h"
 #include "port/likely.h"
 #include "port/port.h"
 #include "rocksdb/db.h"
@@ -49,16 +51,15 @@
 #include "table/table_builder.h"
 #include "util/coding.h"
 #include "util/file_reader_writer.h"
-#include "util/iostats_context_imp.h"
+#include "util/filename.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/random.h"
+#include "util/sst_file_manager_impl.h"
 #include "util/stop_watch.h"
 #include "util/string_util.h"
 #include "util/sync_point.h"
-#include "util/thread_status_util.h"
 
 namespace rocksdb {
 
@@ -593,6 +594,9 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
 
   double read_write_amp = 0.0;
   double write_amp = 0.0;
+  double bytes_read_per_sec = 0;
+  double bytes_written_per_sec = 0;
+
   if (stats.bytes_read_non_output_levels > 0) {
     read_write_amp = (stats.bytes_written + stats.bytes_read_output_level +
                       stats.bytes_read_non_output_levels) /
@@ -600,17 +604,22 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
     write_amp = stats.bytes_written /
                 static_cast<double>(stats.bytes_read_non_output_levels);
   }
-  LogToBuffer(
+  if (stats.micros > 0) {
+    bytes_read_per_sec =
+        (stats.bytes_read_non_output_levels + stats.bytes_read_output_level) /
+        static_cast<double>(stats.micros);
+    bytes_written_per_sec =
+        stats.bytes_written / static_cast<double>(stats.micros);
+  }
+
+  ROCKS_LOG_BUFFER(
       log_buffer_,
       "[%s] compacted to: %s, MB/sec: %.1f rd, %.1f wr, level %d, "
       "files in(%d, %d) out(%d) "
       "MB in(%.1f, %.1f) out(%.1f), read-write-amplify(%.1f) "
       "write-amplify(%.1f) %s, records in: %d, records dropped: %d\n",
-      cfd->GetName().c_str(), vstorage->LevelSummary(&tmp),
-      (stats.bytes_read_non_output_levels + stats.bytes_read_output_level) /
-          static_cast<double>(stats.micros),
-      stats.bytes_written / static_cast<double>(stats.micros),
-      compact_->compaction->output_level(),
+      cfd->GetName().c_str(), vstorage->LevelSummary(&tmp), bytes_read_per_sec,
+      bytes_written_per_sec, compact_->compaction->output_level(),
       stats.num_input_files_in_non_output_levels,
       stats.num_input_files_in_output_level, stats.num_output_files,
       stats.bytes_read_non_output_levels / 1048576.0,
@@ -733,17 +742,28 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   if (start != nullptr) {
     IterKey start_iter;
     start_iter.SetInternalKey(*start, kMaxSequenceNumber, kValueTypeForSeek);
-    input->Seek(start_iter.GetKey());
+    input->Seek(start_iter.GetInternalKey());
   } else {
     input->SeekToFirst();
   }
 
+  // we allow only 1 compaction event listener. Used by blob storage
+  CompactionEventListener* comp_event_listener = nullptr;
+#ifndef ROCKSDB_LITE
+  for (auto& celitr : cfd->ioptions()->listeners) {
+    comp_event_listener = celitr->GetCompactionEventListener();
+    if (comp_event_listener != nullptr) {
+      break;
+    }
+  }
+#endif  // ROCKSDB_LITE
+
   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,
       range_del_agg.get(), sub_compact->compaction, compaction_filter,
-      shutting_down_));
+      comp_event_listener, shutting_down_));
   auto c_iter = sub_compact->c_iter.get();
   c_iter->SeekToFirst();
   if (c_iter->Valid() &&
@@ -994,6 +1014,10 @@ void CompactionJob::RecordDroppedKeys(
     RecordTick(stats_, COMPACTION_RANGE_DEL_DROP_OBSOLETE,
                c_iter_stats.num_range_del_drop_obsolete);
   }
+  if (c_iter_stats.num_optimized_del_drop_obsolete > 0) {
+    RecordTick(stats_, COMPACTION_OPTIMIZED_DEL_DROP_OBSOLETE,
+               c_iter_stats.num_optimized_del_drop_obsolete);
+  }
 }
 
 Status CompactionJob::FinishCompactionOutputFile(
@@ -1011,7 +1035,6 @@ Status CompactionJob::FinishCompactionOutputFile(
   uint64_t output_number = sub_compact->current_output()->meta.fd.GetNumber();
   assert(output_number != 0);
 
-  TableProperties table_properties;
   // Check for iterator errors
   Status s = input_status;
   auto meta = &sub_compact->current_output()->meta;
@@ -1067,10 +1090,32 @@ Status CompactionJob::FinishCompactionOutputFile(
   }
   sub_compact->outfile.reset();
 
+  if (s.ok() && current_entries == 0) {
+    // If there is nothing to output, no necessary to generate a sst file.
+    // This happens when the output level is bottom level, at the same time
+    // the sub_compact output nothing.
+    std::string fname = TableFileName(
+        db_options_.db_paths, meta->fd.GetNumber(), meta->fd.GetPathId());
+    env_->DeleteFile(fname);
+
+    // Also need to remove the file from outputs, or it will be added to the
+    // VersionEdit.
+    assert(!sub_compact->outputs.empty());
+    sub_compact->outputs.pop_back();
+    sub_compact->builder.reset();
+    sub_compact->current_output_file_size = 0;
+    return s;
+  }
+
   ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
   TableProperties tp;
   if (s.ok() && current_entries > 0) {
     // Verify that the table is usable
+    // We set for_compaction to false and don't OptimizeForCompactionTableRead
+    // here because this is a special case after we finish the table building
+    // No matter whether use_direct_io_for_flush_and_compaction is true,
+    // we will regrad this verification as user reads since the goal is
+    // to cache it here for further user reads
     InternalIterator* iter = cfd->table_cache()->NewIterator(
         ReadOptions(), env_options_, cfd->internal_comparator(), meta->fd,
         nullptr /* range_del_agg */, nullptr,
@@ -1091,12 +1136,12 @@ Status CompactionJob::FinishCompactionOutputFile(
       tp = sub_compact->builder->GetTableProperties();
       sub_compact->current_output()->table_properties =
           std::make_shared<TableProperties>(tp);
-      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)" : "");
+      ROCKS_LOG_INFO(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)" : "");
     }
   }
   std::string fname = TableFileName(db_options_.db_paths, meta->fd.GetNumber(),
@@ -1114,12 +1159,22 @@ Status CompactionJob::FinishCompactionOutputFile(
                             meta->fd.GetPathId());
     sfm->OnAddFile(fn);
     if (sfm->IsMaxAllowedSpaceReached()) {
+      // TODO(ajkr): should we return OK() if max space was reached by the final
+      // compaction output file (similarly to how flush works when full)?
+      s = Status::IOError("Max allowed space was reached");
+      TEST_SYNC_POINT(
+          "CompactionJob::FinishCompactionOutputFile:"
+          "MaxAllowedSpaceReached");
       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");
+        Status new_bg_error = s;
+        // may temporarily unlock and lock the mutex.
+        EventHelpers::NotifyOnBackgroundError(
+            cfd->ioptions()->listeners, BackgroundErrorReason::kCompaction,
+            &new_bg_error, db_mutex_);
+        if (!new_bg_error.ok()) {
+          *db_bg_error_ = new_bg_error;
+        }
       }
     }
   }
@@ -1142,17 +1197,16 @@ Status CompactionJob::InstallCompactionResults(
   if (!versions_->VerifyCompactionFileConsistency(compaction)) {
     Compaction::InputLevelSummaryBuffer inputs_summary;
 
-    Log(InfoLogLevel::ERROR_LEVEL, db_options_.info_log,
-        "[%s] [JOB %d] Compaction %s aborted",
-        compaction->column_family_data()->GetName().c_str(), job_id_,
-        compaction->InputLevelSummary(&inputs_summary));
+    ROCKS_LOG_ERROR(db_options_.info_log, "[%s] [JOB %d] Compaction %s aborted",
+                    compaction->column_family_data()->GetName().c_str(),
+                    job_id_, compaction->InputLevelSummary(&inputs_summary));
     return Status::Corruption("Compaction input files inconsistent");
   }
 
   {
     Compaction::InputLevelSummaryBuffer inputs_summary;
-    Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
-        "[%s] [JOB %d] Compacted %s => %" PRIu64 " bytes",
+    ROCKS_LOG_INFO(
+        db_options_.info_log, "[%s] [JOB %d] Compacted %s => %" PRIu64 " bytes",
         compaction->column_family_data()->GetName().c_str(), job_id_,
         compaction->InputLevelSummary(&inputs_summary), compact_->total_bytes);
   }
@@ -1198,9 +1252,14 @@ Status CompactionJob::OpenCompactionOutputFile(
 #endif  // !ROCKSDB_LITE
   // Make the output file
   unique_ptr<WritableFile> writable_file;
-  Status s = NewWritableFile(env_, fname, &writable_file, env_options_);
+  EnvOptions opt_env_opts =
+      env_->OptimizeForCompactionTableWrite(env_options_, db_options_);
+  TEST_SYNC_POINT_CALLBACK("CompactionJob::OpenCompactionOutputFile",
+                           &opt_env_opts.use_direct_writes);
+  Status s = NewWritableFile(env_, fname, &writable_file, opt_env_opts);
   if (!s.ok()) {
-    Log(InfoLogLevel::ERROR_LEVEL, db_options_.info_log,
+    ROCKS_LOG_ERROR(
+        db_options_.info_log,
         "[%s] [JOB %d] OpenCompactionOutputFiles for table #%" PRIu64
         " fails at NewWritableFile with status %s",
         sub_compact->compaction->column_family_data()->GetName().c_str(),
@@ -1230,14 +1289,22 @@ Status CompactionJob::OpenCompactionOutputFile(
   // data is going to be found
   bool skip_filters =
       cfd->ioptions()->optimize_filters_for_hits && bottommost_level_;
+
+  uint64_t output_file_creation_time =
+      sub_compact->compaction->MaxInputFileCreationTime();
+  if (output_file_creation_time == 0) {
+    int64_t _current_time = 0;
+    db_options_.env->GetCurrentTime(&_current_time);  // ignore error
+    output_file_creation_time = static_cast<uint64_t>(_current_time);
+  }
+
   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->compaction->output_level(),
-      &sub_compact->compression_dict,
-      skip_filters));
+      sub_compact->compaction->output_level(), &sub_compact->compression_dict,
+      skip_filters, output_file_creation_time));
   LogFlush(db_options_.info_log);
   return s;
 }
@@ -1376,14 +1443,14 @@ void CompactionJob::LogCompaction() {
   // we're not logging
   if (db_options_.info_log_level <= InfoLogLevel::INFO_LEVEL) {
     Compaction::InputLevelSummaryBuffer inputs_summary;
-    Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
-        "[%s] [JOB %d] Compacting %s, score %.2f", cfd->GetName().c_str(),
-        job_id_, compaction->InputLevelSummary(&inputs_summary),
-        compaction->score());
+    ROCKS_LOG_INFO(
+        db_options_.info_log, "[%s] [JOB %d] Compacting %s, score %.2f",
+        cfd->GetName().c_str(), job_id_,
+        compaction->InputLevelSummary(&inputs_summary), compaction->score());
     char scratch[2345];
     compaction->Summary(scratch, sizeof(scratch));
-    Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
-        "[%s] Compaction start summary: %s\n", cfd->GetName().c_str(), scratch);
+    ROCKS_LOG_INFO(db_options_.info_log, "[%s] Compaction start summary: %s\n",
+                   cfd->GetName().c_str(), scratch);
     // build event logger report
     auto stream = event_logger_->Log();
     stream << "job" << job_id_ << "event"

db/compaction_job.h

@@ -29,6 +29,7 @@
 #include "db/version_edit.h"
 #include "db/write_controller.h"
 #include "db/write_thread.h"
+#include "options/db_options.h"
 #include "port/port.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/compaction_job_stats.h"
@@ -38,7 +39,6 @@
 #include "rocksdb/transaction_log.h"
 #include "table/scoped_arena_iterator.h"
 #include "util/autovector.h"
-#include "util/db_options.h"
 #include "util/event_logger.h"
 #include "util/stop_watch.h"
 #include "util/thread_local.h"

db/compaction_job_stats_test.cc

@@ -23,11 +23,13 @@
 
 #include "db/db_impl.h"
 #include "db/dbformat.h"
-#include "db/filename.h"
 #include "db/job_context.h"
 #include "db/version_set.h"
 #include "db/write_batch_internal.h"
+#include "env/mock_env.h"
 #include "memtable/hash_linklist_rep.h"
+#include "monitoring/statistics.h"
+#include "monitoring/thread_status_util.h"
 #include "port/stack_trace.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
@@ -50,17 +52,15 @@
 #include "table/plain_table_factory.h"
 #include "table/scoped_arena_iterator.h"
 #include "util/compression.h"
+#include "util/filename.h"
 #include "util/hash.h"
 #include "util/logging.h"
-#include "util/mock_env.h"
 #include "util/mutexlock.h"
 #include "util/rate_limiter.h"
-#include "util/statistics.h"
 #include "util/string_util.h"
 #include "util/sync_point.h"
 #include "util/testharness.h"
 #include "util/testutil.h"
-#include "util/thread_status_util.h"
 #include "utilities/merge_operators.h"
 
 #if !defined(IOS_CROSS_COMPILE)
@@ -654,7 +654,6 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
   Options options;
   options.listeners.emplace_back(stats_checker);
   options.create_if_missing = true;
-  options.max_background_flushes = 0;
   // just enough setting to hold off auto-compaction.
   options.level0_file_num_compaction_trigger = kTestScale + 1;
   options.num_levels = 3;
@@ -876,7 +875,6 @@ TEST_P(CompactionJobStatsTest, DeletionStatsTest) {
   Options options;
   options.listeners.emplace_back(stats_checker);
   options.create_if_missing = true;
-  options.max_background_flushes = 0;
   options.level0_file_num_compaction_trigger = kTestScale+1;
   options.num_levels = 3;
   options.compression = kNoCompression;

db/compaction_job_test.cc

@@ -142,6 +142,7 @@ class CompactionJobTest : public testing::Test {
   }
 
   void SetLastSequence(const SequenceNumber sequence_number) {
+    versions_->SetLastToBeWrittenSequence(sequence_number + 1);
     versions_->SetLastSequence(sequence_number + 1);
   }
 
@@ -337,6 +338,24 @@ TEST_F(CompactionJobTest, SimpleDeletion) {
   RunCompaction({files}, expected_results);
 }
 
+TEST_F(CompactionJobTest, OutputNothing) {
+  NewDB();
+
+  auto file1 = mock::MakeMockFile({{KeyStr("a", 1U, kTypeValue), "val"}});
+
+  AddMockFile(file1);
+
+  auto file2 = mock::MakeMockFile({{KeyStr("a", 2U, kTypeDeletion), ""}});
+
+  AddMockFile(file2);
+
+  auto expected_results = mock::MakeMockFile();
+
+  SetLastSequence(4U);
+  auto files = cfd_->current()->storage_info()->LevelFiles(0);
+  RunCompaction({files}, expected_results);
+}
+
 TEST_F(CompactionJobTest, SimpleOverwrite) {
   NewDB();
 

db/compaction_picker.h

@@ -17,10 +17,10 @@
 
 #include "db/compaction.h"
 #include "db/version_set.h"
+#include "options/cf_options.h"
 #include "rocksdb/env.h"
 #include "rocksdb/options.h"
 #include "rocksdb/status.h"
-#include "util/cf_options.h"
 
 namespace rocksdb {
 
@@ -84,15 +84,15 @@ class CompactionPicker {
   void ReleaseCompactionFiles(Compaction* c, Status status);
 
   // Returns true if any one of the specified files are being compacted
-  bool FilesInCompaction(const std::vector<FileMetaData*>& files);
+  bool AreFilesInCompaction(const std::vector<FileMetaData*>& files);
 
   // Takes a list of CompactionInputFiles and returns a (manual) Compaction
   // object.
-  Compaction* 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);
+  Compaction* CompactFiles(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);
 
   // Converts a set of compaction input file numbers into
   // a list of CompactionInputFiles.
@@ -102,12 +102,6 @@ class CompactionPicker {
       const VersionStorageInfo* vstorage,
       const CompactionOptions& compact_options) const;
 
-  // Used in universal compaction when the enabled_trivial_move
-  // option is set. Checks whether there are any overlapping files
-  // in the input. Returns true if the input files are non
-  // overlapping.
-  bool IsInputNonOverlapping(Compaction* c);
-
   // Is there currently a compaction involving level 0 taking place
   bool IsLevel0CompactionInProgress() const {
     return !level0_compactions_in_progress_.empty();
@@ -138,7 +132,6 @@ class CompactionPicker {
   void GetRange(const std::vector<CompactionInputFiles>& inputs,
                 InternalKey* smallest, InternalKey* largest) const;
 
- protected:
   int NumberLevels() const { return ioptions_.num_levels; }
 
   // Add more files to the inputs on "level" to make sure that
@@ -151,14 +144,14 @@ class CompactionPicker {
   // populated.
   //
   // Will return false if it is impossible to apply this compaction.
-  bool ExpandWhileOverlapping(const std::string& cf_name,
+  bool ExpandInputsToCleanCut(const std::string& cf_name,
                               VersionStorageInfo* vstorage,
                               CompactionInputFiles* inputs);
 
   // Returns true if any one of the parent files are being compacted
-  bool RangeInCompaction(VersionStorageInfo* vstorage,
-                         const InternalKey* smallest,
-                         const InternalKey* largest, int level, int* index);
+  bool IsRangeInCompaction(VersionStorageInfo* vstorage,
+                           const InternalKey* smallest,
+                           const InternalKey* largest, int level, int* index);
 
   // Returns true if the key range that `inputs` files cover overlap with the
   // key range of a currently running compaction.
@@ -177,6 +170,20 @@ class CompactionPicker {
                        const CompactionInputFiles& output_level_inputs,
                        std::vector<FileMetaData*>* grandparents);
 
+  // Register this compaction in the set of running compactions
+  void RegisterCompaction(Compaction* c);
+
+  // Remove this compaction from the set of running compactions
+  void UnregisterCompaction(Compaction* c);
+
+  std::set<Compaction*>* level0_compactions_in_progress() {
+    return &level0_compactions_in_progress_;
+  }
+  std::unordered_set<Compaction*>* compactions_in_progress() {
+    return &compactions_in_progress_;
+  }
+
+ protected:
   const ImmutableCFOptions& ioptions_;
 
 // A helper function to SanitizeCompactionInputFiles() that
@@ -187,12 +194,6 @@ class CompactionPicker {
       const ColumnFamilyMetaData& cf_meta, const int output_level) const;
 #endif  // ROCKSDB_LITE
 
-  // Register this compaction in the set of running compactions
-  void RegisterCompaction(Compaction* c);
-
-  // Remove this compaction from the set of running compactions
-  void UnregisterCompaction(Compaction* c);
-
   // Keeps track of all compactions that are running on Level0.
   // Protected by DB mutex
   std::set<Compaction*> level0_compactions_in_progress_;
@@ -216,101 +217,9 @@ class LevelCompactionPicker : public CompactionPicker {
 
   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,
-                            const MutableCFOptions& mutable_cf_options,
-                            int level);
-
- private:
-  // For the specfied level, pick a file that we want to compact.
-  // Returns false if there is no file to compact.
-  // If it returns true, inputs->files.size() will be exactly one.
-  // If level is 0 and there is already a compaction on that level, this
-  // function will return false.
-  bool PickCompactionBySize(VersionStorageInfo* vstorage, int level,
-                            int output_level, CompactionInputFiles* inputs,
-                            int* parent_index, int* base_index);
-
-  // If there is any file marked for compaction, put put it into inputs.
-  // This is still experimental. It will return meaningful results only if
-  // clients call experimental feature SuggestCompactRange()
-  void PickFilesMarkedForCompactionExperimental(const std::string& cf_name,
-                                                VersionStorageInfo* vstorage,
-                                                CompactionInputFiles* inputs,
-                                                int* level, int* output_level);
 };
 
 #ifndef ROCKSDB_LITE
-class UniversalCompactionPicker : public CompactionPicker {
- public:
-  UniversalCompactionPicker(const ImmutableCFOptions& ioptions,
-                            const InternalKeyComparator* icmp)
-      : CompactionPicker(ioptions, icmp) {}
-  virtual Compaction* PickCompaction(const std::string& cf_name,
-                                     const MutableCFOptions& mutable_cf_options,
-                                     VersionStorageInfo* vstorage,
-                                     LogBuffer* log_buffer) override;
-
-  virtual int MaxOutputLevel() const override { return NumberLevels() - 1; }
-
-  virtual bool NeedsCompaction(
-      const VersionStorageInfo* vstorage) const override;
-
- private:
-  struct SortedRun {
-    SortedRun(int _level, FileMetaData* _file, uint64_t _size,
-              uint64_t _compensated_file_size, bool _being_compacted)
-        : level(_level),
-          file(_file),
-          size(_size),
-          compensated_file_size(_compensated_file_size),
-          being_compacted(_being_compacted) {
-      assert(compensated_file_size > 0);
-      assert(level != 0 || file != nullptr);
-    }
-
-    void Dump(char* out_buf, size_t out_buf_size,
-              bool print_path = false) const;
-
-    // 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 level;
-    // `file` Will be null for level > 0. For level = 0, the sorted run is
-    // for this file.
-    FileMetaData* file;
-    // For level > 0, `size` and `compensated_file_size` are sum of sizes all
-    // files in the level. `being_compacted` should be the same for all files
-    // in a non-zero level. Use the value here.
-    uint64_t size;
-    uint64_t compensated_file_size;
-    bool being_compacted;
-  };
-
-  // Pick Universal compaction to limit read amplification
-  Compaction* PickCompactionUniversalReadAmp(
-      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, double score, unsigned int ratio,
-      unsigned int num_files, const std::vector<SortedRun>& sorted_runs,
-      LogBuffer* log_buffer);
-
-  // Pick Universal compaction to limit space amplification.
-  Compaction* PickCompactionUniversalSizeAmp(
-      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, double score,
-      const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer);
-
-  static std::vector<SortedRun> CalculateSortedRuns(
-      const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions);
-
-  // Pick a path ID to place a newly generated file, with its estimated file
-  // size.
-  static uint32_t GetPathId(const ImmutableCFOptions& ioptions,
-                            uint64_t file_size);
-};
-
 class FIFOCompactionPicker : public CompactionPicker {
  public:
   FIFOCompactionPicker(const ImmutableCFOptions& ioptions,
@@ -333,6 +242,17 @@ class FIFOCompactionPicker : public CompactionPicker {
 
   virtual bool NeedsCompaction(
       const VersionStorageInfo* vstorage) const override;
+
+ private:
+  Compaction* PickTTLCompaction(const std::string& cf_name,
+                                const MutableCFOptions& mutable_cf_options,
+                                VersionStorageInfo* version,
+                                LogBuffer* log_buffer);
+
+  Compaction* PickSizeCompaction(const std::string& cf_name,
+                                 const MutableCFOptions& mutable_cf_options,
+                                 VersionStorageInfo* version,
+                                 LogBuffer* log_buffer);
 };
 
 class NullCompactionPicker : public CompactionPicker {

db/compaction_picker_test.cc

@@ -3,11 +3,12 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 
-#include "db/compaction.h"
 #include "db/compaction_picker.h"
 #include <limits>
 #include <string>
 #include <utility>
+#include "db/compaction.h"
+#include "db/compaction_picker_universal.h"
 
 #include "util/logging.h"
 #include "util/string_util.h"
@@ -197,7 +198,7 @@ TEST_F(CompactionPickerTest, LevelMaxScore) {
   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);
+  Add(0, 1U, "150", "200", 1000000U);
   // Level 1 score 1.2
   Add(1, 66U, "150", "200", 6000000U);
   Add(1, 88U, "201", "300", 6000000U);
@@ -399,6 +400,35 @@ TEST_F(CompactionPickerTest, NeedsCompactionUniversal) {
               vstorage_->CompactionScore(0) >= 1);
   }
 }
+
+TEST_F(CompactionPickerTest, CompactionUniversalIngestBehindReservedLevel) {
+  const uint64_t kFileSize = 100000;
+  NewVersionStorage(1, kCompactionStyleUniversal);
+  ioptions_.allow_ingest_behind = true;
+  ioptions_.num_levels = 3;
+  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+  // must return false when there's no files.
+  ASSERT_EQ(universal_compaction_picker.NeedsCompaction(vstorage_.get()),
+            false);
+
+  NewVersionStorage(3, kCompactionStyleUniversal);
+
+  Add(0, 1U, "150", "200", kFileSize, 0, 500, 550);
+  Add(0, 2U, "201", "250", kFileSize, 0, 401, 450);
+  Add(0, 4U, "260", "300", kFileSize, 0, 260, 300);
+  Add(1, 5U, "100", "151", kFileSize, 0, 200, 251);
+  Add(1, 3U, "301", "350", kFileSize, 0, 101, 150);
+  Add(2, 6U, "120", "200", kFileSize, 0, 20, 100);
+
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(
+      universal_compaction_picker.PickCompaction(
+          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+
+  // output level should be the one above the bottom-most
+  ASSERT_EQ(1, compaction->output_level());
+}
 // Tests if the files can be trivially moved in multi level
 // universal compaction when allow_trivial_move option is set
 // In this test as the input files overlaps, they cannot
@@ -482,7 +512,7 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
         kFileSize, 0, i * 100, i * 100 + 99);
     current_size += kFileSize;
     UpdateVersionStorageInfo();
-    ASSERT_EQ(level_compaction_picker.NeedsCompaction(vstorage_.get()),
+    ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()),
               vstorage_->CompactionScore(0) >= 1);
   }
 }
@@ -822,6 +852,80 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys9) {
   ASSERT_EQ(8U, compaction->input(1, 1)->fd.GetNumber());
 }
 
+TEST_F(CompactionPickerTest, OverlappingUserKeys10) {
+  // Locked file encountered when pulling in extra input-level files with same
+  // user keys. Verify we pick the next-best file from the same input level.
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  // file_number 2U is largest and thus first choice. But it overlaps with
+  // file_number 1U which is being compacted. So instead we pick the next-
+  // biggest file, 3U, which is eligible for compaction.
+  Add(1 /* level */, 1U /* file_number */, "100" /* smallest */,
+      "150" /* largest */, 1U /* file_size */);
+  file_map_[1U].first->being_compacted = true;
+  Add(1 /* level */, 2U /* file_number */, "150" /* smallest */,
+      "200" /* largest */, 1000000000U /* file_size */, 0 /* smallest_seq */,
+      0 /* largest_seq */);
+  Add(1 /* level */, 3U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 900000000U /* file_size */);
+  Add(2 /* level */, 4U /* file_number */, "100" /* smallest */,
+      "150" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "151" /* smallest */,
+      "200" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 6U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 1U /* file_size */);
+
+  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(6U, compaction->input(1, 0)->fd.GetNumber());
+}
+
+TEST_F(CompactionPickerTest, OverlappingUserKeys11) {
+  // Locked file encountered when pulling in extra output-level files with same
+  // user keys. Expected to skip that compaction and pick the next-best choice.
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  // score(L1) = 3.7
+  // score(L2) = 1.85
+  // There is no eligible file in L1 to compact since both candidates pull in
+  // file_number 5U, which overlaps with a file pending compaction (6U). The
+  // first eligible compaction is from L2->L3.
+  Add(1 /* level */, 2U /* file_number */, "151" /* smallest */,
+      "200" /* largest */, 1000000000U /* file_size */);
+  Add(1 /* level */, 3U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 4U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 5000000000U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "150" /* smallest */,
+      "201" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 6U /* file_number */, "201" /* smallest */,
+      "249" /* largest */, 1U /* file_size */, 0 /* smallest_seq */,
+      0 /* largest_seq */);
+  file_map_[6U].first->being_compacted = true;
+  Add(3 /* level */, 7U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 1U /* file_size */);
+
+  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(4U, 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;
@@ -920,7 +1024,7 @@ TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri3) {
 TEST_F(CompactionPickerTest, EstimateCompactionBytesNeeded1) {
   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_.level0_file_num_compaction_trigger = 4;
   mutable_cf_options_.max_bytes_for_level_base = 1000;
   mutable_cf_options_.max_bytes_for_level_multiplier = 10;
   NewVersionStorage(num_levels, kCompactionStyleLevel);
@@ -1010,7 +1114,8 @@ TEST_F(CompactionPickerTest, EstimateCompactionBytesNeededDynamicLevel) {
 
   // Set Last level size 50000
   // num_levels - 1 target 5000
-  // num_levels - 2 is base level with taret 500
+  // num_levels - 2 is base level with target 1000 (rounded up to
+  // max_bytes_for_level_base).
   Add(num_levels - 1, 10U, "400", "500", 50000);
 
   Add(0, 1U, "150", "200", 200);
@@ -1019,16 +1124,16 @@ TEST_F(CompactionPickerTest, EstimateCompactionBytesNeededDynamicLevel) {
   Add(0, 5U, "150", "200", 200);
   Add(0, 6U, "150", "200", 200);
   // 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);
+  Add(num_levels - 3, 7U, "400", "500", 550);
+  Add(num_levels - 3, 8U, "600", "700", 550);
   // num_levels - 2 is over target by 1100 + 200
   Add(num_levels - 2, 9U, "150", "200", 5200);
 
   UpdateVersionStorageInfo();
 
-  // Merging to the second last level: (5200 / 1600 + 1) * 1100
+  // Merging to the second last level: (5200 / 2100 + 1) * 1100
   // Merging to the last level: (50000 / 6300 + 1) * 1300
-  ASSERT_EQ(1600u + 4675u + 11617u,
+  ASSERT_EQ(2100u + 3823u + 11617u,
             vstorage_->estimated_compaction_needed_bytes());
 }
 
@@ -1285,6 +1390,49 @@ TEST_F(CompactionPickerTest, IsTrivialMoveOff) {
   ASSERT_FALSE(compaction->IsTrivialMove());
 }
 
+TEST_F(CompactionPickerTest, CacheNextCompactionIndex) {
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  Add(1 /* level */, 1U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 1000000000U /* file_size */);
+  file_map_[1U].first->being_compacted = true;
+  Add(1 /* level */, 2U /* file_number */, "150" /* smallest */,
+      "199" /* largest */, 900000000U /* file_size */);
+  Add(1 /* level */, 3U /* file_number */, "200" /* smallest */,
+      "249" /* largest */, 800000000U /* file_size */);
+  Add(1 /* level */, 4U /* file_number */, "250" /* smallest */,
+      "299" /* largest */, 700000000U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "150" /* smallest */,
+      "199" /* largest */, 1U /* file_size */);
+  file_map_[5U].first->being_compacted = true;
+
+  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_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(0U, compaction->num_input_files(1));
+  ASSERT_EQ(3U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(2, vstorage_->NextCompactionIndex(1 /* level */));
+
+  compaction.reset(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(1U, compaction->num_input_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(0U, compaction->num_input_files(1));
+  ASSERT_EQ(4U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(3, vstorage_->NextCompactionIndex(1 /* level */));
+
+  compaction.reset(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() == nullptr);
+  ASSERT_EQ(4, vstorage_->NextCompactionIndex(1 /* level */));
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/compaction_picker_universal.cc

@@ -0,0 +1,748 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root 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/compaction_picker_universal.h"
+#ifndef ROCKSDB_LITE
+
+#ifndef __STDC_FORMAT_MACROS
+#define __STDC_FORMAT_MACROS
+#endif
+
+#include <inttypes.h>
+#include <limits>
+#include <queue>
+#include <string>
+#include <utility>
+#include "db/column_family.h"
+#include "monitoring/statistics.h"
+#include "util/filename.h"
+#include "util/log_buffer.h"
+#include "util/random.h"
+#include "util/string_util.h"
+#include "util/sync_point.h"
+
+namespace rocksdb {
+namespace {
+// Used in universal compaction when trivial move is enabled.
+// This structure is used for the construction of min heap
+// that contains the file meta data, the level of the file
+// and the index of the file in that level
+
+struct InputFileInfo {
+  InputFileInfo() : f(nullptr) {}
+
+  FileMetaData* f;
+  size_t level;
+  size_t index;
+};
+
+// Used in universal compaction when trivial move is enabled.
+// This comparator is used for the construction of min heap
+// based on the smallest key of the file.
+struct SmallestKeyHeapComparator {
+  explicit SmallestKeyHeapComparator(const Comparator* ucmp) { ucmp_ = ucmp; }
+
+  bool operator()(InputFileInfo i1, InputFileInfo i2) const {
+    return (ucmp_->Compare(i1.f->smallest.user_key(),
+                           i2.f->smallest.user_key()) > 0);
+  }
+
+ private:
+  const Comparator* ucmp_;
+};
+
+typedef std::priority_queue<InputFileInfo, std::vector<InputFileInfo>,
+                            SmallestKeyHeapComparator>
+    SmallestKeyHeap;
+
+// This function creates the heap that is used to find if the files are
+// overlapping during universal compaction when the allow_trivial_move
+// is set.
+SmallestKeyHeap create_level_heap(Compaction* c, const Comparator* ucmp) {
+  SmallestKeyHeap smallest_key_priority_q =
+      SmallestKeyHeap(SmallestKeyHeapComparator(ucmp));
+
+  InputFileInfo input_file;
+
+  for (size_t l = 0; l < c->num_input_levels(); l++) {
+    if (c->num_input_files(l) != 0) {
+      if (l == 0 && c->start_level() == 0) {
+        for (size_t i = 0; i < c->num_input_files(0); i++) {
+          input_file.f = c->input(0, i);
+          input_file.level = 0;
+          input_file.index = i;
+          smallest_key_priority_q.push(std::move(input_file));
+        }
+      } else {
+        input_file.f = c->input(l, 0);
+        input_file.level = l;
+        input_file.index = 0;
+        smallest_key_priority_q.push(std::move(input_file));
+      }
+    }
+  }
+  return smallest_key_priority_q;
+}
+
+#ifndef NDEBUG
+// smallest_seqno and largest_seqno are set iff. `files` is not empty.
+void GetSmallestLargestSeqno(const std::vector<FileMetaData*>& files,
+                             SequenceNumber* smallest_seqno,
+                             SequenceNumber* largest_seqno) {
+  bool is_first = true;
+  for (FileMetaData* f : files) {
+    assert(f->smallest_seqno <= f->largest_seqno);
+    if (is_first) {
+      is_first = false;
+      *smallest_seqno = f->smallest_seqno;
+      *largest_seqno = f->largest_seqno;
+    } else {
+      if (f->smallest_seqno < *smallest_seqno) {
+        *smallest_seqno = f->smallest_seqno;
+      }
+      if (f->largest_seqno > *largest_seqno) {
+        *largest_seqno = f->largest_seqno;
+      }
+    }
+  }
+}
+#endif
+}  // namespace
+
+// Algorithm that checks to see if there are any overlapping
+// files in the input
+bool UniversalCompactionPicker::IsInputFilesNonOverlapping(Compaction* c) {
+  auto comparator = icmp_->user_comparator();
+  int first_iter = 1;
+
+  InputFileInfo prev, curr, next;
+
+  SmallestKeyHeap smallest_key_priority_q =
+      create_level_heap(c, icmp_->user_comparator());
+
+  while (!smallest_key_priority_q.empty()) {
+    curr = smallest_key_priority_q.top();
+    smallest_key_priority_q.pop();
+
+    if (first_iter) {
+      prev = curr;
+      first_iter = 0;
+    } else {
+      if (comparator->Compare(prev.f->largest.user_key(),
+                              curr.f->smallest.user_key()) >= 0) {
+        // found overlapping files, return false
+        return false;
+      }
+      assert(comparator->Compare(curr.f->largest.user_key(),
+                                 prev.f->largest.user_key()) > 0);
+      prev = curr;
+    }
+
+    next.f = nullptr;
+
+    if (curr.level != 0 && curr.index < c->num_input_files(curr.level) - 1) {
+      next.f = c->input(curr.level, curr.index + 1);
+      next.level = curr.level;
+      next.index = curr.index + 1;
+    }
+
+    if (next.f) {
+      smallest_key_priority_q.push(std::move(next));
+    }
+  }
+  return true;
+}
+
+bool UniversalCompactionPicker::NeedsCompaction(
+    const VersionStorageInfo* vstorage) const {
+  const int kLevel0 = 0;
+  return vstorage->CompactionScore(kLevel0) >= 1;
+}
+
+void UniversalCompactionPicker::SortedRun::Dump(char* out_buf,
+                                                size_t out_buf_size,
+                                                bool print_path) const {
+  if (level == 0) {
+    assert(file != nullptr);
+    if (file->fd.GetPathId() == 0 || !print_path) {
+      snprintf(out_buf, out_buf_size, "file %" PRIu64, file->fd.GetNumber());
+    } else {
+      snprintf(out_buf, out_buf_size, "file %" PRIu64
+                                      "(path "
+                                      "%" PRIu32 ")",
+               file->fd.GetNumber(), file->fd.GetPathId());
+    }
+  } else {
+    snprintf(out_buf, out_buf_size, "level %d", level);
+  }
+}
+
+void UniversalCompactionPicker::SortedRun::DumpSizeInfo(
+    char* out_buf, size_t out_buf_size, size_t sorted_run_count) const {
+  if (level == 0) {
+    assert(file != nullptr);
+    snprintf(out_buf, out_buf_size,
+             "file %" PRIu64 "[%" ROCKSDB_PRIszt
+             "] "
+             "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
+             "] "
+             "with size %" PRIu64 " (compensated size %" PRIu64 ")",
+             level, sorted_run_count, size, compensated_file_size);
+  }
+}
+
+std::vector<UniversalCompactionPicker::SortedRun>
+UniversalCompactionPicker::CalculateSortedRuns(
+    const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions) {
+  std::vector<UniversalCompactionPicker::SortedRun> ret;
+  for (FileMetaData* f : vstorage.LevelFiles(0)) {
+    ret.emplace_back(0, f, f->fd.GetFileSize(), f->compensated_file_size,
+                     f->being_compacted);
+  }
+  for (int level = 1; level < vstorage.num_levels(); level++) {
+    uint64_t total_compensated_size = 0U;
+    uint64_t total_size = 0U;
+    bool being_compacted = false;
+    bool is_first = true;
+    for (FileMetaData* f : vstorage.LevelFiles(level)) {
+      total_compensated_size += f->compensated_file_size;
+      total_size += f->fd.GetFileSize();
+      if (ioptions.compaction_options_universal.allow_trivial_move == true) {
+        if (f->being_compacted) {
+          being_compacted = f->being_compacted;
+        }
+      } else {
+        // Compaction always includes all files for a non-zero level, so for a
+        // non-zero level, all the files should share the same being_compacted
+        // value.
+        // This assumption is only valid when
+        // ioptions.compaction_options_universal.allow_trivial_move is false
+        assert(is_first || f->being_compacted == being_compacted);
+      }
+      if (is_first) {
+        being_compacted = f->being_compacted;
+        is_first = false;
+      }
+    }
+    if (total_compensated_size > 0) {
+      ret.emplace_back(level, nullptr, total_size, total_compensated_size,
+                       being_compacted);
+    }
+  }
+  return ret;
+}
+
+// Universal style of compaction. Pick files that are contiguous in
+// time-range to compact.
+//
+Compaction* UniversalCompactionPicker::PickCompaction(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+  const int kLevel0 = 0;
+  double score = vstorage->CompactionScore(kLevel0);
+  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) {
+    ROCKS_LOG_BUFFER(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;
+  ROCKS_LOG_BUFFER_MAX_SZ(
+      log_buffer, 3072,
+      "[%s] Universal: sorted runs files(%" ROCKSDB_PRIszt "): %s\n",
+      cf_name.c_str(), sorted_runs.size(), vstorage->LevelSummary(&tmp));
+
+  // Check for size amplification first.
+  Compaction* c;
+  if ((c = PickCompactionToReduceSizeAmp(cf_name, mutable_cf_options, vstorage,
+                                         score, sorted_runs, log_buffer)) !=
+      nullptr) {
+    ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: compacting for size amp\n",
+                     cf_name.c_str());
+  } else {
+    // Size amplification is within limits. Try reducing read
+    // amplification while maintaining file size ratios.
+    unsigned int ratio = ioptions_.compaction_options_universal.size_ratio;
+
+    if ((c = PickCompactionToReduceSortedRuns(
+             cf_name, mutable_cf_options, vstorage, score, ratio, UINT_MAX,
+             sorted_runs, log_buffer)) != nullptr) {
+      ROCKS_LOG_BUFFER(log_buffer,
+                       "[%s] Universal: compacting for size ratio\n",
+                       cf_name.c_str());
+    } else {
+      // Size amplification and file size ratios are within configured limits.
+      // If max read amplification is exceeding configured limits, then force
+      // compaction without looking at filesize ratios and try to reduce
+      // the number of files to fewer than level0_file_num_compaction_trigger.
+      // This is guaranteed by NeedsCompaction()
+      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 = PickCompactionToReduceSortedRuns(
+                 cf_name, mutable_cf_options, vstorage, score, UINT_MAX,
+                 num_files, sorted_runs, log_buffer)) != nullptr) {
+          ROCKS_LOG_BUFFER(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;
+  }
+
+  if (ioptions_.compaction_options_universal.allow_trivial_move == true) {
+    c->set_is_trivial_move(IsInputFilesNonOverlapping(c));
+  }
+
+// validate that all the chosen files of L0 are non overlapping in time
+#ifndef NDEBUG
+  SequenceNumber prev_smallest_seqno = 0U;
+  bool is_first = true;
+
+  size_t level_index = 0U;
+  if (c->start_level() == 0) {
+    for (auto f : *c->inputs(0)) {
+      assert(f->smallest_seqno <= f->largest_seqno);
+      if (is_first) {
+        is_first = false;
+      }
+      prev_smallest_seqno = f->smallest_seqno;
+    }
+    level_index = 1U;
+  }
+  for (; level_index < c->num_input_levels(); level_index++) {
+    if (c->num_input_files(level_index) != 0) {
+      SequenceNumber smallest_seqno = 0U;
+      SequenceNumber largest_seqno = 0U;
+      GetSmallestLargestSeqno(*(c->inputs(level_index)), &smallest_seqno,
+                              &largest_seqno);
+      if (is_first) {
+        is_first = false;
+      } else if (prev_smallest_seqno > 0) {
+        // A level is considered as the bottommost level if there are
+        // no files in higher levels or if files in higher levels do
+        // not overlap with the files being compacted. Sequence numbers
+        // of files in bottommost level can be set to 0 to help
+        // compression. As a result, the following assert may not hold
+        // if the prev_smallest_seqno is 0.
+        assert(prev_smallest_seqno > largest_seqno);
+      }
+      prev_smallest_seqno = smallest_seqno;
+    }
+  }
+#endif
+  // update statistics
+  MeasureTime(ioptions_.statistics, NUM_FILES_IN_SINGLE_COMPACTION,
+              c->inputs(0)->size());
+
+  RegisterCompaction(c);
+  vstorage->ComputeCompactionScore(ioptions_, mutable_cf_options);
+
+  TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
+                           c);
+  return c;
+}
+
+uint32_t UniversalCompactionPicker::GetPathId(
+    const ImmutableCFOptions& ioptions, uint64_t file_size) {
+  // Two conditions need to be satisfied:
+  // (1) the target path needs to be able to hold the file's size
+  // (2) Total size left in this and previous paths need to be not
+  //     smaller than expected future file size before this new file is
+  //     compacted, which is estimated based on size_ratio.
+  // For example, if now we are compacting files of size (1, 1, 2, 4, 8),
+  // we will make sure the target file, probably with size of 16, will be
+  // placed in a path so that eventually when new files are generated and
+  // compacted to (1, 1, 2, 4, 8, 16), all those files can be stored in or
+  // before the path we chose.
+  //
+  // TODO(sdong): now the case of multiple column families is not
+  // 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;
+  uint32_t p = 0;
+  assert(!ioptions.db_paths.empty());
+  for (; p < ioptions.db_paths.size() - 1; p++) {
+    uint64_t target_size = ioptions.db_paths[p].target_size;
+    if (target_size > file_size &&
+        accumulated_size + (target_size - file_size) > future_size) {
+      return p;
+    }
+    accumulated_size += target_size;
+  }
+  return p;
+}
+
+//
+// Consider compaction files based on their size differences with
+// the next file in time order.
+//
+Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, double score, unsigned int ratio,
+    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;
+  unsigned int max_merge_width =
+      ioptions_.compaction_options_universal.max_merge_width;
+
+  const SortedRun* sr = nullptr;
+  bool done = false;
+  size_t 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);
+  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++) {
+    candidate_count = 0;
+
+    // Skip files that are already being compacted
+    for (sr = nullptr; loop < sorted_runs.size(); loop++) {
+      sr = &sorted_runs[loop];
+
+      if (!sr->being_compacted) {
+        candidate_count = 1;
+        break;
+      }
+      char file_num_buf[kFormatFileNumberBufSize];
+      sr->Dump(file_num_buf, sizeof(file_num_buf));
+      ROCKS_LOG_BUFFER(log_buffer,
+                       "[%s] Universal: %s"
+                       "[%d] being compacted, skipping",
+                       cf_name.c_str(), file_num_buf, loop);
+
+      sr = nullptr;
+    }
+
+    // This file is not being compacted. Consider it as the
+    // first candidate to be compacted.
+    uint64_t candidate_size = sr != nullptr ? sr->compensated_file_size : 0;
+    if (sr != nullptr) {
+      char file_num_buf[kFormatFileNumberBufSize];
+      sr->Dump(file_num_buf, sizeof(file_num_buf), true);
+      ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: Possible candidate %s[%d].",
+                       cf_name.c_str(), file_num_buf, loop);
+    }
+
+    // Check if the succeeding files need compaction.
+    for (size_t i = loop + 1;
+         candidate_count < max_files_to_compact && i < sorted_runs.size();
+         i++) {
+      const SortedRun* succeeding_sr = &sorted_runs[i];
+      if (succeeding_sr->being_compacted) {
+        break;
+      }
+      // Pick files if the total/last candidate file size (increased by the
+      // specified ratio) is still larger than the next candidate file.
+      // candidate_size is the total size of files picked so far with the
+      // default kCompactionStopStyleTotalSize; with
+      // kCompactionStopStyleSimilarSize, it's simply the size of the last
+      // picked file.
+      double sz = candidate_size * (100.0 + ratio) / 100.0;
+      if (sz < static_cast<double>(succeeding_sr->size)) {
+        break;
+      }
+      if (ioptions_.compaction_options_universal.stop_style ==
+          kCompactionStopStyleSimilarSize) {
+        // Similar-size stopping rule: also check the last picked file isn't
+        // far larger than the next candidate file.
+        sz = (succeeding_sr->size * (100.0 + ratio)) / 100.0;
+        if (sz < static_cast<double>(candidate_size)) {
+          // If the small file we've encountered begins a run of similar-size
+          // files, we'll pick them up on a future iteration of the outer
+          // loop. If it's some lonely straggler, it'll eventually get picked
+          // by the last-resort read amp strategy which disregards size ratios.
+          break;
+        }
+        candidate_size = succeeding_sr->compensated_file_size;
+      } else {  // default kCompactionStopStyleTotalSize
+        candidate_size += succeeding_sr->compensated_file_size;
+      }
+      candidate_count++;
+    }
+
+    // Found a series of consecutive files that need compaction.
+    if (candidate_count >= (unsigned int)min_merge_width) {
+      start_index = loop;
+      done = true;
+      break;
+    } else {
+      for (size_t i = loop;
+           i < loop + candidate_count && i < sorted_runs.size(); i++) {
+        const SortedRun* skipping_sr = &sorted_runs[i];
+        char file_num_buf[256];
+        skipping_sr->DumpSizeInfo(file_num_buf, sizeof(file_num_buf), loop);
+        ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: Skipping %s",
+                         cf_name.c_str(), file_num_buf);
+      }
+    }
+  }
+  if (!done || candidate_count <= 1) {
+    return nullptr;
+  }
+  size_t first_index_after = start_index + candidate_count;
+  // Compression is enabled if files compacted earlier already reached
+  // size ratio of compression.
+  bool enable_compression = true;
+  int ratio_to_compress =
+      ioptions_.compaction_options_universal.compression_size_percent;
+  if (ratio_to_compress >= 0) {
+    uint64_t total_size = 0;
+    for (auto& sorted_run : sorted_runs) {
+      total_size += sorted_run.compensated_file_size;
+    }
+
+    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) {
+        enable_compression = false;
+        break;
+      }
+    }
+  }
+
+  uint64_t estimated_total_size = 0;
+  for (unsigned int i = 0; i < first_index_after; i++) {
+    estimated_total_size += sorted_runs[i].size;
+  }
+  uint32_t path_id = GetPathId(ioptions_, estimated_total_size);
+  int start_level = sorted_runs[start_index].level;
+  int output_level;
+  if (first_index_after == sorted_runs.size()) {
+    output_level = vstorage->num_levels() - 1;
+  } else if (sorted_runs[first_index_after].level == 0) {
+    output_level = 0;
+  } else {
+    output_level = sorted_runs[first_index_after].level - 1;
+  }
+
+  // last level is reserved for the files ingested behind
+  if (ioptions_.allow_ingest_behind &&
+      (output_level == vstorage->num_levels() - 1)) {
+    assert(output_level > 1);
+    output_level--;
+  }
+
+  std::vector<CompactionInputFiles> inputs(vstorage->num_levels());
+  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++) {
+    auto& picking_sr = sorted_runs[i];
+    if (picking_sr.level == 0) {
+      FileMetaData* picking_file = picking_sr.file;
+      inputs[0].files.push_back(picking_file);
+    } else {
+      auto& files = inputs[picking_sr.level - start_level].files;
+      for (auto* f : vstorage->LevelFiles(picking_sr.level)) {
+        files.push_back(f);
+      }
+    }
+    char file_num_buf[256];
+    picking_sr.DumpSizeInfo(file_num_buf, sizeof(file_num_buf), i);
+    ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: Picking %s", cf_name.c_str(),
+                     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, ioptions_, 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);
+}
+
+// Look at overall size amplification. If size amplification
+// exceeeds the configured value, then do a compaction
+// of the candidate files all the way upto the earliest
+// base file (overrides configured values of file-size ratios,
+// min_merge_width and max_merge_width).
+//
+Compaction* UniversalCompactionPicker::PickCompactionToReduceSizeAmp(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, double score,
+    const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer) {
+  // percentage flexibility while reducing size amplification
+  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;
+  const SortedRun* sr = nullptr;
+
+  // Skip files that are already being compacted
+  for (size_t 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.
+      break;
+    }
+    char file_num_buf[kFormatFileNumberBufSize];
+    sr->Dump(file_num_buf, sizeof(file_num_buf), true);
+    ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: skipping %s[%d] compacted %s",
+                     cf_name.c_str(), file_num_buf, loop,
+                     " cannot be a candidate to reduce size amp.\n");
+    sr = nullptr;
+  }
+
+  if (sr == nullptr) {
+    return nullptr;  // no candidate files
+  }
+  {
+    char file_num_buf[kFormatFileNumberBufSize];
+    sr->Dump(file_num_buf, sizeof(file_num_buf), true);
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] Universal: First candidate %s[%" ROCKSDB_PRIszt "] %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++) {
+    sr = &sorted_runs[loop];
+    if (sr->being_compacted) {
+      char file_num_buf[kFormatFileNumberBufSize];
+      sr->Dump(file_num_buf, sizeof(file_num_buf), true);
+      ROCKS_LOG_BUFFER(
+          log_buffer, "[%s] Universal: Possible candidate %s[%d] %s",
+          cf_name.c_str(), file_num_buf, start_index,
+          " is already being compacted. No size amp reduction possible.\n");
+      return nullptr;
+    }
+    candidate_size += sr->compensated_file_size;
+    candidate_count++;
+  }
+  if (candidate_count == 0) {
+    return nullptr;
+  }
+
+  // size of earliest file
+  uint64_t earliest_file_size = sorted_runs.back().size;
+
+  // size amplification = percentage of additional size
+  if (candidate_size * 100 < ratio * earliest_file_size) {
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] Universal: size amp not needed. newer-files-total-size %" PRIu64
+        " earliest-file-size %" PRIu64,
+        cf_name.c_str(), candidate_size, earliest_file_size);
+    return nullptr;
+  } else {
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] Universal: size amp needed. newer-files-total-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++) {
+    estimated_total_size += sorted_runs[loop].size;
+  }
+  uint32_t path_id = GetPathId(ioptions_, estimated_total_size);
+  int start_level = sorted_runs[start_index].level;
+
+  std::vector<CompactionInputFiles> inputs(vstorage->num_levels());
+  for (size_t i = 0; i < inputs.size(); ++i) {
+    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++) {
+    auto& picking_sr = sorted_runs[loop];
+    if (picking_sr.level == 0) {
+      FileMetaData* f = picking_sr.file;
+      inputs[0].files.push_back(f);
+    } else {
+      auto& files = inputs[picking_sr.level - start_level].files;
+      for (auto* f : vstorage->LevelFiles(picking_sr.level)) {
+        files.push_back(f);
+      }
+    }
+    char file_num_buf[256];
+    picking_sr.DumpSizeInfo(file_num_buf, sizeof(file_num_buf), loop);
+    ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: size amp picking %s",
+                     cf_name.c_str(), file_num_buf);
+  }
+
+  // output files at the bottom most level, unless it's reserved
+  int output_level = vstorage->num_levels() - 1;
+  // last level is reserved for the files ingested behind
+  if (ioptions_.allow_ingest_behind) {
+    assert(output_level > 1);
+    output_level--;
+  }
+
+  return new Compaction(
+      vstorage, ioptions_, mutable_cf_options, std::move(inputs),
+      output_level, mutable_cf_options.MaxFileSizeForLevel(output_level),
+      /* max_grandparent_overlap_bytes */ LLONG_MAX, path_id,
+      GetCompressionType(ioptions_, vstorage, mutable_cf_options,
+                         output_level, 1),
+      /* grandparents */ {}, /* is manual */ false, score,
+      false /* deletion_compaction */,
+      CompactionReason::kUniversalSizeAmplification);
+}
+}  // namespace rocksdb
+
+#endif  // !ROCKSDB_LITE

db/compaction_picker_universal.h

@@ -0,0 +1,91 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root 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.
+
+#pragma once
+#ifndef ROCKSDB_LITE
+
+#include "db/compaction_picker.h"
+
+namespace rocksdb {
+class UniversalCompactionPicker : public CompactionPicker {
+ public:
+  UniversalCompactionPicker(const ImmutableCFOptions& ioptions,
+                            const InternalKeyComparator* icmp)
+      : CompactionPicker(ioptions, icmp) {}
+  virtual Compaction* PickCompaction(const std::string& cf_name,
+                                     const MutableCFOptions& mutable_cf_options,
+                                     VersionStorageInfo* vstorage,
+                                     LogBuffer* log_buffer) override;
+
+  virtual int MaxOutputLevel() const override { return NumberLevels() - 1; }
+
+  virtual bool NeedsCompaction(
+      const VersionStorageInfo* vstorage) const override;
+
+ private:
+  struct SortedRun {
+    SortedRun(int _level, FileMetaData* _file, uint64_t _size,
+              uint64_t _compensated_file_size, bool _being_compacted)
+        : level(_level),
+          file(_file),
+          size(_size),
+          compensated_file_size(_compensated_file_size),
+          being_compacted(_being_compacted) {
+      assert(compensated_file_size > 0);
+      assert(level != 0 || file != nullptr);
+    }
+
+    void Dump(char* out_buf, size_t out_buf_size,
+              bool print_path = false) const;
+
+    // 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 level;
+    // `file` Will be null for level > 0. For level = 0, the sorted run is
+    // for this file.
+    FileMetaData* file;
+    // For level > 0, `size` and `compensated_file_size` are sum of sizes all
+    // files in the level. `being_compacted` should be the same for all files
+    // in a non-zero level. Use the value here.
+    uint64_t size;
+    uint64_t compensated_file_size;
+    bool being_compacted;
+  };
+
+  // Pick Universal compaction to limit read amplification
+  Compaction* PickCompactionToReduceSortedRuns(
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      VersionStorageInfo* vstorage, double score, unsigned int ratio,
+      unsigned int num_files, const std::vector<SortedRun>& sorted_runs,
+      LogBuffer* log_buffer);
+
+  // Pick Universal compaction to limit space amplification.
+  Compaction* PickCompactionToReduceSizeAmp(
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      VersionStorageInfo* vstorage, double score,
+      const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer);
+
+  // Used in universal compaction when the enabled_trivial_move
+  // option is set. Checks whether there are any overlapping files
+  // in the input. Returns true if the input files are non
+  // overlapping.
+  bool IsInputFilesNonOverlapping(Compaction* c);
+
+  static std::vector<SortedRun> CalculateSortedRuns(
+      const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions);
+
+  // Pick a path ID to place a newly generated file, with its estimated file
+  // size.
+  static uint32_t GetPathId(const ImmutableCFOptions& ioptions,
+                            uint64_t file_size);
+};
+}  // namespace rocksdb
+#endif  // !ROCKSDB_LITE

db/convenience.cc

@@ -9,19 +9,50 @@
 #include "rocksdb/convenience.h"
 
 #include "db/db_impl.h"
+#include "util/cast_util.h"
 
 namespace rocksdb {
 
 void CancelAllBackgroundWork(DB* db, bool wait) {
-  (dynamic_cast<DBImpl*>(db))->CancelAllBackgroundWork(wait);
+  (static_cast_with_check<DBImpl, DB>(db->GetRootDB()))
+      ->CancelAllBackgroundWork(wait);
 }
 
 Status DeleteFilesInRange(DB* db, ColumnFamilyHandle* column_family,
                           const Slice* begin, const Slice* end) {
-  return (dynamic_cast<DBImpl*>(db))
+  return (static_cast_with_check<DBImpl, DB>(db->GetRootDB()))
       ->DeleteFilesInRange(column_family, begin, end);
 }
 
+Status VerifySstFileChecksum(const Options& options,
+                             const EnvOptions& env_options,
+                             const std::string& file_path) {
+  unique_ptr<RandomAccessFile> file;
+  uint64_t file_size;
+  InternalKeyComparator internal_comparator(options.comparator);
+  ImmutableCFOptions ioptions(options);
+
+  Status s = ioptions.env->NewRandomAccessFile(file_path, &file, env_options);
+  if (s.ok()) {
+    s = ioptions.env->GetFileSize(file_path, &file_size);
+  } else {
+    return s;
+  }
+  unique_ptr<TableReader> table_reader;
+  std::unique_ptr<RandomAccessFileReader> file_reader(
+      new RandomAccessFileReader(std::move(file), file_path));
+  s = ioptions.table_factory->NewTableReader(
+      TableReaderOptions(ioptions, env_options, internal_comparator,
+                         false /* skip_filters */, -1 /* level */),
+      std::move(file_reader), file_size, &table_reader,
+      false /* prefetch_index_and_filter_in_cache */);
+  if (!s.ok()) {
+    return s;
+  }
+  s = table_reader->VerifyChecksum();
+  return s;
+}
+
 }  // namespace rocksdb
 
 #endif  // ROCKSDB_LITE

db/corruption_test.cc

@@ -11,20 +11,21 @@
 
 #include "rocksdb/db.h"
 
-#include <inttypes.h>
 #include <errno.h>
 #include <fcntl.h>
+#include <inttypes.h>
 #include <sys/stat.h>
 #include <sys/types.h>
+#include "db/db_impl.h"
+#include "db/log_format.h"
+#include "db/version_set.h"
 #include "rocksdb/cache.h"
+#include "rocksdb/convenience.h"
 #include "rocksdb/env.h"
 #include "rocksdb/table.h"
 #include "rocksdb/write_batch.h"
-#include "db/db_impl.h"
-#include "db/filename.h"
-#include "db/log_format.h"
-#include "db/version_set.h"
-#include "util/logging.h"
+#include "util/filename.h"
+#include "util/string_util.h"
 #include "util/testharness.h"
 #include "util/testutil.h"
 
@@ -152,7 +153,7 @@ class CorruptionTest : public testing::Test {
     struct stat sbuf;
     if (stat(fname.c_str(), &sbuf) != 0) {
       const char* msg = strerror(errno);
-      ASSERT_TRUE(false) << fname << ": " << msg;
+      FAIL() << fname << ": " << msg;
     }
 
     if (offset < 0) {
@@ -179,6 +180,9 @@ class CorruptionTest : public testing::Test {
     }
     s = WriteStringToFile(Env::Default(), contents, fname);
     ASSERT_TRUE(s.ok()) << s.ToString();
+    Options options;
+    EnvOptions env_options;
+    ASSERT_NOK(VerifySstFileChecksum(options, env_options, fname));
   }
 
   void Corrupt(FileType filetype, int offset, int bytes_to_corrupt) {
@@ -213,7 +217,7 @@ class CorruptionTest : public testing::Test {
         return;
       }
     }
-    ASSERT_TRUE(false) << "no file found at level";
+    FAIL() << "no file found at level";
   }
 
 
@@ -312,6 +316,7 @@ TEST_F(CorruptionTest, TableFile) {
 
   Corrupt(kTableFile, 100, 1);
   Check(99, 99);
+  ASSERT_NOK(dbi->VerifyChecksum());
 }
 
 TEST_F(CorruptionTest, TableFileIndexData) {
@@ -327,9 +332,11 @@ TEST_F(CorruptionTest, TableFileIndexData) {
   // corrupt an index block of an entire file
   Corrupt(kTableFile, -2000, 500);
   Reopen();
+  dbi = reinterpret_cast<DBImpl*>(db_);
   // one full file should be readable, since only one was corrupted
   // the other file should be fully non-readable, since index was corrupted
   Check(5000, 5000);
+  ASSERT_NOK(dbi->VerifyChecksum());
 }
 
 TEST_F(CorruptionTest, MissingDescriptor) {
@@ -389,10 +396,12 @@ TEST_F(CorruptionTest, CompactionInputError) {
 
   Corrupt(kTableFile, 100, 1);
   Check(9, 9);
+  ASSERT_NOK(dbi->VerifyChecksum());
 
   // Force compactions by writing lots of values
   Build(10000);
   Check(10000, 10000);
+  ASSERT_NOK(dbi->VerifyChecksum());
 }
 
 TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
@@ -424,6 +433,7 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
 
   CorruptTableFileAtLevel(0, 100, 1);
   Check(9, 9);
+  ASSERT_NOK(dbi->VerifyChecksum());
 
   // Write must eventually fail because of corrupted table
   Status s;
@@ -445,6 +455,7 @@ TEST_F(CorruptionTest, UnrelatedKeys) {
   DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
   dbi->TEST_FlushMemTable();
   Corrupt(kTableFile, 100, 1);
+  ASSERT_NOK(dbi->VerifyChecksum());
 
   std::string tmp1, tmp2;
   ASSERT_OK(db_->Put(WriteOptions(), Key(1000, &tmp1), Value(1000, &tmp2)));

db/cuckoo_table_db_test.cc

@@ -8,9 +8,10 @@
 #include "db/db_impl.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
-#include "table/meta_blocks.h"
 #include "table/cuckoo_table_factory.h"
 #include "table/cuckoo_table_reader.h"
+#include "table/meta_blocks.h"
+#include "util/string_util.h"
 #include "util/testharness.h"
 #include "util/testutil.h"
 
@@ -319,8 +320,14 @@ TEST_F(CuckooTableDBTest, AdaptiveTable) {
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {
-  ::testing::InitGoogleTest(&argc, argv);
-  return RUN_ALL_TESTS();
+  if (rocksdb::port::kLittleEndian) {
+    ::testing::InitGoogleTest(&argc, argv);
+    return RUN_ALL_TESTS();
+  }
+  else {
+    fprintf(stderr, "SKIPPED as Cuckoo table doesn't support Big Endian\n");
+    return 0;
+  }
 }
 
 #else

db/db_basic_test.cc

@@ -175,6 +175,8 @@ TEST_F(DBBasicTest, LevelLimitReopen) {
   int i = 0;
   while (NumTableFilesAtLevel(2, 1) == 0) {
     ASSERT_OK(Put(1, Key(i++), value));
+    dbfull()->TEST_WaitForFlushMemTable();
+    dbfull()->TEST_WaitForCompact();
   }
 
   options.num_levels = 1;
@@ -358,15 +360,15 @@ TEST_F(DBBasicTest, FLUSH) {
     WriteOptions writeOpt = WriteOptions();
     writeOpt.disableWAL = true;
     SetPerfLevel(kEnableTime);
-    ;
     ASSERT_OK(dbfull()->Put(writeOpt, handles_[1], "foo", "v1"));
     // this will now also flush the last 2 writes
     ASSERT_OK(Flush(1));
     ASSERT_OK(dbfull()->Put(writeOpt, handles_[1], "bar", "v1"));
 
-    perf_context.Reset();
+    get_perf_context()->Reset();
     Get(1, "foo");
-    ASSERT_TRUE((int)perf_context.get_from_output_files_time > 0);
+    ASSERT_TRUE((int)get_perf_context()->get_from_output_files_time > 0);
+    ASSERT_EQ(2, (int)get_perf_context()->get_read_bytes);
 
     ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
     ASSERT_EQ("v1", Get(1, "foo"));
@@ -379,9 +381,9 @@ TEST_F(DBBasicTest, FLUSH) {
 
     ReopenWithColumnFamilies({"default", "pikachu"}, CurrentOptions());
     ASSERT_EQ("v2", Get(1, "bar"));
-    perf_context.Reset();
+    get_perf_context()->Reset();
     ASSERT_EQ("v2", Get(1, "foo"));
-    ASSERT_TRUE((int)perf_context.get_from_output_files_time > 0);
+    ASSERT_TRUE((int)get_perf_context()->get_from_output_files_time > 0);
 
     writeOpt.disableWAL = false;
     ASSERT_OK(dbfull()->Put(writeOpt, handles_[1], "bar", "v3"));
@@ -723,6 +725,7 @@ TEST_F(DBBasicTest, FlushOneColumnFamily) {
 TEST_F(DBBasicTest, MultiGetSimple) {
   do {
     CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
+    SetPerfLevel(kEnableCount);
     ASSERT_OK(Put(1, "k1", "v1"));
     ASSERT_OK(Put(1, "k2", "v2"));
     ASSERT_OK(Put(1, "k3", "v3"));
@@ -736,12 +739,15 @@ TEST_F(DBBasicTest, MultiGetSimple) {
     std::vector<std::string> values(20, "Temporary data to be overwritten");
     std::vector<ColumnFamilyHandle*> cfs(keys.size(), handles_[1]);
 
+    get_perf_context()->Reset();
     std::vector<Status> s = db_->MultiGet(ReadOptions(), cfs, keys, &values);
     ASSERT_EQ(values.size(), keys.size());
     ASSERT_EQ(values[0], "v1");
     ASSERT_EQ(values[1], "v2");
     ASSERT_EQ(values[2], "v3");
     ASSERT_EQ(values[4], "v5");
+    // four kv pairs * two bytes per value
+    ASSERT_EQ(8, (int)get_perf_context()->multiget_read_bytes);
 
     ASSERT_OK(s[0]);
     ASSERT_OK(s[1]);
@@ -749,6 +755,7 @@ TEST_F(DBBasicTest, MultiGetSimple) {
     ASSERT_TRUE(s[3].IsNotFound());
     ASSERT_OK(s[4]);
     ASSERT_TRUE(s[5].IsNotFound());
+    SetPerfLevel(kDisable);
   } while (ChangeCompactOptions());
 }
 
@@ -785,36 +792,30 @@ TEST_F(DBBasicTest, MultiGetEmpty) {
 TEST_F(DBBasicTest, ChecksumTest) {
   BlockBasedTableOptions table_options;
   Options options = CurrentOptions();
+  // change when new checksum type added
+  int max_checksum = static_cast<int>(kxxHash);
+  const int kNumPerFile = 2;
 
-  table_options.checksum = kCRC32c;
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-  Reopen(options);
-  ASSERT_OK(Put("a", "b"));
-  ASSERT_OK(Put("c", "d"));
-  ASSERT_OK(Flush());  // table with crc checksum
+  // generate one table with each type of checksum
+  for (int i = 0; i <= max_checksum; ++i) {
+    table_options.checksum = static_cast<ChecksumType>(i);
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+    Reopen(options);
+    for (int j = 0; j < kNumPerFile; ++j) {
+      ASSERT_OK(Put(Key(i * kNumPerFile + j), Key(i * kNumPerFile + j)));
+    }
+    ASSERT_OK(Flush());
+  }
 
-  table_options.checksum = kxxHash;
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-  Reopen(options);
-  ASSERT_OK(Put("e", "f"));
-  ASSERT_OK(Put("g", "h"));
-  ASSERT_OK(Flush());  // table with xxhash checksum
-
-  table_options.checksum = kCRC32c;
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-  Reopen(options);
-  ASSERT_EQ("b", Get("a"));
-  ASSERT_EQ("d", Get("c"));
-  ASSERT_EQ("f", Get("e"));
-  ASSERT_EQ("h", Get("g"));
-
-  table_options.checksum = kCRC32c;
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-  Reopen(options);
-  ASSERT_EQ("b", Get("a"));
-  ASSERT_EQ("d", Get("c"));
-  ASSERT_EQ("f", Get("e"));
-  ASSERT_EQ("h", Get("g"));
+  // verify data with each type of checksum
+  for (int i = 0; i <= kxxHash; ++i) {
+    table_options.checksum = static_cast<ChecksumType>(i);
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+    Reopen(options);
+    for (int j = 0; j < (max_checksum + 1) * kNumPerFile; ++j) {
+      ASSERT_EQ(Key(j), Get(Key(j)));
+    }
+  }
 }
 
 // On Windows you can have either memory mapped file or a file
@@ -822,6 +823,9 @@ TEST_F(DBBasicTest, ChecksumTest) {
 // sense to run
 #ifndef OS_WIN
 TEST_F(DBBasicTest, MmapAndBufferOptions) {
+  if (!IsMemoryMappedAccessSupported()) {
+    return;
+  }
   Options options = CurrentOptions();
 
   options.use_direct_reads = true;

db/db_block_cache_test.cc

@@ -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 <cstdlib>
+#include "cache/lru_cache.h"
 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/lru_cache.h"
 
 namespace rocksdb {
 
@@ -497,7 +497,7 @@ TEST_F(DBBlockCacheTest, CompressedCache) {
         options.compression = kNoCompression;
         break;
       default:
-        ASSERT_TRUE(false);
+        FAIL();
     }
     CreateAndReopenWithCF({"pikachu"}, options);
     // default column family doesn't have block cache
@@ -560,7 +560,7 @@ TEST_F(DBBlockCacheTest, CompressedCache) {
         ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT), 0);
         break;
       default:
-        ASSERT_TRUE(false);
+        FAIL();
     }
 
     options.create_if_missing = true;

db/db_bloom_filter_test.cc

@@ -20,17 +20,23 @@ class DBBloomFilterTest : public DBTestBase {
   DBBloomFilterTest() : DBTestBase("/db_bloom_filter_test") {}
 };
 
-class DBBloomFilterTestWithParam : public DBTestBase,
-                                   public testing::WithParamInterface<bool> {
+class DBBloomFilterTestWithParam
+    : public DBTestBase,
+      public testing::WithParamInterface<std::tuple<bool, bool>> {
+  //                             public testing::WithParamInterface<bool> {
  protected:
   bool use_block_based_filter_;
+  bool partition_filters_;
 
  public:
   DBBloomFilterTestWithParam() : DBTestBase("/db_bloom_filter_tests") {}
 
   ~DBBloomFilterTestWithParam() {}
 
-  void SetUp() override { use_block_based_filter_ = GetParam(); }
+  void SetUp() override {
+    use_block_based_filter_ = std::get<0>(GetParam());
+    partition_filters_ = std::get<1>(GetParam());
+  }
 };
 
 // KeyMayExist can lead to a few false positives, but not false negatives.
@@ -43,7 +49,17 @@ TEST_P(DBBloomFilterTestWithParam, KeyMayExist) {
     anon::OptionsOverride options_override;
     options_override.filter_policy.reset(
         NewBloomFilterPolicy(20, use_block_based_filter_));
+    options_override.partition_filters = partition_filters_;
+    options_override.metadata_block_size = 32;
     Options options = CurrentOptions(options_override);
+    if (partition_filters_ &&
+        static_cast<BlockBasedTableOptions*>(
+            options.table_factory->GetOptions())
+                ->index_type != BlockBasedTableOptions::kTwoLevelIndexSearch) {
+      // In the current implementation partitioned filters depend on partitioned
+      // indexes
+      continue;
+    }
     options.statistics = rocksdb::CreateDBStatistics();
     CreateAndReopenWithCF({"pikachu"}, options);
 
@@ -102,192 +118,204 @@ TEST_P(DBBloomFilterTestWithParam, KeyMayExist) {
 }
 
 TEST_F(DBBloomFilterTest, GetFilterByPrefixBloom) {
-  Options options = last_options_;
-  options.prefix_extractor.reset(NewFixedPrefixTransform(8));
-  options.statistics = rocksdb::CreateDBStatistics();
-  BlockBasedTableOptions bbto;
-  bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
-  bbto.whole_key_filtering = false;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  DestroyAndReopen(options);
+  for (bool partition_filters : {true, false}) {
+    Options options = last_options_;
+    options.prefix_extractor.reset(NewFixedPrefixTransform(8));
+    options.statistics = rocksdb::CreateDBStatistics();
+    BlockBasedTableOptions bbto;
+    bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
+    if (partition_filters) {
+      bbto.partition_filters = true;
+      bbto.index_type = BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+    }
+    bbto.whole_key_filtering = false;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    DestroyAndReopen(options);
 
-  WriteOptions wo;
-  ReadOptions ro;
-  FlushOptions fo;
-  fo.wait = true;
-  std::string value;
+    WriteOptions wo;
+    ReadOptions ro;
+    FlushOptions fo;
+    fo.wait = true;
+    std::string value;
 
-  ASSERT_OK(dbfull()->Put(wo, "barbarbar", "foo"));
-  ASSERT_OK(dbfull()->Put(wo, "barbarbar2", "foo2"));
-  ASSERT_OK(dbfull()->Put(wo, "foofoofoo", "bar"));
+    ASSERT_OK(dbfull()->Put(wo, "barbarbar", "foo"));
+    ASSERT_OK(dbfull()->Put(wo, "barbarbar2", "foo2"));
+    ASSERT_OK(dbfull()->Put(wo, "foofoofoo", "bar"));
 
-  dbfull()->Flush(fo);
+    dbfull()->Flush(fo);
 
-  ASSERT_EQ("foo", Get("barbarbar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-  ASSERT_EQ("foo2", Get("barbarbar2"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-  ASSERT_EQ("NOT_FOUND", Get("barbarbar3"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    ASSERT_EQ("foo", Get("barbarbar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    ASSERT_EQ("foo2", Get("barbarbar2"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    ASSERT_EQ("NOT_FOUND", Get("barbarbar3"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
 
-  ASSERT_EQ("NOT_FOUND", Get("barfoofoo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("barfoofoo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
 
-  ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
 
-  ro.total_order_seek = true;
-  ASSERT_TRUE(db_->Get(ro, "foobarbar", &value).IsNotFound());
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+    ro.total_order_seek = true;
+    ASSERT_TRUE(db_->Get(ro, "foobarbar", &value).IsNotFound());
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+  }
 }
 
 TEST_F(DBBloomFilterTest, WholeKeyFilterProp) {
-  Options options = last_options_;
-  options.prefix_extractor.reset(NewFixedPrefixTransform(3));
-  options.statistics = rocksdb::CreateDBStatistics();
+  for (bool partition_filters : {true, false}) {
+    Options options = last_options_;
+    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
+    options.statistics = rocksdb::CreateDBStatistics();
 
-  BlockBasedTableOptions bbto;
-  bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
-  bbto.whole_key_filtering = false;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  DestroyAndReopen(options);
+    BlockBasedTableOptions bbto;
+    bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
+    bbto.whole_key_filtering = false;
+    if (partition_filters) {
+      bbto.partition_filters = true;
+      bbto.index_type = BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+    }
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    DestroyAndReopen(options);
 
-  WriteOptions wo;
-  ReadOptions ro;
-  FlushOptions fo;
-  fo.wait = true;
-  std::string value;
+    WriteOptions wo;
+    ReadOptions ro;
+    FlushOptions fo;
+    fo.wait = true;
+    std::string value;
 
-  ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
-  // Needs insert some keys to make sure files are not filtered out by key
-  // ranges.
-  ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
-  ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
-  dbfull()->Flush(fo);
+    ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
+    // Needs insert some keys to make sure files are not filtered out by key
+    // ranges.
+    ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
+    ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
+    dbfull()->Flush(fo);
 
-  Reopen(options);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    Reopen(options);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
 
-  // Reopen with whole key filtering enabled and prefix extractor
-  // NULL. Bloom filter should be off for both of whole key and
-  // prefix bloom.
-  bbto.whole_key_filtering = true;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  options.prefix_extractor.reset();
-  Reopen(options);
+    // Reopen with whole key filtering enabled and prefix extractor
+    // NULL. Bloom filter should be off for both of whole key and
+    // prefix bloom.
+    bbto.whole_key_filtering = true;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    options.prefix_extractor.reset();
+    Reopen(options);
 
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  // Write DB with only full key filtering.
-  ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
-  // Needs insert some keys to make sure files are not filtered out by key
-  // ranges.
-  ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
-  ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
-  db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    // Write DB with only full key filtering.
+    ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
+    // Needs insert some keys to make sure files are not filtered out by key
+    // ranges.
+    ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
+    ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
+    db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
 
-  // Reopen with both of whole key off and prefix extractor enabled.
-  // Still no bloom filter should be used.
-  options.prefix_extractor.reset(NewFixedPrefixTransform(3));
-  bbto.whole_key_filtering = false;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  Reopen(options);
+    // Reopen with both of whole key off and prefix extractor enabled.
+    // Still no bloom filter should be used.
+    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
+    bbto.whole_key_filtering = false;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    Reopen(options);
 
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
 
-  // Try to create a DB with mixed files:
-  ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
-  // Needs insert some keys to make sure files are not filtered out by key
-  // ranges.
-  ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
-  ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
-  db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+    // Try to create a DB with mixed files:
+    ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
+    // Needs insert some keys to make sure files are not filtered out by key
+    // ranges.
+    ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
+    ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
+    db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
 
-  options.prefix_extractor.reset();
-  bbto.whole_key_filtering = true;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  Reopen(options);
+    options.prefix_extractor.reset();
+    bbto.whole_key_filtering = true;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    Reopen(options);
 
-  // Try to create a DB with mixed files.
-  ASSERT_OK(dbfull()->Put(wo, "barfoo", "bar"));
-  // In this case needs insert some keys to make sure files are
-  // not filtered out by key ranges.
-  ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
-  ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
-  Flush();
+    // Try to create a DB with mixed files.
+    ASSERT_OK(dbfull()->Put(wo, "barfoo", "bar"));
+    // In this case needs insert some keys to make sure files are
+    // not filtered out by key ranges.
+    ASSERT_OK(dbfull()->Put(wo, "aaa", ""));
+    ASSERT_OK(dbfull()->Put(wo, "zzz", ""));
+    Flush();
 
-  // Now we have two files:
-  // File 1: An older file with prefix bloom.
-  // File 2: A newer file with whole bloom filter.
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
-  ASSERT_EQ("bar", Get("barfoo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    // Now we have two files:
+    // File 1: An older file with prefix bloom.
+    // File 2: A newer file with whole bloom filter.
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    ASSERT_EQ("bar", Get("barfoo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
 
-  // Reopen with the same setting: only whole key is used
-  Reopen(options);
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 5);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 6);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
-  ASSERT_EQ("bar", Get("barfoo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    // Reopen with the same setting: only whole key is used
+    Reopen(options);
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 5);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 6);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    ASSERT_EQ("bar", Get("barfoo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
 
-  // Restart with both filters are allowed
-  options.prefix_extractor.reset(NewFixedPrefixTransform(3));
-  bbto.whole_key_filtering = true;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  Reopen(options);
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
-  // File 1 will has it filtered out.
-  // File 2 will not, as prefix `foo` exists in the file.
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 8);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 10);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
-  ASSERT_EQ("bar", Get("barfoo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    // Restart with both filters are allowed
+    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
+    bbto.whole_key_filtering = true;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    Reopen(options);
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    // File 1 will has it filtered out.
+    // File 2 will not, as prefix `foo` exists in the file.
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 8);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 10);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    ASSERT_EQ("bar", Get("barfoo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
 
-  // Restart with only prefix bloom is allowed.
-  options.prefix_extractor.reset(NewFixedPrefixTransform(3));
-  bbto.whole_key_filtering = false;
-  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
-  Reopen(options);
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
-  ASSERT_EQ("NOT_FOUND", Get("foo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
-  ASSERT_EQ("NOT_FOUND", Get("bar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
-  ASSERT_EQ("foo", Get("foobar"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
-  ASSERT_EQ("bar", Get("barfoo"));
-  ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    // Restart with only prefix bloom is allowed.
+    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
+    bbto.whole_key_filtering = false;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+    Reopen(options);
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    ASSERT_EQ("NOT_FOUND", Get("foo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    ASSERT_EQ("NOT_FOUND", Get("bar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    ASSERT_EQ("foo", Get("foobar"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    ASSERT_EQ("bar", Get("barfoo"));
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+  }
 }
 
 TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
@@ -301,6 +329,12 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
     table_options.no_block_cache = true;
     table_options.filter_policy.reset(
         NewBloomFilterPolicy(10, use_block_based_filter_));
+    table_options.partition_filters = partition_filters_;
+    if (partition_filters_) {
+      table_options.index_type =
+          BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+    }
+    table_options.metadata_block_size = 32;
     options.table_factory.reset(NewBlockBasedTableFactory(table_options));
 
     CreateAndReopenWithCF({"pikachu"}, options);
@@ -327,7 +361,13 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
     int reads = env_->random_read_counter_.Read();
     fprintf(stderr, "%d present => %d reads\n", N, reads);
     ASSERT_GE(reads, N);
-    ASSERT_LE(reads, N + 2 * N / 100);
+    if (partition_filters_) {
+      // Without block cache, we read an extra partition filter per each
+      // level*read and a partition index per each read
+      ASSERT_LE(reads, 4 * N + 2 * N / 100);
+    } else {
+      ASSERT_LE(reads, N + 2 * N / 100);
+    }
 
     // Lookup present keys.  Should rarely read from either sstable.
     env_->random_read_counter_.Reset();
@@ -336,7 +376,13 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
     }
     reads = env_->random_read_counter_.Read();
     fprintf(stderr, "%d missing => %d reads\n", N, reads);
-    ASSERT_LE(reads, 3 * N / 100);
+    if (partition_filters_) {
+      // With partitioned filter we read one extra filter per level per each
+      // missed read.
+      ASSERT_LE(reads, 2 * N + 3 * N / 100);
+    } else {
+      ASSERT_LE(reads, 3 * N / 100);
+    }
 
     env_->delay_sstable_sync_.store(false, std::memory_order_release);
     Close();
@@ -344,7 +390,9 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
 }
 
 INSTANTIATE_TEST_CASE_P(DBBloomFilterTestWithParam, DBBloomFilterTestWithParam,
-                        ::testing::Bool());
+                        ::testing::Values(std::make_tuple(true, false),
+                                          std::make_tuple(false, true),
+                                          std::make_tuple(false, false)));
 
 TEST_F(DBBloomFilterTest, BloomFilterRate) {
   while (ChangeFilterOptions()) {
@@ -401,28 +449,13 @@ TEST_F(DBBloomFilterTest, BloomFilterCompatibility) {
     ASSERT_EQ(Key(i), Get(1, Key(i)));
   }
   ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-}
 
-TEST_F(DBBloomFilterTest, BloomFilterReverseCompatibility) {
-  Options options = CurrentOptions();
-  options.statistics = rocksdb::CreateDBStatistics();
-  BlockBasedTableOptions table_options;
+  // Check db with partitioned full filter
+  table_options.partition_filters = true;
+  table_options.index_type =
+      BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
   table_options.filter_policy.reset(NewBloomFilterPolicy(10, false));
   options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-
-  // Create with full filter
-  CreateAndReopenWithCF({"pikachu"}, options);
-
-  const int maxKey = 10000;
-  for (int i = 0; i < maxKey; i++) {
-    ASSERT_OK(Put(1, Key(i), Key(i)));
-  }
-  ASSERT_OK(Put(1, Key(maxKey + 55555), Key(maxKey + 55555)));
-  Flush(1);
-
-  // Check db with block_based filter
-  table_options.filter_policy.reset(NewBloomFilterPolicy(10, true));
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   ReopenWithColumnFamilies({"default", "pikachu"}, options);
 
   // Check if they can be found
@@ -432,6 +465,43 @@ TEST_F(DBBloomFilterTest, BloomFilterReverseCompatibility) {
   ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
 }
 
+TEST_F(DBBloomFilterTest, BloomFilterReverseCompatibility) {
+  for (bool partition_filters : {true, false}) {
+    Options options = CurrentOptions();
+    options.statistics = rocksdb::CreateDBStatistics();
+    BlockBasedTableOptions table_options;
+    if (partition_filters) {
+      table_options.partition_filters = true;
+      table_options.index_type =
+          BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+    }
+    table_options.filter_policy.reset(NewBloomFilterPolicy(10, false));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+    DestroyAndReopen(options);
+
+    // Create with full filter
+    CreateAndReopenWithCF({"pikachu"}, options);
+
+    const int maxKey = 10000;
+    for (int i = 0; i < maxKey; i++) {
+      ASSERT_OK(Put(1, Key(i), Key(i)));
+    }
+    ASSERT_OK(Put(1, Key(maxKey + 55555), Key(maxKey + 55555)));
+    Flush(1);
+
+    // Check db with block_based filter
+    table_options.filter_policy.reset(NewBloomFilterPolicy(10, true));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+    ReopenWithColumnFamilies({"default", "pikachu"}, options);
+
+    // Check if they can be found
+    for (int i = 0; i < maxKey; i++) {
+      ASSERT_EQ(Key(i), Get(1, Key(i)));
+    }
+    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+  }
+}
+
 namespace {
 // A wrapped bloom over default FilterPolicy
 class WrappedBloom : public FilterPolicy {
@@ -586,11 +656,12 @@ TEST_F(DBBloomFilterTest, PrefixExtractorBlockFilter) {
 #ifndef ROCKSDB_LITE
 class BloomStatsTestWithParam
     : public DBBloomFilterTest,
-      public testing::WithParamInterface<std::tuple<bool, bool>> {
+      public testing::WithParamInterface<std::tuple<bool, bool, bool>> {
  public:
   BloomStatsTestWithParam() {
     use_block_table_ = std::get<0>(GetParam());
     use_block_based_builder_ = std::get<1>(GetParam());
+    partition_filters_ = std::get<2>(GetParam());
 
     options_.create_if_missing = true;
     options_.prefix_extractor.reset(rocksdb::NewFixedPrefixTransform(4));
@@ -599,20 +670,28 @@ class BloomStatsTestWithParam
     if (use_block_table_) {
       BlockBasedTableOptions table_options;
       table_options.hash_index_allow_collision = false;
+      if (partition_filters_) {
+        assert(!use_block_based_builder_);
+        table_options.partition_filters = partition_filters_;
+        table_options.index_type =
+            BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+      }
       table_options.filter_policy.reset(
           NewBloomFilterPolicy(10, use_block_based_builder_));
       options_.table_factory.reset(NewBlockBasedTableFactory(table_options));
     } else {
+      assert(!partition_filters_);  // not supported in plain table
       PlainTableOptions table_options;
       options_.table_factory.reset(NewPlainTableFactory(table_options));
     }
+    options_.env = env_;
 
-    perf_context.Reset();
+    get_perf_context()->Reset();
     DestroyAndReopen(options_);
   }
 
   ~BloomStatsTestWithParam() {
-    perf_context.Reset();
+    get_perf_context()->Reset();
     Destroy(options_);
   }
 
@@ -622,6 +701,7 @@ class BloomStatsTestWithParam
 
   bool use_block_table_;
   bool use_block_based_builder_;
+  bool partition_filters_;
   Options options_;
 };
 
@@ -644,33 +724,33 @@ TEST_P(BloomStatsTestWithParam, BloomStatsTest) {
 
   // check memtable bloom stats
   ASSERT_EQ(value1, Get(key1));
-  ASSERT_EQ(1, perf_context.bloom_memtable_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_hit_count);
   ASSERT_EQ(value3, Get(key3));
-  ASSERT_EQ(2, perf_context.bloom_memtable_hit_count);
-  ASSERT_EQ(0, perf_context.bloom_memtable_miss_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_memtable_hit_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_memtable_miss_count);
 
   ASSERT_EQ("NOT_FOUND", Get(key2));
-  ASSERT_EQ(1, perf_context.bloom_memtable_miss_count);
-  ASSERT_EQ(2, perf_context.bloom_memtable_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_miss_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_memtable_hit_count);
 
   // sanity checks
-  ASSERT_EQ(0, perf_context.bloom_sst_hit_count);
-  ASSERT_EQ(0, perf_context.bloom_sst_miss_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_sst_hit_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_sst_miss_count);
 
   Flush();
 
   // sanity checks
-  ASSERT_EQ(0, perf_context.bloom_sst_hit_count);
-  ASSERT_EQ(0, perf_context.bloom_sst_miss_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_sst_hit_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_sst_miss_count);
 
   // check SST bloom stats
   ASSERT_EQ(value1, Get(key1));
-  ASSERT_EQ(1, perf_context.bloom_sst_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_sst_hit_count);
   ASSERT_EQ(value3, Get(key3));
-  ASSERT_EQ(2, perf_context.bloom_sst_hit_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_sst_hit_count);
 
   ASSERT_EQ("NOT_FOUND", Get(key2));
-  ASSERT_EQ(1, perf_context.bloom_sst_miss_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_sst_miss_count);
 }
 
 // Same scenario as in BloomStatsTest but using an iterator
@@ -691,21 +771,21 @@ TEST_P(BloomStatsTestWithParam, BloomStatsTestWithIter) {
   ASSERT_OK(iter->status());
   ASSERT_TRUE(iter->Valid());
   ASSERT_EQ(value1, iter->value().ToString());
-  ASSERT_EQ(1, perf_context.bloom_memtable_hit_count);
-  ASSERT_EQ(0, perf_context.bloom_memtable_miss_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_hit_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_memtable_miss_count);
 
   iter->Seek(key3);
   ASSERT_OK(iter->status());
   ASSERT_TRUE(iter->Valid());
   ASSERT_EQ(value3, iter->value().ToString());
-  ASSERT_EQ(2, perf_context.bloom_memtable_hit_count);
-  ASSERT_EQ(0, perf_context.bloom_memtable_miss_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_memtable_hit_count);
+  ASSERT_EQ(0, get_perf_context()->bloom_memtable_miss_count);
 
   iter->Seek(key2);
   ASSERT_OK(iter->status());
   ASSERT_TRUE(!iter->Valid());
-  ASSERT_EQ(1, perf_context.bloom_memtable_miss_count);
-  ASSERT_EQ(2, perf_context.bloom_memtable_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_miss_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_memtable_hit_count);
 
   Flush();
 
@@ -716,25 +796,27 @@ TEST_P(BloomStatsTestWithParam, BloomStatsTestWithIter) {
   ASSERT_OK(iter->status());
   ASSERT_TRUE(iter->Valid());
   ASSERT_EQ(value1, iter->value().ToString());
-  ASSERT_EQ(1, perf_context.bloom_sst_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_sst_hit_count);
 
   iter->Seek(key3);
   ASSERT_OK(iter->status());
   ASSERT_TRUE(iter->Valid());
   ASSERT_EQ(value3, iter->value().ToString());
-  ASSERT_EQ(2, perf_context.bloom_sst_hit_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_sst_hit_count);
 
   iter->Seek(key2);
   ASSERT_OK(iter->status());
   ASSERT_TRUE(!iter->Valid());
-  ASSERT_EQ(1, perf_context.bloom_sst_miss_count);
-  ASSERT_EQ(2, perf_context.bloom_sst_hit_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_sst_miss_count);
+  ASSERT_EQ(2, get_perf_context()->bloom_sst_hit_count);
 }
 
 INSTANTIATE_TEST_CASE_P(BloomStatsTestWithParam, BloomStatsTestWithParam,
-                        ::testing::Values(std::make_tuple(true, true),
-                                          std::make_tuple(true, false),
-                                          std::make_tuple(false, false)));
+                        ::testing::Values(std::make_tuple(true, true, false),
+                                          std::make_tuple(true, false, false),
+                                          std::make_tuple(true, false, true),
+                                          std::make_tuple(false, false,
+                                                          false)));
 
 namespace {
 void PrefixScanInit(DBBloomFilterTest* dbtest) {

db/db_compaction_filter_test.cc

@@ -626,7 +626,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
       iter->Next();
     }
     ASSERT_EQ(total, 700);
-    ASSERT_EQ(count, 2);
+    ASSERT_EQ(count, 1);
   }
 }
 #endif  // ROCKSDB_LITE
@@ -778,7 +778,7 @@ TEST_F(DBTestCompactionFilter, SkipUntil) {
 
   cfilter_skips = 0;
   ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-  // Numberof skips in tables: 2, 3, 3, 3.
+  // Number of skips in tables: 2, 3, 3, 3.
   ASSERT_EQ(11, cfilter_skips);
 
   for (int table = 0; table < 4; ++table) {
@@ -799,6 +799,43 @@ TEST_F(DBTestCompactionFilter, SkipUntil) {
   }
 }
 
+TEST_F(DBTestCompactionFilter, SkipUntilWithBloomFilter) {
+  BlockBasedTableOptions table_options;
+  table_options.whole_key_filtering = false;
+  table_options.filter_policy.reset(NewBloomFilterPolicy(100, false));
+
+  Options options = CurrentOptions();
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  options.prefix_extractor.reset(NewCappedPrefixTransform(9));
+  options.compaction_filter_factory = std::make_shared<SkipEvenFilterFactory>();
+  options.disable_auto_compactions = true;
+  options.create_if_missing = true;
+  DestroyAndReopen(options);
+
+  Put("0000000010", "v10");
+  Put("0000000020", "v20"); // skipped
+  Put("0000000050", "v50");
+  Flush();
+
+  cfilter_skips = 0;
+  EXPECT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
+  EXPECT_EQ(1, cfilter_skips);
+
+  Status s;
+  std::string val;
+
+  s = db_->Get(ReadOptions(), "0000000010", &val);
+  ASSERT_OK(s);
+  EXPECT_EQ("v10", val);
+
+  s = db_->Get(ReadOptions(), "0000000020", &val);
+  EXPECT_TRUE(s.IsNotFound());
+
+  s = db_->Get(ReadOptions(), "0000000050", &val);
+  ASSERT_OK(s);
+  EXPECT_EQ("v50", val);
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/db_compaction_test.cc

@@ -40,6 +40,12 @@ class DBCompactionTestWithParam
   bool exclusive_manual_compaction_;
 };
 
+class DBCompactionDirectIOTest : public DBCompactionTest,
+                                 public ::testing::WithParamInterface<bool> {
+ public:
+  DBCompactionDirectIOTest() : DBCompactionTest() {}
+};
+
 namespace {
 
 class FlushedFileCollector : public EventListener {
@@ -513,7 +519,6 @@ TEST_F(DBCompactionTest, BGCompactionsAllowed) {
   options.level0_file_num_compaction_trigger = 2;
   options.level0_slowdown_writes_trigger = 20;
   options.soft_pending_compaction_bytes_limit = 1 << 30;  // Infinitely large
-  options.base_background_compactions = 1;
   options.max_background_compactions = 3;
   options.memtable_factory.reset(new SpecialSkipListFactory(kNumKeysPerFile));
 
@@ -772,6 +777,41 @@ TEST_F(DBCompactionTest, ZeroSeqIdCompaction) {
   ASSERT_OK(Put("", ""));
 }
 
+TEST_F(DBCompactionTest, ManualCompactionUnknownOutputSize) {
+  // github issue #2249
+  Options options = CurrentOptions();
+  options.compaction_style = kCompactionStyleLevel;
+  options.level0_file_num_compaction_trigger = 3;
+  DestroyAndReopen(options);
+
+  // create two files in l1 that we can compact
+  for (int i = 0; i < 2; ++i) {
+    for (int j = 0; j < options.level0_file_num_compaction_trigger; j++) {
+      // make l0 files' ranges overlap to avoid trivial move
+      Put(std::to_string(2 * i), std::string(1, 'A'));
+      Put(std::to_string(2 * i + 1), std::string(1, 'A'));
+      Flush();
+      dbfull()->TEST_WaitForFlushMemTable();
+    }
+    dbfull()->TEST_WaitForCompact();
+    ASSERT_EQ(NumTableFilesAtLevel(0, 0), 0);
+    ASSERT_EQ(NumTableFilesAtLevel(1, 0), i + 1);
+  }
+
+  ColumnFamilyMetaData cf_meta;
+  dbfull()->GetColumnFamilyMetaData(dbfull()->DefaultColumnFamily(), &cf_meta);
+  ASSERT_EQ(2, cf_meta.levels[1].files.size());
+  std::vector<std::string> input_filenames;
+  for (const auto& sst_file : cf_meta.levels[1].files) {
+    input_filenames.push_back(sst_file.name);
+  }
+
+  // note CompactionOptions::output_file_size_limit is unset.
+  CompactionOptions compact_opt;
+  compact_opt.compression = kNoCompression;
+  dbfull()->CompactFiles(compact_opt, input_filenames, 1);
+}
+
 // Check that writes done during a memtable compaction are recovered
 // if the database is shutdown during the memtable compaction.
 TEST_F(DBCompactionTest, RecoverDuringMemtableCompaction) {
@@ -1152,7 +1192,8 @@ TEST_P(DBCompactionTestWithParam, ManualCompactionPartial) {
   }
 }
 
-TEST_F(DBCompactionTest, ManualPartialFill) {
+// Disable as the test is flaky.
+TEST_F(DBCompactionTest, DISABLED_ManualPartialFill) {
   int32_t trivial_move = 0;
   int32_t non_trivial_move = 0;
   rocksdb::SyncPoint::GetInstance()->SetCallBack(
@@ -1186,6 +1227,9 @@ TEST_F(DBCompactionTest, ManualPartialFill) {
   options.max_background_compactions = 3;
 
   DestroyAndReopen(options);
+  // make sure all background compaction jobs can be scheduled
+  auto stop_token =
+      dbfull()->TEST_write_controler().GetCompactionPressureToken();
   int32_t value_size = 10 * 1024;  // 10 KB
 
   // Add 2 non-overlapping files
@@ -1843,6 +1887,47 @@ TEST_F(DBCompactionTest, L0_CompactionBug_Issue44_b) {
   } while (ChangeCompactOptions());
 }
 
+TEST_F(DBCompactionTest, ManualAutoRace) {
+  CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
+  rocksdb::SyncPoint::GetInstance()->LoadDependency(
+      {{"DBImpl::BGWorkCompaction", "DBCompactionTest::ManualAutoRace:1"},
+       {"DBImpl::RunManualCompaction:WaitScheduled",
+        "BackgroundCallCompaction:0"}});
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  Put(1, "foo", "");
+  Put(1, "bar", "");
+  Flush(1);
+  Put(1, "foo", "");
+  Put(1, "bar", "");
+  // Generate four files in CF 0, which should trigger an auto compaction
+  Put("foo", "");
+  Put("bar", "");
+  Flush();
+  Put("foo", "");
+  Put("bar", "");
+  Flush();
+  Put("foo", "");
+  Put("bar", "");
+  Flush();
+  Put("foo", "");
+  Put("bar", "");
+  Flush();
+
+  // The auto compaction is scheduled but waited until here
+  TEST_SYNC_POINT("DBCompactionTest::ManualAutoRace:1");
+  // The auto compaction will wait until the manual compaction is registerd
+  // before processing so that it will be cancelled.
+  dbfull()->CompactRange(CompactRangeOptions(), handles_[1], nullptr, nullptr);
+  ASSERT_EQ("0,1", FilesPerLevel(1));
+
+  // Eventually the cancelled compaction will be rescheduled and executed.
+  dbfull()->TEST_WaitForCompact();
+  ASSERT_EQ("0,1", FilesPerLevel(0));
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+
 TEST_P(DBCompactionTestWithParam, ManualCompaction) {
   Options options = CurrentOptions();
   options.max_subcompactions = max_subcompactions_;
@@ -1892,7 +1977,6 @@ TEST_P(DBCompactionTestWithParam, ManualCompaction) {
 
     if (iter == 0) {
       options = CurrentOptions();
-      options.max_background_flushes = 0;
       options.num_levels = 3;
       options.create_if_missing = true;
       options.statistics = rocksdb::CreateDBStatistics();
@@ -2338,15 +2422,12 @@ TEST_F(DBCompactionTest, SanitizeCompactionOptionsTest) {
   options.hard_pending_compaction_bytes_limit = 100;
   options.create_if_missing = true;
   DestroyAndReopen(options);
-  ASSERT_EQ(5, db_->GetOptions().base_background_compactions);
   ASSERT_EQ(100, db_->GetOptions().soft_pending_compaction_bytes_limit);
 
-  options.base_background_compactions = 4;
   options.max_background_compactions = 3;
   options.soft_pending_compaction_bytes_limit = 200;
   options.hard_pending_compaction_bytes_limit = 150;
   DestroyAndReopen(options);
-  ASSERT_EQ(3, db_->GetOptions().base_background_compactions);
   ASSERT_EQ(150, db_->GetOptions().soft_pending_compaction_bytes_limit);
 }
 
@@ -2492,12 +2573,204 @@ TEST_P(DBCompactionTestWithParam, ForceBottommostLevelCompaction) {
   rocksdb::SyncPoint::GetInstance()->DisableProcessing();
 }
 
+TEST_P(DBCompactionTestWithParam, IntraL0Compaction) {
+  Options options = CurrentOptions();
+  options.compression = kNoCompression;
+  options.level0_file_num_compaction_trigger = 5;
+  options.max_background_compactions = 2;
+  options.max_subcompactions = max_subcompactions_;
+  DestroyAndReopen(options);
+
+  const size_t kValueSize = 1 << 20;
+  Random rnd(301);
+  std::string value(RandomString(&rnd, kValueSize));
+
+  rocksdb::SyncPoint::GetInstance()->LoadDependency(
+      {{"LevelCompactionPicker::PickCompactionBySize:0",
+        "CompactionJob::Run():Start"}});
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  // index:   0   1   2   3   4   5   6   7   8   9
+  // size:  1MB 1MB 1MB 1MB 1MB 2MB 1MB 1MB 1MB 1MB
+  // score:                     1.5 1.3 1.5 2.0 inf
+  //
+  // Files 0-4 will be included in an L0->L1 compaction.
+  //
+  // L0->L0 will be triggered since the sync points guarantee compaction to base
+  // level is still blocked when files 5-9 trigger another compaction.
+  //
+  // Files 6-9 are the longest span of available files for which
+  // work-per-deleted-file decreases (see "score" row above).
+  for (int i = 0; i < 10; ++i) {
+    ASSERT_OK(Put(Key(0), ""));  // prevents trivial move
+    if (i == 5) {
+      ASSERT_OK(Put(Key(i + 1), value + value));
+    } else {
+      ASSERT_OK(Put(Key(i + 1), value));
+    }
+    ASSERT_OK(Flush());
+  }
+  dbfull()->TEST_WaitForCompact();
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  std::vector<std::vector<FileMetaData>> level_to_files;
+  dbfull()->TEST_GetFilesMetaData(dbfull()->DefaultColumnFamily(),
+                                  &level_to_files);
+  ASSERT_GE(level_to_files.size(), 2);  // at least L0 and L1
+  // L0 has the 2MB file (not compacted) and 4MB file (output of L0->L0)
+  ASSERT_EQ(2, level_to_files[0].size());
+  ASSERT_GT(level_to_files[1].size(), 0);
+  for (int i = 0; i < 2; ++i) {
+    ASSERT_GE(level_to_files[0][i].fd.file_size, 1 << 21);
+  }
+}
+
+TEST_P(DBCompactionTestWithParam, IntraL0CompactionDoesNotObsoleteDeletions) {
+  // regression test for issue #2722: L0->L0 compaction can resurrect deleted
+  // keys from older L0 files if L1+ files' key-ranges do not include the key.
+  Options options = CurrentOptions();
+  options.compression = kNoCompression;
+  options.level0_file_num_compaction_trigger = 5;
+  options.max_background_compactions = 2;
+  options.max_subcompactions = max_subcompactions_;
+  DestroyAndReopen(options);
+
+  const size_t kValueSize = 1 << 20;
+  Random rnd(301);
+  std::string value(RandomString(&rnd, kValueSize));
+
+  rocksdb::SyncPoint::GetInstance()->LoadDependency(
+      {{"LevelCompactionPicker::PickCompactionBySize:0",
+        "CompactionJob::Run():Start"}});
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  // index:   0   1   2   3   4    5    6   7   8   9
+  // size:  1MB 1MB 1MB 1MB 1MB  1MB  1MB 1MB 1MB 1MB
+  // score:                     1.25 1.33 1.5 2.0 inf
+  //
+  // Files 0-4 will be included in an L0->L1 compaction.
+  //
+  // L0->L0 will be triggered since the sync points guarantee compaction to base
+  // level is still blocked when files 5-9 trigger another compaction. All files
+  // 5-9 are included in the L0->L0 due to work-per-deleted file decreasing.
+  //
+  // Put a key-value in files 0-4. Delete that key in files 5-9. Verify the
+  // L0->L0 preserves the deletion such that the key remains deleted.
+  for (int i = 0; i < 10; ++i) {
+    // key 0 serves both to prevent trivial move and as the key we want to
+    // verify is not resurrected by L0->L0 compaction.
+    if (i < 5) {
+      ASSERT_OK(Put(Key(0), ""));
+    } else {
+      ASSERT_OK(Delete(Key(0)));
+    }
+    ASSERT_OK(Put(Key(i + 1), value));
+    ASSERT_OK(Flush());
+  }
+  dbfull()->TEST_WaitForCompact();
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  std::vector<std::vector<FileMetaData>> level_to_files;
+  dbfull()->TEST_GetFilesMetaData(dbfull()->DefaultColumnFamily(),
+                                  &level_to_files);
+  ASSERT_GE(level_to_files.size(), 2);  // at least L0 and L1
+  // L0 has a single output file from L0->L0
+  ASSERT_EQ(1, level_to_files[0].size());
+  ASSERT_GT(level_to_files[1].size(), 0);
+  ASSERT_GE(level_to_files[0][0].fd.file_size, 1 << 22);
+
+  ReadOptions roptions;
+  std::string result;
+  ASSERT_TRUE(db_->Get(roptions, Key(0), &result).IsNotFound());
+}
+
+TEST_F(DBCompactionTest, OptimizedDeletionObsoleting) {
+  // Deletions can be dropped when compacted to non-last level if they fall
+  // outside the lower-level files' key-ranges.
+  const int kNumL0Files = 4;
+  Options options = CurrentOptions();
+  options.level0_file_num_compaction_trigger = kNumL0Files;
+  options.statistics = rocksdb::CreateDBStatistics();
+  DestroyAndReopen(options);
+
+  // put key 1 and 3 in separate L1, L2 files.
+  // So key 0, 2, and 4+ fall outside these levels' key-ranges.
+  for (int level = 2; level >= 1; --level) {
+    for (int i = 0; i < 2; ++i) {
+      Put(Key(2 * i + 1), "val");
+      Flush();
+    }
+    MoveFilesToLevel(level);
+    ASSERT_EQ(2, NumTableFilesAtLevel(level));
+  }
+
+  // Delete keys in range [1, 4]. These L0 files will be compacted with L1:
+  // - Tombstones for keys 2 and 4 can be dropped early.
+  // - Tombstones for keys 1 and 3 must be kept due to L2 files' key-ranges.
+  for (int i = 0; i < kNumL0Files; ++i) {
+    Put(Key(0), "val");  // sentinel to prevent trivial move
+    Delete(Key(i + 1));
+    Flush();
+  }
+  dbfull()->TEST_WaitForCompact();
+
+  for (int i = 0; i < kNumL0Files; ++i) {
+    std::string value;
+    ASSERT_TRUE(db_->Get(ReadOptions(), Key(i + 1), &value).IsNotFound());
+  }
+  ASSERT_EQ(2, options.statistics->getTickerCount(
+                   COMPACTION_OPTIMIZED_DEL_DROP_OBSOLETE));
+  ASSERT_EQ(2,
+            options.statistics->getTickerCount(COMPACTION_KEY_DROP_OBSOLETE));
+}
+
 INSTANTIATE_TEST_CASE_P(DBCompactionTestWithParam, DBCompactionTestWithParam,
                         ::testing::Values(std::make_tuple(1, true),
                                           std::make_tuple(1, false),
                                           std::make_tuple(4, true),
                                           std::make_tuple(4, false)));
 
+TEST_P(DBCompactionDirectIOTest, DirectIO) {
+  Options options = CurrentOptions();
+  Destroy(options);
+  options.create_if_missing = true;
+  options.disable_auto_compactions = true;
+  options.use_direct_io_for_flush_and_compaction = GetParam();
+  options.env = new MockEnv(Env::Default());
+  Reopen(options);
+  bool readahead = false;
+  SyncPoint::GetInstance()->SetCallBack(
+      "TableCache::NewIterator:for_compaction", [&](void* arg) {
+        bool* use_direct_reads = static_cast<bool*>(arg);
+        ASSERT_EQ(*use_direct_reads,
+                  options.use_direct_io_for_flush_and_compaction);
+      });
+  SyncPoint::GetInstance()->SetCallBack(
+      "CompactionJob::OpenCompactionOutputFile", [&](void* arg) {
+        bool* use_direct_writes = static_cast<bool*>(arg);
+        ASSERT_EQ(*use_direct_writes,
+                  options.use_direct_io_for_flush_and_compaction);
+      });
+  if (options.use_direct_io_for_flush_and_compaction) {
+    SyncPoint::GetInstance()->SetCallBack(
+        "SanitizeOptions:direct_io", [&](void* arg) {
+          readahead = true;
+        });
+  }
+  SyncPoint::GetInstance()->EnableProcessing();
+  CreateAndReopenWithCF({"pikachu"}, options);
+  MakeTables(3, "p", "q", 1);
+  ASSERT_EQ("1,1,1", FilesPerLevel(1));
+  Compact(1, "p1", "p9");
+  ASSERT_FALSE(readahead ^ options.use_direct_io_for_flush_and_compaction);
+  ASSERT_EQ("0,0,1", FilesPerLevel(1));
+  Destroy(options);
+  delete options.env;
+}
+
+INSTANTIATE_TEST_CASE_P(DBCompactionDirectIOTest, DBCompactionDirectIOTest,
+                        testing::Bool());
+
 class CompactionPriTest : public DBTestBase,
                           public testing::WithParamInterface<uint32_t> {
  public:

db/db_dynamic_level_test.cc

@@ -57,7 +57,6 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase) {
       Options options;
       options.env = env.get();
       options.create_if_missing = true;
-      options.db_write_buffer_size = 2048;
       options.write_buffer_size = 2048;
       options.max_write_buffer_number = 2;
       options.level0_file_num_compaction_trigger = 2;
@@ -75,6 +74,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase) {
       options.compression_per_level[0] = kNoCompression;
       options.compression_per_level[1] = kLZ4Compression;
       options.compression_per_level[2] = kSnappyCompression;
+      options.env = env_;
 
       DestroyAndReopen(options);
 
@@ -126,7 +126,6 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
 
   Options options = CurrentOptions();
   options.create_if_missing = true;
-  options.db_write_buffer_size = 204800;
   options.write_buffer_size = 20480;
   options.max_write_buffer_number = 2;
   options.level0_file_num_compaction_trigger = 2;
@@ -283,7 +282,6 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesCompactRange) {
 
   Options options = CurrentOptions();
   options.create_if_missing = true;
-  options.db_write_buffer_size = 2048;
   options.write_buffer_size = 2048;
   options.max_write_buffer_number = 2;
   options.level0_file_num_compaction_trigger = 2;
@@ -361,7 +359,6 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesCompactRange) {
 TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBaseInc) {
   Options options = CurrentOptions();
   options.create_if_missing = true;
-  options.db_write_buffer_size = 2048;
   options.write_buffer_size = 2048;
   options.max_write_buffer_number = 2;
   options.level0_file_num_compaction_trigger = 2;
@@ -374,6 +371,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBaseInc) {
   options.soft_rate_limit = 1.1;
   options.max_background_compactions = 2;
   options.num_levels = 5;
+  options.max_compaction_bytes = 100000000;
 
   DestroyAndReopen(options);
 
@@ -413,7 +411,6 @@ TEST_F(DBTestDynamicLevel, DISABLED_MigrateToDynamicLevelMaxBytesBase) {
 
   Options options;
   options.create_if_missing = true;
-  options.db_write_buffer_size = 2048;
   options.write_buffer_size = 2048;
   options.max_write_buffer_number = 8;
   options.level0_file_num_compaction_trigger = 4;

db/db_encryption_test.cc

@@ -0,0 +1,94 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+#include "db/db_test_util.h"
+#include "port/stack_trace.h"
+#include "rocksdb/perf_context.h"
+#if !defined(ROCKSDB_LITE)
+#include "util/sync_point.h"
+#endif
+#include <iostream>
+#include <string>
+
+namespace rocksdb {
+
+class DBEncryptionTest : public DBTestBase {
+ public:
+  DBEncryptionTest() : DBTestBase("/db_encryption_test") {}
+};
+
+#ifndef ROCKSDB_LITE
+
+TEST_F(DBEncryptionTest, CheckEncrypted) {
+  ASSERT_OK(Put("foo567", "v1.fetdq"));
+  ASSERT_OK(Put("bar123", "v2.dfgkjdfghsd"));
+  Close();
+
+  // Open all files and look for the values we've put in there.
+  // They should not be found if encrypted, otherwise
+  // they should be found.
+  std::vector<std::string> fileNames;
+  auto status = env_->GetChildren(dbname_, &fileNames);
+  ASSERT_OK(status);
+
+  auto defaultEnv = Env::Default();
+  int hits = 0;
+  for (auto it = fileNames.begin() ; it != fileNames.end(); ++it) {
+    if ((*it == "..") || (*it == ".")) {
+      continue;
+    }
+    auto filePath = dbname_ + "/" + *it;
+    unique_ptr<SequentialFile> seqFile;
+    auto envOptions = EnvOptions(CurrentOptions());
+    status = defaultEnv->NewSequentialFile(filePath, &seqFile, envOptions);
+    ASSERT_OK(status);
+
+    uint64_t fileSize;
+    status = defaultEnv->GetFileSize(filePath, &fileSize);
+    ASSERT_OK(status);
+
+    std::string scratch;
+    scratch.reserve(fileSize);
+    Slice data;
+    status = seqFile->Read(fileSize, &data, (char*)scratch.data());
+    ASSERT_OK(status);
+
+    if (data.ToString().find("foo567") != std::string::npos) {
+      hits++; 
+      //std::cout << "Hit in " << filePath << "\n";
+    }
+    if (data.ToString().find("v1.fetdq") != std::string::npos) {
+      hits++; 
+      //std::cout << "Hit in " << filePath << "\n";
+    }
+    if (data.ToString().find("bar123") != std::string::npos) {
+      hits++; 
+      //std::cout << "Hit in " << filePath << "\n";
+    }
+    if (data.ToString().find("v2.dfgkjdfghsd") != std::string::npos) {
+      hits++; 
+      //std::cout << "Hit in " << filePath << "\n";
+    }
+    if (data.ToString().find("dfgk") != std::string::npos) {
+      hits++; 
+      //std::cout << "Hit in " << filePath << "\n";
+    }
+  }
+  if (encrypted_env_) {
+    ASSERT_EQ(hits, 0);
+  } else {
+    ASSERT_GE(hits, 4);
+  }
+}
+
+#endif // ROCKSDB_LITE
+
+}  // namespace rocksdb
+
+int main(int argc, char** argv) {
+  rocksdb::port::InstallStackTraceHandler();
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

db/db_filesnapshot.cc

@@ -15,13 +15,13 @@
 #include <algorithm>
 #include <string>
 #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 "util/filename.h"
 #include "util/mutexlock.h"
 #include "util/sync_point.h"
 
@@ -31,12 +31,11 @@ Status DBImpl::DisableFileDeletions() {
   InstrumentedMutexLock l(&mutex_);
   ++disable_delete_obsolete_files_;
   if (disable_delete_obsolete_files_ == 1) {
-    Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-        "File Deletions Disabled");
+    ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Disabled");
   } else {
-    Log(InfoLogLevel::WARN_LEVEL, immutable_db_options_.info_log,
-        "File Deletions Disabled, but already disabled. Counter: %d",
-        disable_delete_obsolete_files_);
+    ROCKS_LOG_WARN(immutable_db_options_.info_log,
+                   "File Deletions Disabled, but already disabled. Counter: %d",
+                   disable_delete_obsolete_files_);
   }
   return Status::OK();
 }
@@ -55,12 +54,12 @@ Status DBImpl::EnableFileDeletions(bool force) {
       --disable_delete_obsolete_files_;
     }
     if (disable_delete_obsolete_files_ == 0)  {
-      Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-          "File Deletions Enabled");
+      ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Enabled");
       should_purge_files = true;
       FindObsoleteFiles(&job_context, true);
     } else {
-      Log(InfoLogLevel::WARN_LEVEL, immutable_db_options_.info_log,
+      ROCKS_LOG_WARN(
+          immutable_db_options_.info_log,
           "File Deletions Enable, but not really enabled. Counter: %d",
           disable_delete_obsolete_files_);
     }
@@ -106,8 +105,8 @@ Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
 
     if (!status.ok()) {
       mutex_.Unlock();
-      Log(InfoLogLevel::ERROR_LEVEL, immutable_db_options_.info_log,
-          "Cannot Flush data %s\n", status.ToString().c_str());
+      ROCKS_LOG_ERROR(immutable_db_options_.info_log, "Cannot Flush data %s\n",
+                      status.ToString().c_str());
       return status;
     }
   }

db/db_flush_test.cc

@@ -19,12 +19,19 @@ class DBFlushTest : public DBTestBase {
   DBFlushTest() : DBTestBase("/db_flush_test") {}
 };
 
+class DBFlushDirectIOTest : public DBFlushTest,
+                            public ::testing::WithParamInterface<bool> {
+ public:
+  DBFlushDirectIOTest() : DBFlushTest() {}
+};
+
 // 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;
+  options.env = env_;
   Reopen(options);
   FlushOptions no_wait;
   no_wait.wait = false;
@@ -50,7 +57,7 @@ TEST_F(DBFlushTest, FlushWhileWritingManifest) {
 
 TEST_F(DBFlushTest, SyncFail) {
   std::unique_ptr<FaultInjectionTestEnv> fault_injection_env(
-      new FaultInjectionTestEnv(Env::Default()));
+      new FaultInjectionTestEnv(env_));
   Options options;
   options.disable_auto_compactions = true;
   options.env = fault_injection_env.get();
@@ -82,6 +89,70 @@ TEST_F(DBFlushTest, SyncFail) {
   Destroy(options);
 }
 
+TEST_F(DBFlushTest, FlushInLowPriThreadPool) {
+  // Verify setting an empty high-pri (flush) thread pool causes flushes to be
+  // scheduled in the low-pri (compaction) thread pool.
+  Options options = CurrentOptions();
+  options.level0_file_num_compaction_trigger = 4;
+  options.memtable_factory.reset(new SpecialSkipListFactory(1));
+  Reopen(options);
+  env_->SetBackgroundThreads(0, Env::HIGH);
+
+  std::thread::id tid;
+  int num_flushes = 0, num_compactions = 0;
+  SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::BGWorkFlush", [&](void* arg) {
+        if (tid == std::thread::id()) {
+          tid = std::this_thread::get_id();
+        } else {
+          ASSERT_EQ(tid, std::this_thread::get_id());
+        }
+        ++num_flushes;
+      });
+  SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::BGWorkCompaction", [&](void* arg) {
+        ASSERT_EQ(tid, std::this_thread::get_id());
+        ++num_compactions;
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(Put("key", "val"));
+  for (int i = 0; i < 4; ++i) {
+    ASSERT_OK(Put("key", "val"));
+    dbfull()->TEST_WaitForFlushMemTable();
+  }
+  dbfull()->TEST_WaitForCompact();
+  ASSERT_EQ(4, num_flushes);
+  ASSERT_EQ(1, num_compactions);
+}
+
+TEST_P(DBFlushDirectIOTest, DirectIO) {
+  Options options;
+  options.create_if_missing = true;
+  options.disable_auto_compactions = true;
+  options.max_background_flushes = 2;
+  options.use_direct_io_for_flush_and_compaction = GetParam();
+  options.env = new MockEnv(Env::Default());
+  SyncPoint::GetInstance()->SetCallBack(
+      "BuildTable:create_file", [&](void* arg) {
+        bool* use_direct_writes = static_cast<bool*>(arg);
+        ASSERT_EQ(*use_direct_writes,
+                  options.use_direct_io_for_flush_and_compaction);
+      });
+
+  SyncPoint::GetInstance()->EnableProcessing();
+  Reopen(options);
+  ASSERT_OK(Put("foo", "v"));
+  FlushOptions flush_options;
+  flush_options.wait = true;
+  ASSERT_OK(dbfull()->Flush(flush_options));
+  Destroy(options);
+  delete options.env;
+}
+
+INSTANTIATE_TEST_CASE_P(DBFlushDirectIOTest, DBFlushDirectIOTest,
+                        testing::Bool());
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/db_impl.h

@@ -34,18 +34,19 @@
 #include "db/write_controller.h"
 #include "db/write_thread.h"
 #include "memtable_list.h"
+#include "monitoring/instrumented_mutex.h"
+#include "options/db_options.h"
 #include "port/port.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
+#include "rocksdb/status.h"
 #include "rocksdb/transaction_log.h"
 #include "rocksdb/write_buffer_manager.h"
 #include "table/scoped_arena_iterator.h"
 #include "util/autovector.h"
-#include "util/db_options.h"
 #include "util/event_logger.h"
 #include "util/hash.h"
-#include "util/instrumented_mutex.h"
 #include "util/stop_watch.h"
 #include "util/thread_local.h"
 
@@ -91,7 +92,7 @@ class DBImpl : public DB {
   using DB::Get;
   virtual Status Get(const ReadOptions& options,
                      ColumnFamilyHandle* column_family, const Slice& key,
-                     std::string* value) override;
+                     PinnableSlice* value) override;
   using DB::MultiGet;
   virtual std::vector<Status> MultiGet(
       const ReadOptions& options,
@@ -99,10 +100,19 @@ class DBImpl : public DB {
       const std::vector<Slice>& keys,
       std::vector<std::string>* values) override;
 
-  virtual Status CreateColumnFamily(const ColumnFamilyOptions& options,
+  virtual Status CreateColumnFamily(const ColumnFamilyOptions& cf_options,
                                     const std::string& column_family,
                                     ColumnFamilyHandle** handle) override;
+  virtual Status CreateColumnFamilies(
+      const ColumnFamilyOptions& cf_options,
+      const std::vector<std::string>& column_family_names,
+      std::vector<ColumnFamilyHandle*>* handles) override;
+  virtual Status CreateColumnFamilies(
+      const std::vector<ColumnFamilyDescriptor>& column_families,
+      std::vector<ColumnFamilyHandle*>* handles) override;
   virtual Status DropColumnFamily(ColumnFamilyHandle* column_family) override;
+  virtual Status DropColumnFamilies(
+      const std::vector<ColumnFamilyHandle*>& column_families) override;
 
   // Returns false if key doesn't exist in the database and true if it may.
   // If value_found is not passed in as null, then return the value if found in
@@ -187,11 +197,16 @@ class DBImpl : public DB {
   using DB::Flush;
   virtual Status Flush(const FlushOptions& options,
                        ColumnFamilyHandle* column_family) override;
+  virtual Status FlushWAL(bool sync) override;
   virtual Status SyncWAL() override;
 
   virtual SequenceNumber GetLatestSequenceNumber() const override;
 
+  bool HasActiveSnapshotLaterThanSN(SequenceNumber sn);
+
 #ifndef ROCKSDB_LITE
+  using DB::ResetStats;
+  virtual Status ResetStats() override;
   virtual Status DisableFileDeletions() override;
   virtual Status EnableFileDeletions(bool force) override;
   virtual int IsFileDeletionsEnabled() const;
@@ -220,11 +235,11 @@ class DBImpl : public DB {
       ColumnFamilyHandle* column_family,
       ColumnFamilyMetaData* metadata) override;
 
-  // experimental API
   Status SuggestCompactRange(ColumnFamilyHandle* column_family,
-                             const Slice* begin, const Slice* end);
+                             const Slice* begin, const Slice* end) override;
 
-  Status PromoteL0(ColumnFamilyHandle* column_family, int target_level);
+  Status PromoteL0(ColumnFamilyHandle* column_family,
+                   int target_level) override;
 
   // Similar to Write() but will call the callback once on the single write
   // thread to determine whether it is safe to perform the write.
@@ -278,6 +293,8 @@ class DBImpl : public DB {
       const std::vector<std::string>& external_files,
       const IngestExternalFileOptions& ingestion_options) override;
 
+  virtual Status VerifyChecksum() override;
+
 #endif  // ROCKSDB_LITE
 
   // Similar to GetSnapshot(), but also lets the db know that this snapshot
@@ -314,7 +331,7 @@ class DBImpl : public DB {
                            ColumnFamilyHandle* column_family = nullptr,
                            bool disallow_trivial_move = false);
 
-  void TEST_MaybeFlushColumnFamilies();
+  void TEST_HandleWALFull();
 
   bool TEST_UnableToFlushOldestLog() {
     return unable_to_flush_oldest_log_;
@@ -373,9 +390,9 @@ class DBImpl : public DB {
   Status TEST_GetAllImmutableCFOptions(
       std::unordered_map<std::string, const ImmutableCFOptions*>* iopts_map);
 
-  // Return the lastest MutableCFOptions of of a column family
+  // Return the lastest MutableCFOptions of a column family
   Status TEST_GetLatestMutableCFOptions(ColumnFamilyHandle* column_family,
-                                        MutableCFOptions* mutable_cf_opitons);
+                                        MutableCFOptions* mutable_cf_options);
 
   Cache* TEST_table_cache() { return table_cache_.get(); }
 
@@ -385,12 +402,21 @@ class DBImpl : public DB {
   uint64_t TEST_FindMinPrepLogReferencedByMemTable();
 
   int TEST_BGCompactionsAllowed() const;
+  int TEST_BGFlushesAllowed() const;
 
 #endif  // NDEBUG
 
-  // Return maximum background compaction allowed to be scheduled based on
-  // compaction status.
-  int BGCompactionsAllowed() const;
+  struct BGJobLimits {
+    int max_flushes;
+    int max_compactions;
+  };
+  // Returns maximum background flushes and compactions allowed to be scheduled
+  BGJobLimits GetBGJobLimits() const;
+  // Need a static version that can be called during SanitizeOptions().
+  static BGJobLimits GetBGJobLimits(int max_background_flushes,
+                                    int max_background_compactions,
+                                    int max_background_jobs,
+                                    bool parallelize_compactions);
 
   // move logs pending closing from job_context to the DB queue and
   // schedule a purge
@@ -451,6 +477,9 @@ class DBImpl : public DB {
   // mutex is released.
   ColumnFamilyHandle* GetColumnFamilyHandle(uint32_t column_family_id);
 
+  // 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() {
@@ -467,6 +496,12 @@ class DBImpl : public DB {
 
   const WriteController& write_controller() { return write_controller_; }
 
+  InternalIterator* NewInternalIterator(const ReadOptions&,
+                                        ColumnFamilyData* cfd,
+                                        SuperVersion* super_version,
+                                        Arena* arena,
+                                        RangeDelAggregator* range_del_agg);
+
   // hollow transactions shell used for recovery.
   // these will then be passed to TransactionDB so that
   // locks can be reacquired before writing can resume.
@@ -525,6 +560,7 @@ class DBImpl : public DB {
   void AddToLogsToFreeQueue(log::Writer* log_writer) {
     logs_to_free_queue_.push_back(log_writer);
   }
+  InstrumentedMutex* mutex() { return &mutex_; }
 
   Status NewDB();
 
@@ -539,16 +575,11 @@ class DBImpl : public DB {
   std::unordered_map<std::string, RecoveredTransaction*>
       recovered_transactions_;
 
-  InternalIterator* NewInternalIterator(const ReadOptions&,
-                                        ColumnFamilyData* cfd,
-                                        SuperVersion* super_version,
-                                        Arena* arena,
-                                        RangeDelAggregator* range_del_agg);
-
   // Except in DB::Open(), WriteOptionsFile can only be called when:
-  // 1. WriteThread::Writer::EnterUnbatched() is used.
-  // 2. db_mutex is held
-  Status WriteOptionsFile();
+  // Persist options to options file.
+  // If need_mutex_lock = false, the method will lock DB mutex.
+  // If need_enter_write_thread = false, the method will enter write thread.
+  Status WriteOptionsFile(bool need_mutex_lock, bool need_enter_write_thread);
 
   // The following two functions can only be called when:
   // 1. WriteThread::Writer::EnterUnbatched() is used.
@@ -556,6 +587,10 @@ class DBImpl : public DB {
   Status RenameTempFileToOptionsFile(const std::string& file_name);
   Status DeleteObsoleteOptionsFiles();
 
+  void NotifyOnFlushBegin(ColumnFamilyData* cfd, FileMetaData* file_meta,
+                          const MutableCFOptions& mutable_cf_options,
+                          int job_id, TableProperties prop);
+
   void NotifyOnFlushCompleted(ColumnFamilyData* cfd, FileMetaData* file_meta,
                               const MutableCFOptions& mutable_cf_options,
                               int job_id, TableProperties prop);
@@ -581,7 +616,18 @@ class DBImpl : public DB {
   Status WriteImpl(const WriteOptions& options, WriteBatch* updates,
                    WriteCallback* callback = nullptr,
                    uint64_t* log_used = nullptr, uint64_t log_ref = 0,
-                   bool disable_memtable = false);
+                   bool disable_memtable = false, uint64_t* seq_used = nullptr);
+
+  Status PipelinedWriteImpl(const WriteOptions& options, WriteBatch* updates,
+                            WriteCallback* callback = nullptr,
+                            uint64_t* log_used = nullptr, uint64_t log_ref = 0,
+                            bool disable_memtable = false,
+                            uint64_t* seq_used = nullptr);
+
+  Status WriteImplWALOnly(const WriteOptions& options, WriteBatch* updates,
+                          WriteCallback* callback = nullptr,
+                          uint64_t* log_used = nullptr, uint64_t log_ref = 0,
+                          uint64_t* seq_used = nullptr);
 
   uint64_t FindMinLogContainingOutstandingPrep();
   uint64_t FindMinPrepLogReferencedByMemTable();
@@ -589,7 +635,9 @@ class DBImpl : public DB {
  private:
   friend class DB;
   friend class InternalStats;
-  friend class TransactionImpl;
+  friend class PessimisticTransaction;
+  friend class WriteCommittedTxn;
+  friend class WritePreparedTxn;
 #ifndef ROCKSDB_LITE
   friend class ForwardIterator;
 #endif
@@ -600,8 +648,21 @@ class DBImpl : public DB {
 #endif
   struct CompactionState;
 
-  struct WriteContext;
+  struct WriteContext {
+    autovector<SuperVersion*> superversions_to_free_;
+    autovector<MemTable*> memtables_to_free_;
 
+    ~WriteContext() {
+      for (auto& sv : superversions_to_free_) {
+        delete sv;
+      }
+      for (auto& m : memtables_to_free_) {
+        delete m;
+      }
+    }
+  };
+
+  struct PrepickedCompaction;
   struct PurgeFileInfo;
 
   // Recover the descriptor from persistent storage.  May do a significant
@@ -615,6 +676,12 @@ class DBImpl : public DB {
 
   const Status CreateArchivalDirectory();
 
+  Status CreateColumnFamilyImpl(const ColumnFamilyOptions& cf_options,
+                                const std::string& cf_name,
+                                ColumnFamilyHandle** handle);
+
+  Status DropColumnFamilyImpl(ColumnFamilyHandle* column_family);
+
   // Delete any unneeded files and stale in-memory entries.
   void DeleteObsoleteFiles();
   // Delete obsolete files and log status and information of file deletion
@@ -666,6 +733,9 @@ class DBImpl : public DB {
   //            `num_bytes` going through.
   Status DelayWrite(uint64_t num_bytes, const WriteOptions& write_options);
 
+  Status ThrottleLowPriWritesIfNeeded(const WriteOptions& write_options,
+                                      WriteBatch* my_batch);
+
   Status ScheduleFlushes(WriteContext* context);
 
   Status SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context);
@@ -677,6 +747,37 @@ class DBImpl : public DB {
   // Wait for memtable flushed
   Status WaitForFlushMemTable(ColumnFamilyData* cfd);
 
+  // REQUIRES: mutex locked
+  Status HandleWALFull(WriteContext* write_context);
+
+  // REQUIRES: mutex locked
+  Status HandleWriteBufferFull(WriteContext* write_context);
+
+  // REQUIRES: mutex locked
+  Status PreprocessWrite(const WriteOptions& write_options, bool* need_log_sync,
+                         WriteContext* write_context);
+
+  WriteBatch* MergeBatch(const WriteThread::WriteGroup& write_group,
+                         WriteBatch* tmp_batch, size_t* write_with_wal);
+
+  Status WriteToWAL(const WriteBatch& merged_batch, log::Writer* log_writer,
+                    uint64_t* log_used, uint64_t* log_size);
+
+  Status WriteToWAL(const WriteThread::WriteGroup& write_group,
+                    log::Writer* log_writer, uint64_t* log_used,
+                    bool need_log_sync, bool need_log_dir_sync,
+                    SequenceNumber sequence);
+
+  Status ConcurrentWriteToWAL(const WriteThread::WriteGroup& write_group,
+                              uint64_t* log_used, SequenceNumber* last_sequence,
+                              int total_count);
+
+  // Used by WriteImpl to update bg_error_ if paranoid check is enabled.
+  void WriteCallbackStatusCheck(const Status& status);
+
+  // Used by WriteImpl to update bg_error_ in case of memtable insert error.
+  void MemTableInsertStatusCheck(const Status& memtable_insert_status);
+
 #ifndef ROCKSDB_LITE
 
   Status CompactFilesImpl(const CompactionOptions& compact_options,
@@ -703,14 +804,19 @@ class DBImpl : public DB {
   void SchedulePendingPurge(std::string fname, FileType type, uint64_t number,
                             uint32_t path_id, int job_id);
   static void BGWorkCompaction(void* arg);
+  // Runs a pre-chosen universal compaction involving bottom level in a
+  // separate, bottom-pri thread pool.
+  static void BGWorkBottomCompaction(void* arg);
   static void BGWorkFlush(void* db);
   static void BGWorkPurge(void* arg);
   static void UnscheduleCallback(void* arg);
-  void BackgroundCallCompaction(void* arg);
+  void BackgroundCallCompaction(PrepickedCompaction* prepicked_compaction,
+                                Env::Priority bg_thread_pri);
   void BackgroundCallFlush();
   void BackgroundCallPurge();
   Status BackgroundCompaction(bool* madeProgress, JobContext* job_context,
-                              LogBuffer* log_buffer, void* m = 0);
+                              LogBuffer* log_buffer,
+                              PrepickedCompaction* prepicked_compaction);
   Status BackgroundFlush(bool* madeProgress, JobContext* job_context,
                          LogBuffer* log_buffer);
 
@@ -740,12 +846,6 @@ class DBImpl : public DB {
 
   const Snapshot* GetSnapshotImpl(bool is_write_conflict_boundary);
 
-  // Persist RocksDB options under the single write thread
-  // REQUIRES: mutex locked
-  Status PersistOptions();
-
-  void MaybeFlushColumnFamilies();
-
   uint64_t GetMaxTotalWalSize() const;
 
   // table_cache_ provides its own synchronization
@@ -754,11 +854,16 @@ class DBImpl : public DB {
   // 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_;
+  // In addition to mutex_, log_write_mutex_ protected writes to logs_ and
+  // logfile_number_. With concurrent_prepare it also protects alive_log_files_,
+  // and log_empty_. Refer to the definition of each variable below for more
+  // details.
+  InstrumentedMutex log_write_mutex_;
   // State below is protected by mutex_
+  // With concurrent_prepare enabled, some of the variables that accessed during
+  // WriteToWAL need different synchronization: log_empty_, alive_log_files_,
+  // logs_, logfile_number_. Refer to the definition of each variable below for
+  // more description.
   mutable InstrumentedMutex mutex_;
 
   std::atomic<bool> shutting_down_;
@@ -772,10 +877,20 @@ class DBImpl : public DB {
   // * whenever there is an error in background purge, flush or compaction
   // * whenever num_running_ingest_file_ goes to 0.
   InstrumentedCondVar bg_cv_;
+  // Writes are protected by locking both mutex_ and log_write_mutex_, and reads
+  // must be under either mutex_ or log_write_mutex_. Since after ::Open,
+  // logfile_number_ is currently updated only in write_thread_, it can be read
+  // from the same write_thread_ without any locks.
   uint64_t logfile_number_;
   std::deque<uint64_t>
       log_recycle_files;  // a list of log files that we can recycle
   bool log_dir_synced_;
+  // Without concurrent_prepare, read and writes to log_empty_ are protected by
+  // mutex_. Since it is currently updated/read only in write_thread_, it can be
+  // accessed from the same write_thread_ without any locks. With
+  // concurrent_prepare writes, where it can be updated in different threads,
+  // read and writes are protected by log_write_mutex_ instead. This is to avoid
+  // expesnive mutex_ lock during WAL write, which update log_empty_.
   bool log_empty_;
   ColumnFamilyHandleImpl* default_cf_handle_;
   InternalStats* default_cf_internal_stats_;
@@ -810,18 +925,30 @@ class DBImpl : public DB {
     // true for some prefix of logs_
     bool getting_synced = false;
   };
+  // Without concurrent_prepare, read and writes to alive_log_files_ are
+  // protected by mutex_. However since back() is never popped, and push_back()
+  // is done only from write_thread_, the same thread can access the item
+  // reffered by back() without mutex_. With concurrent_prepare_, writes
+  // are protected by locking both mutex_ and log_write_mutex_, and reads must
+  // be under either mutex_ or log_write_mutex_.
   std::deque<LogFileNumberSize> alive_log_files_;
   // Log files that aren't fully synced, and the current log file.
   // Synchronization:
-  //  - push_back() is done from write thread with locked mutex_,
-  //  - pop_front() is done from any thread with locked mutex_,
+  //  - push_back() is done from write_thread_ with locked mutex_ and
+  //  log_write_mutex_
+  //  - pop_front() is done from any thread with locked mutex_ and
+  //  log_write_mutex_
+  //  - reads are done with either locked mutex_ or log_write_mutex_
   //  - back() and items with getting_synced=true are not popped,
-  //  - it follows that write thread with unlocked mutex_ can safely access
-  //    back() and items with getting_synced=true.
+  //  - The same thread that sets getting_synced=true will reset it.
+  //  - it follows that the object referred by back() can be safely read from
+  //  the write_thread_ without using mutex
+  //  - it follows that the items with getting_synced=true can be safely read
+  //  from the same thread that has set getting_synced=true
   std::deque<LogWriterNumber> logs_;
   // Signaled when getting_synced becomes false for some of the logs_.
   InstrumentedCondVar log_sync_cv_;
-  uint64_t total_log_size_;
+  std::atomic<uint64_t> total_log_size_;
   // only used for dynamically adjusting max_total_wal_size. it is a sum of
   // [write_buffer_size * max_write_buffer_number] over all column families
   uint64_t max_total_in_memory_state_;
@@ -866,13 +993,19 @@ class DBImpl : public DB {
   WriteBufferManager* write_buffer_manager_;
 
   WriteThread write_thread_;
-
   WriteBatch tmp_batch_;
+  // The write thread when the writers have no memtable write. This will be used
+  // in 2PC to batch the prepares separately from the serial commit.
+  WriteThread nonmem_write_thread_;
 
   WriteController write_controller_;
 
+  unique_ptr<RateLimiter> low_pri_write_rate_limiter_;
+
   // Size of the last batch group. In slowdown mode, next write needs to
   // sleep if it uses up the quota.
+  // Note: This is to protect memtable and compaction. If the batch only writes
+  // to the WAL its size need not to be included in this.
   uint64_t last_batch_group_size_;
 
   FlushScheduler flush_scheduler_;
@@ -936,6 +1069,10 @@ class DBImpl : public DB {
   int unscheduled_flushes_;
   int unscheduled_compactions_;
 
+  // count how many background compactions are running or have been scheduled in
+  // the BOTTOM pool
+  int bg_bottom_compaction_scheduled_;
+
   // count how many background compactions are running or have been scheduled
   int bg_compaction_scheduled_;
 
@@ -952,7 +1089,7 @@ class DBImpl : public DB {
   int bg_purge_scheduled_;
 
   // Information for a manual compaction
-  struct ManualCompaction {
+  struct ManualCompactionState {
     ColumnFamilyData* cfd;
     int input_level;
     int output_level;
@@ -968,13 +1105,21 @@ class DBImpl : public DB {
     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 PrepickedCompaction {
+    // background compaction takes ownership of `compaction`.
+    Compaction* compaction;
+    // caller retains ownership of `manual_compaction_state` as it is reused
+    // across background compactions.
+    ManualCompactionState* manual_compaction_state;  // nullptr if non-manual
+  };
+  std::deque<ManualCompactionState*> manual_compaction_dequeue_;
 
   struct CompactionArg {
+    // caller retains ownership of `db`.
     DBImpl* db;
-    ManualCompaction* m;
+    // background compaction takes ownership of `prepicked_compaction`.
+    PrepickedCompaction* prepicked_compaction;
   };
 
   // Have we encountered a background error in paranoid mode?
@@ -1044,7 +1189,7 @@ class DBImpl : public DB {
   // Indicate DB was opened successfully
   bool opened_successfully_;
 
-  // minmum log number still containing prepared data.
+  // minimum 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
@@ -1057,7 +1202,7 @@ class DBImpl : public DB {
   // 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
+  // when trying to determine the minimum 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
@@ -1070,13 +1215,6 @@ class DBImpl : public DB {
   DBImpl(const DBImpl&);
   void operator=(const DBImpl&);
 
-  // Return the earliest snapshot where seqno is visible.
-  // Store the snapshot right before that, if any, in prev_snapshot
-  inline SequenceNumber findEarliestVisibleSnapshot(
-    SequenceNumber in,
-    std::vector<SequenceNumber>& snapshots,
-    SequenceNumber* prev_snapshot);
-
   // Background threads call this function, which is just a wrapper around
   // the InstallSuperVersion() function. Background threads carry
   // job_context which can have new_superversion already
@@ -1106,7 +1244,7 @@ class DBImpl : public DB {
   // Function that Get and KeyMayExist call with no_io true or false
   // Note: 'value_found' from KeyMayExist propagates here
   Status GetImpl(const ReadOptions& options, ColumnFamilyHandle* column_family,
-                 const Slice& key, std::string* value,
+                 const Slice& key, PinnableSlice* value,
                  bool* value_found = nullptr);
 
   bool GetIntPropertyInternal(ColumnFamilyData* cfd,
@@ -1115,13 +1253,18 @@ class DBImpl : public DB {
 
   bool HasPendingManualCompaction();
   bool HasExclusiveManualCompaction();
-  void AddManualCompaction(ManualCompaction* m);
-  void RemoveManualCompaction(ManualCompaction* m);
-  bool ShouldntRunManualCompaction(ManualCompaction* m);
+  void AddManualCompaction(ManualCompactionState* m);
+  void RemoveManualCompaction(ManualCompactionState* m);
+  bool ShouldntRunManualCompaction(ManualCompactionState* m);
   bool HaveManualCompaction(ColumnFamilyData* cfd);
-  bool MCOverlap(ManualCompaction* m, ManualCompaction* m1);
+  bool MCOverlap(ManualCompactionState* m, ManualCompactionState* m1);
 
   size_t GetWalPreallocateBlockSize(uint64_t write_buffer_size) const;
+
+  // When set, we use a seprate queue for writes that dont write to memtable. In
+  // 2PC these are the writes at Prepare phase.
+  const bool concurrent_prepare_;
+  const bool manual_wal_flush_;
 };
 
 extern Options SanitizeOptions(const std::string& db,
@@ -1129,6 +1272,10 @@ extern Options SanitizeOptions(const std::string& db,
 
 extern DBOptions SanitizeOptions(const std::string& db, const DBOptions& src);
 
+extern CompressionType GetCompressionFlush(
+    const ImmutableCFOptions& ioptions,
+    const MutableCFOptions& mutable_cf_options);
+
 // Fix user-supplied options to be reasonable
 template <class T, class V>
 static void ClipToRange(T* ptr, V minvalue, V maxvalue) {

db/db_impl_debug.cc

@@ -10,7 +10,7 @@
 #ifndef NDEBUG
 
 #include "db/db_impl.h"
-#include "util/thread_status_updater.h"
+#include "monitoring/thread_status_updater.h"
 
 namespace rocksdb {
 
@@ -19,9 +19,10 @@ uint64_t DBImpl::TEST_GetLevel0TotalSize() {
   return default_cf_handle_->cfd()->current()->storage_info()->NumLevelBytes(0);
 }
 
-void DBImpl::TEST_MaybeFlushColumnFamilies() {
+void DBImpl::TEST_HandleWALFull() {
+  WriteContext write_context;
   InstrumentedMutexLock l(&mutex_);
-  MaybeFlushColumnFamilies();
+  HandleWALFull(&write_context);
 }
 
 int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
@@ -111,7 +112,9 @@ Status DBImpl::TEST_WaitForCompact() {
   // OR flush to finish.
 
   InstrumentedMutexLock l(&mutex_);
-  while ((bg_compaction_scheduled_ || bg_flush_scheduled_) && bg_error_.ok()) {
+  while ((bg_bottom_compaction_scheduled_ || bg_compaction_scheduled_ ||
+          bg_flush_scheduled_) &&
+         bg_error_.ok()) {
     bg_cv_.Wait();
   }
   return bg_error_;
@@ -185,7 +188,12 @@ Status DBImpl::TEST_GetLatestMutableCFOptions(
 
 int DBImpl::TEST_BGCompactionsAllowed() const {
   InstrumentedMutexLock l(&mutex_);
-  return BGCompactionsAllowed();
+  return GetBGJobLimits().max_compactions;
+}
+
+int DBImpl::TEST_BGFlushesAllowed() const {
+  InstrumentedMutexLock l(&mutex_);
+  return GetBGJobLimits().max_flushes;
 }
 
 }  // namespace rocksdb

db/db_impl_experimental.cc

@@ -61,8 +61,8 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
   assert(column_family);
 
   if (target_level < 1) {
-    Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-        "PromoteL0 FAILED. Invalid target level %d\n", target_level);
+    ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                   "PromoteL0 FAILED. Invalid target level %d\n", target_level);
     return Status::InvalidArgument("Invalid target level");
   }
 
@@ -75,8 +75,9 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
     const auto* vstorage = cfd->current()->storage_info();
 
     if (target_level >= vstorage->num_levels()) {
-      Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-          "PromoteL0 FAILED. Target level %d does not exist\n", target_level);
+      ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                     "PromoteL0 FAILED. Target level %d does not exist\n",
+                     target_level);
       job_context.Clean();
       return Status::InvalidArgument("Target level does not exist");
     }
@@ -94,9 +95,9 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
     for (size_t i = 0; i < l0_files.size(); ++i) {
       auto f = l0_files[i];
       if (f->being_compacted) {
-        Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-            "PromoteL0 FAILED. File %" PRIu64 " being compacted\n",
-            f->fd.GetNumber());
+        ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                       "PromoteL0 FAILED. File %" PRIu64 " being compacted\n",
+                       f->fd.GetNumber());
         job_context.Clean();
         return Status::InvalidArgument("PromoteL0 called during L0 compaction");
       }
@@ -104,10 +105,10 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
       if (i == 0) continue;
       auto prev_f = l0_files[i - 1];
       if (icmp->Compare(prev_f->largest, f->smallest) >= 0) {
-        Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-            "PromoteL0 FAILED. Files %" PRIu64 " and %" PRIu64
-            " have overlapping ranges\n",
-            prev_f->fd.GetNumber(), f->fd.GetNumber());
+        ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                       "PromoteL0 FAILED. Files %" PRIu64 " and %" PRIu64
+                       " have overlapping ranges\n",
+                       prev_f->fd.GetNumber(), f->fd.GetNumber());
         job_context.Clean();
         return Status::InvalidArgument("L0 has overlapping files");
       }
@@ -116,8 +117,8 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
     // Check that all levels up to target_level are empty.
     for (int level = 1; level <= target_level; ++level) {
       if (vstorage->NumLevelFiles(level) > 0) {
-        Log(InfoLogLevel::INFO_LEVEL, immutable_db_options_.info_log,
-            "PromoteL0 FAILED. Level %d not empty\n", level);
+        ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                       "PromoteL0 FAILED. Level %d not empty\n", level);
         job_context.Clean();
         return Status::InvalidArgument(
             "All levels up to target_level "

db/db_impl_files.cc

@@ -0,0 +1,548 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root 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_impl.h"
+
+#ifndef __STDC_FORMAT_MACROS
+#define __STDC_FORMAT_MACROS
+#endif
+#include <inttypes.h>
+#include "db/event_helpers.h"
+#include "util/file_util.h"
+#include "util/sst_file_manager_impl.h"
+
+
+namespace rocksdb {
+uint64_t DBImpl::FindMinPrepLogReferencedByMemTable() {
+  if (!allow_2pc()) {
+    return 0;
+  }
+
+  uint64_t min_log = 0;
+
+  // we must look through the memtables for two phase transactions
+  // that have been committed but not yet flushed
+  for (auto loop_cfd : *versions_->GetColumnFamilySet()) {
+    if (loop_cfd->IsDropped()) {
+      continue;
+    }
+
+    auto log = loop_cfd->imm()->GetMinLogContainingPrepSection();
+
+    if (log > 0 && (min_log == 0 || log < min_log)) {
+      min_log = log;
+    }
+
+    log = loop_cfd->mem()->GetMinLogContainingPrepSection();
+
+    if (log > 0 && (min_log == 0 || log < min_log)) {
+      min_log = log;
+    }
+  }
+
+  return min_log;
+}
+
+void DBImpl::MarkLogAsHavingPrepSectionFlushed(uint64_t log) {
+  assert(log != 0);
+  std::lock_guard<std::mutex> lock(prep_heap_mutex_);
+  auto it = prepared_section_completed_.find(log);
+  assert(it != prepared_section_completed_.end());
+  it->second += 1;
+}
+
+void DBImpl::MarkLogAsContainingPrepSection(uint64_t log) {
+  assert(log != 0);
+  std::lock_guard<std::mutex> lock(prep_heap_mutex_);
+  min_log_with_prep_.push(log);
+  auto it = prepared_section_completed_.find(log);
+  if (it == prepared_section_completed_.end()) {
+    prepared_section_completed_[log] = 0;
+  }
+}
+
+uint64_t DBImpl::FindMinLogContainingOutstandingPrep() {
+
+  if (!allow_2pc()) {
+    return 0;
+  }
+
+  std::lock_guard<std::mutex> lock(prep_heap_mutex_);
+  uint64_t min_log = 0;
+
+  // first we look in the prepared heap where we keep
+  // track of transactions that have been prepared (written to WAL)
+  // but not yet committed.
+  while (!min_log_with_prep_.empty()) {
+    min_log = min_log_with_prep_.top();
+
+    auto it = prepared_section_completed_.find(min_log);
+
+    // value was marked as 'deleted' from heap
+    if (it != prepared_section_completed_.end() && it->second > 0) {
+      it->second -= 1;
+      min_log_with_prep_.pop();
+
+      // back to squere one...
+      min_log = 0;
+      continue;
+    } else {
+      // found a valid value
+      break;
+    }
+  }
+
+  return min_log;
+}
+
+uint64_t DBImpl::MinLogNumberToKeep() {
+  uint64_t log_number = versions_->MinLogNumber();
+
+  if (allow_2pc()) {
+    // if are 2pc we must consider logs containing prepared
+    // sections of outstanding transactions.
+    //
+    // We must check min logs with outstanding prep before we check
+    // logs referneces by memtables because a log referenced by the
+    // first data structure could transition to the second under us.
+    //
+    // TODO(horuff): iterating over all column families under db mutex.
+    // should find more optimial solution
+    auto min_log_in_prep_heap = FindMinLogContainingOutstandingPrep();
+
+    if (min_log_in_prep_heap != 0 && min_log_in_prep_heap < log_number) {
+      log_number = min_log_in_prep_heap;
+    }
+
+    auto min_log_refed_by_mem = FindMinPrepLogReferencedByMemTable();
+
+    if (min_log_refed_by_mem != 0 && min_log_refed_by_mem < log_number) {
+      log_number = min_log_refed_by_mem;
+    }
+  }
+  return log_number;
+}
+
+// * Returns the list of live files in 'sst_live'
+// If it's doing full scan:
+// * Returns the list of all files in the filesystem in
+// 'full_scan_candidate_files'.
+// Otherwise, gets obsolete files from VersionSet.
+// no_full_scan = true -- never do the full scan using GetChildren()
+// force = false -- don't force the full scan, except every
+//  mutable_db_options_.delete_obsolete_files_period_micros
+// force = true -- force the full scan
+void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
+                               bool no_full_scan) {
+  mutex_.AssertHeld();
+
+  // if deletion is disabled, do nothing
+  if (disable_delete_obsolete_files_ > 0) {
+    return;
+  }
+
+  bool doing_the_full_scan = false;
+
+  // logic for figurint out if we're doing the full scan
+  if (no_full_scan) {
+    doing_the_full_scan = false;
+  } else if (force ||
+             mutable_db_options_.delete_obsolete_files_period_micros == 0) {
+    doing_the_full_scan = true;
+  } else {
+    const uint64_t now_micros = env_->NowMicros();
+    if ((delete_obsolete_files_last_run_ +
+         mutable_db_options_.delete_obsolete_files_period_micros) <
+        now_micros) {
+      doing_the_full_scan = true;
+      delete_obsolete_files_last_run_ = now_micros;
+    }
+  }
+
+  // don't delete files that might be currently written to from compaction
+  // threads
+  // Since job_context->min_pending_output is set, until file scan finishes,
+  // mutex_ cannot be released. Otherwise, we might see no min_pending_output
+  // here but later find newer generated unfinalized files while scannint.
+  if (!pending_outputs_.empty()) {
+    job_context->min_pending_output = *pending_outputs_.begin();
+  } else {
+    // delete all of them
+    job_context->min_pending_output = std::numeric_limits<uint64_t>::max();
+  }
+
+  // Get obsolete files.  This function will also update the list of
+  // pending files in VersionSet().
+  versions_->GetObsoleteFiles(&job_context->sst_delete_files,
+                              &job_context->manifest_delete_files,
+                              job_context->min_pending_output);
+
+  // store the current filenum, lognum, etc
+  job_context->manifest_file_number = versions_->manifest_file_number();
+  job_context->pending_manifest_file_number =
+      versions_->pending_manifest_file_number();
+  job_context->log_number = MinLogNumberToKeep();
+
+  job_context->prev_log_number = versions_->prev_log_number();
+
+  versions_->AddLiveFiles(&job_context->sst_live);
+  if (doing_the_full_scan) {
+    for (size_t path_id = 0; path_id < immutable_db_options_.db_paths.size();
+         path_id++) {
+      // set of all files in the directory. We'll exclude files that are still
+      // alive in the subsequent processings.
+      std::vector<std::string> files;
+      env_->GetChildren(immutable_db_options_.db_paths[path_id].path,
+                        &files);  // Ignore errors
+      for (std::string file : files) {
+        // TODO(icanadi) clean up this mess to avoid having one-off "/" prefixes
+        job_context->full_scan_candidate_files.emplace_back(
+            "/" + file, static_cast<uint32_t>(path_id));
+      }
+    }
+
+    // Add log files in wal_dir
+    if (immutable_db_options_.wal_dir != dbname_) {
+      std::vector<std::string> log_files;
+      env_->GetChildren(immutable_db_options_.wal_dir,
+                        &log_files);  // Ignore errors
+      for (std::string log_file : log_files) {
+        job_context->full_scan_candidate_files.emplace_back(log_file, 0);
+      }
+    }
+    // Add info log files in db_log_dir
+    if (!immutable_db_options_.db_log_dir.empty() &&
+        immutable_db_options_.db_log_dir != dbname_) {
+      std::vector<std::string> info_log_files;
+      // Ignore errors
+      env_->GetChildren(immutable_db_options_.db_log_dir, &info_log_files);
+      for (std::string log_file : info_log_files) {
+        job_context->full_scan_candidate_files.emplace_back(log_file, 0);
+      }
+    }
+  }
+
+  // logs_ is empty when called during recovery, in which case there can't yet
+  // be any tracked obsolete logs
+  if (!alive_log_files_.empty() && !logs_.empty()) {
+    uint64_t min_log_number = job_context->log_number;
+    size_t num_alive_log_files = alive_log_files_.size();
+    // find newly obsoleted log files
+    while (alive_log_files_.begin()->number < min_log_number) {
+      auto& earliest = *alive_log_files_.begin();
+      if (immutable_db_options_.recycle_log_file_num >
+          log_recycle_files.size()) {
+        ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                       "adding log %" PRIu64 " to recycle list\n",
+                       earliest.number);
+        log_recycle_files.push_back(earliest.number);
+      } else {
+        job_context->log_delete_files.push_back(earliest.number);
+      }
+      if (job_context->size_log_to_delete == 0) {
+        job_context->prev_total_log_size = total_log_size_;
+        job_context->num_alive_log_files = num_alive_log_files;
+      }
+      job_context->size_log_to_delete += earliest.size;
+      total_log_size_ -= earliest.size;
+      if (concurrent_prepare_) {
+        log_write_mutex_.Lock();
+      }
+      alive_log_files_.pop_front();
+      if (concurrent_prepare_) {
+        log_write_mutex_.Unlock();
+      }
+      // Current log should always stay alive since it can't have
+      // number < MinLogNumber().
+      assert(alive_log_files_.size());
+    }
+    while (!logs_.empty() && logs_.front().number < min_log_number) {
+      auto& log = logs_.front();
+      if (log.getting_synced) {
+        log_sync_cv_.Wait();
+        // logs_ could have changed while we were waiting.
+        continue;
+      }
+      logs_to_free_.push_back(log.ReleaseWriter());
+      {
+        InstrumentedMutexLock wl(&log_write_mutex_);
+        logs_.pop_front();
+      }
+    }
+    // Current log cannot be obsolete.
+    assert(!logs_.empty());
+  }
+
+  // We're just cleaning up for DB::Write().
+  assert(job_context->logs_to_free.empty());
+  job_context->logs_to_free = logs_to_free_;
+  job_context->log_recycle_files.assign(log_recycle_files.begin(),
+                                        log_recycle_files.end());
+  logs_to_free_.clear();
+}
+
+namespace {
+bool CompareCandidateFile(const JobContext::CandidateFileInfo& first,
+                          const JobContext::CandidateFileInfo& second) {
+  if (first.file_name > second.file_name) {
+    return true;
+  } else if (first.file_name < second.file_name) {
+    return false;
+  } else {
+    return (first.path_id > second.path_id);
+  }
+}
+};  // namespace
+
+// Delete obsolete files and log status and information of file deletion
+void DBImpl::DeleteObsoleteFileImpl(Status file_deletion_status, int job_id,
+                                    const std::string& fname, FileType type,
+                                    uint64_t number, uint32_t path_id) {
+  if (type == kTableFile) {
+    file_deletion_status =
+        DeleteSSTFile(&immutable_db_options_, fname, path_id);
+  } else {
+    file_deletion_status = env_->DeleteFile(fname);
+  }
+  if (file_deletion_status.ok()) {
+    ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
+                    "[JOB %d] Delete %s type=%d #%" PRIu64 " -- %s\n", job_id,
+                    fname.c_str(), type, number,
+                    file_deletion_status.ToString().c_str());
+  } else if (env_->FileExists(fname).IsNotFound()) {
+    ROCKS_LOG_INFO(
+        immutable_db_options_.info_log,
+        "[JOB %d] Tried to delete a non-existing file %s type=%d #%" PRIu64
+        " -- %s\n",
+        job_id, fname.c_str(), type, number,
+        file_deletion_status.ToString().c_str());
+  } else {
+    ROCKS_LOG_ERROR(immutable_db_options_.info_log,
+                    "[JOB %d] Failed to delete %s type=%d #%" PRIu64 " -- %s\n",
+                    job_id, fname.c_str(), type, number,
+                    file_deletion_status.ToString().c_str());
+  }
+  if (type == kTableFile) {
+    EventHelpers::LogAndNotifyTableFileDeletion(
+        &event_logger_, job_id, number, fname, file_deletion_status, GetName(),
+        immutable_db_options_.listeners);
+  }
+}
+
+// Diffs the files listed in filenames and those that do not
+// 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 DBImpl::PurgeObsoleteFiles(const JobContext& state, bool schedule_only) {
+  // we'd better have sth to delete
+  assert(state.HaveSomethingToDelete());
+
+  // this checks if FindObsoleteFiles() was run before. If not, don't do
+  // PurgeObsoleteFiles(). If FindObsoleteFiles() was run, we need to also
+  // run PurgeObsoleteFiles(), even if disable_delete_obsolete_files_ is true
+  if (state.manifest_file_number == 0) {
+    return;
+  }
+
+  // Now, convert live list to an unordered map, WITHOUT mutex held;
+  // set is slow.
+  std::unordered_map<uint64_t, const FileDescriptor*> sst_live_map;
+  for (const FileDescriptor& fd : state.sst_live) {
+    sst_live_map[fd.GetNumber()] = &fd;
+  }
+  std::unordered_set<uint64_t> log_recycle_files_set(
+      state.log_recycle_files.begin(), state.log_recycle_files.end());
+
+  auto candidate_files = state.full_scan_candidate_files;
+  candidate_files.reserve(
+      candidate_files.size() + state.sst_delete_files.size() +
+      state.log_delete_files.size() + state.manifest_delete_files.size());
+  // We may ignore the dbname when generating the file names.
+  const char* kDumbDbName = "";
+  for (auto file : state.sst_delete_files) {
+    candidate_files.emplace_back(
+        MakeTableFileName(kDumbDbName, file->fd.GetNumber()),
+        file->fd.GetPathId());
+    if (file->table_reader_handle) {
+      table_cache_->Release(file->table_reader_handle);
+    }
+    delete file;
+  }
+
+  for (auto file_num : state.log_delete_files) {
+    if (file_num > 0) {
+      candidate_files.emplace_back(LogFileName(kDumbDbName, file_num), 0);
+    }
+  }
+  for (const auto& filename : state.manifest_delete_files) {
+    candidate_files.emplace_back(filename, 0);
+  }
+
+  // dedup state.candidate_files so we don't try to delete the same
+  // file twice
+  std::sort(candidate_files.begin(), candidate_files.end(),
+            CompareCandidateFile);
+  candidate_files.erase(
+      std::unique(candidate_files.begin(), candidate_files.end()),
+      candidate_files.end());
+
+  if (state.prev_total_log_size > 0) {
+    ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                   "[JOB %d] Try to delete WAL files size %" PRIu64
+                   ", prev total WAL file size %" PRIu64
+                   ", number of live WAL files %" ROCKSDB_PRIszt ".\n",
+                   state.job_id, state.size_log_to_delete,
+                   state.prev_total_log_size, state.num_alive_log_files);
+  }
+
+  std::vector<std::string> old_info_log_files;
+  InfoLogPrefix info_log_prefix(!immutable_db_options_.db_log_dir.empty(),
+                                dbname_);
+  for (const auto& candidate_file : candidate_files) {
+    std::string to_delete = candidate_file.file_name;
+    uint32_t path_id = candidate_file.path_id;
+    uint64_t number;
+    FileType type;
+    // Ignore file if we cannot recognize it.
+    if (!ParseFileName(to_delete, &number, info_log_prefix.prefix, &type)) {
+      continue;
+    }
+
+    bool keep = true;
+    switch (type) {
+      case kLogFile:
+        keep = ((number >= state.log_number) ||
+                (number == state.prev_log_number) ||
+                (log_recycle_files_set.find(number) !=
+                 log_recycle_files_set.end()));
+        break;
+      case kDescriptorFile:
+        // Keep my manifest file, and any newer incarnations'
+        // (can happen during manifest roll)
+        keep = (number >= state.manifest_file_number);
+        break;
+      case kTableFile:
+        // If the second condition is not there, this makes
+        // DontDeletePendingOutputs fail
+        keep = (sst_live_map.find(number) != sst_live_map.end()) ||
+               number >= state.min_pending_output;
+        break;
+      case kTempFile:
+        // Any temp files that are currently being written to must
+        // be recorded in pending_outputs_, which is inserted into "live".
+        // Also, SetCurrentFile creates a temp file when writing out new
+        // manifest, which is equal to state.pending_manifest_file_number. We
+        // should not delete that file
+        //
+        // TODO(yhchiang): carefully modify the third condition to safely
+        //                 remove the temp options files.
+        keep = (sst_live_map.find(number) != sst_live_map.end()) ||
+               (number == state.pending_manifest_file_number) ||
+               (to_delete.find(kOptionsFileNamePrefix) != std::string::npos);
+        break;
+      case kInfoLogFile:
+        keep = true;
+        if (number != 0) {
+          old_info_log_files.push_back(to_delete);
+        }
+        break;
+      case kCurrentFile:
+      case kDBLockFile:
+      case kIdentityFile:
+      case kMetaDatabase:
+      case kOptionsFile:
+      case kBlobFile:
+        keep = true;
+        break;
+    }
+
+    if (keep) {
+      continue;
+    }
+
+    std::string fname;
+    if (type == kTableFile) {
+      // evict from cache
+      TableCache::Evict(table_cache_.get(), number);
+      fname = TableFileName(immutable_db_options_.db_paths, number, path_id);
+    } else {
+      fname = ((type == kLogFile) ? immutable_db_options_.wal_dir : dbname_) +
+              "/" + to_delete;
+    }
+
+#ifndef ROCKSDB_LITE
+    if (type == kLogFile && (immutable_db_options_.wal_ttl_seconds > 0 ||
+                             immutable_db_options_.wal_size_limit_mb > 0)) {
+      wal_manager_.ArchiveWALFile(fname, number);
+      continue;
+    }
+#endif  // !ROCKSDB_LITE
+
+    Status file_deletion_status;
+    if (schedule_only) {
+      InstrumentedMutexLock guard_lock(&mutex_);
+      SchedulePendingPurge(fname, type, number, path_id, state.job_id);
+    } else {
+      DeleteObsoleteFileImpl(file_deletion_status, state.job_id, fname, type,
+                             number, path_id);
+    }
+  }
+
+  // Delete old info log files.
+  size_t old_info_log_file_count = old_info_log_files.size();
+  if (old_info_log_file_count != 0 &&
+      old_info_log_file_count >= immutable_db_options_.keep_log_file_num) {
+    std::sort(old_info_log_files.begin(), old_info_log_files.end());
+    size_t end =
+        old_info_log_file_count - immutable_db_options_.keep_log_file_num;
+    for (unsigned int i = 0; i <= end; i++) {
+      std::string& to_delete = old_info_log_files.at(i);
+      std::string full_path_to_delete =
+          (immutable_db_options_.db_log_dir.empty()
+               ? dbname_
+               : immutable_db_options_.db_log_dir) +
+          "/" + to_delete;
+      ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                     "[JOB %d] Delete info log file %s\n", state.job_id,
+                     full_path_to_delete.c_str());
+      Status s = env_->DeleteFile(full_path_to_delete);
+      if (!s.ok()) {
+        if (env_->FileExists(full_path_to_delete).IsNotFound()) {
+          ROCKS_LOG_INFO(
+              immutable_db_options_.info_log,
+              "[JOB %d] Tried to delete non-existing info log file %s FAILED "
+              "-- %s\n",
+              state.job_id, to_delete.c_str(), s.ToString().c_str());
+        } else {
+          ROCKS_LOG_ERROR(immutable_db_options_.info_log,
+                          "[JOB %d] Delete info log file %s FAILED -- %s\n",
+                          state.job_id, to_delete.c_str(),
+                          s.ToString().c_str());
+        }
+      }
+    }
+  }
+#ifndef ROCKSDB_LITE
+  wal_manager_.PurgeObsoleteWALFiles();
+#endif  // ROCKSDB_LITE
+  LogFlush(immutable_db_options_.info_log);
+}
+
+void DBImpl::DeleteObsoleteFiles() {
+  mutex_.AssertHeld();
+  JobContext job_context(next_job_id_.fetch_add(1));
+  FindObsoleteFiles(&job_context, true);
+
+  mutex_.Unlock();
+  if (job_context.HaveSomethingToDelete()) {
+    PurgeObsoleteFiles(job_context);
+  }
+  job_context.Clean();
+  mutex_.Lock();
+}
+}  // namespace rocksdb

db/db_impl_readonly.cc

@@ -3,7 +3,6 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 
-
 #include "db/db_impl_readonly.h"
 
 #include "db/compacted_db_impl.h"
@@ -11,7 +10,7 @@
 #include "db/db_iter.h"
 #include "db/merge_context.h"
 #include "db/range_del_aggregator.h"
-#include "util/perf_context_imp.h"
+#include "monitoring/perf_context_imp.h"
 
 namespace rocksdb {
 
@@ -20,8 +19,8 @@ namespace rocksdb {
 DBImplReadOnly::DBImplReadOnly(const DBOptions& db_options,
                                const std::string& dbname)
     : DBImpl(db_options, dbname) {
-  Log(INFO_LEVEL, immutable_db_options_.info_log,
-      "Opening the db in read only mode");
+  ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                 "Opening the db in read only mode");
   LogFlush(immutable_db_options_.info_log);
 }
 
@@ -31,7 +30,8 @@ DBImplReadOnly::~DBImplReadOnly() {
 // Implementations of the DB interface
 Status DBImplReadOnly::Get(const ReadOptions& read_options,
                            ColumnFamilyHandle* column_family, const Slice& key,
-                           std::string* value) {
+                           PinnableSlice* pinnable_val) {
+  assert(pinnable_val != nullptr);
   Status s;
   SequenceNumber snapshot = versions_->LastSequence();
   auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
@@ -40,12 +40,13 @@ Status DBImplReadOnly::Get(const ReadOptions& read_options,
   MergeContext merge_context;
   RangeDelAggregator range_del_agg(cfd->internal_comparator(), snapshot);
   LookupKey lkey(key, snapshot);
-  if (super_version->mem->Get(lkey, value, &s, &merge_context, &range_del_agg,
-                              read_options)) {
+  if (super_version->mem->Get(lkey, pinnable_val->GetSelf(), &s, &merge_context,
+                              &range_del_agg, read_options)) {
+    pinnable_val->PinSelf();
   } else {
     PERF_TIMER_GUARD(get_from_output_files_time);
-    super_version->current->Get(read_options, lkey, value, &s, &merge_context,
-                                &range_del_agg);
+    super_version->current->Get(read_options, lkey, pinnable_val, &s,
+                                &merge_context, &range_del_agg);
   }
   return s;
 }
@@ -57,7 +58,7 @@ Iterator* DBImplReadOnly::NewIterator(const ReadOptions& read_options,
   SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
   SequenceNumber latest_snapshot = versions_->LastSequence();
   auto db_iter = NewArenaWrappedDbIterator(
-      env_, *cfd->ioptions(), cfd->user_comparator(),
+      env_, read_options, *cfd->ioptions(),
       (read_options.snapshot != nullptr
            ? reinterpret_cast<const SnapshotImpl*>(read_options.snapshot)
                  ->number_
@@ -86,7 +87,7 @@ Status DBImplReadOnly::NewIterators(
     auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(cfh)->cfd();
     auto* sv = cfd->GetSuperVersion()->Ref();
     auto* db_iter = NewArenaWrappedDbIterator(
-        env_, *cfd->ioptions(), cfd->user_comparator(),
+        env_, read_options, *cfd->ioptions(),
         (read_options.snapshot != nullptr
              ? reinterpret_cast<const SnapshotImpl*>(read_options.snapshot)
                    ->number_

db/db_impl_readonly.h

@@ -22,7 +22,7 @@ class DBImplReadOnly : public DBImpl {
   using DB::Get;
   virtual Status Get(const ReadOptions& options,
                      ColumnFamilyHandle* column_family, const Slice& key,
-                     std::string* value) override;
+                     PinnableSlice* value) override;
 
   // TODO: Implement ReadOnly MultiGet?
 

db/db_info_dumper.cc

@@ -15,8 +15,8 @@
 #include <algorithm>
 #include <vector>
 
-#include "db/filename.h"
 #include "rocksdb/env.h"
+#include "util/filename.h"
 
 namespace rocksdb {
 

db/db_info_dumper.h

@@ -6,7 +6,7 @@
 
 #include <string>
 
-#include "util/db_options.h"
+#include "options/db_options.h"
 
 namespace rocksdb {
 void DumpDBFileSummary(const ImmutableDBOptions& options,

db/db_io_failure_test.cc

@@ -204,7 +204,6 @@ TEST_F(DBIOFailureTest, ManifestWriteError) {
     ASSERT_EQ("bar2", Get("foo2"));
   }
 }
-#endif  // ROCKSDB_LITE
 
 TEST_F(DBIOFailureTest, PutFailsParanoid) {
   // Test the following:
@@ -250,7 +249,316 @@ TEST_F(DBIOFailureTest, PutFailsParanoid) {
   // the next put should NOT fail
   ASSERT_TRUE(s.ok());
 }
+#if !(defined NDEBUG) || !defined(OS_WIN)
+TEST_F(DBIOFailureTest, FlushSstRangeSyncError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.write_buffer_size = 256 * 1024 * 1024;
+  options.writable_file_max_buffer_size = 128 * 1024;
+  options.bytes_per_sync = 128 * 1024;
+  options.level0_file_num_compaction_trigger = 4;
+  options.memtable_factory.reset(new SpecialSkipListFactory(10));
+  BlockBasedTableOptions table_options;
+  table_options.filter_policy.reset(NewBloomFilterPolicy(10));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
 
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+
+  std::atomic<int> range_sync_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::RangeSync", [&](void* arg) {
+        if (range_sync_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("range sync dummy error");
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  Random rnd(301);
+  std::string rnd_str =
+      RandomString(&rnd, static_cast<int>(options.bytes_per_sync / 2));
+  std::string rnd_str_512kb = RandomString(&rnd, 512 * 1024);
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  // First 1MB doesn't get range synced
+  ASSERT_OK(Put(1, "foo0_0", rnd_str_512kb));
+  ASSERT_OK(Put(1, "foo0_1", rnd_str_512kb));
+  ASSERT_OK(Put(1, "foo1_1", rnd_str));
+  ASSERT_OK(Put(1, "foo1_2", rnd_str));
+  ASSERT_OK(Put(1, "foo1_3", rnd_str));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  ASSERT_OK(Put(1, "foo3_1", rnd_str));
+  ASSERT_OK(Put(1, "foo3_2", rnd_str));
+  ASSERT_OK(Put(1, "foo3_3", rnd_str));
+  ASSERT_OK(Put(1, "foo4", "bar"));
+  dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
+
+  // Following writes should fail as flush failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar", Get(1, "foo"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  ASSERT_GE(1, range_sync_called.load());
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar", Get(1, "foo"));
+}
+
+TEST_F(DBIOFailureTest, CompactSstRangeSyncError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.write_buffer_size = 256 * 1024 * 1024;
+  options.writable_file_max_buffer_size = 128 * 1024;
+  options.bytes_per_sync = 128 * 1024;
+  options.level0_file_num_compaction_trigger = 2;
+  options.target_file_size_base = 256 * 1024 * 1024;
+  options.disable_auto_compactions = true;
+  BlockBasedTableOptions table_options;
+  table_options.filter_policy.reset(NewBloomFilterPolicy(10));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+
+  Random rnd(301);
+  std::string rnd_str =
+      RandomString(&rnd, static_cast<int>(options.bytes_per_sync / 2));
+  std::string rnd_str_512kb = RandomString(&rnd, 512 * 1024);
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  // First 1MB doesn't get range synced
+  ASSERT_OK(Put(1, "foo0_0", rnd_str_512kb));
+  ASSERT_OK(Put(1, "foo0_1", rnd_str_512kb));
+  ASSERT_OK(Put(1, "foo1_1", rnd_str));
+  ASSERT_OK(Put(1, "foo1_2", rnd_str));
+  ASSERT_OK(Put(1, "foo1_3", rnd_str));
+  Flush(1);
+  ASSERT_OK(Put(1, "foo", "bar"));
+  ASSERT_OK(Put(1, "foo3_1", rnd_str));
+  ASSERT_OK(Put(1, "foo3_2", rnd_str));
+  ASSERT_OK(Put(1, "foo3_3", rnd_str));
+  ASSERT_OK(Put(1, "foo4", "bar"));
+  Flush(1);
+  dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
+
+  std::atomic<int> range_sync_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::RangeSync", [&](void* arg) {
+        if (range_sync_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("range sync dummy error");
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(dbfull()->SetOptions(handles_[1],
+                                 {
+                                     {"disable_auto_compactions", "false"},
+                                 }));
+  dbfull()->TEST_WaitForCompact();
+
+  // Following writes should fail as flush failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar", Get(1, "foo"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  ASSERT_GE(1, range_sync_called.load());
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar", Get(1, "foo"));
+}
+
+TEST_F(DBIOFailureTest, FlushSstCloseError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.level0_file_num_compaction_trigger = 4;
+  options.memtable_factory.reset(new SpecialSkipListFactory(2));
+
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+  std::atomic<int> close_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::Close", [&](void* arg) {
+        if (close_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("close dummy error");
+        }
+      });
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  ASSERT_OK(Put(1, "foo1", "bar1"));
+  ASSERT_OK(Put(1, "foo", "bar2"));
+  dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
+
+  // Following writes should fail as flush failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar2", Get(1, "foo"));
+  ASSERT_EQ("bar1", Get(1, "foo1"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar2", Get(1, "foo"));
+  ASSERT_EQ("bar1", Get(1, "foo1"));
+}
+
+TEST_F(DBIOFailureTest, CompactionSstCloseError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.level0_file_num_compaction_trigger = 2;
+  options.disable_auto_compactions = true;
+
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  ASSERT_OK(Put(1, "foo", "bar2"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  ASSERT_OK(Put(1, "foo", "bar3"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  dbfull()->TEST_WaitForCompact();
+
+  std::atomic<int> close_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::Close", [&](void* arg) {
+        if (close_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("close dummy error");
+        }
+      });
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  ASSERT_OK(dbfull()->SetOptions(handles_[1],
+                                 {
+                                     {"disable_auto_compactions", "false"},
+                                 }));
+  dbfull()->TEST_WaitForCompact();
+
+  // Following writes should fail as compaction failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar3", Get(1, "foo"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar3", Get(1, "foo"));
+}
+
+TEST_F(DBIOFailureTest, FlushSstSyncError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.use_fsync = false;
+  options.level0_file_num_compaction_trigger = 4;
+  options.memtable_factory.reset(new SpecialSkipListFactory(2));
+
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+  std::atomic<int> sync_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::Sync", [&](void* arg) {
+        if (sync_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("sync dummy error");
+        }
+      });
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  ASSERT_OK(Put(1, "foo1", "bar1"));
+  ASSERT_OK(Put(1, "foo", "bar2"));
+  dbfull()->TEST_WaitForFlushMemTable(handles_[1]);
+
+  // Following writes should fail as flush failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar2", Get(1, "foo"));
+  ASSERT_EQ("bar1", Get(1, "foo1"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar2", Get(1, "foo"));
+  ASSERT_EQ("bar1", Get(1, "foo1"));
+}
+
+TEST_F(DBIOFailureTest, CompactionSstSyncError) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.error_if_exists = false;
+  options.paranoid_checks = true;
+  options.level0_file_num_compaction_trigger = 2;
+  options.disable_auto_compactions = true;
+  options.use_fsync = false;
+
+  DestroyAndReopen(options);
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Status s;
+
+  ASSERT_OK(Put(1, "foo", "bar"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  ASSERT_OK(Put(1, "foo", "bar2"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  ASSERT_OK(Put(1, "foo", "bar3"));
+  ASSERT_OK(Put(1, "foo2", "bar"));
+  Flush(1);
+  dbfull()->TEST_WaitForCompact();
+
+  std::atomic<int> sync_called(0);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SpecialEnv::SStableFile::Sync", [&](void* arg) {
+        if (sync_called.fetch_add(1) == 0) {
+          Status* st = static_cast<Status*>(arg);
+          *st = Status::IOError("close dummy error");
+        }
+      });
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  ASSERT_OK(dbfull()->SetOptions(handles_[1],
+                                 {
+                                     {"disable_auto_compactions", "false"},
+                                 }));
+  dbfull()->TEST_WaitForCompact();
+
+  // Following writes should fail as compaction failed.
+  ASSERT_NOK(Put(1, "foo2", "bar3"));
+  ASSERT_EQ("bar3", Get(1, "foo"));
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  ASSERT_EQ("bar3", Get(1, "foo"));
+}
+#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif  // ROCKSDB_LITE
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/db_iter.cc

@@ -14,10 +14,10 @@
 #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 "monitoring/perf_context_imp.h"
 #include "port/port.h"
 #include "rocksdb/env.h"
 #include "rocksdb/iterator.h"
@@ -25,9 +25,9 @@
 #include "rocksdb/options.h"
 #include "table/internal_iterator.h"
 #include "util/arena.h"
+#include "util/filename.h"
 #include "util/logging.h"
 #include "util/mutexlock.h"
-#include "util/perf_context_imp.h"
 #include "util/string_util.h"
 
 namespace rocksdb {
@@ -86,6 +86,7 @@ class DBIter: public Iterator {
       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_);
+      PERF_COUNTER_ADD(iter_read_bytes, bytes_read_);
       ResetCounters();
     }
 
@@ -101,33 +102,31 @@ class DBIter: public Iterator {
     uint64_t bytes_read_;
   };
 
-  DBIter(Env* env, const ImmutableCFOptions& ioptions, const Comparator* cmp,
+  DBIter(Env* _env, const ReadOptions& read_options,
+         const ImmutableCFOptions& cf_options, 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,
-         bool total_order_seek = false)
+         uint64_t max_sequential_skip_in_iterations)
       : arena_mode_(arena_mode),
-        env_(env),
-        logger_(ioptions.info_log),
+        env_(_env),
+        logger_(cf_options.info_log),
         user_comparator_(cmp),
-        merge_operator_(ioptions.merge_operator),
+        merge_operator_(cf_options.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),
-        total_order_seek_(total_order_seek),
-        range_del_agg_(ioptions.internal_comparator, s,
+        statistics_(cf_options.statistics),
+        iterate_upper_bound_(read_options.iterate_upper_bound),
+        prefix_same_as_start_(read_options.prefix_same_as_start),
+        pin_thru_lifetime_(read_options.pin_data),
+        total_order_seek_(read_options.total_order_seek),
+        range_del_agg_(cf_options.internal_comparator, s,
                        true /* collapse_deletions */) {
     RecordTick(statistics_, NO_ITERATORS);
-    prefix_extractor_ = ioptions.prefix_extractor;
+    prefix_extractor_ = cf_options.prefix_extractor;
     max_skip_ = max_sequential_skip_in_iterations;
+    max_skippable_internal_keys_ = read_options.max_skippable_internal_keys;
     if (pin_thru_lifetime_) {
       pinned_iters_mgr_.StartPinning();
     }
@@ -160,7 +159,7 @@ class DBIter: public Iterator {
   virtual bool Valid() const override { return valid_; }
   virtual Slice key() const override {
     assert(valid_);
-    return saved_key_.GetKey();
+    return saved_key_.GetUserKey();
   }
   virtual Slice value() const override {
     assert(valid_);
@@ -189,10 +188,7 @@ class DBIter: public Iterator {
     }
     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();
+      return iter_->GetProperty(prop_name, prop);
     } else if (prop_name == "rocksdb.iterator.is-key-pinned") {
       if (valid_) {
         *prop = (pin_thru_lifetime_ && saved_key_.IsKeyPinned()) ? "1" : "0";
@@ -210,6 +206,9 @@ class DBIter: public Iterator {
   virtual void SeekForPrev(const Slice& target) override;
   virtual void SeekToFirst() override;
   virtual void SeekToLast() override;
+  Env* env() { return env_; }
+  void set_sequence(uint64_t s) { sequence_ = s; }
+  void set_valid(bool v) { valid_ = v; }
 
  private:
   void ReverseToForward();
@@ -224,6 +223,7 @@ class DBIter: public Iterator {
   void FindNextUserEntryInternal(bool skipping, bool prefix_check);
   bool ParseKey(ParsedInternalKey* key);
   void MergeValuesNewToOld();
+  bool TooManyInternalKeysSkipped(bool increment = true);
 
   // Temporarily pin the blocks that we encounter until ReleaseTempPinnedData()
   // is called
@@ -249,6 +249,10 @@ class DBIter: public Iterator {
     }
   }
 
+  inline void ResetInternalKeysSkippedCounter() {
+    num_internal_keys_skipped_ = 0;
+  }
+
   const SliceTransform* prefix_extractor_;
   bool arena_mode_;
   Env* const env_;
@@ -256,7 +260,7 @@ class DBIter: public Iterator {
   const Comparator* const user_comparator_;
   const MergeOperator* const merge_operator_;
   InternalIterator* iter_;
-  SequenceNumber const sequence_;
+  SequenceNumber sequence_;
 
   Status status_;
   IterKey saved_key_;
@@ -268,7 +272,8 @@ class DBIter: public Iterator {
   // for prefix seek mode to support prev()
   Statistics* statistics_;
   uint64_t max_skip_;
-  uint64_t version_number_;
+  uint64_t max_skippable_internal_keys_;
+  uint64_t num_internal_keys_skipped_;
   const Slice* iterate_upper_bound_;
   IterKey prefix_start_buf_;
   Slice prefix_start_key_;
@@ -291,9 +296,8 @@ class DBIter: public Iterator {
 inline bool DBIter::ParseKey(ParsedInternalKey* ikey) {
   if (!ParseInternalKey(iter_->key(), ikey)) {
     status_ = Status::Corruption("corrupted internal key in DBIter");
-    Log(InfoLogLevel::ERROR_LEVEL,
-        logger_, "corrupted internal key in DBIter: %s",
-        iter_->key().ToString(true).c_str());
+    ROCKS_LOG_ERROR(logger_, "corrupted internal key in DBIter: %s",
+                    iter_->key().ToString(true).c_str());
     return false;
   } else {
     return true;
@@ -305,6 +309,7 @@ void DBIter::Next() {
 
   // Release temporarily pinned blocks from last operation
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   if (direction_ == kReverse) {
     ReverseToForward();
   } else if (iter_->Valid() && !current_entry_is_merged_) {
@@ -391,9 +396,14 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
       break;
     }
 
+    if (TooManyInternalKeysSkipped()) {
+      return;
+    }
+
     if (ikey.sequence <= sequence_) {
       if (skipping &&
-          user_comparator_->Compare(ikey.user_key, saved_key_.GetKey()) <= 0) {
+          user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) <=
+              0) {
         num_skipped++;  // skip this entry
         PERF_COUNTER_ADD(internal_key_skipped_count, 1);
       } else {
@@ -403,14 +413,14 @@ 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(
+            saved_key_.SetUserKey(
                 ikey.user_key,
                 !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
             skipping = true;
             PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
             break;
           case kTypeValue:
-            saved_key_.SetKey(
+            saved_key_.SetUserKey(
                 ikey.user_key,
                 !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
             if (range_del_agg_.ShouldDelete(
@@ -427,7 +437,7 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
             }
             break;
           case kTypeMerge:
-            saved_key_.SetKey(
+            saved_key_.SetUserKey(
                 ikey.user_key,
                 !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
             if (range_del_agg_.ShouldDelete(
@@ -458,12 +468,13 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
 
       // Here saved_key_ may contain some old key, or the default empty key, or
       // key assigned by some random other method. We don't care.
-      if (user_comparator_->Compare(ikey.user_key, saved_key_.GetKey()) <= 0) {
+      if (user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) <=
+          0) {
         num_skipped++;
       } else {
-        saved_key_.SetKey(
-                ikey.user_key,
-                !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+        saved_key_.SetUserKey(
+            ikey.user_key,
+            !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
         skipping = false;
         num_skipped = 0;
       }
@@ -478,8 +489,8 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
         // We're looking for the next user-key but all we see are the same
         // user-key with decreasing sequence numbers. Fast forward to
         // sequence number 0 and type deletion (the smallest type).
-        AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetKey(), 0,
-                                                       kTypeDeletion));
+        AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetUserKey(),
+                                                       0, kTypeDeletion));
         // Don't set skipping = false because we may still see more user-keys
         // equal to saved_key_.
       } else {
@@ -489,7 +500,7 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
         // many times since our snapshot was taken, not the case when a lot of
         // different keys were inserted after our snapshot was taken.
         AppendInternalKey(&last_key,
-                          ParsedInternalKey(saved_key_.GetKey(), sequence_,
+                          ParsedInternalKey(saved_key_.GetUserKey(), sequence_,
                                             kValueTypeForSeek));
       }
       iter_->Seek(last_key);
@@ -509,8 +520,7 @@ void DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
 //       iter_ points to the next entry (or invalid)
 void DBIter::MergeValuesNewToOld() {
   if (!merge_operator_) {
-    Log(InfoLogLevel::ERROR_LEVEL,
-        logger_, "Options::merge_operator is null.");
+    ROCKS_LOG_ERROR(logger_, "Options::merge_operator is null.");
     status_ = Status::InvalidArgument("merge_operator_ must be set.");
     valid_ = false;
     return;
@@ -531,7 +541,7 @@ void DBIter::MergeValuesNewToOld() {
       continue;
     }
 
-    if (!user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+    if (!user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
       // hit the next user key, stop right here
       break;
     } else if (kTypeDeletion == ikey.type || kTypeSingleDeletion == ikey.type ||
@@ -549,7 +559,7 @@ void DBIter::MergeValuesNewToOld() {
       const Slice val = iter_->value();
       s = MergeHelper::TimedFullMerge(
           merge_operator_, ikey.user_key, &val, merge_context_.GetOperands(),
-          &saved_value_, logger_, statistics_, env_, &pinned_value_);
+          &saved_value_, logger_, statistics_, env_, &pinned_value_, true);
       if (!s.ok()) {
         status_ = s;
       }
@@ -571,9 +581,10 @@ void DBIter::MergeValuesNewToOld() {
   // a deletion marker.
   // feed null as the existing value to the merge operator, such that
   // client can differentiate this scenario and do things accordingly.
-  s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(), nullptr,
-                                  merge_context_.GetOperands(), &saved_value_,
-                                  logger_, statistics_, env_, &pinned_value_);
+  s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetUserKey(),
+                                  nullptr, merge_context_.GetOperands(),
+                                  &saved_value_, logger_, statistics_, env_,
+                                  &pinned_value_, true);
   if (!s.ok()) {
     status_ = s;
   }
@@ -582,6 +593,7 @@ void DBIter::MergeValuesNewToOld() {
 void DBIter::Prev() {
   assert(valid_);
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   if (direction_ == kForward) {
     ReverseToBackward();
   }
@@ -599,8 +611,8 @@ void DBIter::ReverseToForward() {
   if (prefix_extractor_ != nullptr && !total_order_seek_) {
     IterKey last_key;
     last_key.SetInternalKey(ParsedInternalKey(
-        saved_key_.GetKey(), kMaxSequenceNumber, kValueTypeForSeek));
-    iter_->Seek(last_key.GetKey());
+        saved_key_.GetUserKey(), kMaxSequenceNumber, kValueTypeForSeek));
+    iter_->Seek(last_key.GetInternalKey());
   }
   FindNextUserKey();
   direction_ = kForward;
@@ -613,9 +625,9 @@ void DBIter::ReverseToForward() {
 void DBIter::ReverseToBackward() {
   if (prefix_extractor_ != nullptr && !total_order_seek_) {
     IterKey last_key;
-    last_key.SetInternalKey(
-        ParsedInternalKey(saved_key_.GetKey(), 0, kValueTypeForSeekForPrev));
-    iter_->SeekForPrev(last_key.GetKey());
+    last_key.SetInternalKey(ParsedInternalKey(saved_key_.GetUserKey(), 0,
+                                              kValueTypeForSeekForPrev));
+    iter_->SeekForPrev(last_key.GetInternalKey());
   }
   if (current_entry_is_merged_) {
     // Not placed in the same key. Need to call Prev() until finding the
@@ -627,7 +639,9 @@ void DBIter::ReverseToBackward() {
     ParsedInternalKey ikey;
     FindParseableKey(&ikey, kReverse);
     while (iter_->Valid() &&
-           user_comparator_->Compare(ikey.user_key, saved_key_.GetKey()) > 0) {
+           user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) >
+               0) {
+      assert(ikey.sequence != kMaxSequenceNumber);
       if (ikey.sequence > sequence_) {
         PERF_COUNTER_ADD(internal_recent_skipped_count, 1);
       } else {
@@ -641,7 +655,8 @@ void DBIter::ReverseToBackward() {
   if (iter_->Valid()) {
     ParsedInternalKey ikey;
     assert(ParseKey(&ikey));
-    assert(user_comparator_->Compare(ikey.user_key, saved_key_.GetKey()) <= 0);
+    assert(user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) <=
+           0);
   }
 #endif
 
@@ -658,29 +673,36 @@ void DBIter::PrevInternal() {
   ParsedInternalKey ikey;
 
   while (iter_->Valid()) {
-    saved_key_.SetKey(ExtractUserKey(iter_->key()),
-                      !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+    saved_key_.SetUserKey(
+        ExtractUserKey(iter_->key()),
+        !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+
     if (FindValueForCurrentKey()) {
       valid_ = true;
       if (!iter_->Valid()) {
         return;
       }
       FindParseableKey(&ikey, kReverse);
-      if (user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+      if (user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
         FindPrevUserKey();
       }
       if (valid_ && prefix_extractor_ && prefix_same_as_start_ &&
-          prefix_extractor_->Transform(saved_key_.GetKey())
+          prefix_extractor_->Transform(saved_key_.GetUserKey())
                   .compare(prefix_start_key_) != 0) {
         valid_ = false;
       }
       return;
     }
+
+    if (TooManyInternalKeysSkipped(false)) {
+      return;
+    }
+
     if (!iter_->Valid()) {
       break;
     }
     FindParseableKey(&ikey, kReverse);
-    if (user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+    if (user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
       FindPrevUserKey();
     }
   }
@@ -710,7 +732,11 @@ bool DBIter::FindValueForCurrentKey() {
   TempPinData();
   size_t num_skipped = 0;
   while (iter_->Valid() && ikey.sequence <= sequence_ &&
-         user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+         user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
+    if (TooManyInternalKeysSkipped()) {
+      return false;
+    }
+
     // We iterate too much: let's use Seek() to avoid too much key comparisons
     if (num_skipped >= max_skip_) {
       return FindValueForCurrentKeyUsingSeek();
@@ -757,7 +783,7 @@ bool DBIter::FindValueForCurrentKey() {
     }
 
     PERF_COUNTER_ADD(internal_key_skipped_count, 1);
-    assert(user_comparator_->Equal(ikey.user_key, saved_key_.GetKey()));
+    assert(user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey()));
     iter_->Prev();
     ++num_skipped;
     FindParseableKey(&ikey, kReverse);
@@ -775,16 +801,16 @@ bool DBIter::FindValueForCurrentKey() {
       if (last_not_merge_type == kTypeDeletion ||
           last_not_merge_type == kTypeSingleDeletion ||
           last_not_merge_type == kTypeRangeDeletion) {
-        s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(),
-                                        nullptr, merge_context_.GetOperands(),
-                                        &saved_value_, logger_, statistics_,
-                                        env_, &pinned_value_);
+        s = MergeHelper::TimedFullMerge(
+            merge_operator_, saved_key_.GetUserKey(), nullptr,
+            merge_context_.GetOperands(), &saved_value_, logger_, statistics_,
+            env_, &pinned_value_, true);
       } else {
         assert(last_not_merge_type == kTypeValue);
         s = MergeHelper::TimedFullMerge(
-            merge_operator_, saved_key_.GetKey(), &pinned_value_,
+            merge_operator_, saved_key_.GetUserKey(), &pinned_value_,
             merge_context_.GetOperands(), &saved_value_, logger_, statistics_,
-            env_, &pinned_value_);
+            env_, &pinned_value_, true);
       }
       break;
     case kTypeValue:
@@ -808,8 +834,8 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
   // FindValueForCurrentKeyUsingSeek()
   assert(pinned_iters_mgr_.PinningEnabled());
   std::string last_key;
-  AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetKey(), sequence_,
-                                                 kValueTypeForSeek));
+  AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetUserKey(),
+                                                 sequence_, kValueTypeForSeek));
   iter_->Seek(last_key);
   RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
 
@@ -836,7 +862,7 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
   merge_context_.Clear();
   while (
       iter_->Valid() &&
-      user_comparator_->Equal(ikey.user_key, saved_key_.GetKey()) &&
+      user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey()) &&
       ikey.type == kTypeMerge &&
       !range_del_agg_.ShouldDelete(
           ikey, RangeDelAggregator::RangePositioningMode::kBackwardTraversal)) {
@@ -849,17 +875,17 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
 
   Status s;
   if (!iter_->Valid() ||
-      !user_comparator_->Equal(ikey.user_key, saved_key_.GetKey()) ||
+      !user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey()) ||
       ikey.type == kTypeDeletion || ikey.type == kTypeSingleDeletion ||
       range_del_agg_.ShouldDelete(
           ikey, RangeDelAggregator::RangePositioningMode::kBackwardTraversal)) {
-    s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(),
+    s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetUserKey(),
                                     nullptr, merge_context_.GetOperands(),
                                     &saved_value_, logger_, statistics_, env_,
-                                    &pinned_value_);
+                                    &pinned_value_, true);
     // Make iter_ valid and point to saved_key_
     if (!iter_->Valid() ||
-        !user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+        !user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
       iter_->Seek(last_key);
       RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
     }
@@ -871,9 +897,10 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
   }
 
   const Slice& val = iter_->value();
-  s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetKey(), &val,
-                                  merge_context_.GetOperands(), &saved_value_,
-                                  logger_, statistics_, env_, &pinned_value_);
+  s = MergeHelper::TimedFullMerge(merge_operator_, saved_key_.GetUserKey(),
+                                  &val, merge_context_.GetOperands(),
+                                  &saved_value_, logger_, statistics_, env_,
+                                  &pinned_value_, true);
   valid_ = true;
   if (!s.ok()) {
     status_ = s;
@@ -892,7 +919,7 @@ void DBIter::FindNextUserKey() {
   ParsedInternalKey ikey;
   FindParseableKey(&ikey, kForward);
   while (iter_->Valid() &&
-         !user_comparator_->Equal(ikey.user_key, saved_key_.GetKey())) {
+         !user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
     iter_->Next();
     FindParseableKey(&ikey, kForward);
   }
@@ -907,21 +934,27 @@ void DBIter::FindPrevUserKey() {
   ParsedInternalKey ikey;
   FindParseableKey(&ikey, kReverse);
   int cmp;
-  while (iter_->Valid() && ((cmp = user_comparator_->Compare(
-                                 ikey.user_key, saved_key_.GetKey())) == 0 ||
-                            (cmp > 0 && ikey.sequence > sequence_))) {
+  while (iter_->Valid() &&
+         ((cmp = user_comparator_->Compare(ikey.user_key,
+                                           saved_key_.GetUserKey())) == 0 ||
+          (cmp > 0 && ikey.sequence > sequence_))) {
+    if (TooManyInternalKeysSkipped()) {
+      return;
+    }
+
     if (cmp == 0) {
       if (num_skipped >= max_skip_) {
         num_skipped = 0;
         IterKey last_key;
         last_key.SetInternalKey(ParsedInternalKey(
-            saved_key_.GetKey(), kMaxSequenceNumber, kValueTypeForSeek));
-        iter_->Seek(last_key.GetKey());
+            saved_key_.GetUserKey(), kMaxSequenceNumber, kValueTypeForSeek));
+        iter_->Seek(last_key.GetInternalKey());
         RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
       } else {
         ++num_skipped;
       }
     }
+    assert(ikey.sequence != kMaxSequenceNumber);
     if (ikey.sequence > sequence_) {
       PERF_COUNTER_ADD(internal_recent_skipped_count, 1);
     } else {
@@ -932,6 +965,18 @@ void DBIter::FindPrevUserKey() {
   }
 }
 
+bool DBIter::TooManyInternalKeysSkipped(bool increment) {
+  if ((max_skippable_internal_keys_ > 0) &&
+      (num_internal_keys_skipped_ > max_skippable_internal_keys_)) {
+    valid_ = false;
+    status_ = Status::Incomplete("Too many internal keys skipped.");
+    return true;
+  } else if (increment) {
+    num_internal_keys_skipped_++;
+  }
+  return false;
+}
+
 // Skip all unparseable keys
 void DBIter::FindParseableKey(ParsedInternalKey* ikey, Direction direction) {
   while (iter_->Valid() && !ParseKey(ikey)) {
@@ -946,12 +991,13 @@ void DBIter::FindParseableKey(ParsedInternalKey* ikey, Direction direction) {
 void DBIter::Seek(const Slice& target) {
   StopWatch sw(env_, statistics_, DB_SEEK);
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   saved_key_.Clear();
   saved_key_.SetInternalKey(target, sequence_);
 
   {
     PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->Seek(saved_key_.GetKey());
+    iter_->Seek(saved_key_.GetInternalKey());
     range_del_agg_.InvalidateTombstoneMapPositions();
   }
   RecordTick(statistics_, NUMBER_DB_SEEK);
@@ -961,9 +1007,6 @@ void DBIter::Seek(const Slice& target) {
     }
     direction_ = kForward;
     ClearSavedValue();
-    // convert the InternalKey to UserKey in saved_key_ before
-    // passed to FindNextUserEntry
-    saved_key_.Reserve(saved_key_.Size() - 8);
     FindNextUserEntry(false /* not skipping */, prefix_same_as_start_);
     if (!valid_) {
       prefix_start_key_.clear();
@@ -972,6 +1015,7 @@ void DBIter::Seek(const Slice& target) {
       if (valid_) {
         RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
         RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
+        PERF_COUNTER_ADD(iter_read_bytes, key().size() + value().size());
       }
     }
   } else {
@@ -979,14 +1023,15 @@ void DBIter::Seek(const Slice& target) {
   }
 
   if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
-    prefix_start_buf_.SetKey(prefix_start_key_);
-    prefix_start_key_ = prefix_start_buf_.GetKey();
+    prefix_start_buf_.SetUserKey(prefix_start_key_);
+    prefix_start_key_ = prefix_start_buf_.GetUserKey();
   }
 }
 
 void DBIter::SeekForPrev(const Slice& target) {
   StopWatch sw(env_, statistics_, DB_SEEK);
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   saved_key_.Clear();
   // now saved_key is used to store internal key.
   saved_key_.SetInternalKey(target, 0 /* sequence_number */,
@@ -994,7 +1039,7 @@ void DBIter::SeekForPrev(const Slice& target) {
 
   {
     PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->SeekForPrev(saved_key_.GetKey());
+    iter_->SeekForPrev(saved_key_.GetInternalKey());
     range_del_agg_.InvalidateTombstoneMapPositions();
   }
 
@@ -1013,14 +1058,15 @@ void DBIter::SeekForPrev(const Slice& target) {
       if (valid_) {
         RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
         RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
+        PERF_COUNTER_ADD(iter_read_bytes, key().size() + value().size());
       }
     }
   } 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();
+    prefix_start_buf_.SetUserKey(prefix_start_key_);
+    prefix_start_key_ = prefix_start_buf_.GetUserKey();
   }
 }
 
@@ -1032,6 +1078,7 @@ void DBIter::SeekToFirst() {
   }
   direction_ = kForward;
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   ClearSavedValue();
 
   {
@@ -1042,21 +1089,24 @@ void DBIter::SeekToFirst() {
 
   RecordTick(statistics_, NUMBER_DB_SEEK);
   if (iter_->Valid()) {
-    saved_key_.SetKey(ExtractUserKey(iter_->key()),
-                      !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+    saved_key_.SetUserKey(
+        ExtractUserKey(iter_->key()),
+        !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
     FindNextUserEntry(false /* not skipping */, false /* no prefix check */);
     if (statistics_ != nullptr) {
       if (valid_) {
         RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
         RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
+        PERF_COUNTER_ADD(iter_read_bytes, key().size() + value().size());
       }
     }
   } 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();
+    prefix_start_buf_.SetUserKey(
+        prefix_extractor_->Transform(saved_key_.GetUserKey()));
+    prefix_start_key_ = prefix_start_buf_.GetUserKey();
   }
 }
 
@@ -1068,6 +1118,7 @@ void DBIter::SeekToLast() {
   }
   direction_ = kReverse;
   ReleaseTempPinnedData();
+  ResetInternalKeysSkippedCounter();
   ClearSavedValue();
 
   {
@@ -1094,31 +1145,30 @@ void DBIter::SeekToLast() {
     if (valid_) {
       RecordTick(statistics_, NUMBER_DB_SEEK_FOUND);
       RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
+      PERF_COUNTER_ADD(iter_read_bytes, 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();
+    prefix_start_buf_.SetUserKey(
+        prefix_extractor_->Transform(saved_key_.GetUserKey()));
+    prefix_start_key_ = prefix_start_buf_.GetUserKey();
   }
 }
 
-Iterator* NewDBIterator(
-    Env* env, const ImmutableCFOptions& ioptions,
-    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,
-    bool prefix_same_as_start, bool pin_data, bool total_order_seek) {
-  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, total_order_seek);
+Iterator* NewDBIterator(Env* env, const ReadOptions& read_options,
+                        const ImmutableCFOptions& cf_options,
+                        const Comparator* user_key_comparator,
+                        InternalIterator* internal_iter,
+                        const SequenceNumber& sequence,
+                        uint64_t max_sequential_skip_in_iterations) {
+  DBIter* db_iter = new DBIter(env, read_options, cf_options,
+                               user_key_comparator, internal_iter, sequence,
+                               false, max_sequential_skip_in_iterations);
   return db_iter;
 }
 
 ArenaWrappedDBIter::~ArenaWrappedDBIter() { db_iter_->~DBIter(); }
 
-void ArenaWrappedDBIter::SetDBIter(DBIter* iter) { db_iter_ = iter; }
-
 RangeDelAggregator* ArenaWrappedDBIter::GetRangeDelAggregator() {
   return db_iter_->GetRangeDelAggregator();
 }
@@ -1143,28 +1193,67 @@ 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) {
+  if (prop_name == "rocksdb.iterator.super-version-number") {
+    // First try to pass the value returned from inner iterator.
+    if (!db_iter_->GetProperty(prop_name, prop).ok()) {
+      *prop = ToString(sv_number_);
+    }
+    return Status::OK();
+  }
   return db_iter_->GetProperty(prop_name, prop);
 }
-void ArenaWrappedDBIter::RegisterCleanup(CleanupFunction function, void* arg1,
-                                         void* arg2) {
-  db_iter_->RegisterCleanup(function, arg1, arg2);
+
+void ArenaWrappedDBIter::Init(Env* env, const ReadOptions& read_options,
+                              const ImmutableCFOptions& cf_options,
+                              const SequenceNumber& sequence,
+                              uint64_t max_sequential_skip_in_iteration,
+                              uint64_t version_number) {
+  auto mem = arena_.AllocateAligned(sizeof(DBIter));
+  db_iter_ = new (mem)
+      DBIter(env, read_options, cf_options, cf_options.user_comparator, nullptr,
+             sequence, true, max_sequential_skip_in_iteration);
+  sv_number_ = version_number;
+}
+
+Status ArenaWrappedDBIter::Refresh() {
+  if (cfd_ == nullptr || db_impl_ == nullptr) {
+    return Status::NotSupported("Creating renew iterator is not allowed.");
+  }
+  assert(db_iter_ != nullptr);
+  SequenceNumber latest_seq = db_impl_->GetLatestSequenceNumber();
+  uint64_t cur_sv_number = cfd_->GetSuperVersionNumber();
+  if (sv_number_ != cur_sv_number) {
+    Env* env = db_iter_->env();
+    db_iter_->~DBIter();
+    arena_.~Arena();
+    new (&arena_) Arena();
+
+    SuperVersion* sv = cfd_->GetReferencedSuperVersion(db_impl_->mutex());
+    Init(env, read_options_, *(cfd_->ioptions()), latest_seq,
+         sv->mutable_cf_options.max_sequential_skip_in_iterations,
+         cur_sv_number);
+
+    InternalIterator* internal_iter = db_impl_->NewInternalIterator(
+        read_options_, cfd_, sv, &arena_, db_iter_->GetRangeDelAggregator());
+    SetIterUnderDBIter(internal_iter);
+  } else {
+    db_iter_->set_sequence(latest_seq);
+    db_iter_->set_valid(false);
+  }
+  return Status::OK();
 }
 
 ArenaWrappedDBIter* NewArenaWrappedDbIterator(
-    Env* env, const ImmutableCFOptions& ioptions,
-    const Comparator* user_key_comparator, const SequenceNumber& sequence,
+    Env* env, const ReadOptions& read_options,
+    const ImmutableCFOptions& cf_options, 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,
-    bool total_order_seek) {
+    DBImpl* db_impl, ColumnFamilyData* cfd) {
   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, total_order_seek);
-
-  iter->SetDBIter(db_iter);
+  iter->Init(env, read_options, cf_options, sequence,
+             max_sequential_skip_in_iterations, version_number);
+  if (db_impl != nullptr && cfd != nullptr) {
+    iter->StoreRefreshInfo(read_options, db_impl, cfd);
+  }
 
   return iter;
 }

db/db_iter.h

@@ -10,13 +10,14 @@
 #pragma once
 #include <stdint.h>
 #include <string>
+#include "db/db_impl.h"
 #include "db/dbformat.h"
 #include "db/range_del_aggregator.h"
+#include "options/cf_options.h"
 #include "rocksdb/db.h"
 #include "rocksdb/iterator.h"
 #include "util/arena.h"
 #include "util/autovector.h"
-#include "util/cf_options.h"
 
 namespace rocksdb {
 
@@ -27,13 +28,12 @@ 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,
-    bool total_order_seek = false);
+extern Iterator* NewDBIterator(Env* env, const ReadOptions& read_options,
+                               const ImmutableCFOptions& cf_options,
+                               const Comparator* user_key_comparator,
+                               InternalIterator* internal_iter,
+                               const SequenceNumber& sequence,
+                               uint64_t max_sequential_skip_in_iterations);
 
 // 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
@@ -49,10 +49,6 @@ class ArenaWrappedDBIter : public Iterator {
   virtual Arena* GetArena() { return &arena_; }
   virtual RangeDelAggregator* GetRangeDelAggregator();
 
-  // Set the DB Iterator to be wrapped
-
-  virtual void SetDBIter(DBIter* iter);
-
   // Set the internal iterator wrapped inside the DB Iterator. Usually it is
   // a merging iterator.
   virtual void SetIterUnderDBIter(InternalIterator* iter);
@@ -66,22 +62,39 @@ class ArenaWrappedDBIter : public Iterator {
   virtual Slice key() const override;
   virtual Slice value() const override;
   virtual Status status() const override;
+  virtual Status Refresh() override;
 
-  void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
   virtual Status GetProperty(std::string prop_name, std::string* prop) override;
 
+  void Init(Env* env, const ReadOptions& read_options,
+            const ImmutableCFOptions& cf_options,
+            const SequenceNumber& sequence,
+            uint64_t max_sequential_skip_in_iterations,
+            uint64_t version_number);
+
+  void StoreRefreshInfo(const ReadOptions& read_options, DBImpl* db_impl,
+                        ColumnFamilyData* cfd) {
+    read_options_ = read_options;
+    db_impl_ = db_impl;
+    cfd_ = cfd;
+  }
+
  private:
   DBIter* db_iter_;
   Arena arena_;
+  uint64_t sv_number_;
+  ColumnFamilyData* cfd_ = nullptr;
+  DBImpl* db_impl_ = nullptr;
+  ReadOptions read_options_;
 };
 
 // Generate the arena wrapped iterator class.
+// `db_impl` and `cfd` are used for reneweal. If left null, renewal will not
+// be supported.
 extern ArenaWrappedDBIter* NewArenaWrappedDbIterator(
-    Env* env, const ImmutableCFOptions& options,
-    const Comparator* user_key_comparator, const SequenceNumber& sequence,
+    Env* env, const ReadOptions& read_options,
+    const ImmutableCFOptions& cf_options, 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,
-    bool total_order_seek = false);
+    DBImpl* db_impl = nullptr, ColumnFamilyData* cfd = nullptr);
 
 }  // namespace rocksdb

db/db_iterator_test.cc

@@ -22,6 +22,34 @@ class DBIteratorTest : public DBTestBase {
   DBIteratorTest() : DBTestBase("/db_iterator_test") {}
 };
 
+class FlushBlockEveryKeyPolicy : public FlushBlockPolicy {
+ public:
+  virtual bool Update(const Slice& key, const Slice& value) override {
+    if (!start_) {
+      start_ = true;
+      return false;
+    }
+    return true;
+  }
+ private:
+  bool start_ = false;
+};
+
+class FlushBlockEveryKeyPolicyFactory : public FlushBlockPolicyFactory {
+ public:
+  explicit FlushBlockEveryKeyPolicyFactory() {}
+
+  const char* Name() const override {
+    return "FlushBlockEveryKeyPolicyFactory";
+  }
+
+  FlushBlockPolicy* NewFlushBlockPolicy(
+    const BlockBasedTableOptions& table_options,
+    const BlockBuilder& data_block_builder) const override {
+    return new FlushBlockEveryKeyPolicy;
+  }
+};
+
 TEST_F(DBIteratorTest, IteratorProperty) {
   // The test needs to be changed if kPersistedTier is supported in iterator.
   Options options = CurrentOptions();
@@ -944,11 +972,11 @@ TEST_F(DBIteratorTest, DBIteratorBoundTest) {
     ASSERT_TRUE(iter->Valid());
     ASSERT_EQ(iter->key().compare(("b1")), 0);
 
-    perf_context.Reset();
+    get_perf_context()->Reset();
     iter->Next();
 
     ASSERT_TRUE(iter->Valid());
-    ASSERT_EQ(static_cast<int>(perf_context.internal_delete_skipped_count), 2);
+    ASSERT_EQ(static_cast<int>(get_perf_context()->internal_delete_skipped_count), 2);
 
     // now testing with iterate_bound
     Slice prefix("c");
@@ -956,7 +984,7 @@ TEST_F(DBIteratorTest, DBIteratorBoundTest) {
 
     iter.reset(db_->NewIterator(ro));
 
-    perf_context.Reset();
+    get_perf_context()->Reset();
 
     iter->Seek("b");
     ASSERT_TRUE(iter->Valid());
@@ -971,16 +999,61 @@ TEST_F(DBIteratorTest, DBIteratorBoundTest) {
     // even though the key is deleted
     // hence internal_delete_skipped_count should be 0
     ASSERT_TRUE(!iter->Valid());
-    ASSERT_EQ(static_cast<int>(perf_context.internal_delete_skipped_count), 0);
+    ASSERT_EQ(static_cast<int>(get_perf_context()->internal_delete_skipped_count), 0);
   }
 }
 
+TEST_F(DBIteratorTest, DBIteratorBoundOptimizationTest) {
+  int upper_bound_hits = 0;
+  Options options = CurrentOptions();
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "BlockBasedTable::BlockEntryIteratorState::KeyReachedUpperBound",
+      [&upper_bound_hits](void* arg) {
+        assert(arg != nullptr);
+        upper_bound_hits += (*static_cast<bool*>(arg) ? 1 : 0);
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  options.env = env_;
+  options.create_if_missing = true;
+  options.prefix_extractor = nullptr;
+  BlockBasedTableOptions table_options;
+  table_options.flush_block_policy_factory =
+    std::make_shared<FlushBlockEveryKeyPolicyFactory>();
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+
+  DestroyAndReopen(options);
+  ASSERT_OK(Put("foo1", "bar1"));
+  ASSERT_OK(Put("foo2", "bar2"));
+  ASSERT_OK(Put("foo4", "bar4"));
+  ASSERT_OK(Flush());
+
+  Slice ub("foo3");
+  ReadOptions ro;
+  ro.iterate_upper_bound = &ub;
+
+  std::unique_ptr<Iterator> iter(db_->NewIterator(ro));
+
+  iter->Seek("foo");
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("foo1")), 0);
+  ASSERT_EQ(upper_bound_hits, 0);
+
+  iter->Next();
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("foo2")), 0);
+  ASSERT_EQ(upper_bound_hits, 0);
+
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+  ASSERT_EQ(upper_bound_hits, 1);
+}
 // TODO(3.13): fix the issue of Seek() + Prev() which might not necessary
 //             return the biggest key which is smaller than the seek key.
 TEST_F(DBIteratorTest, PrevAfterAndNextAfterMerge) {
   Options options;
   options.create_if_missing = true;
   options.merge_operator = MergeOperators::CreatePutOperator();
+  options.env = env_;
   DestroyAndReopen(options);
 
   // write three entries with different keys using Merge()
@@ -1646,12 +1719,15 @@ TEST_F(DBIteratorTest, IteratorWithLocalStatistics) {
 
   std::vector<port::Thread> threads;
   std::function<void()> reader_func_next = [&]() {
+    SetPerfLevel(kEnableCount);
+    get_perf_context()->Reset();
     Iterator* iter = db_->NewIterator(ReadOptions());
 
     iter->SeekToFirst();
     // Seek will bump ITER_BYTES_READ
-    total_bytes += iter->key().size();
-    total_bytes += iter->value().size();
+    uint64_t bytes = 0;
+    bytes += iter->key().size();
+    bytes += iter->value().size();
     while (true) {
       iter->Next();
       total_next++;
@@ -1660,20 +1736,25 @@ TEST_F(DBIteratorTest, IteratorWithLocalStatistics) {
         break;
       }
       total_next_found++;
-      total_bytes += iter->key().size();
-      total_bytes += iter->value().size();
+      bytes += iter->key().size();
+      bytes += iter->value().size();
     }
 
     delete iter;
+    ASSERT_EQ(bytes, get_perf_context()->iter_read_bytes);
+    SetPerfLevel(kDisable);
+    total_bytes += bytes;
   };
 
   std::function<void()> reader_func_prev = [&]() {
+    SetPerfLevel(kEnableCount);
     Iterator* iter = db_->NewIterator(ReadOptions());
 
     iter->SeekToLast();
     // Seek will bump ITER_BYTES_READ
-    total_bytes += iter->key().size();
-    total_bytes += iter->value().size();
+    uint64_t bytes = 0;
+    bytes += iter->key().size();
+    bytes += iter->value().size();
     while (true) {
       iter->Prev();
       total_prev++;
@@ -1682,11 +1763,14 @@ TEST_F(DBIteratorTest, IteratorWithLocalStatistics) {
         break;
       }
       total_prev_found++;
-      total_bytes += iter->key().size();
-      total_bytes += iter->value().size();
+      bytes += iter->key().size();
+      bytes += iter->value().size();
     }
 
     delete iter;
+    ASSERT_EQ(bytes, get_perf_context()->iter_read_bytes);
+    SetPerfLevel(kDisable);
+    total_bytes += bytes;
   };
 
   for (int i = 0; i < 10; i++) {
@@ -1813,7 +1897,7 @@ TEST_F(DBIteratorTest, DBIteratorSkipRecentDuplicatesTest) {
 #endif
 
   // Seek iterator to a smaller key.
-  perf_context.Reset();
+  get_perf_context()->Reset();
   iter->Seek("a");
   ASSERT_TRUE(iter->Valid());
   EXPECT_EQ("b", iter->key().ToString());
@@ -1821,21 +1905,80 @@ TEST_F(DBIteratorTest, DBIteratorSkipRecentDuplicatesTest) {
 
   // Check that the seek didn't do too much work.
   // Checks are not tight, just make sure that everything is well below 100.
-  EXPECT_LT(perf_context.internal_key_skipped_count, 4);
-  EXPECT_LT(perf_context.internal_recent_skipped_count, 8);
-  EXPECT_LT(perf_context.seek_on_memtable_count, 10);
-  EXPECT_LT(perf_context.next_on_memtable_count, 10);
-  EXPECT_LT(perf_context.prev_on_memtable_count, 10);
+  EXPECT_LT(get_perf_context()->internal_key_skipped_count, 4);
+  EXPECT_LT(get_perf_context()->internal_recent_skipped_count, 8);
+  EXPECT_LT(get_perf_context()->seek_on_memtable_count, 10);
+  EXPECT_LT(get_perf_context()->next_on_memtable_count, 10);
+  EXPECT_LT(get_perf_context()->prev_on_memtable_count, 10);
 
   // Check that iterator did something like what we expect.
-  EXPECT_EQ(perf_context.internal_delete_skipped_count, 0);
-  EXPECT_EQ(perf_context.internal_merge_count, 0);
-  EXPECT_GE(perf_context.internal_recent_skipped_count, 2);
-  EXPECT_GE(perf_context.seek_on_memtable_count, 2);
+  EXPECT_EQ(get_perf_context()->internal_delete_skipped_count, 0);
+  EXPECT_EQ(get_perf_context()->internal_merge_count, 0);
+  EXPECT_GE(get_perf_context()->internal_recent_skipped_count, 2);
+  EXPECT_GE(get_perf_context()->seek_on_memtable_count, 2);
   EXPECT_EQ(1, options.statistics->getTickerCount(
                  NUMBER_OF_RESEEKS_IN_ITERATION));
 }
 
+TEST_F(DBIteratorTest, Refresh) {
+  ASSERT_OK(Put("x", "y"));
+
+  std::unique_ptr<Iterator> iter(db_->NewIterator(ReadOptions()));
+  iter->Seek(Slice("a"));
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("x")), 0);
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+
+  ASSERT_OK(Put("c", "d"));
+
+  iter->Seek(Slice("a"));
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("x")), 0);
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+
+  iter->Refresh();
+
+  iter->Seek(Slice("a"));
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("c")), 0);
+  iter->Next();
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("x")), 0);
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+
+  dbfull()->Flush(FlushOptions());
+
+  ASSERT_OK(Put("m", "n"));
+
+  iter->Seek(Slice("a"));
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("c")), 0);
+  iter->Next();
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("x")), 0);
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+
+  iter->Refresh();
+
+  iter->Seek(Slice("a"));
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("c")), 0);
+  iter->Next();
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("m")), 0);
+  iter->Next();
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(iter->key().compare(Slice("x")), 0);
+  iter->Next();
+  ASSERT_FALSE(iter->Valid());
+
+  iter.reset();
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {