Disable error as warning

The previuos change that use gcc-5 set USE_SSE to wrong flag by mistake. Fix it.

.travis.yml

@@ -34,7 +34,7 @@ before_script:
 # as EnvPosixTest::AllocateTest expects within the Travis OpenVZ environment.
 script:
   - if [[ "${TRAVIS_OS_NAME}" == 'linux' ]]; then OPT=-DTRAVIS CLANG_FORMAT_DIFF=/tmp/clang-format-diff.py make format || true; fi
-  - OPT=-DTRAVIS V=1 make -j4 check && OPT=-DTRAVIS V=1 make clean jclean rocksdbjava jtest
+  - OPT=-DTRAVIS V=1 make -j4 check && OPT=-DTRAVIS V=1 make clean jclean rocksdbjava jtest && make clean && OPT="-DTRAVIS -DROCKSDB_LITE" V=1 make -j4 check
 
 notifications:
     email:

CMakeLists.txt

@@ -14,10 +14,15 @@
 # 3. Run cmake to generate project files for Windows, add more options to enable required third-party libraries.
 #    See thirdparty.inc for more information.
 #        sample command: cmake -G "Visual Studio 12 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 ..
-# 4. Then build the project in debug mode (you may want to add /m:<N> flag to run msbuild in <N> parallel threads)
-#        msbuild ALL_BUILD.vcxproj
+# 4. Then build the project in debug mode (you may want to add /m[:<N>] flag to run msbuild in <N> parallel threads
+#                                          or simply /m ot use all avail cores)
+#        msbuild rocksdb.sln
+#
+#        rocksdb.sln build features exclusions of test only code in Release. If you build ALL_BUILD then everything
+#        will be attempted but test only code does not build in Release mode.
+#
 # 5. And release mode (/m[:<N>] is also supported)
-#        msbuild ALL_BUILD.vcxproj /p:Configuration=Release
+#        msbuild rocksdb.sln /p:Configuration=Release
 #
 
 cmake_minimum_required(VERSION 2.6)
@@ -31,7 +36,14 @@ string(REGEX REPLACE "(..)/(..)/..(..).*" "\\1/\\2/\\3" DATE ${DATE})
 string(REGEX REPLACE "(..):(.....).*" " \\1:\\2" TIME ${TIME})
 string(CONCAT GIT_DATE_TIME ${DATE} ${TIME})
 
-execute_process(COMMAND $ENV{COMSPEC} " /C git rev-parse HEAD 2>nil" OUTPUT_VARIABLE GIT_SHA)
+find_package(Git)
+
+if (GIT_FOUND AND EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
+    execute_process(COMMAND $ENV{COMSPEC} /C ${GIT_EXECUTABLE} -C ${CMAKE_CURRENT_SOURCE_DIR} rev-parse HEAD OUTPUT_VARIABLE GIT_SHA)
+else()
+    set(GIT_SHA 0)
+endif()
+
 string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA ${GIT_SHA})
 
 set(BUILD_VERSION_CC ${CMAKE_CURRENT_SOURCE_DIR}/util/build_version.cc)
@@ -46,7 +58,7 @@ add_custom_command(OUTPUT ${BUILD_VERSION_CC}
 add_custom_target(GenerateBuildVersion DEPENDS ${BUILD_VERSION_CC})
 
 set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo  /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /WX /wd4018 /wd4100 /wd4101 /wd4127 /wd4189 /wd4200 /wd4244 /wd4267 /wd4296 /wd4305 /wd4307 /wd4309 /wd4512 /wd4701 /wd4702 /wd4800 /wd4804 /wd4996")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /WX /wd4127 /wd4244 /wd4267 /wd4800 /wd4804 /wd4996")
 
 # Used to run CI build and tests so we can run faster
 set(OPTIMIZE_DEBUG_DEFAULT 0)        # Debug build is unoptimized by default use -DOPTDBG=1 to optimize
@@ -83,6 +95,7 @@ set(LIBS ${ROCKSDB_LIBS} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
 
 add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
 
+# Main library source code
 set(SOURCES
         db/builder.cc
         db/c.cc
@@ -130,6 +143,9 @@ set(SOURCES
         db/write_batch_base.cc
         db/write_controller.cc
         db/write_thread.cc
+        memtable/hash_cuckoo_rep.cc
+        memtable/hash_linklist_rep.cc
+        memtable/hash_skiplist_rep.cc
         port/stack_trace.cc
         port/win/env_win.cc
         port/win/port_win.cc
@@ -163,6 +179,7 @@ set(SOURCES
         table/plain_table_reader.cc
         table/table_properties.cc
         table/two_level_iterator.cc
+        tools/sst_dump_tool.cc
         tools/dump/db_dump_tool.cc
         util/arena.cc
         util/auto_roll_logger.cc
@@ -175,7 +192,6 @@ set(SOURCES
         util/crc32c.cc
         util/db_info_dumper.cc
         util/delete_scheduler_impl.cc
-        util/db_test_util.cc
         util/dynamic_bloom.cc
         util/env.cc
         util/env_hdfs.cc
@@ -184,14 +200,11 @@ set(SOURCES
         util/file_reader_writer.cc
         util/filter_policy.cc
         util/hash.cc
-        util/hash_cuckoo_rep.cc
-        util/hash_linklist_rep.cc
-        util/hash_skiplist_rep.cc
         util/histogram.cc
         util/instrumented_mutex.cc
         util/iostats_context.cc
-        util/ldb_cmd.cc
-        util/ldb_tool.cc
+        tools/ldb_cmd.cc
+        tools/ldb_tool.cc
         util/logging.cc
         util/log_buffer.cc
         util/memenv.cc
@@ -202,12 +215,12 @@ set(SOURCES
         util/options_builder.cc
         util/options_helper.cc
         util/options_parser.cc
+        util/options_sanity_check.cc
         util/perf_context.cc
         util/perf_level.cc
         util/rate_limiter.cc
         util/skiplistrep.cc
         util/slice.cc
-        util/sst_dump_tool.cc
         util/statistics.cc
         util/status.cc
         util/status_message.cc
@@ -218,7 +231,6 @@ set(SOURCES
         util/thread_local.cc
         util/thread_status_impl.cc
         util/thread_status_updater.cc
-        util/thread_status_updater_debug.cc
         util/thread_status_util.cc
         util/thread_status_util_debug.cc
         util/vectorrep.cc
@@ -232,6 +244,7 @@ set(SOURCES
         utilities/flashcache/flashcache.cc
         utilities/geodb/geodb_impl.cc
         utilities/leveldb_options/leveldb_options.cc
+        utilities/memory/memory_util.cc
         utilities/merge_operators/string_append/stringappend.cc
         utilities/merge_operators/string_append/stringappend2.cc
         utilities/merge_operators/put.cc
@@ -252,6 +265,13 @@ set(SOURCES
         utilities/write_batch_with_index/write_batch_with_index_internal.cc
 )
 
+# For test util library that is build only in DEBUG mode
+# and linked to tests. Add test only code that is not #ifdefed for Release here.
+set(TESTUTIL_SOURCE
+    db/db_test_util.cc
+    util/thread_status_updater_debug.cc
+)
+
 add_library(rocksdblib${ARTIFACT_SUFFIX} ${SOURCES})
 set_target_properties(rocksdblib${ARTIFACT_SUFFIX} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/rocksdblib${ARTIFACT_SUFFIX}.pdb")
 add_dependencies(rocksdblib${ARTIFACT_SUFFIX} GenerateBuildVersion)
@@ -266,6 +286,7 @@ set(APPS
         db/memtablerep_bench.cc
         table/table_reader_bench.cc
         tools/db_stress.cc
+        tools/write_stress.cc
         tools/db_repl_stress.cc
         tools/sst_dump.cc
         tools/dump/rocksdb_dump.cc
@@ -294,6 +315,7 @@ set(TESTS
         db/db_universal_compaction_test.cc
         db/db_wal_test.cc
         db/db_tailing_iter_test.cc
+        db/db_table_properties_test.cc
         db/dbformat_test.cc
         db/deletefile_test.cc
         db/fault_injection_test.cc
@@ -302,6 +324,7 @@ set(TESTS
         db/flush_job_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
@@ -326,7 +349,9 @@ set(TESTS
         table/merger_test.cc
         table/table_test.cc
         tools/db_sanity_test.cc
+        tools/ldb_cmd_test.cc
         tools/reduce_levels_test.cc
+        tools/sst_dump_test.cc
         util/arena_test.cc
         util/autovector_test.cc
         util/auto_roll_logger_test.cc
@@ -341,14 +366,11 @@ set(TESTS
         util/file_reader_writer_test.cc
         util/heap_test.cc
         util/histogram_test.cc
-        util/ldb_cmd_test.cc
-        util/manual_compaction_test.cc
         util/memenv_test.cc
         util/mock_env_test.cc
         util/options_test.cc
         util/rate_limiter_test.cc
         util/slice_transform_test.cc
-        util/sst_dump_test.cc
         util/thread_list_test.cc
         util/thread_local_test.cc
         utilities/backupable/backupable_db_test.cc
@@ -356,6 +378,7 @@ set(TESTS
         utilities/document/document_db_test.cc
         utilities/document/json_document_test.cc
         utilities/geodb/geodb_test.cc
+        utilities/memory/memory_test.cc
         utilities/merge_operators/string_append/stringappend_test.cc
         utilities/redis/redis_lists_test.cc
         utilities/spatialdb/spatial_db_test.cc
@@ -366,7 +389,7 @@ set(TESTS
         utilities/write_batch_with_index/write_batch_with_index_test.cc
 )
 
-set(EXES ${APPS} ${TESTS})
+set(EXES ${APPS})
 
 foreach(sourcefile ${EXES})
     string(REPLACE ".cc" "" exename ${sourcefile})
@@ -375,12 +398,42 @@ foreach(sourcefile ${EXES})
     target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS})
 endforeach(sourcefile ${EXES})
 
-# C executables must link to a shared object
-set(C_EXES ${C_TESTS})
+# test utilities are only build in debug
+set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
+add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
+set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
+set_target_properties(${TESTUTILLIB}
+      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+      )
 
-foreach(sourcefile ${C_EXES})
+# Tests are excluded from Release builds
+set(TEST_EXES ${TESTS})
+
+foreach(sourcefile ${TEST_EXES})
+    string(REPLACE ".cc" "" exename ${sourcefile})
+    string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
+    add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
+    set_target_properties(${exename}${ARTIFACT_SUFFIX}
+      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+      )
+    target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS} testutillib${ARTIFACT_SUFFIX})
+endforeach(sourcefile ${TEST_EXES})
+
+# C executables must link to a shared object
+set(C_TEST_EXES ${C_TESTS})
+
+foreach(sourcefile ${C_TEST_EXES})
     string(REPLACE ".c" "" exename ${sourcefile})
     string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
     add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
-    target_link_libraries(${exename}${ARTIFACT_SUFFIX} rocksdb${ARTIFACT_SUFFIX})
-endforeach(sourcefile ${C_TESTS})
+    set_target_properties(${exename}${ARTIFACT_SUFFIX}
+      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+      )
+    target_link_libraries(${exename}${ARTIFACT_SUFFIX} rocksdb${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX})
+endforeach(sourcefile ${C_TEST_EXES})

HISTORY.md

@@ -1,5 +1,16 @@
 # Rocksdb Change Log
 
+## 4.2.0 (11/9/2015)
+### New Features
+* Introduce CreateLoggerFromOptions(), this function create a Logger for provided DBOptions. 
+* Add GetAggregatedIntProperty(), which returns the sum of the GetIntProperty of all the column families. 
+* Add MemoryUtil in rocksdb/utilities/memory.h.  It currently offers a way to get the memory usage by type from a list rocksdb instances.
+### Public API Changes
+* CompactionFilter::Context includes information of Column Family ID
+* The need-compaction hint given by TablePropertiesCollector::NeedCompact() will be persistent and recoverable after DB recovery. This introduces a breaking format change. If you use this experimental feature, including NewCompactOnDeletionCollectorFactory() in the new version, you may not be able to directly downgrade the DB back to version 4.0 or lower.
+* TablePropertiesCollectorFactory::CreateTablePropertiesCollector() now takes an option Context, containing the information of column family ID for the file being written.
+* Remove DefaultCompactionFilterFactory.
+
 ## 4.1.0 (10/8/2015)
 ### New Features
 * Added single delete operation as a more efficient way to delete keys that have not been overwritten.

INSTALL.md

@@ -1,19 +1,24 @@
 ## Compilation
 
+**Important**: If you plan to run RocksDB in production, don't compile using default 
+`make` or `make all`. That will compile RocksDB in debug mode, which is much slower
+than release mode.
+
 RocksDB's library should be able to compile without any dependency installed,
 although we recommend installing some compression libraries (see below).
 We do depend on newer gcc/clang with C++11 support.
 
 There are few options when compiling RocksDB:
 
-* [recommended] `make static_lib` will compile librocksdb.a, RocksDB static library.
+* [recommended] `make static_lib` will compile librocksdb.a, RocksDB static library. Compiles static library in release mode.
 
-* `make shared_lib` will compile librocksdb.so, RocksDB shared library.
+* `make shared_lib` will compile librocksdb.so, RocksDB shared library. Compiles shared library in release mode.
 
-* `make check` will compile and run all the unit tests
+* `make check` will compile and run all the unit tests. `make check` will compile RocksDB in debug mode.
 
 * `make all` will compile our static library, and all our tools and unit tests. Our tools
-depend on gflags. You will need to have gflags installed to run `make all`.
+depend on gflags. You will need to have gflags installed to run `make all`. This will compile RocksDB in debug mode. Don't
+use binaries compiled by `make all` in production.
 
 * By default the binary we produce is optimized for the platform you're compiling on
 (-march=native). If you want to build a portable binary, add 'PORTABLE=1' before
@@ -82,4 +87,5 @@ your make commands, like this: `PORTABLE=1 make static_lib`
   * Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define two important pre-processing macros: `ROCKSDB_LITE` and `IOS_CROSS_COMPILE`.
 
 * **Windows**:
+  * For building with MS Visual Studio 13 you will need Update 4 installed.
   * Read and follow the instructions at CMakeLists.txt

Makefile

@@ -33,12 +33,22 @@ quoted_perl_command = $(subst ','\'',$(perl_command))
 # with debug level 0. To compile with level 0, run `make shared_lib`,
 # `make install-shared`, `make static_lib`, `make install-static` or
 # `make install`
-DEBUG_LEVEL=1
+
+# Set the default DEBUG_LEVEL to 1
+DEBUG_LEVEL?=1
 
 ifeq ($(MAKECMDGOALS),dbg)
 	DEBUG_LEVEL=2
 endif
 
+ifeq ($(MAKECMDGOALS),clean)
+	DEBUG_LEVEL=0
+endif
+
+ifeq ($(MAKECMDGOALS),release)
+	DEBUG_LEVEL=0
+endif
+
 ifeq ($(MAKECMDGOALS),shared_lib)
 	DEBUG_LEVEL=0
 endif
@@ -84,6 +94,8 @@ endif
 ifeq ($(DEBUG_LEVEL),0)
 OPT += -DNDEBUG
 DISABLE_WARNING_AS_ERROR=1
+else
+$(warning Warning: Compiling in debug mode. Don't use the resulting binary in production)
 endif
 
 #-----------------------------------------------
@@ -232,6 +244,7 @@ TESTS = \
 	db_tailing_iter_test \
 	db_universal_compaction_test \
 	db_wal_test \
+	db_table_properties_test \
 	block_hash_index_test \
 	autovector_test \
 	column_family_test \
@@ -263,6 +276,7 @@ TESTS = \
 	mock_env_test \
 	memtable_list_test \
 	merge_helper_test \
+	memory_test \
 	merge_test \
 	merger_test \
 	redis_test \
@@ -319,6 +333,7 @@ TOOLS = \
 	sst_dump \
 	db_sanity_test \
 	db_stress \
+	write_stress \
 	ldb \
 	db_repl_stress \
 	rocksdb_dump \
@@ -326,10 +341,14 @@ TOOLS = \
 
 BENCHMARKS = db_bench table_reader_bench cache_bench memtablerep_bench
 
-# The library name is configurable since we are maintaining libraries of both
-# debug/release mode.
+# if user didn't config LIBNAME, set the default
 ifeq ($(LIBNAME),)
+# we should only run rocksdb in production with DEBUG_LEVEL 0
+ifeq ($(DEBUG_LEVEL),0)
         LIBNAME=librocksdb
+else
+        LIBNAME=librocksdb_debug
+endif
 endif
 LIBRARY = ${LIBNAME}.a
 
@@ -375,7 +394,7 @@ $(SHARED3): $(SHARED4)
 endif
 
 $(SHARED4):
-	$(CXX) $(PLATFORM_SHARED_LDFLAGS)$(SHARED3) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) $(LIB_SOURCES) \
+	$(CXX) $(PLATFORM_SHARED_LDFLAGS)$(SHARED3) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) $(LIB_SOURCES) $(TOOL_SOURCES) \
 		$(LDFLAGS) -o $@
 
 endif  # PLATFORM_SHARED_EXT
@@ -399,7 +418,7 @@ dbg: $(LIBRARY) $(BENCHMARKS) tools $(TESTS)
 # creates static library and programs
 release:
 	$(MAKE) clean
-	OPT="-DNDEBUG -O2" $(MAKE) static_lib tools db_bench
+	DEBUG_LEVEL=0 $(MAKE) static_lib tools db_bench
 
 coverage:
 	$(MAKE) clean
@@ -548,8 +567,10 @@ check: all
 	      echo "===== Running $$t"; ./$$t || exit 1; done;          \
 	fi
 	rm -rf $(TMPD)
+ifeq ($(filter -DROCKSDB_LITE,$(OPT)),)
 	python tools/ldb_test.py
 	sh tools/rocksdb_dump_test.sh
+endif
 
 check_some: $(SUBSET) ldb_tests
 	for t in $(SUBSET); do echo "===== Running $$t"; ./$$t || exit 1; done
@@ -561,12 +582,12 @@ ldb_tests: ldb
 crash_test: whitebox_crash_test blackbox_crash_test
 
 blackbox_crash_test: db_stress
-	python -u tools/db_crashtest.py -s
-	python -u tools/db_crashtest.py
+	python -u tools/db_crashtest.py --simple blackbox 
+	python -u tools/db_crashtest.py blackbox
 
 whitebox_crash_test: db_stress
-	python -u tools/db_crashtest2.py -s
-	python -u tools/db_crashtest2.py
+	python -u tools/db_crashtest.py --simple whitebox
+	python -u tools/db_crashtest.py whitebox
 
 asan_check:
 	$(MAKE) clean
@@ -607,7 +628,7 @@ unity.a: unity.o
 	$(AM_V_at)$(AR) $(ARFLAGS) $@ unity.o
 
 # try compiling db_test with unity
-unity_test: db/db_test.o util/db_test_util.o $(TESTHARNESS) unity.a
+unity_test: db/db_test.o db/db_test_util.o $(TESTHARNESS) unity.a
 	$(AM_LINK)
 	./unity_test
 
@@ -653,6 +674,9 @@ block_hash_index_test: table/block_hash_index_test.o $(LIBOBJECTS) $(TESTHARNESS
 db_stress: tools/db_stress.o $(LIBOBJECTS) $(TESTUTIL)
 	$(AM_LINK)
 
+write_stress: tools/write_stress.o $(LIBOBJECTS) $(TESTUTIL)
+	$(AM_LINK)
+
 db_sanity_test: tools/db_sanity_test.o $(LIBOBJECTS) $(TESTUTIL)
 	$(AM_LINK)
 
@@ -707,34 +731,37 @@ crc32c_test: util/crc32c_test.o $(LIBOBJECTS) $(TESTHARNESS)
 slice_transform_test: util/slice_transform_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_test: db/db_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_test: db/db_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_log_iter_test: db/db_log_iter_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_log_iter_test: db/db_log_iter_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_compaction_filter_test: db/db_compaction_filter_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_compaction_filter_test: db/db_compaction_filter_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_compaction_test: db/db_compaction_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_compaction_test: db/db_compaction_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_dynamic_level_test: db/db_dynamic_level_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_dynamic_level_test: db/db_dynamic_level_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_inplace_update_test: db/db_inplace_update_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_inplace_update_test: db/db_inplace_update_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_tailing_iter_test: db/db_tailing_iter_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_tailing_iter_test: db/db_tailing_iter_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 db_iter_test: db/db_iter_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_universal_compaction_test: db/db_universal_compaction_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_universal_compaction_test: db/db_universal_compaction_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-db_wal_test: db/db_wal_test.o util/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+db_wal_test: db/db_wal_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
+db_table_properties_test: db/db_table_properties_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 log_write_bench: util/log_write_bench.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -860,6 +887,9 @@ write_controller_test: db/write_controller_test.o $(LIBOBJECTS) $(TESTHARNESS)
 merge_helper_test: db/merge_helper_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
+memory_test: utilities/memory/memory_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(AM_LINK)
+
 merge_test: db/merge_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
@@ -902,7 +932,7 @@ options_test: util/options_test.o $(LIBOBJECTS) $(TESTHARNESS)
 event_logger_test: util/event_logger_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-sst_dump_test: util/sst_dump_test.o $(LIBOBJECTS) $(TESTHARNESS)
+sst_dump_test: tools/sst_dump_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 memenv_test : util/memenv_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -914,7 +944,7 @@ optimistic_transaction_test: utilities/transactions/optimistic_transaction_test.
 mock_env_test : util/mock_env_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
-manual_compaction_test: util/manual_compaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
+manual_compaction_test: db/manual_compaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 filelock_test: util/filelock_test.o $(LIBOBJECTS) $(TESTHARNESS)
@@ -938,7 +968,7 @@ transaction_test: utilities/transactions/transaction_test.o $(LIBOBJECTS) $(TEST
 sst_dump: tools/sst_dump.o $(LIBOBJECTS)
 	$(AM_LINK)
 
-ldb_cmd_test: util/ldb_cmd_test.o $(LIBOBJECTS) $(TESTHARNESS)
+ldb_cmd_test: tools/ldb_cmd_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)
 
 ldb: tools/ldb.o $(LIBOBJECTS)
@@ -1037,8 +1067,6 @@ liblz4.a:
 java_static_libobjects = $(patsubst %,jls/%,$(LIBOBJECTS))
 CLEAN_FILES += jls
 
-$(java_libobjects): jl/%.o: %.cc
-	$(AM_V_CC)mkdir -p $(@D) && $(CXX) $(CXXFLAGS) -fPIC -c $< -o $@ $(COVERAGEFLAGS)
 JAVA_STATIC_FLAGS = -DZLIB -DBZIP2 -DSNAPPY -DLZ4
 JAVA_STATIC_INCLUDES = -I./zlib-1.2.8 -I./bzip2-1.0.6 -I./snappy-1.1.1 -I./lz4-r127/lib
 

appveyor.yml

@@ -5,7 +5,7 @@ before_build:
 - cmake -G "Visual Studio 12 Win64" ..
 - cd ..
 build:
-  project: build\ALL_BUILD.vcxproj
+  project: build\rocksdb.sln
   parallel: true
   verbosity: minimal
 test: off

appveyordailytests.yml

@@ -5,7 +5,7 @@ before_build:
 - cmake -G "Visual Studio 12 Win64" -DOPTDBG=1 ..
 - cd ..
 build:
-  project: build\ALL_BUILD.vcxproj
+  project: build\rocksdb.sln
   parallel: true
   verbosity: minimal
 test:

arcanist_util/cpp_linter/FbcodeCppLinter.php

@@ -88,6 +88,9 @@ class FbcodeCppLinter extends ArcanistLinter {
   }
 
   private function getCppLintOutput($path) {
+     if (!array_key_exists($path, $this->rawLintOutput)) {
+       return array();
+     }
     list($output) = $this->rawLintOutput[$path];
 
     $msgs = array();

build_tools/build_detect_platform

@@ -45,7 +45,7 @@ fi
 # we depend on C++11
 PLATFORM_CXXFLAGS="-std=c++11"
 # we currently depend on POSIX platform
-COMMON_FLAGS="-DROCKSDB_PLATFORM_POSIX"
+COMMON_FLAGS="-DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX"
 
 # Default to fbcode gcc on internal fb machines
 if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then

build_tools/fbcode_config.sh

@@ -45,6 +45,14 @@ if test -z $PIC_BUILD; then
   CFLAGS+=" -DLZ4"
 fi
 
+ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+if test -z $PIC_BUILD; then
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
+else
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
+fi
+CFLAGS+=" -DZSTD"
+
 # location of gflags headers and libraries
 GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
 if test -z $PIC_BUILD; then
@@ -83,7 +91,7 @@ export USE_SSE=1
 BINUTILS="$BINUTILS_BASE/bin"
 AR="$BINUTILS/ar"
 
-DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
+DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
 
 STDLIBS="-L $GCC_BASE/lib64"
 
@@ -127,7 +135,7 @@ CFLAGS+=" $DEPS_INCLUDE"
 CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_SUPPORT_THREAD_LOCAL -DHAVE_SSE42"
 CXXFLAGS+=" $CFLAGS"
 
-EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
+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+=" $LIBUNWIND"
@@ -137,7 +145,7 @@ EXEC_LDFLAGS+=" -ldl"
 
 PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
 
-EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $GFLAGS_LIBS $TBB_LIBS"
+EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS"
 
 VALGRIND_VER="$VALGRIND_BASE/bin/"
 

build_tools/fbcode_config4.8.1.sh

@@ -30,9 +30,9 @@ LZ4_REV=065ec7e38fe83329031f6668c43bef83eff5808b
 LZ4_INCLUDE=" -I /mnt/gvfs/third-party2/lz4/$LZ4_REV/r108/gcc-4.8.1-glibc-2.17/c3f970a/include"
 LZ4_LIBS=" /mnt/gvfs/third-party2/lz4/$LZ4_REV/r108/gcc-4.8.1-glibc-2.17/c3f970a/lib/liblz4.a"
 
-ZSTD_REV=8df2d01673ae6afcc8c8d16fec862b2d67ecc1e9
-ZSTD_INCLUDE=" -I /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.1.1/gcc-4.8.1-glibc-2.17/c3f970a/include"
-ZSTD_LIBS=" /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.1.1/gcc-4.8.1-glibc-2.17/c3f970a/lib/libzstd.a"
+ZSTD_REV=810b81b4705def5243e998b54701f3c504e4009e
+ZSTD_INCLUDE=" -I /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.4.2/gcc-4.8.1-glibc-2.17/c3f970a/include"
+ZSTD_LIBS=" /mnt/gvfs/third-party2/zstd/$ZSTD_REV/0.4.2/gcc-4.8.1-glibc-2.17/c3f970a/lib/libzstd.a"
 
 # location of gflags headers and libraries
 GFLAGS_INCLUDE=" -I /mnt/gvfs/third-party2/gflags/1ad047a6e6f6673991918ecadc670868205a243a/1.6/gcc-4.8.1-glibc-2.17/c3f970a/include/"
@@ -91,7 +91,7 @@ else
 fi
 
 CFLAGS+=" $DEPS_INCLUDE"
-CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE"
 CFLAGS+=" -DSNAPPY -DGFLAGS=google -DZLIB -DBZIP2 -DLZ4 -DZSTD -DNUMA"
 CXXFLAGS+=" $CFLAGS"
 

build_tools/rocksdb-lego-determinator

@@ -68,6 +68,14 @@ CLANG="USE_CLANG=1"
 LITE="OPT=-DROCKSDB_LITE"
 TSAN="COMPILE_WITH_TSAN=1"
 DISABLE_JEMALLOC="DISABLE_JEMALLOC=1"
+PARSER="'parser':'egrep \'Failure|^#|Abort\''"
+
+ARTIFACTS=" 'artifacts': [
+    {
+        'name':'database',
+        'paths':[ '/dev/shm/rocksdb' ],
+    }
+]"
 
 #
 # A mechanism to disable tests temporarily
@@ -100,13 +108,15 @@ PARALLEL_UNIT_TEST_COMMANDS="[
             {
                 'name':'Build and test RocksDB debug version',
                 'shell':'$DEBUG make -j$(nproc) all && $SHM make check > /dev/null 2>&1 || cat t/log-*',
-                'user':'root'
+                '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 || cat t/log-*',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -125,7 +135,8 @@ UNIT_TEST_COMMANDS="[
             {
                 'name':'Build and test RocksDB debug version',
                 'shell':'$SHM $DEBUG make J=1 check',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -144,7 +155,8 @@ UNIT_TEST_COMMANDS_481="[
             {
                 'name':'Build and test RocksDB debug version',
                 'shell':'$SHM $GCC_481 $DEBUG make J=1 check',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -163,7 +175,8 @@ CLANG_UNIT_TEST_COMMANDS="[
             {
                 'name':'Build and test RocksDB debug',
                 'shell':'$CLANG $SHM $DEBUG make J=1 check',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -182,7 +195,8 @@ CLANG_ANALYZE_COMMANDS="[
             {
                 'name':'RocksDB build and analyze',
                 'shell':'$CLANG $SHM $DEBUG make J=1 analyze',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -201,7 +215,8 @@ CODE_COV_COMMANDS="[
             {
                 'name':'Build, test and collect code coverage info',
                 'shell':'$SHM $DEBUG make J=1 coverage',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -220,7 +235,8 @@ UNITY_COMMANDS="[
             {
                 'name':'Build, test unity test',
                 'shell':'$SHM $DEBUG V=1 make J=1 unity_test',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -239,7 +255,28 @@ LITE_BUILD_COMMANDS="[
             {
                 'name':'Build RocksDB debug version',
                 'shell':'$LITE $DEBUG make J=1 static_lib',
-                'user':'root'
+                'user':'root',
+                $PARSER
+            },
+        ],
+        $REPORT
+    }
+]"
+
+#
+# RocksDB lite tests
+#
+LITE_UNIT_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb Lite Unit Test',
+        'oncall':'$ONCALL',
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build RocksDB debug version',
+                'shell':'$SHM $LITE $DEBUG make J=1 check',
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -259,19 +296,44 @@ STRESS_CRASH_TEST_COMMANDS="[
             {
                 'name':'Build and run RocksDB debug stress tests',
                 'shell':'$SHM $DEBUG make J=1 db_stress',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
             {
                 'name':'Build and run RocksDB debug crash tests',
                 'timeout': 86400,
                 'shell':'$SHM $DEBUG make J=1 crash_test',
-                'user':'root'
+                'user':'root',
+                $PARSER
             }
         ],
+        $ARTIFACTS,
         $REPORT
     }
 ]"
 
+# RocksDB write stress test.
+# We run on disk device on purpose (i.e. no $SHM)
+# because we want to add some randomness to fsync commands
+WRITE_STRESS_COMMANDS="[
+    {
+        'name':'Rocksdb Write Stress Test',
+        'oncall':'$ONCALL',
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB write stress tests',
+                'shell':'make write_stress && python tools/write_stress_runner.py --runtime_sec=3600 --db=/tmp/rocksdb_write_stress',
+                'user':'root',
+                $PARSER
+            }
+        ],
+        'artifacts': [{'name': 'database', 'paths': ['/tmp/rocksdb_write_stress']}],
+        $REPORT
+    }
+]"
+
+
 #
 # RocksDB test under address sanitizer
 #
@@ -284,7 +346,8 @@ ASAN_TEST_COMMANDS="[
             {
                 'name':'Test RocksDB debug under ASAN',
                 'shell':'set -o pipefail && $SHM $ASAN $DEBUG make J=1 asan_check |& /usr/facebook/ops/scripts/asan_symbolize.py -d',
-                'user':'root'
+                'user':'root',
+                $PARSER
             }
         ],
         $REPORT
@@ -305,7 +368,8 @@ ASAN_CRASH_TEST_COMMANDS="[
                 'name':'Build and run RocksDB debug asan_crash_test',
                 'timeout': 86400,
                 'shell':'$SHM $DEBUG make J=1 asan_crash_test',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -324,7 +388,8 @@ VALGRIND_TEST_COMMANDS="[
             {
                 'name':'Run RocksDB debug unit tests',
                 'shell':'$DISABLE_JEMALLOC $SHM $DEBUG make valgrind_check',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -343,7 +408,8 @@ TSAN_UNIT_TEST_COMMANDS="[
             {
                 'name':'Run RocksDB debug unit test',
                 'shell':'set -o pipefail && $SHM $DEBUG $TSAN make J=1 check',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -364,13 +430,16 @@ TSAN_CRASH_TEST_COMMANDS="[
                 'name':'Compile and run',
                 'timeout': 86400,
                 'shell':'set -o pipefail && $SHM $DEBUG $TSAN make J=1 crash_test',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
     }
 ]"
 
+TSAN_CRASH_TEST_COMMANDS=$DISABLE_COMMANDS
+
 #
 # RocksDB format compatible
 #
@@ -425,7 +494,8 @@ FORMAT_COMPATIBLE_COMMANDS="[
             {
                 'name':'Run RocksDB debug unit test',
                 'shell':'build_tools/rocksdb-lego-determinator run_format_compatible',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -457,7 +527,8 @@ NO_COMPRESSION_COMMANDS="[
             {
                 'name':'Run RocksDB debug unit test',
                 'shell':'build_tools/rocksdb-lego-determinator run_no_compression',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -513,7 +584,8 @@ REGRESSION_COMMANDS="[
             {
                 'name':'Make and run script',
                 'shell':'build_tools/rocksdb-lego-determinator run_regression',
-                'user':'root'
+                'user':'root',
+                $PARSER
             },
         ],
         $REPORT
@@ -545,9 +617,15 @@ case $1 in
   lite)
     echo $LITE_BUILD_COMMANDS
     ;;
+  lite_test)
+    echo $LITE_UNIT_TEST_COMMANDS
+    ;;
   stress_crash)
     echo $STRESS_CRASH_TEST_COMMANDS
     ;;
+  write_stress)
+    echo $WRITE_STRESS_COMMANDS
+    ;;
   asan)
     echo $ASAN_TEST_COMMANDS
     ;;

build_tools/run_ci_db_test.ps1

@@ -8,7 +8,10 @@ Param(
   [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|tests
+  [string]$Run = "db_test",  # Run db_test|tests
+   # Number of async tasks that would run concurrently. Recommend a number below 64.
+   # However, CPU utlization really depends on the storage media. Recommend ram based disk.
+  [int]$Concurrency = 62
 )
 
 # Folders and commands must be fullpath to run assuming
@@ -122,10 +125,17 @@ function Normalize-DbTests($HashTable) {
 function Discover-TestBinaries($HashTable) {
 
     $Exclusions = @("db_test*", "db_sanity_test*")
-    $p = -join ($BinariesFolder, "*_test*.exe")
+    if($EnableJE) {
+        $p = -join ($BinariesFolder, "*_test_je.exe")
+    } else {
+        $p = -join ($BinariesFolder, "*_test.exe")
+    }
 
     dir -Path $p -Exclude $Exclusions | ForEach-Object {
        $t = ($_.Name) -replace '.exe$', ''
+       if($ExcludeTests.Contains($t)) {
+            continue
+       }
        $test_log = -join ($t, ".log")
        $HashTable.Add($t, $test_log)
     }
@@ -137,6 +147,9 @@ if($Run -ceq "db_test") {
     Normalize-DbTests -HashTable $TestToLog
 } elseif($Run -ceq "tests") {
     Discover-TestBinaries -HashTable $TestToLog
+} else {
+    Write-Warning "Invalid -Run option value"
+    exit 2
 }
 
 
@@ -159,40 +172,48 @@ $JobToLog = @{}
 # Test limiting factor here
 $count = 0
 
-ForEach($k in $TestToLog.keys) {
-
-    Write-Host "Starting $k"
-    $log_path = -join ($LogFolder, ($TestToLog.$k))
-
-    if($Run -ceq "db_test") {
-        $job = Start-Job -Name $k -ScriptBlock $InvokeTestCase -ArgumentList @($db_test,$k,$log_path)
-    } else {
-        [string]$Exe =  -Join ($BinariesFolder, $k)
-        $job = Start-Job -Name $k -ScriptBlock $InvokeTestAsync -ArgumentList @($exe,$log_path)
-    }
-
-    $JobToLog.Add($job, $log_path)
-
-    # Limiting trial runs
-    if(($Limit -gt 0) -and (++$count -ge $Limit)) {
-         break
-    }
-}
-
 [bool]$success = $true;
 
 # Wait for all to finish and get the results
-while($JobToLog.Count -gt 0) {
+while(($JobToLog.Count -gt 0) -or
+      ($TestToLog.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 $Concurrency) -and
+          (($TestToLog.Count -gt 0) -and
+          (($Limit -lt 0) -or ($count -lt $Limit)))) {
+
+
+        # We only need the first key
+        foreach($key in $TestToLog.keys) {
+            $k = $key
+            break
+        }
+
+        Write-Host "Starting $k"
+        $log_path = -join ($LogFolder, ($TestToLog.$k))
+
+        if($Run -ceq "db_test") {
+          $job = Start-Job -Name $k -ScriptBlock $InvokeTestCase -ArgumentList @($db_test,$k,$log_path)
+        } else {
+          [string]$Exe =  -Join ($BinariesFolder, $k)
+           $job = Start-Job -Name $k -ScriptBlock $InvokeTestAsync -ArgumentList @($exe,$log_path)
+        }
+
+        $JobToLog.Add($job, $log_path)
+        $TestToLog.Remove($k)
+
+        ++$count
+    }
+
+    if($JobToLog.Count -lt 1) {
+      break
+    }
 
     $jobs = @()
     foreach($k in $JobToLog.Keys) { $jobs += $k }
 
-<#
-    if(!$success) {
-        break
-    }
-#>
-
     $completed = Wait-Job -Job $jobs -Any
     $log = $JobToLog[$completed]
     $JobToLog.Remove($completed)

db/builder.cc

@@ -26,6 +26,7 @@
 #include "rocksdb/options.h"
 #include "rocksdb/table.h"
 #include "table/block_based_table_builder.h"
+#include "table/internal_iterator.h"
 #include "util/file_reader_writer.h"
 #include "util/iostats_context_imp.h"
 #include "util/stop_watch.h"
@@ -40,22 +41,25 @@ TableBuilder* NewTableBuilder(
     const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
-    WritableFileWriter* file, const CompressionType compression_type,
+    uint32_t column_family_id, WritableFileWriter* file,
+    const CompressionType compression_type,
     const CompressionOptions& compression_opts, const bool skip_filters) {
   return ioptions.table_factory->NewTableBuilder(
       TableBuilderOptions(ioptions, internal_comparator,
                           int_tbl_prop_collector_factories, compression_type,
                           compression_opts, skip_filters),
-      file);
+      column_family_id, file);
 }
 
 Status BuildTable(
     const std::string& dbname, Env* env, const ImmutableCFOptions& ioptions,
-    const EnvOptions& env_options, TableCache* table_cache, Iterator* iter,
-    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
+    const EnvOptions& env_options, TableCache* table_cache,
+    InternalIterator* iter, FileMetaData* meta,
+    const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
-    std::vector<SequenceNumber> snapshots, const CompressionType compression,
+    uint32_t column_family_id, std::vector<SequenceNumber> snapshots,
+    const CompressionType compression,
     const CompressionOptions& compression_opts, bool paranoid_file_checks,
     InternalStats* internal_stats, const Env::IOPriority io_priority,
     TableProperties* table_properties) {
@@ -72,7 +76,7 @@ Status BuildTable(
     unique_ptr<WritableFileWriter> file_writer;
     {
       unique_ptr<WritableFile> file;
-      s = env->NewWritableFile(fname, &file, env_options);
+      s = NewWritableFile(env, fname, &file, env_options);
       if (!s.ok()) {
         return s;
       }
@@ -82,7 +86,7 @@ Status BuildTable(
 
       builder = NewTableBuilder(
           ioptions, internal_comparator, int_tbl_prop_collector_factories,
-          file_writer.get(), compression, compression_opts);
+          column_family_id, file_writer.get(), compression, compression_opts);
     }
 
     MergeHelper merge(env, internal_comparator.user_comparator(),
@@ -139,7 +143,7 @@ Status BuildTable(
 
     if (s.ok() && !empty) {
       // Verify that the table is usable
-      std::unique_ptr<Iterator> it(table_cache->NewIterator(
+      std::unique_ptr<InternalIterator> it(table_cache->NewIterator(
           ReadOptions(), env_options, internal_comparator, meta->fd, nullptr,
           (internal_stats == nullptr) ? nullptr
                                       : internal_stats->GetFileReadHist(0),

db/builder.h

@@ -31,13 +31,15 @@ class VersionEdit;
 class TableBuilder;
 class WritableFileWriter;
 class InternalStats;
+class InternalIterator;
 
 TableBuilder* NewTableBuilder(
     const ImmutableCFOptions& options,
     const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
-    WritableFileWriter* file, const CompressionType compression_type,
+    uint32_t column_family_id, WritableFileWriter* file,
+    const CompressionType compression_type,
     const CompressionOptions& compression_opts,
     const bool skip_filters = false);
 
@@ -48,11 +50,13 @@ TableBuilder* NewTableBuilder(
 // zero, and no Table file will be produced.
 extern Status BuildTable(
     const std::string& dbname, Env* env, const ImmutableCFOptions& options,
-    const EnvOptions& env_options, TableCache* table_cache, Iterator* iter,
-    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
+    const EnvOptions& env_options, TableCache* table_cache,
+    InternalIterator* iter, FileMetaData* meta,
+    const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
-    std::vector<SequenceNumber> snapshots, const CompressionType compression,
+    uint32_t column_family_id, std::vector<SequenceNumber> snapshots,
+    const CompressionType compression,
     const CompressionOptions& compression_opts, bool paranoid_file_checks,
     InternalStats* internal_stats,
     const Env::IOPriority io_priority = Env::IO_HIGH,

db/c.cc

@@ -1288,6 +1288,11 @@ void rocksdb_block_based_options_set_cache_index_and_filter_blocks(
   options->rep.cache_index_and_filter_blocks = v;
 }
 
+void rocksdb_block_based_options_set_skip_table_builder_flush(
+    rocksdb_block_based_table_options_t* options, unsigned char v) {
+  options->rep.skip_table_builder_flush = v;
+}
+
 void rocksdb_options_set_block_based_table_factory(
     rocksdb_options_t *opt,
     rocksdb_block_based_table_options_t* table_options) {
@@ -1687,6 +1692,11 @@ void rocksdb_options_set_keep_log_file_num(rocksdb_options_t* opt, size_t v) {
   opt->rep.keep_log_file_num = v;
 }
 
+void rocksdb_options_set_recycle_log_file_num(rocksdb_options_t* opt,
+                                              size_t v) {
+  opt->rep.recycle_log_file_num = v;
+}
+
 void rocksdb_options_set_soft_rate_limit(rocksdb_options_t* opt, double v) {
   opt->rep.soft_rate_limit = v;
 }

db/column_family.cc

@@ -27,9 +27,9 @@
 #include "db/version_set.h"
 #include "db/write_controller.h"
 #include "db/writebuffer.h"
+#include "memtable/hash_skiplist_rep.h"
 #include "util/autovector.h"
 #include "util/compression.h"
-#include "util/hash_skiplist_rep.h"
 #include "util/options_helper.h"
 #include "util/thread_status_util.h"
 #include "util/xfunc.h"

db/column_family_test.cc

@@ -23,8 +23,6 @@
 #include "util/sync_point.h"
 #include "utilities/merge_operators.h"
 
-#if !(defined NDEBUG) || !defined(OS_WIN)
-
 namespace rocksdb {
 
 namespace {
@@ -1262,13 +1260,8 @@ TEST_F(ColumnFamilyTest, FlushAndDropRaceCondition) {
 }
 
 }  // namespace rocksdb
-#endif
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();
-#else
-  return 0;
-#endif
 }

db/compaction.cc

@@ -439,6 +439,7 @@ std::unique_ptr<CompactionFilter> Compaction::CreateCompactionFilter() const {
   CompactionFilter::Context context;
   context.is_full_compaction = is_full_compaction_;
   context.is_manual_compaction = is_manual_compaction_;
+  context.column_family_id = cfd_->GetID();
   return cfd_->ioptions()->compaction_filter_factory->CreateCompactionFilter(
       context);
 }

db/compaction.h

@@ -210,6 +210,14 @@ class Compaction {
       int output_level, VersionStorageInfo* vstorage,
       const std::vector<CompactionInputFiles>& inputs);
 
+  TablePropertiesCollection GetOutputTableProperties() const {
+    return output_table_properties_;
+  }
+
+  void SetOutputTableProperties(TablePropertiesCollection tp) {
+    output_table_properties_ = std::move(tp);
+  }
+
  private:
   // mark (or clear) all files that are being compacted
   void MarkFilesBeingCompacted(bool mark_as_compacted);
@@ -273,6 +281,9 @@ class Compaction {
 
   // Does input compression match the output compression?
   bool InputCompressionMatchesOutput() const;
+
+  // table properties of output files
+  TablePropertiesCollection output_table_properties_;
 };
 
 // Utility function

db/compaction_iterator.cc

@@ -6,11 +6,12 @@
 //  of patent rights can be found in the PATENTS file in the same directory.
 
 #include "db/compaction_iterator.h"
+#include "table/internal_iterator.h"
 
 namespace rocksdb {
 
 CompactionIterator::CompactionIterator(
-    Iterator* input, const Comparator* cmp, MergeHelper* merge_helper,
+    InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
     SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
     Env* env, bool expect_valid_internal_key, Compaction* compaction,
     const CompactionFilter* compaction_filter, LogBuffer* log_buffer)

db/compaction_iterator.h

@@ -37,7 +37,7 @@ struct CompactionIteratorStats {
 
 class CompactionIterator {
  public:
-  CompactionIterator(Iterator* input, const Comparator* cmp,
+  CompactionIterator(InternalIterator* input, const Comparator* cmp,
                      MergeHelper* merge_helper, SequenceNumber last_sequence,
                      std::vector<SequenceNumber>* snapshots, Env* env,
                      bool expect_valid_internal_key,
@@ -84,7 +84,7 @@ class CompactionIterator {
   inline SequenceNumber findEarliestVisibleSnapshot(
       SequenceNumber in, SequenceNumber* prev_snapshot);
 
-  Iterator* input_;
+  InternalIterator* input_;
   const Comparator* cmp_;
   MergeHelper* merge_helper_;
   const std::vector<SequenceNumber>* snapshots_;

db/compaction_job.cc

@@ -77,6 +77,7 @@ struct CompactionJob::SubcompactionState {
   struct Output {
     FileMetaData meta;
     bool finished;
+    std::shared_ptr<const TableProperties> table_properties;
   };
 
   // State kept for output being generated
@@ -487,6 +488,16 @@ Status CompactionJob::Run() {
     }
   }
 
+  TablePropertiesCollection tp;
+  for (const auto& state : compact_->sub_compact_states) {
+    for (const auto& output : state.outputs) {
+      auto fn = TableFileName(db_options_.db_paths, output.meta.fd.GetNumber(),
+                              output.meta.fd.GetPathId());
+      tp[fn] = output.table_properties;
+    }
+  }
+  compact_->compaction->SetOutputTableProperties(std::move(tp));
+
   // Finish up all book-keeping to unify the subcompaction results
   AggregateStatistics();
   UpdateCompactionStats();
@@ -574,7 +585,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options,
 
 void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   assert(sub_compact != nullptr);
-  std::unique_ptr<Iterator> input(
+  std::unique_ptr<InternalIterator> input(
       versions_->MakeInputIterator(sub_compact->compaction));
 
   AutoThreadOperationStageUpdater stage_updater(
@@ -590,10 +601,10 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   if (measure_io_stats_) {
     prev_perf_level = GetPerfLevel();
     SetPerfLevel(PerfLevel::kEnableTime);
-    prev_write_nanos = iostats_context.write_nanos;
-    prev_fsync_nanos = iostats_context.fsync_nanos;
-    prev_range_sync_nanos = iostats_context.range_sync_nanos;
-    prev_prepare_write_nanos = iostats_context.prepare_write_nanos;
+    prev_write_nanos = IOSTATS(write_nanos);
+    prev_fsync_nanos = IOSTATS(fsync_nanos);
+    prev_range_sync_nanos = IOSTATS(range_sync_nanos);
+    prev_prepare_write_nanos = IOSTATS(prepare_write_nanos);
   }
 
   ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
@@ -717,13 +728,13 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
 
   if (measure_io_stats_) {
     sub_compact->compaction_job_stats.file_write_nanos +=
-        iostats_context.write_nanos - prev_write_nanos;
+        IOSTATS(write_nanos) - prev_write_nanos;
     sub_compact->compaction_job_stats.file_fsync_nanos +=
-        iostats_context.fsync_nanos - prev_fsync_nanos;
+        IOSTATS(fsync_nanos) - prev_fsync_nanos;
     sub_compact->compaction_job_stats.file_range_sync_nanos +=
-        iostats_context.range_sync_nanos - prev_range_sync_nanos;
+        IOSTATS(range_sync_nanos) - prev_range_sync_nanos;
     sub_compact->compaction_job_stats.file_prepare_write_nanos +=
-        iostats_context.prepare_write_nanos - prev_prepare_write_nanos;
+        IOSTATS(prepare_write_nanos) - prev_prepare_write_nanos;
     if (prev_perf_level != PerfLevel::kEnableTime) {
       SetPerfLevel(prev_perf_level);
     }
@@ -800,7 +811,7 @@ Status CompactionJob::FinishCompactionOutputFile(
   if (s.ok() && current_entries > 0) {
     // Verify that the table is usable
     ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
-    Iterator* iter = cfd->table_cache()->NewIterator(
+    InternalIterator* iter = cfd->table_cache()->NewIterator(
         ReadOptions(), env_options_, cfd->internal_comparator(), meta->fd,
         nullptr, cfd->internal_stats()->GetFileReadHist(
                      compact_->compaction->output_level()),
@@ -814,7 +825,10 @@ Status CompactionJob::FinishCompactionOutputFile(
 
     delete iter;
     if (s.ok()) {
-      TableFileCreationInfo info(sub_compact->builder->GetTableProperties());
+      auto tp = sub_compact->builder->GetTableProperties();
+      sub_compact->current_output()->table_properties =
+          std::make_shared<TableProperties>(tp);
+      TableFileCreationInfo info(std::move(tp));
       info.db_name = dbname_;
       info.cf_name = cfd->GetName();
       info.file_path =
@@ -897,7 +911,7 @@ Status CompactionJob::OpenCompactionOutputFile(
   unique_ptr<WritableFile> writable_file;
   std::string fname = TableFileName(db_options_.db_paths, file_number,
                                     sub_compact->compaction->output_path_id());
-  Status s = env_->NewWritableFile(fname, &writable_file, env_options_);
+  Status s = NewWritableFile(env_, fname, &writable_file, env_options_);
   if (!s.ok()) {
     Log(InfoLogLevel::ERROR_LEVEL, db_options_.info_log,
         "[%s] [JOB %d] OpenCompactionOutputFiles for table #%" PRIu64
@@ -927,8 +941,8 @@ Status CompactionJob::OpenCompactionOutputFile(
       cfd->ioptions()->optimize_filters_for_hits && bottommost_level_;
   sub_compact->builder.reset(NewTableBuilder(
       *cfd->ioptions(), cfd->internal_comparator(),
-      cfd->int_tbl_prop_collector_factories(), sub_compact->outfile.get(),
-      sub_compact->compaction->output_compression(),
+      cfd->int_tbl_prop_collector_factories(), cfd->GetID(),
+      sub_compact->outfile.get(), sub_compact->compaction->output_compression(),
       cfd->ioptions()->compression_opts, skip_filters));
   LogFlush(db_options_.info_log);
   return s;

db/compaction_job.h

@@ -35,9 +35,9 @@
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/transaction_log.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/autovector.h"
 #include "util/event_logger.h"
-#include "util/scoped_arena_iterator.h"
 #include "util/stop_watch.h"
 #include "util/thread_local.h"
 

db/compaction_job_stats_test.cc

@@ -27,6 +27,7 @@
 #include "db/job_context.h"
 #include "db/version_set.h"
 #include "db/write_batch_internal.h"
+#include "memtable/hash_linklist_rep.h"
 #include "port/stack_trace.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
@@ -47,14 +48,13 @@
 #include "table/block_based_table_factory.h"
 #include "table/mock_table.h"
 #include "table/plain_table_factory.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/compression.h"
 #include "util/hash.h"
-#include "util/hash_linklist_rep.h"
 #include "util/logging.h"
 #include "util/mock_env.h"
 #include "util/mutexlock.h"
 #include "util/rate_limiter.h"
-#include "util/scoped_arena_iterator.h"
 #include "util/statistics.h"
 #include "util/string_util.h"
 #include "util/sync_point.h"
@@ -64,7 +64,7 @@
 #include "util/xfunc.h"
 #include "utilities/merge_operators.h"
 
-#if !defined(IOS_CROSS_COMPILE) && (!defined(NDEBUG) || !defined(OS_WIN))
+#if !defined(IOS_CROSS_COMPILE)
 #ifndef ROCKSDB_LITE
 namespace rocksdb {
 

db/compaction_job_test.cc

@@ -3,6 +3,8 @@
 //  LICENSE file in the root directory of this source tree. An additional grant
 //  of patent rights can be found in the PATENTS file in the same directory.
 
+#ifndef ROCKSDB_LITE
+
 #include <algorithm>
 #include <map>
 #include <string>
@@ -196,7 +198,7 @@ class CompactionJobTest : public testing::Test {
     unique_ptr<WritableFileWriter> file_writer(
         new WritableFileWriter(std::move(file), env_options_));
     {
-      log::Writer log(std::move(file_writer));
+      log::Writer log(std::move(file_writer), 0, false);
       std::string record;
       new_db.EncodeTo(&record);
       s = log.AddRecord(record);
@@ -695,3 +697,14 @@ int main(int argc, char** argv) {
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();
 }
+
+#else
+#include <stdio.h>
+
+int main(int argc, char** argv) {
+  fprintf(stderr,
+          "SKIPPED as CompactionJobStats is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // ROCKSDB_LITE

db/comparator_db_test.cc

@@ -7,10 +7,11 @@
 #include <map>
 #include <string>
 
+#include "memtable/stl_wrappers.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
 #include "util/hash.h"
-#include "util/stl_wrappers.h"
+#include "util/kv_map.h"
 #include "util/string_util.h"
 #include "util/testharness.h"
 #include "util/testutil.h"

db/db_bench.cc

@@ -315,7 +315,7 @@ DEFINE_int32(compaction_style, (int32_t) rocksdb::Options().compaction_style,
              "style of compaction: level-based vs universal");
 
 static rocksdb::CompactionPri FLAGS_compaction_pri_e;
-DEFINE_int32(compaction_pri, (int32_t)rocksdb::Options().compaction_style,
+DEFINE_int32(compaction_pri, (int32_t)rocksdb::Options().compaction_pri,
              "priority of files to compaction: by size or by data age");
 
 DEFINE_int32(universal_size_ratio, 0,
@@ -373,6 +373,15 @@ DEFINE_int32(new_table_reader_for_compaction_inputs, true,
 
 DEFINE_int32(compaction_readahead_size, 0, "Compaction readahead size");
 
+DEFINE_int32(random_access_max_buffer_size, 1024 * 1024,
+             "Maximum windows randomaccess buffer size");
+
+DEFINE_int32(writable_file_max_buffer_size, 1024 * 1024, 
+             "Maximum write buffer for Writeable File");
+
+DEFINE_int32(skip_table_builder_flush, false, "Skip flushing block in "
+             "table builder ");
+
 DEFINE_int32(bloom_bits, -1, "Bloom filter bits per key. Negative means"
              " use default settings.");
 DEFINE_int32(memtable_bloom_bits, 0, "Bloom filter bits per key for memtable. "
@@ -486,6 +495,7 @@ DEFINE_int32(deletepercent, 2, "Percentage of deletes out of reads/writes/"
 DEFINE_uint64(delete_obsolete_files_period_micros, 0,
               "Ignored. Left here for backward compatibility");
 
+#ifndef ROCKSDB_LITE
 DEFINE_bool(optimistic_transaction_db, false,
             "Open a OptimisticTransactionDB instance. "
             "Required for randomtransaction benchmark.");
@@ -509,6 +519,7 @@ DEFINE_int32(transaction_sleep, 0,
 DEFINE_uint64(transaction_lock_timeout, 100,
               "If using a transaction_db, specifies the lock wait timeout in"
               " milliseconds before failing a transaction waiting on a lock");
+#endif  // ROCKSDB_LITE
 
 DEFINE_bool(compaction_measure_io_stats, false,
             "Measure times spents on I/Os while in compactions. ");
@@ -624,7 +635,7 @@ DEFINE_double(soft_rate_limit, 0.0, "");
 
 DEFINE_double(hard_rate_limit, 0.0, "DEPRECATED");
 
-DEFINE_uint64(hard_pending_compaction_bytes_limit, 128u * 1024 * 1024 * 1024,
+DEFINE_uint64(hard_pending_compaction_bytes_limit, 128ull * 1024 * 1024 * 1024,
               "Stop writes if pending compaction bytes exceed this number");
 
 DEFINE_uint64(delayed_write_rate, 2097152u,
@@ -645,7 +656,9 @@ DEFINE_int32(max_grandparent_overlap_factor, 10, "Control maximum bytes of "
              "overlaps in grandparent (i.e., level+2) before we stop building a"
              " single file in a level->level+1 compaction.");
 
+#ifndef ROCKSDB_LITE
 DEFINE_bool(readonly, false, "Run read only benchmarks.");
+#endif  // ROCKSDB_LITE
 
 DEFINE_bool(disable_auto_compactions, false, "Do not auto trigger compactions");
 
@@ -983,7 +996,9 @@ static void AppendWithSpace(std::string* str, Slice msg) {
 struct DBWithColumnFamilies {
   std::vector<ColumnFamilyHandle*> cfh;
   DB* db;
+#ifndef ROCKSDB_LITE
   OptimisticTransactionDB* opt_txn_db;
+#endif  // ROCKSDB_LITE
   std::atomic<size_t> num_created;  // Need to be updated after all the
                                     // new entries in cfh are set.
   size_t num_hot;  // Number of column families to be queried at each moment.
@@ -991,7 +1006,12 @@ struct DBWithColumnFamilies {
                    // Column families will be created and used to be queried.
   port::Mutex create_cf_mutex;  // Only one thread can execute CreateNewCf()
 
-  DBWithColumnFamilies() : db(nullptr), opt_txn_db(nullptr) {
+  DBWithColumnFamilies()
+      : db(nullptr)
+#ifndef ROCKSDB_LITE
+        , opt_txn_db(nullptr)
+#endif  // ROCKSDB_LITE
+  {
     cfh.clear();
     num_created = 0;
     num_hot = 0;
@@ -1000,7 +1020,9 @@ struct DBWithColumnFamilies {
   DBWithColumnFamilies(const DBWithColumnFamilies& other)
       : cfh(other.cfh),
         db(other.db),
+#ifndef ROCKSDB_LITE
         opt_txn_db(other.opt_txn_db),
+#endif  // ROCKSDB_LITE
         num_created(other.num_created.load()),
         num_hot(other.num_hot) {}
 
@@ -1008,13 +1030,18 @@ struct DBWithColumnFamilies {
     std::for_each(cfh.begin(), cfh.end(),
                   [](ColumnFamilyHandle* cfhi) { delete cfhi; });
     cfh.clear();
+#ifndef ROCKSDB_LITE
     if (opt_txn_db) {
       delete opt_txn_db;
       opt_txn_db = nullptr;
     } else {
       delete db;
+      db = nullptr;
     }
+#else
+    delete db;
     db = nullptr;
+#endif  // ROCKSDB_LITE
   }
 
   ColumnFamilyHandle* GetCfh(int64_t rand_num) {
@@ -1940,9 +1967,11 @@ class Benchmark {
         method = &Benchmark::Compress;
       } else if (name == "uncompress") {
         method = &Benchmark::Uncompress;
+#ifndef ROCKSDB_LITE
       } else if (name == "randomtransaction") {
         method = &Benchmark::RandomTransaction;
         post_process_method = &Benchmark::RandomTransactionVerify;
+#endif  // ROCKSDB_LITE
       } else if (name == "randomreplacekeys") {
         fresh_db = true;
         method = &Benchmark::RandomReplaceKeys;
@@ -2275,6 +2304,8 @@ class Benchmark {
     options.new_table_reader_for_compaction_inputs =
         FLAGS_new_table_reader_for_compaction_inputs;
     options.compaction_readahead_size = FLAGS_compaction_readahead_size;
+    options.random_access_max_buffer_size = FLAGS_random_access_max_buffer_size;
+    options.writable_file_max_buffer_size = FLAGS_writable_file_max_buffer_size;
     options.statistics = dbstats;
     if (FLAGS_enable_io_prio) {
       FLAGS_env->LowerThreadPoolIOPriority(Env::LOW);
@@ -2417,6 +2448,8 @@ class Benchmark {
       block_based_options.block_size = FLAGS_block_size;
       block_based_options.block_restart_interval = FLAGS_block_restart_interval;
       block_based_options.filter_policy = filter_policy_;
+      block_based_options.skip_table_builder_flush = 
+          FLAGS_skip_table_builder_flush;
       block_based_options.format_version = 2;
       options.table_factory.reset(
           NewBlockBasedTableFactory(block_based_options));
@@ -2518,10 +2551,12 @@ class Benchmark {
           NewGenericRateLimiter(FLAGS_rate_limiter_bytes_per_sec));
     }
 
+#ifndef ROCKSDB_LITE
     if (FLAGS_readonly && FLAGS_transaction_db) {
       fprintf(stderr, "Cannot use readonly flag with transaction_db\n");
       exit(1);
     }
+#endif  // ROCKSDB_LITE
 
     if (FLAGS_num_multi_db <= 1) {
       OpenDb(options, FLAGS_db, &db_);
@@ -2554,6 +2589,7 @@ class Benchmark {
         column_families.push_back(ColumnFamilyDescriptor(
               ColumnFamilyName(i), ColumnFamilyOptions(options)));
       }
+#ifndef ROCKSDB_LITE
       if (FLAGS_readonly) {
         s = DB::OpenForReadOnly(options, db_name, column_families,
             &db->cfh, &db->db);
@@ -2574,9 +2610,13 @@ class Benchmark {
       } else {
         s = DB::Open(options, db_name, column_families, &db->cfh, &db->db);
       }
+#else
+      s = DB::Open(options, db_name, column_families, &db->cfh, &db->db);
+#endif  // ROCKSDB_LITE
       db->cfh.resize(FLAGS_num_column_families);
       db->num_created = num_hot;
       db->num_hot = num_hot;
+#ifndef ROCKSDB_LITE
     } else if (FLAGS_readonly) {
       s = DB::OpenForReadOnly(options, db_name, &db->db);
     } else if (FLAGS_optimistic_transaction_db) {
@@ -2591,7 +2631,7 @@ class Benchmark {
       if (s.ok()) {
         db->db = ptr;
       }
-
+#endif  // ROCKSDB_LITE
     } else {
       s = DB::Open(options, db_name, &db->db);
     }
@@ -3636,6 +3676,7 @@ class Benchmark {
     }
   }
 
+#ifndef ROCKSDB_LITE
   // This benchmark stress tests Transactions.  For a given --duration (or
   // total number of --writes, a Transaction will perform a read-modify-write
   // to increment the value of a key in each of N(--transaction-sets) sets of
@@ -3868,6 +3909,7 @@ class Benchmark {
 
     fprintf(stdout, "RandomTransactionVerify Success!\n");
   }
+#endif  // ROCKSDB_LITE
 
   // Writes and deletes random keys without overwriting keys.
   //

db/db_compaction_filter_test.cc

@@ -7,8 +7,8 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
 
 namespace rocksdb {
 
@@ -97,8 +97,11 @@ class ChangeFilter : public CompactionFilter {
 
 class KeepFilterFactory : public CompactionFilterFactory {
  public:
-  explicit KeepFilterFactory(bool check_context = false)
-      : check_context_(check_context) {}
+  explicit KeepFilterFactory(bool check_context = false,
+                             bool check_context_cf_id = false)
+      : check_context_(check_context),
+        check_context_cf_id_(check_context_cf_id),
+        compaction_filter_created_(false) {}
 
   virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
       const CompactionFilter::Context& context) override {
@@ -106,13 +109,22 @@ class KeepFilterFactory : public CompactionFilterFactory {
       EXPECT_EQ(expect_full_compaction_.load(), context.is_full_compaction);
       EXPECT_EQ(expect_manual_compaction_.load(), context.is_manual_compaction);
     }
+    if (check_context_cf_id_) {
+      EXPECT_EQ(expect_cf_id_.load(), context.column_family_id);
+    }
+    compaction_filter_created_ = true;
     return std::unique_ptr<CompactionFilter>(new KeepFilter());
   }
 
+  bool compaction_filter_created() const { return compaction_filter_created_; }
+
   virtual const char* Name() const override { return "KeepFilterFactory"; }
   bool check_context_;
+  bool check_context_cf_id_;
   std::atomic_bool expect_full_compaction_;
   std::atomic_bool expect_manual_compaction_;
+  std::atomic<uint32_t> expect_cf_id_;
+  bool compaction_filter_created_;
 };
 
 class DeleteFilterFactory : public CompactionFilterFactory {
@@ -174,6 +186,7 @@ class ChangeFilterFactory : public CompactionFilterFactory {
   virtual const char* Name() const override { return "ChangeFilterFactory"; }
 };
 
+#ifndef ROCKSDB_LITE
 TEST_F(DBTestCompactionFilter, CompactionFilter) {
   Options options = CurrentOptions();
   options.max_open_files = -1;
@@ -216,7 +229,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
   Arena arena;
   {
     ScopedArenaIterator iter(
-        dbfull()->TEST_NewInternalIterator(&arena, handles_[1]));
+        dbfull()->NewInternalIterator(&arena, handles_[1]));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
@@ -304,7 +317,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
   count = 0;
   {
     ScopedArenaIterator iter(
-        dbfull()->TEST_NewInternalIterator(&arena, handles_[1]));
+        dbfull()->NewInternalIterator(&arena, handles_[1]));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
@@ -350,6 +363,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterDeletesAll) {
 
   delete itr;
 }
+#endif  // ROCKSDB_LITE
 
 TEST_F(DBTestCompactionFilter, CompactionFilterWithValueChange) {
   do {
@@ -481,8 +495,9 @@ TEST_F(DBTestCompactionFilter, CompactionFilterWithMergeOperator) {
   ASSERT_EQ(newvalue, four);
 }
 
+#ifndef ROCKSDB_LITE
 TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
-  KeepFilterFactory* filter = new KeepFilterFactory();
+  KeepFilterFactory* filter = new KeepFilterFactory(true, true);
 
   Options options = CurrentOptions();
   options.compaction_style = kCompactionStyleUniversal;
@@ -504,22 +519,24 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
     // be triggered.
     num_keys_per_file /= 2;
   }
+  dbfull()->TEST_WaitForCompact();
 
   // Force a manual compaction
   cfilter_count = 0;
   filter->expect_manual_compaction_.store(true);
-  filter->expect_full_compaction_.store(false);  // Manual compaction always
-                                                 // set this flag.
+  filter->expect_full_compaction_.store(true);
+  filter->expect_cf_id_.store(0);
   dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
   ASSERT_EQ(cfilter_count, 700);
   ASSERT_EQ(NumSortedRuns(0), 1);
+  ASSERT_TRUE(filter->compaction_filter_created());
 
   // Verify total number of keys is correct after manual compaction.
   {
     int count = 0;
     int total = 0;
     Arena arena;
-    ScopedArenaIterator iter(dbfull()->TEST_NewInternalIterator(&arena));
+    ScopedArenaIterator iter(dbfull()->NewInternalIterator(&arena));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
@@ -536,7 +553,38 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
     ASSERT_EQ(count, 1);
   }
 }
+#endif  // ROCKSDB_LITE
 
+TEST_F(DBTestCompactionFilter, CompactionFilterContextCfId) {
+  KeepFilterFactory* filter = new KeepFilterFactory(false, true);
+  filter->expect_cf_id_.store(1);
+
+  Options options = CurrentOptions();
+  options.compaction_filter_factory.reset(filter);
+  options.compression = kNoCompression;
+  options.level0_file_num_compaction_trigger = 2;
+  CreateAndReopenWithCF({"pikachu"}, options);
+
+  int num_keys_per_file = 400;
+  for (int j = 0; j < 3; j++) {
+    // Write several keys.
+    const std::string value(10, 'x');
+    for (int i = 0; i < num_keys_per_file; i++) {
+      char key[100];
+      snprintf(key, sizeof(key), "B%08d%02d", i, j);
+      Put(1, key, value);
+    }
+    Flush(1);
+    // Make sure next file is much smaller so automatic compaction will not
+    // be triggered.
+    num_keys_per_file /= 2;
+  }
+  dbfull()->TEST_WaitForCompact();
+
+  ASSERT_TRUE(filter->compaction_filter_created());
+}
+
+#ifndef ROCKSDB_LITE
 // Compaction filters should only be applied to records that are newer than the
 // latest snapshot. This test inserts records and applies a delete filter.
 TEST_F(DBTestCompactionFilter, CompactionFilterSnapshot) {
@@ -572,15 +620,12 @@ TEST_F(DBTestCompactionFilter, CompactionFilterSnapshot) {
   ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_EQ(0U, CountLiveFiles());
 }
+#endif  // ROCKSDB_LITE
 
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();
-#else
-  return 0;
-#endif
 }

db/db_compaction_test.cc

@@ -7,15 +7,14 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
 #include "rocksdb/experimental.h"
-#include "util/db_test_util.h"
 #include "util/sync_point.h"
 namespace rocksdb {
 
 // SYNC_POINT is not supported in released Windows mode.
-#if !(defined NDEBUG) || !defined(OS_WIN)
-
+#if !defined(ROCKSDB_LITE)
 
 class DBCompactionTest : public DBTestBase {
  public:
@@ -1228,6 +1227,7 @@ TEST_F(DBCompactionTest, L0_CompactionBug_Issue44_b) {
 TEST_P(DBCompactionTestWithParam, ManualCompaction) {
   Options options = CurrentOptions();
   options.max_subcompactions = max_subcompactions_;
+  options.statistics = rocksdb::CreateDBStatistics();
   CreateAndReopenWithCF({"pikachu"}, options);
 
   // iter - 0 with 7 levels
@@ -1259,7 +1259,14 @@ TEST_P(DBCompactionTestWithParam, ManualCompaction) {
     // Compact all
     MakeTables(1, "a", "z", 1);
     ASSERT_EQ("1,0,2", FilesPerLevel(1));
+
+    uint64_t prev_block_cache_add =
+        options.statistics->getTickerCount(BLOCK_CACHE_ADD);
     db_->CompactRange(CompactRangeOptions(), handles_[1], nullptr, nullptr);
+    // Verify manual compaction doesn't fill block cache
+    ASSERT_EQ(prev_block_cache_add,
+              options.statistics->getTickerCount(BLOCK_CACHE_ADD));
+
     ASSERT_EQ("0,0,1", FilesPerLevel(1));
 
     if (iter == 0) {
@@ -1267,6 +1274,7 @@ TEST_P(DBCompactionTestWithParam, ManualCompaction) {
       options.max_background_flushes = 0;
       options.num_levels = 3;
       options.create_if_missing = true;
+      options.statistics = rocksdb::CreateDBStatistics();
       DestroyAndReopen(options);
       CreateAndReopenWithCF({"pikachu"}, options);
     }
@@ -1844,11 +1852,11 @@ TEST_P(DBCompactionTestWithParam, ForceBottommostLevelCompaction) {
 
 INSTANTIATE_TEST_CASE_P(DBCompactionTestWithParam, DBCompactionTestWithParam,
                         ::testing::Values(1, 4));
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif // !defined(ROCKSDB_LITE)
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();

db/db_dynamic_level_test.cc

@@ -10,10 +10,10 @@
 // Introduction of SyncPoint effectively disabled building and running this test
 // in Release build.
 // which is a pity, it is a good test
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
 
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
 
 namespace rocksdb {
 class DBTestDynamicLevel : public DBTestBase {
@@ -484,10 +484,10 @@ TEST_F(DBTestDynamicLevel, MigrateToDynamicLevelMaxBytesBase) {
 }
 }  // namespace rocksdb
 
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif  // !defined(ROCKSDB_LITE)
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();

db/db_impl.cc

@@ -50,6 +50,8 @@
 #include "db/write_batch_internal.h"
 #include "db/write_callback.h"
 #include "db/writebuffer.h"
+#include "memtable/hash_linklist_rep.h"
+#include "memtable/hash_skiplist_rep.h"
 #include "port/likely.h"
 #include "port/port.h"
 #include "rocksdb/cache.h"
@@ -63,6 +65,7 @@
 #include "rocksdb/status.h"
 #include "rocksdb/table.h"
 #include "rocksdb/version.h"
+#include "rocksdb/wal_filter.h"
 #include "table/block.h"
 #include "table/block_based_table_factory.h"
 #include "table/merger.h"
@@ -77,8 +80,6 @@
 #include "util/db_info_dumper.h"
 #include "util/file_reader_writer.h"
 #include "util/file_util.h"
-#include "util/hash_linklist_rep.h"
-#include "util/hash_skiplist_rep.h"
 #include "util/iostats_context_imp.h"
 #include "util/log_buffer.h"
 #include "util/logging.h"
@@ -132,8 +133,7 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
   }
 
   if (result.info_log == nullptr) {
-    Status s = CreateLoggerFromOptions(dbname, result.db_log_dir, src.env,
-                                       result, &result.info_log);
+    Status s = CreateLoggerFromOptions(dbname, result, &result.info_log);
     if (!s.ok()) {
       // No place suitable for logging
       result.info_log = nullptr;
@@ -150,6 +150,10 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
     }
   }
 
+  if (result.WAL_ttl_seconds > 0 || result.WAL_size_limit_MB > 0) {
+    result.recycle_log_file_num = false;
+  }
+
   if (result.wal_dir.empty()) {
     // Use dbname as default
     result.wal_dir = dbname;
@@ -245,8 +249,10 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname)
       unscheduled_flushes_(0),
       unscheduled_compactions_(0),
       bg_compaction_scheduled_(0),
+      num_running_compactions_(0),
       bg_manual_only_(0),
       bg_flush_scheduled_(0),
+      num_running_flushes_(0),
       manual_compaction_(nullptr),
       disable_delete_obsolete_files_(0),
       delete_obsolete_files_next_run_(
@@ -261,6 +267,7 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname)
 #endif  // ROCKSDB_LITE
       event_logger_(db_options_.info_log.get()),
       bg_work_paused_(0),
+      bg_compaction_paused_(0),
       refitting_level_(false),
       opened_successfully_(false) {
   env_->GetAbsolutePath(dbname, &db_absolute_path_);
@@ -407,14 +414,14 @@ Status DBImpl::NewDB() {
   {
     unique_ptr<WritableFile> file;
     EnvOptions env_options = env_->OptimizeForManifestWrite(env_options_);
-    s = env_->NewWritableFile(manifest, &file, env_options);
+    s = NewWritableFile(env_, manifest, &file, env_options);
     if (!s.ok()) {
       return s;
     }
     file->SetPreallocationBlockSize(db_options_.manifest_preallocation_size);
     unique_ptr<WritableFileWriter> file_writer(
         new WritableFileWriter(std::move(file), env_options));
-    log::Writer log(std::move(file_writer));
+    log::Writer log(std::move(file_writer), 0, false);
     std::string record;
     new_db.EncodeTo(&record);
     s = log.AddRecord(record);
@@ -599,7 +606,13 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
     // find newly obsoleted log files
     while (alive_log_files_.begin()->number < min_log_number) {
       auto& earliest = *alive_log_files_.begin();
-      job_context->log_delete_files.push_back(earliest.number);
+      if (db_options_.recycle_log_file_num > log_recycle_files.size()) {
+        Log(InfoLogLevel::INFO_LEVEL, 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);
+      }
       total_log_size_ -= earliest.size;
       alive_log_files_.pop_front();
       // Current log should always stay alive since it can't have
@@ -1107,28 +1120,13 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
     // paranoid_checks==false so that corruptions cause entire commits
     // to be skipped instead of propagating bad information (like overly
     // large sequence numbers).
-    log::Reader reader(std::move(file_reader), &reporter, true /*checksum*/,
-                       0 /*initial_offset*/);
+    log::Reader reader(db_options_.info_log, std::move(file_reader), &reporter,
+                       true /*checksum*/, 0 /*initial_offset*/, log_number);
     Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
         "Recovering log #%" PRIu64 " mode %d skip-recovery %d", log_number,
         db_options_.wal_recovery_mode, !continue_replay_log);
 
     // Determine if we should tolerate incomplete records at the tail end of the
-    // log
-    bool report_eof_inconsistency;
-    if (db_options_.wal_recovery_mode ==
-        WALRecoveryMode::kAbsoluteConsistency) {
-      // in clean shutdown we don't expect any error in the log files
-      report_eof_inconsistency = true;
-    } else {
-      // for other modes ignore only incomplete records in the last log file
-      // which is presumably due to write in progress during restart
-      report_eof_inconsistency = false;
-
-      // TODO krad: Evaluate if we need to move to a more strict mode where we
-      // restrict the inconsistency to only the last log
-    }
-
     // Read all the records and add to a memtable
     std::string scratch;
     Slice record;
@@ -1143,9 +1141,10 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
       }
     }
 
-    while (continue_replay_log &&
-           reader.ReadRecord(&record, &scratch, report_eof_inconsistency) &&
-           status.ok()) {
+    while (
+        continue_replay_log &&
+        reader.ReadRecord(&record, &scratch, db_options_.wal_recovery_mode) &&
+        status.ok()) {
       if (record.size() < 12) {
         reporter.Corruption(record.size(),
                             Status::Corruption("log record too small"));
@@ -1153,6 +1152,80 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
       }
       WriteBatchInternal::SetContents(&batch, record);
 
+#ifndef ROCKSDB_LITE
+      if (db_options_.wal_filter != nullptr) {
+        WriteBatch new_batch;
+        bool batch_changed = false;
+
+        WalFilter::WalProcessingOption wal_processing_option =
+            db_options_.wal_filter->LogRecord(batch, &new_batch,
+                                              &batch_changed);
+
+        switch (wal_processing_option) {
+          case WalFilter::WalProcessingOption::kContinueProcessing:
+            // do nothing, proceeed normally
+            break;
+          case WalFilter::WalProcessingOption::kIgnoreCurrentRecord:
+            // skip current record
+            continue;
+          case WalFilter::WalProcessingOption::kStopReplay:
+            // skip current record and stop replay
+            continue_replay_log = false;
+            continue;
+          case WalFilter::WalProcessingOption::kCorruptedRecord: {
+            status = Status::Corruption("Corruption reported by Wal Filter ",
+                                        db_options_.wal_filter->Name());
+            MaybeIgnoreError(&status);
+            if (!status.ok()) {
+              reporter.Corruption(record.size(), status);
+              continue;
+            }
+            break;
+          }
+          default: {
+            assert(false);  // unhandled case
+            status = Status::NotSupported(
+                "Unknown WalProcessingOption returned"
+                " by Wal Filter ",
+                db_options_.wal_filter->Name());
+            MaybeIgnoreError(&status);
+            if (!status.ok()) {
+              return status;
+            } else {
+              // Ignore the error with current record processing.
+              continue;
+            }
+          }
+        }
+
+        if (batch_changed) {
+          // Make sure that the count in the new batch is
+          // within the orignal count.
+          int new_count = WriteBatchInternal::Count(&new_batch);
+          int original_count = WriteBatchInternal::Count(&batch);
+          if (new_count > original_count) {
+            Log(InfoLogLevel::FATAL_LEVEL, db_options_.info_log,
+                "Recovering log #%" PRIu64
+                " mode %d log filter %s returned "
+                "more records (%d) than original (%d) which is not allowed. "
+                "Aborting recovery.",
+                log_number, db_options_.wal_recovery_mode,
+                db_options_.wal_filter->Name(), new_count, original_count);
+            status = Status::NotSupported(
+                "More than original # of records "
+                "returned by Wal Filter ",
+                db_options_.wal_filter->Name());
+            return status;
+          }
+          // Set the same sequence number in the new_batch
+          // as the original batch.
+          WriteBatchInternal::SetSequence(&new_batch,
+                                          WriteBatchInternal::Sequence(&batch));
+          batch = new_batch;
+        }
+      }
+#endif  // ROCKSDB_LITE
+
       // If column family was not found, it might mean that the WAL write
       // batch references to the column family that was dropped after the
       // insert. We don't want to fail the whole write batch in that case --
@@ -1317,8 +1390,8 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
       s = BuildTable(
           dbname_, env_, *cfd->ioptions(), env_options_, cfd->table_cache(),
           iter.get(), &meta, cfd->internal_comparator(),
-          cfd->int_tbl_prop_collector_factories(), snapshots_.GetAll(),
-          GetCompressionFlush(*cfd->ioptions()),
+          cfd->int_tbl_prop_collector_factories(), cfd->GetID(),
+          snapshots_.GetAll(), GetCompressionFlush(*cfd->ioptions()),
           cfd->ioptions()->compression_opts, paranoid_file_checks,
           cfd->internal_stats(), Env::IO_HIGH, &info.table_properties);
       LogFlush(db_options_.info_log);
@@ -1412,15 +1485,16 @@ Status DBImpl::FlushMemTableToOutputFile(
   if (s.ok()) {
     // may temporarily unlock and lock the mutex.
     NotifyOnFlushCompleted(cfd, &file_meta, mutable_cf_options,
-                           job_context->job_id);
+                           job_context->job_id, flush_job.GetTableProperties());
   }
 #endif  // ROCKSDB_LITE
   return s;
 }
 
-void DBImpl::NotifyOnFlushCompleted(
-    ColumnFamilyData* cfd, FileMetaData* file_meta,
-    const MutableCFOptions& mutable_cf_options, int job_id) {
+void DBImpl::NotifyOnFlushCompleted(ColumnFamilyData* cfd,
+                                    FileMetaData* file_meta,
+                                    const MutableCFOptions& mutable_cf_options,
+                                    int job_id, TableProperties prop) {
 #ifndef ROCKSDB_LITE
   if (db_options_.listeners.size() == 0U) {
     return;
@@ -1450,6 +1524,7 @@ void DBImpl::NotifyOnFlushCompleted(
     info.triggered_writes_stop = triggered_writes_stop;
     info.smallest_seqno = file_meta->smallest_seqno;
     info.largest_seqno = file_meta->largest_seqno;
+    info.table_properties = prop;
     for (auto listener : db_options_.listeners) {
       listener->OnFlushCompleted(this, info);
     }
@@ -1755,17 +1830,22 @@ Status DBImpl::CompactFilesImpl(
 
 Status DBImpl::PauseBackgroundWork() {
   InstrumentedMutexLock guard_lock(&mutex_);
-  bg_work_paused_++;
+  bg_compaction_paused_++;
   while (bg_compaction_scheduled_ > 0 || bg_flush_scheduled_ > 0) {
     bg_cv_.Wait();
   }
+  bg_work_paused_++;
   return Status::OK();
 }
 
 Status DBImpl::ContinueBackgroundWork() {
   InstrumentedMutexLock guard_lock(&mutex_);
   assert(bg_work_paused_ > 0);
+  assert(bg_compaction_paused_ > 0);
+  bg_compaction_paused_--;
   bg_work_paused_--;
+  // It's sufficient to check just bg_work_paused_ here since
+  // bg_work_paused_ is always no greater than bg_compaction_paused_
   if (bg_work_paused_ == 0) {
     MaybeScheduleFlushOrCompaction();
   }
@@ -1795,12 +1875,19 @@ void DBImpl::NotifyOnCompactionCompleted(
     info.base_input_level = c->start_level();
     info.output_level = c->output_level();
     info.stats = compaction_job_stats;
+    info.table_properties = c->GetOutputTableProperties();
     for (size_t i = 0; i < c->num_input_levels(); ++i) {
       for (const auto fmd : *c->inputs(i)) {
-        info.input_files.push_back(
-            TableFileName(db_options_.db_paths,
-                          fmd->fd.GetNumber(),
-                          fmd->fd.GetPathId()));
+        auto fn = TableFileName(db_options_.db_paths, fmd->fd.GetNumber(),
+                                fmd->fd.GetPathId());
+        info.input_files.push_back(fn);
+        if (info.table_properties.count(fn) == 0) {
+          std::shared_ptr<const TableProperties> tp;
+          auto s = cfd->current()->GetTableProperties(&tp, fmd, &fn);
+          if (s.ok()) {
+            info.table_properties[fn] = tp;
+          }
+        }
       }
     }
     for (const auto newf : c->edit()->GetNewFiles()) {
@@ -2158,6 +2245,23 @@ Status DBImpl::RunManualCompaction(ColumnFamilyData* cfd, int input_level,
   return manual.status;
 }
 
+InternalIterator* DBImpl::NewInternalIterator(
+    Arena* arena, ColumnFamilyHandle* column_family) {
+  ColumnFamilyData* cfd;
+  if (column_family == nullptr) {
+    cfd = default_cf_handle_->cfd();
+  } else {
+    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+    cfd = cfh->cfd();
+  }
+
+  mutex_.Lock();
+  SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
+  mutex_.Unlock();
+  ReadOptions roptions;
+  return NewInternalIterator(roptions, cfd, super_version, arena);
+}
+
 Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
                              const FlushOptions& flush_options) {
   Status s;
@@ -2241,7 +2345,12 @@ void DBImpl::MaybeScheduleFlushOrCompaction() {
     }
   }
 
-  if (bg_manual_only_) {
+  if (bg_compaction_paused_ > 0) {
+    // we paused the background compaction
+    return;
+  }
+
+  if (bg_manual_only_) { 
     // only manual compactions are allowed to run. don't schedule automatic
     // compactions
     return;
@@ -2377,6 +2486,7 @@ void DBImpl::BackgroundCallFlush() {
   LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, db_options_.info_log.get());
   {
     InstrumentedMutexLock l(&mutex_);
+    num_running_flushes_++;
 
     auto pending_outputs_inserted_elem =
         CaptureCurrentFileNumberInPendingOutputs();
@@ -2422,6 +2532,8 @@ void DBImpl::BackgroundCallFlush() {
       mutex_.Lock();
     }
 
+    assert(num_running_flushes_ > 0);
+    num_running_flushes_--;
     bg_flush_scheduled_--;
     // See if there's more work to be done
     MaybeScheduleFlushOrCompaction();
@@ -2442,6 +2554,7 @@ void DBImpl::BackgroundCallCompaction() {
   LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, db_options_.info_log.get());
   {
     InstrumentedMutexLock l(&mutex_);
+    num_running_compactions_++;
 
     auto pending_outputs_inserted_elem =
         CaptureCurrentFileNumberInPendingOutputs();
@@ -2490,6 +2603,8 @@ void DBImpl::BackgroundCallCompaction() {
       mutex_.Lock();
     }
 
+    assert(num_running_compactions_ > 0);
+    num_running_compactions_--;
     bg_compaction_scheduled_--;
 
     versions_->GetColumnFamilySet()->FreeDeadColumnFamilies();
@@ -2835,11 +2950,11 @@ static void CleanupIteratorState(void* arg1, void* arg2) {
 }
 }  // namespace
 
-Iterator* DBImpl::NewInternalIterator(const ReadOptions& read_options,
-                                      ColumnFamilyData* cfd,
-                                      SuperVersion* super_version,
-                                      Arena* arena) {
-  Iterator* internal_iter;
+InternalIterator* DBImpl::NewInternalIterator(const ReadOptions& read_options,
+                                              ColumnFamilyData* cfd,
+                                              SuperVersion* super_version,
+                                              Arena* arena) {
+  InternalIterator* internal_iter;
   assert(arena != nullptr);
   // Need to create internal iterator from the arena.
   MergeIteratorBuilder merge_iter_builder(&cfd->internal_comparator(), arena);
@@ -3138,7 +3253,8 @@ Status DBImpl::AddFile(ColumnFamilyHandle* column_family,
   file_info.num_entries = table_reader->GetTableProperties()->num_entries;
 
   ParsedInternalKey key;
-  std::unique_ptr<Iterator> iter(table_reader->NewIterator(ReadOptions()));
+  std::unique_ptr<InternalIterator> iter(
+      table_reader->NewIterator(ReadOptions()));
 
   // Get first (smallest) key from file
   iter->SeekToFirst();
@@ -3169,10 +3285,6 @@ Status DBImpl::AddFile(ColumnFamilyHandle* column_family,
   auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
   ColumnFamilyData* cfd = cfh->cfd();
 
-  if (cfd->NumberLevels() <= 1) {
-    return Status::NotSupported(
-        "AddFile requires a database with at least 2 levels");
-  }
   if (file_info->version != 1) {
     return Status::InvalidArgument("Generated table version is not supported");
   }
@@ -3219,48 +3331,44 @@ Status DBImpl::AddFile(ColumnFamilyHandle* column_family,
     WriteThread::Writer w;
     write_thread_.EnterUnbatched(&w, &mutex_);
 
-    // Make sure memtables are empty
-    if (!cfd->mem()->IsEmpty() || cfd->imm()->NumNotFlushed() > 0) {
-      // Cannot add the file since the keys in memtable
-      // will hide the keys in file
-      status = Status::NotSupported("Memtable is not empty");
+    if (!snapshots_.empty()) {
+      // Check that no snapshots are being held
+      status =
+          Status::NotSupported("Cannot add a file while holding snapshots");
     }
 
-    // Make sure last sequence number is 0, if there are existing files then
-    // they should have sequence number = 0
-    if (status.ok() && versions_->LastSequence() > 0) {
-      status = Status::NotSupported("Last Sequence number is not zero");
-    }
-
-    auto* vstorage = cfd->current()->storage_info();
     if (status.ok()) {
-      // Make sure that the key range in the file we will add does not overlap
-      // with previously added files
-      Slice smallest_user_key = meta.smallest.user_key();
-      Slice largest_user_key = meta.largest.user_key();
-      for (int level = 0; level < vstorage->num_non_empty_levels(); level++) {
-        if (vstorage->OverlapInLevel(level, &smallest_user_key,
-                                     &largest_user_key)) {
-          status = Status::NotSupported("Cannot add overlapping files");
-          break;
+      // Verify that added file key range dont overlap with any keys in DB
+      SuperVersion* sv = cfd->GetSuperVersion()->Ref();
+      Arena arena;
+      ReadOptions ro;
+      ro.total_order_seek = true;
+      ScopedArenaIterator iter(NewInternalIterator(ro, cfd, sv, &arena));
+
+      InternalKey range_start(file_info->smallest_key, kMaxSequenceNumber,
+                              kTypeValue);
+      iter->Seek(range_start.Encode());
+      status = iter->status();
+
+      if (status.ok() && iter->Valid()) {
+        ParsedInternalKey seek_result;
+        if (ParseInternalKey(iter->key(), &seek_result)) {
+          auto* vstorage = cfd->current()->storage_info();
+          if (vstorage->InternalComparator()->user_comparator()->Compare(
+                  seek_result.user_key, file_info->largest_key) <= 0) {
+            status = Status::NotSupported("Cannot add overlapping range");
+          }
+        } else {
+          status = Status::Corruption("DB have corrupted keys");
         }
       }
     }
 
     if (status.ok()) {
-      // We add the file to the last level
-      int target_level = cfd->NumberLevels() - 1;
-      if (cfd->ioptions()->level_compaction_dynamic_level_bytes == false) {
-        // If we are using dynamic level compaction we add the file to
-        // last level with files
-        target_level = vstorage->num_non_empty_levels() - 1;
-        if (target_level <= 0) {
-          target_level = 1;
-        }
-      }
+      // Add file to L0
       VersionEdit edit;
       edit.SetColumnFamily(cfd->GetID());
-      edit.AddFile(target_level, meta.fd.GetNumber(), meta.fd.GetPathId(),
+      edit.AddFile(0, meta.fd.GetNumber(), meta.fd.GetPathId(),
                    meta.fd.GetFileSize(), meta.smallest, meta.largest,
                    meta.smallest_seqno, meta.largest_seqno,
                    meta.marked_for_compaction);
@@ -3476,10 +3584,11 @@ Iterator* DBImpl::NewIterator(const ReadOptions& read_options,
 #else
     SuperVersion* sv = cfd->GetReferencedSuperVersion(&mutex_);
     auto iter = new ForwardIterator(this, read_options, cfd, sv);
-    return NewDBIterator(env_, *cfd->ioptions(), cfd->user_comparator(), iter,
+    return NewDBIterator(
+        env_, *cfd->ioptions(), cfd->user_comparator(), iter,
         kMaxSequenceNumber,
         sv->mutable_cf_options.max_sequential_skip_in_iterations,
-        read_options.iterate_upper_bound);
+        read_options.iterate_upper_bound, read_options.prefix_same_as_start);
 #endif
   } else {
     SequenceNumber latest_snapshot = versions_->LastSequence();
@@ -3534,11 +3643,11 @@ Iterator* DBImpl::NewIterator(const ReadOptions& read_options,
     // likely that any iterator pointer is close to the iterator it points to so
     // that they are likely to be in the same cache line and/or page.
     ArenaWrappedDBIter* db_iter = NewArenaWrappedDbIterator(
-        env_, *cfd->ioptions(), cfd->user_comparator(),
-        snapshot, sv->mutable_cf_options.max_sequential_skip_in_iterations,
-        read_options.iterate_upper_bound);
+        env_, *cfd->ioptions(), cfd->user_comparator(), snapshot,
+        sv->mutable_cf_options.max_sequential_skip_in_iterations,
+        read_options.iterate_upper_bound, read_options.prefix_same_as_start);
 
-    Iterator* internal_iter =
+    InternalIterator* internal_iter =
         NewInternalIterator(read_options, cfd, sv, db_iter->GetArena());
     db_iter->SetIterUnderDBIter(internal_iter);
 
@@ -3605,8 +3714,8 @@ Status DBImpl::NewIterators(
       ArenaWrappedDBIter* db_iter = NewArenaWrappedDbIterator(
           env_, *cfd->ioptions(), cfd->user_comparator(), snapshot,
           sv->mutable_cf_options.max_sequential_skip_in_iterations);
-      Iterator* internal_iter = NewInternalIterator(
-          read_options, cfd, sv, db_iter->GetArena());
+      InternalIterator* internal_iter =
+          NewInternalIterator(read_options, cfd, sv, db_iter->GetArena());
       db_iter->SetIterUnderDBIter(internal_iter);
       iterators->push_back(db_iter);
     }
@@ -4036,6 +4145,12 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
   // Do this without holding the dbmutex lock.
   assert(versions_->prev_log_number() == 0);
   bool creating_new_log = !log_empty_;
+  uint64_t recycle_log_number = 0;
+  if (creating_new_log && db_options_.recycle_log_file_num &&
+      !log_recycle_files.empty()) {
+    recycle_log_number = log_recycle_files.front();
+    log_recycle_files.pop_front();
+  }
   uint64_t new_log_number =
       creating_new_log ? versions_->NewFileNumber() : logfile_number_;
   SuperVersion* new_superversion = nullptr;
@@ -4046,17 +4161,27 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
     if (creating_new_log) {
       EnvOptions opt_env_opt =
           env_->OptimizeForLogWrite(env_options_, db_options_);
-      s = env_->NewWritableFile(
-          LogFileName(db_options_.wal_dir, new_log_number), &lfile,
-          opt_env_opt);
+      if (recycle_log_number) {
+        Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,
+            "reusing log %" PRIu64 " from recycle list\n", recycle_log_number);
+        s = env_->ReuseWritableFile(
+            LogFileName(db_options_.wal_dir, new_log_number),
+            LogFileName(db_options_.wal_dir, recycle_log_number), &lfile,
+            opt_env_opt);
+      } else {
+        s = NewWritableFile(env_,
+                            LogFileName(db_options_.wal_dir, new_log_number),
+                            &lfile, opt_env_opt);
+      }
       if (s.ok()) {
         // Our final size should be less than write_buffer_size
         // (compression, etc) but err on the side of caution.
-        lfile->SetPreallocationBlockSize(
-            1.1 * mutable_cf_options.write_buffer_size);
+        lfile->SetPreallocationBlockSize(1.1 *
+                                         mutable_cf_options.write_buffer_size);
         unique_ptr<WritableFileWriter> file_writer(
             new WritableFileWriter(std::move(lfile), opt_env_opt));
-        new_log = new log::Writer(std::move(file_writer));
+        new_log = new log::Writer(std::move(file_writer), new_log_number,
+                                  db_options_.recycle_log_file_num > 0);
       }
     }
 
@@ -4125,6 +4250,29 @@ Status DBImpl::GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
 
   return s;
 }
+
+Status DBImpl::GetPropertiesOfTablesInRange(ColumnFamilyHandle* column_family,
+                                            const Range* range, std::size_t n,
+                                            TablePropertiesCollection* props) {
+  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfd = cfh->cfd();
+
+  // Increment the ref count
+  mutex_.Lock();
+  auto version = cfd->current();
+  version->Ref();
+  mutex_.Unlock();
+
+  auto s = version->GetPropertiesOfTablesInRange(range, n, props);
+
+  // Decrement the ref count
+  mutex_.Lock();
+  version->Unref();
+  mutex_.Unlock();
+
+  return s;
+}
+
 #endif  // ROCKSDB_LITE
 
 const std::string& DBImpl::GetName() const {
@@ -4150,17 +4298,16 @@ bool DBImpl::GetProperty(ColumnFamilyHandle* column_family,
       GetPropertyType(property, &is_int_property, &need_out_of_mutex);
 
   value->clear();
+  auto cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family)->cfd();
   if (is_int_property) {
     uint64_t int_value;
-    bool ret_value = GetIntPropertyInternal(column_family, property_type,
-                                            need_out_of_mutex, &int_value);
+    bool ret_value = GetIntPropertyInternal(
+        cfd, property_type, need_out_of_mutex, false, &int_value);
     if (ret_value) {
       *value = ToString(int_value);
     }
     return ret_value;
   } else {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
-    auto cfd = cfh->cfd();
     InstrumentedMutexLock l(&mutex_);
     return cfd->internal_stats()->GetStringProperty(property_type, property,
                                                     value);
@@ -4176,31 +4323,70 @@ bool DBImpl::GetIntProperty(ColumnFamilyHandle* column_family,
   if (!is_int_property) {
     return false;
   }
-  return GetIntPropertyInternal(column_family, property_type, need_out_of_mutex,
+  auto cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family)->cfd();
+  return GetIntPropertyInternal(cfd, property_type, need_out_of_mutex, false,
                                 value);
 }
 
-bool DBImpl::GetIntPropertyInternal(ColumnFamilyHandle* column_family,
+bool DBImpl::GetIntPropertyInternal(ColumnFamilyData* cfd,
                                     DBPropertyType property_type,
-                                    bool need_out_of_mutex, uint64_t* value) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
-  auto cfd = cfh->cfd();
-
+                                    bool need_out_of_mutex, bool is_locked,
+                                    uint64_t* value) {
   if (!need_out_of_mutex) {
-    InstrumentedMutexLock l(&mutex_);
-    return cfd->internal_stats()->GetIntProperty(property_type, value, this);
+    if (is_locked) {
+      mutex_.AssertHeld();
+      return cfd->internal_stats()->GetIntProperty(property_type, value, this);
+    } else {
+      InstrumentedMutexLock l(&mutex_);
+      return cfd->internal_stats()->GetIntProperty(property_type, value, this);
+    }
   } else {
-    SuperVersion* sv = GetAndRefSuperVersion(cfd);
+    SuperVersion* sv = nullptr;
+    if (!is_locked) {
+      sv = GetAndRefSuperVersion(cfd);
+    } else {
+      sv = cfd->GetSuperVersion();
+    }
 
     bool ret = cfd->internal_stats()->GetIntPropertyOutOfMutex(
         property_type, sv->current, value);
 
-    ReturnAndCleanupSuperVersion(cfd, sv);
+    if (!is_locked) {
+      ReturnAndCleanupSuperVersion(cfd, sv);
+    }
 
     return ret;
   }
 }
 
+bool DBImpl::GetAggregatedIntProperty(const Slice& property,
+                                      uint64_t* aggregated_value) {
+  bool need_out_of_mutex;
+  bool is_int_property;
+  DBPropertyType property_type =
+      GetPropertyType(property, &is_int_property, &need_out_of_mutex);
+  if (!is_int_property) {
+    return false;
+  }
+
+  uint64_t sum = 0;
+  {
+    // Needs mutex to protect the list of column families.
+    InstrumentedMutexLock l(&mutex_);
+    uint64_t value;
+    for (auto* cfd : *versions_->GetColumnFamilySet()) {
+      if (GetIntPropertyInternal(cfd, property_type, need_out_of_mutex, true,
+                                 &value)) {
+        sum += value;
+      } else {
+        return false;
+      }
+    }
+  }
+  *aggregated_value = sum;
+  return true;
+}
+
 SuperVersion* DBImpl::GetAndRefSuperVersion(ColumnFamilyData* cfd) {
   // TODO(ljin): consider using GetReferencedSuperVersion() directly
   return cfd->GetThreadLocalSuperVersion(&mutex_);
@@ -4664,16 +4850,18 @@ Status DB::Open(const DBOptions& db_options, const std::string& dbname,
     EnvOptions soptions(db_options);
     EnvOptions opt_env_options =
         impl->db_options_.env->OptimizeForLogWrite(soptions, impl->db_options_);
-    s = impl->db_options_.env->NewWritableFile(
-        LogFileName(impl->db_options_.wal_dir, new_log_number), &lfile,
-        opt_env_options);
+    s = NewWritableFile(impl->db_options_.env,
+                        LogFileName(impl->db_options_.wal_dir, new_log_number),
+                        &lfile, opt_env_options);
     if (s.ok()) {
       lfile->SetPreallocationBlockSize(1.1 * max_write_buffer_size);
       impl->logfile_number_ = new_log_number;
       unique_ptr<WritableFileWriter> file_writer(
           new WritableFileWriter(std::move(lfile), opt_env_options));
-      impl->logs_.emplace_back(new_log_number,
-                               new log::Writer(std::move(file_writer)));
+      impl->logs_.emplace_back(
+          new_log_number,
+          new log::Writer(std::move(file_writer), new_log_number,
+                          impl->db_options_.recycle_log_file_num > 0));
 
       // set column family handles
       for (auto cf : column_families) {

db/db_impl.h

@@ -36,11 +36,11 @@
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/transaction_log.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/autovector.h"
 #include "util/event_logger.h"
 #include "util/hash.h"
 #include "util/instrumented_mutex.h"
-#include "util/scoped_arena_iterator.h"
 #include "util/stop_watch.h"
 #include "util/thread_local.h"
 
@@ -122,6 +122,9 @@ class DBImpl : public DB {
   using DB::GetIntProperty;
   virtual bool GetIntProperty(ColumnFamilyHandle* column_family,
                               const Slice& property, uint64_t* value) override;
+  using DB::GetAggregatedIntProperty;
+  virtual bool GetAggregatedIntProperty(const Slice& property,
+                                        uint64_t* aggregated_value) override;
   using DB::GetApproximateSizes;
   virtual void GetApproximateSizes(ColumnFamilyHandle* column_family,
                                    const Range* range, int n, uint64_t* sizes,
@@ -248,7 +251,13 @@ class DBImpl : public DB {
                              const Slice* begin, const Slice* end,
                              bool disallow_trivial_move = false);
 
-#ifndef ROCKSDB_LITE
+  // Return an internal iterator over the current state of the database.
+  // The keys of this iterator are internal keys (see format.h).
+  // The returned iterator should be deleted when no longer needed.
+  InternalIterator* NewInternalIterator(
+      Arena* arena, ColumnFamilyHandle* column_family = nullptr);
+
+#ifndef NDEBUG
   // Extra methods (for testing) that are not in the public DB interface
   // Implemented in db_impl_debug.cc
 
@@ -266,12 +275,6 @@ class DBImpl : public DB {
   // Wait for any compaction
   Status TEST_WaitForCompact();
 
-  // Return an internal iterator over the current state of the database.
-  // The keys of this iterator are internal keys (see format.h).
-  // The returned iterator should be deleted when no longer needed.
-  Iterator* TEST_NewInternalIterator(
-      Arena* arena, ColumnFamilyHandle* column_family = nullptr);
-
   // Return the maximum overlapping data (in bytes) at next level for any
   // file at a level >= 1.
   int64_t TEST_MaxNextLevelOverlappingBytes(ColumnFamilyHandle* column_family =
@@ -305,7 +308,13 @@ class DBImpl : public DB {
 
   uint64_t TEST_LogfileNumber();
 
-#endif  // ROCKSDB_LITE
+  // Returns column family name to ImmutableCFOptions map.
+  Status TEST_GetAllImmutableCFOptions(
+      std::unordered_map<std::string, const ImmutableCFOptions*>* iopts_map);
+
+  Cache* TEST_table_cache() { return table_cache_.get(); }
+
+#endif  // NDEBUG
 
   // Returns the list of live files in 'live' and the list
   // of all files in the filesystem in 'candidate_files'.
@@ -363,6 +372,20 @@ class DBImpl : public DB {
   // Same as above, should called without mutex held and not on write thread.
   ColumnFamilyHandle* GetColumnFamilyHandleUnlocked(uint32_t column_family_id);
 
+  // Returns the number of currently running flushes.
+  // REQUIREMENT: mutex_ must be held when calling this function.
+  int num_running_flushes() {
+    mutex_.AssertHeld();
+    return num_running_flushes_;
+  }
+
+  // Returns the number of currently running compactions.
+  // REQUIREMENT: mutex_ must be held when calling this function.
+  int num_running_compactions() {
+    mutex_.AssertHeld();
+    return num_running_compactions_;
+  }
+
  protected:
   Env* const env_;
   const std::string dbname_;
@@ -370,12 +393,14 @@ class DBImpl : public DB {
   const DBOptions db_options_;
   Statistics* stats_;
 
-  Iterator* NewInternalIterator(const ReadOptions&, ColumnFamilyData* cfd,
-                                SuperVersion* super_version, Arena* arena);
+  InternalIterator* NewInternalIterator(const ReadOptions&,
+                                        ColumnFamilyData* cfd,
+                                        SuperVersion* super_version,
+                                        Arena* arena);
 
   void NotifyOnFlushCompleted(ColumnFamilyData* cfd, FileMetaData* file_meta,
                               const MutableCFOptions& mutable_cf_options,
-                              int job_id);
+                              int job_id, TableProperties prop);
 
   void NotifyOnCompactionCompleted(ColumnFamilyData* cfd,
                                    Compaction *c, const Status &st,
@@ -540,6 +565,8 @@ class DBImpl : public DB {
   // * whenever there is an error in background flush or compaction
   InstrumentedCondVar bg_cv_;
   uint64_t logfile_number_;
+  std::deque<uint64_t>
+      log_recycle_files;  // a list of log files that we can recycle
   bool log_dir_synced_;
   bool log_empty_;
   ColumnFamilyHandleImpl* default_cf_handle_;
@@ -685,6 +712,9 @@ class DBImpl : public DB {
   // count how many background compactions are running or have been scheduled
   int bg_compaction_scheduled_;
 
+  // stores the number of compactions are currently running
+  int num_running_compactions_;
+
   // If non-zero, MaybeScheduleFlushOrCompaction() will only schedule manual
   // compactions (if manual_compaction_ is not null). This mechanism enables
   // manual compactions to wait until all other compactions are finished.
@@ -693,6 +723,9 @@ class DBImpl : public DB {
   // number of background memtable flush jobs, submitted to the HIGH pool
   int bg_flush_scheduled_;
 
+  // stores the number of flushes are currently running
+  int num_running_flushes_;
+
   // Information for a manual compaction
   struct ManualCompaction {
     ColumnFamilyData* cfd;
@@ -748,9 +781,12 @@ class DBImpl : public DB {
   // Unified interface for logging events
   EventLogger event_logger_;
 
-  // A value of >0 temporarily disables scheduling of background work
+  // A value of > 0 temporarily disables scheduling of background work
   int bg_work_paused_;
 
+  // A value of > 0 temporarily disables scheduling of background compaction
+  int bg_compaction_paused_;
+
   // Guard against multiple concurrent refitting
   bool refitting_level_;
 
@@ -788,6 +824,10 @@ class DBImpl : public DB {
   virtual Status GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
                                           TablePropertiesCollection* props)
       override;
+  virtual Status GetPropertiesOfTablesInRange(
+      ColumnFamilyHandle* column_family, const Range* range, std::size_t n,
+      TablePropertiesCollection* props) override;
+
 #endif  // ROCKSDB_LITE
 
   // Function that Get and KeyMayExist call with no_io true or false
@@ -796,9 +836,10 @@ class DBImpl : public DB {
                  const Slice& key, std::string* value,
                  bool* value_found = nullptr);
 
-  bool GetIntPropertyInternal(ColumnFamilyHandle* column_family,
+  bool GetIntPropertyInternal(ColumnFamilyData* cfd,
                               DBPropertyType property_type,
-                              bool need_out_of_mutex, uint64_t* value);
+                              bool need_out_of_mutex, bool is_locked,
+                              uint64_t* value);
 };
 
 // Sanitize db options.  The caller should delete result.info_log if

db/db_impl_debug.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.
 
-#ifndef ROCKSDB_LITE
+#ifndef NDEBUG
 
 #include "db/db_impl.h"
 #include "util/thread_status_updater.h"
@@ -19,23 +19,6 @@ uint64_t DBImpl::TEST_GetLevel0TotalSize() {
   return default_cf_handle_->cfd()->current()->storage_info()->NumLevelBytes(0);
 }
 
-Iterator* DBImpl::TEST_NewInternalIterator(Arena* arena,
-                                           ColumnFamilyHandle* column_family) {
-  ColumnFamilyData* cfd;
-  if (column_family == nullptr) {
-    cfd = default_cf_handle_->cfd();
-  } else {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
-    cfd = cfh->cfd();
-  }
-
-  mutex_.Lock();
-  SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
-  mutex_.Unlock();
-  ReadOptions roptions;
-  return NewInternalIterator(roptions, cfd, super_version, arena);
-}
-
 int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
     ColumnFamilyHandle* column_family) {
   ColumnFamilyData* cfd;
@@ -152,5 +135,24 @@ uint64_t DBImpl::TEST_LogfileNumber() {
   return logfile_number_;
 }
 
+Status DBImpl::TEST_GetAllImmutableCFOptions(
+    std::unordered_map<std::string, const ImmutableCFOptions*>* iopts_map) {
+  std::vector<std::string> cf_names;
+  std::vector<const ImmutableCFOptions*> iopts;
+  {
+    InstrumentedMutexLock l(&mutex_);
+    for (auto cfd : *versions_->GetColumnFamilySet()) {
+      cf_names.push_back(cfd->GetName());
+      iopts.push_back(cfd->ioptions());
+    }
+  }
+  iopts_map->clear();
+  for (size_t i = 0; i < cf_names.size(); ++i) {
+    iopts_map->insert({cf_names[i], iopts[i]});
+  }
+
+  return Status::OK();
+}
+
 }  // namespace rocksdb
-#endif  // ROCKSDB_LITE
+#endif  // NDEBUG

db/db_inplace_update_test.cc

@@ -6,8 +6,8 @@
 // Copyright (c) 2011 The LevelDB Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
 
 namespace rocksdb {
 

db/db_iter.cc

@@ -15,11 +15,12 @@
 
 #include "db/filename.h"
 #include "db/dbformat.h"
+#include "port/port.h"
 #include "rocksdb/env.h"
 #include "rocksdb/options.h"
 #include "rocksdb/iterator.h"
 #include "rocksdb/merge_operator.h"
-#include "port/port.h"
+#include "table/internal_iterator.h"
 #include "util/arena.h"
 #include "util/logging.h"
 #include "util/mutexlock.h"
@@ -58,10 +59,11 @@ class DBIter: public Iterator {
     kReverse
   };
 
-  DBIter(Env* env, const ImmutableCFOptions& ioptions,
-         const Comparator* cmp, Iterator* iter, SequenceNumber s,
-         bool arena_mode, uint64_t max_sequential_skip_in_iterations,
-         const Slice* iterate_upper_bound = nullptr)
+  DBIter(Env* env, const ImmutableCFOptions& ioptions, const Comparator* cmp,
+         InternalIterator* iter, SequenceNumber s, bool arena_mode,
+         uint64_t max_sequential_skip_in_iterations,
+         const Slice* iterate_upper_bound = nullptr,
+         bool prefix_same_as_start = false)
       : arena_mode_(arena_mode),
         env_(env),
         logger_(ioptions.info_log),
@@ -73,7 +75,8 @@ class DBIter: public Iterator {
         valid_(false),
         current_entry_is_merged_(false),
         statistics_(ioptions.statistics),
-        iterate_upper_bound_(iterate_upper_bound) {
+        iterate_upper_bound_(iterate_upper_bound),
+        prefix_same_as_start_(prefix_same_as_start) {
     RecordTick(statistics_, NO_ITERATORS);
     prefix_extractor_ = ioptions.prefix_extractor;
     max_skip_ = max_sequential_skip_in_iterations;
@@ -83,10 +86,10 @@ class DBIter: public Iterator {
     if (!arena_mode_) {
       delete iter_;
     } else {
-      iter_->~Iterator();
+      iter_->~InternalIterator();
     }
   }
-  virtual void SetIter(Iterator* iter) {
+  virtual void SetIter(InternalIterator* iter) {
     assert(iter_ == nullptr);
     iter_ = iter;
   }
@@ -142,7 +145,7 @@ class DBIter: public Iterator {
   Logger* logger_;
   const Comparator* const user_comparator_;
   const MergeOperator* const user_merge_operator_;
-  Iterator* iter_;
+  InternalIterator* iter_;
   SequenceNumber const sequence_;
 
   Status status_;
@@ -154,6 +157,8 @@ class DBIter: public Iterator {
   Statistics* statistics_;
   uint64_t max_skip_;
   const Slice* iterate_upper_bound_;
+  IterKey prefix_start_;
+  bool prefix_same_as_start_;
 
   // No copying allowed
   DBIter(const DBIter&);
@@ -196,6 +201,11 @@ void DBIter::Next() {
       RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
     }
   }
+  if (valid_ && prefix_extractor_ && prefix_same_as_start_ &&
+      prefix_extractor_->Transform(saved_key_.GetKey())
+              .compare(prefix_start_.GetKey()) != 0) {
+    valid_ = false;
+  }
 }
 
 // PRE: saved_key_ has the current user key if skipping
@@ -366,6 +376,11 @@ void DBIter::Prev() {
       RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
     }
   }
+  if (valid_ && prefix_extractor_ && prefix_same_as_start_ &&
+      prefix_extractor_->Transform(saved_key_.GetKey())
+              .compare(prefix_start_.GetKey()) != 0) {
+    valid_ = false;
+  }
 }
 
 void DBIter::ReverseToBackward() {
@@ -667,6 +682,9 @@ void DBIter::Seek(const Slice& target) {
   } else {
     valid_ = false;
   }
+  if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
+    prefix_start_.SetKey(prefix_extractor_->Transform(target));
+  }
 }
 
 void DBIter::SeekToFirst() {
@@ -695,6 +713,9 @@ void DBIter::SeekToFirst() {
   } else {
     valid_ = false;
   }
+  if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
+    prefix_start_.SetKey(prefix_extractor_->Transform(saved_key_.GetKey()));
+  }
 }
 
 void DBIter::SeekToLast() {
@@ -740,24 +761,28 @@ void DBIter::SeekToLast() {
       RecordTick(statistics_, ITER_BYTES_READ, key().size() + value().size());
     }
   }
+  if (valid_ && prefix_extractor_ && prefix_same_as_start_) {
+    prefix_start_.SetKey(prefix_extractor_->Transform(saved_key_.GetKey()));
+  }
 }
 
 Iterator* NewDBIterator(Env* env, const ImmutableCFOptions& ioptions,
                         const Comparator* user_key_comparator,
-                        Iterator* internal_iter,
+                        InternalIterator* internal_iter,
                         const SequenceNumber& sequence,
                         uint64_t max_sequential_skip_in_iterations,
-                        const Slice* iterate_upper_bound) {
+                        const Slice* iterate_upper_bound,
+                        bool prefix_same_as_start) {
   return new DBIter(env, ioptions, user_key_comparator, internal_iter, sequence,
                     false, max_sequential_skip_in_iterations,
-                    iterate_upper_bound);
+                    iterate_upper_bound, prefix_same_as_start);
 }
 
 ArenaWrappedDBIter::~ArenaWrappedDBIter() { db_iter_->~DBIter(); }
 
 void ArenaWrappedDBIter::SetDBIter(DBIter* iter) { db_iter_ = iter; }
 
-void ArenaWrappedDBIter::SetIterUnderDBIter(Iterator* iter) {
+void ArenaWrappedDBIter::SetIterUnderDBIter(InternalIterator* iter) {
   static_cast<DBIter*>(db_iter_)->SetIter(iter);
 }
 
@@ -779,16 +804,16 @@ void ArenaWrappedDBIter::RegisterCleanup(CleanupFunction function, void* arg1,
 
 ArenaWrappedDBIter* NewArenaWrappedDbIterator(
     Env* env, const ImmutableCFOptions& ioptions,
-    const Comparator* user_key_comparator,
-    const SequenceNumber& sequence,
+    const Comparator* user_key_comparator, const SequenceNumber& sequence,
     uint64_t max_sequential_skip_in_iterations,
-    const Slice* iterate_upper_bound) {
+    const Slice* iterate_upper_bound, bool prefix_same_as_start) {
   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,
-      iterate_upper_bound);
+  DBIter* db_iter =
+      new (mem) DBIter(env, ioptions, user_key_comparator, nullptr, sequence,
+                       true, max_sequential_skip_in_iterations,
+                       iterate_upper_bound, prefix_same_as_start);
 
   iter->SetDBIter(db_iter);
 

db/db_iter.h

@@ -10,6 +10,7 @@
 #pragma once
 #include <stdint.h>
 #include "rocksdb/db.h"
+#include "rocksdb/iterator.h"
 #include "db/dbformat.h"
 #include "util/arena.h"
 #include "util/autovector.h"
@@ -18,18 +19,18 @@ namespace rocksdb {
 
 class Arena;
 class DBIter;
+class InternalIterator;
 
 // Return a new iterator that converts internal keys (yielded by
 // "*internal_iter") that were live at the specified "sequence" number
 // into appropriate user keys.
-extern Iterator* NewDBIterator(
-    Env* env,
-    const ImmutableCFOptions& options,
-    const Comparator *user_key_comparator,
-    Iterator* internal_iter,
-    const SequenceNumber& sequence,
-    uint64_t max_sequential_skip_in_iterations,
-    const Slice* iterate_upper_bound = nullptr);
+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,
+                               const Slice* iterate_upper_bound = nullptr,
+                               bool prefix_same_as_start = false);
 
 // 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
@@ -50,7 +51,7 @@ class ArenaWrappedDBIter : public Iterator {
 
   // Set the internal iterator wrapped inside the DB Iterator. Usually it is
   // a merging iterator.
-  virtual void SetIterUnderDBIter(Iterator* iter);
+  virtual void SetIterUnderDBIter(InternalIterator* iter);
   virtual bool Valid() const override;
   virtual void SeekToFirst() override;
   virtual void SeekToLast() override;
@@ -60,6 +61,7 @@ class ArenaWrappedDBIter : public Iterator {
   virtual Slice key() const override;
   virtual Slice value() const override;
   virtual Status status() const override;
+
   void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
 
  private:
@@ -70,8 +72,9 @@ class ArenaWrappedDBIter : public Iterator {
 // Generate the arena wrapped iterator class.
 extern ArenaWrappedDBIter* NewArenaWrappedDbIterator(
     Env* env, const ImmutableCFOptions& options,
-    const Comparator* user_key_comparator,
-    const SequenceNumber& sequence, uint64_t max_sequential_skip_in_iterations,
-    const Slice* iterate_upper_bound = nullptr);
+    const Comparator* user_key_comparator, const SequenceNumber& sequence,
+    uint64_t max_sequential_skip_in_iterations,
+    const Slice* iterate_upper_bound = nullptr,
+    bool prefix_same_as_start = false);
 
 }  // namespace rocksdb

db/db_iter_test.cc

@@ -29,7 +29,7 @@ static uint64_t TestGetTickerCount(const Options& options,
   return options.statistics->getTickerCount(ticker_type);
 }
 
-class TestIterator : public Iterator {
+class TestIterator : public InternalIterator {
  public:
   explicit TestIterator(const Comparator* comparator)
       : initialized_(false),
@@ -1864,11 +1864,12 @@ class DBIterWithMergeIterTest : public testing::Test {
     internal_iter2_->Add("d", kTypeValue, "7", 3u);
     internal_iter2_->Finish();
 
-    std::vector<Iterator*> child_iters;
+    std::vector<InternalIterator*> child_iters;
     child_iters.push_back(internal_iter1_);
     child_iters.push_back(internal_iter2_);
     InternalKeyComparator icomp(BytewiseComparator());
-    Iterator* merge_iter = NewMergingIterator(&icomp_, &child_iters[0], 2u);
+    InternalIterator* merge_iter =
+        NewMergingIterator(&icomp_, &child_iters[0], 2u);
 
     db_iter_.reset(NewDBIterator(env_, ImmutableCFOptions(options_),
                                  BytewiseComparator(), merge_iter,
@@ -1942,8 +1943,6 @@ TEST_F(DBIterWithMergeIterTest, InnerMergeIterator2) {
   ASSERT_EQ(db_iter_->value().ToString(), "4");
 }
 
-#if !(defined NDEBUG) || !defined(OS_WIN)
-
 TEST_F(DBIterWithMergeIterTest, InnerMergeIteratorDataRace1) {
   // Test Prev() when one child iterator is at its end but more rows
   // are added.
@@ -2294,7 +2293,6 @@ TEST_F(DBIterWithMergeIterTest, InnerMergeIteratorDataRace8) {
 
   rocksdb::SyncPoint::GetInstance()->DisableProcessing();
 }
-#endif // #if !(defined NDEBUG) || !defined(OS_WIN)
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/db_log_iter_test.cc

@@ -10,10 +10,10 @@
 // Introduction of SyncPoint effectively disabled building and running this test
 // in Release build.
 // which is a pity, it is a good test
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
 
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
 
 namespace rocksdb {
 
@@ -277,10 +277,10 @@ TEST_F(DBTestXactLogIterator, TransactionLogIteratorBlobs) {
 }
 }  // namespace rocksdb
 
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif  // !defined(ROCKSDB_LITE)
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();

db/db_table_properties_test.cc

@@ -0,0 +1,213 @@
+//  Copyright (c) 2013, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under the BSD-style license found in the
+//  LICENSE file in the root directory of this source tree. An additional grant
+//  of patent rights can be found in the PATENTS file in the same directory.
+//
+// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file. See the AUTHORS file for names of contributors.
+
+#include <unordered_set>
+#include <vector>
+
+#include "db/db_test_util.h"
+#include "port/stack_trace.h"
+#include "rocksdb/db.h"
+#include "util/testharness.h"
+#include "util/testutil.h"
+
+#ifndef ROCKSDB_LITE
+
+namespace rocksdb {
+
+// A helper function that ensures the table properties returned in
+// `GetPropertiesOfAllTablesTest` is correct.
+// This test assumes entries size is different for each of the tables.
+namespace {
+
+void VerifyTableProperties(DB* db, uint64_t expected_entries_size) {
+  TablePropertiesCollection props;
+  ASSERT_OK(db->GetPropertiesOfAllTables(&props));
+
+  ASSERT_EQ(4U, props.size());
+  std::unordered_set<uint64_t> unique_entries;
+
+  // Indirect test
+  uint64_t sum = 0;
+  for (const auto& item : props) {
+    unique_entries.insert(item.second->num_entries);
+    sum += item.second->num_entries;
+  }
+
+  ASSERT_EQ(props.size(), unique_entries.size());
+  ASSERT_EQ(expected_entries_size, sum);
+}
+}  // namespace
+
+class DBTablePropertiesTest : public DBTestBase {
+ public:
+  DBTablePropertiesTest() : DBTestBase("/db_table_properties_test") {}
+  TablePropertiesCollection TestGetPropertiesOfTablesInRange(
+      std::vector<Range> ranges, std::size_t* num_properties = nullptr,
+      std::size_t* num_files = nullptr);
+};
+
+TEST_F(DBTablePropertiesTest, GetPropertiesOfAllTablesTest) {
+  Options options = CurrentOptions();
+  options.level0_file_num_compaction_trigger = 8;
+  Reopen(options);
+  // Create 4 tables
+  for (int table = 0; table < 4; ++table) {
+    for (int i = 0; i < 10 + table; ++i) {
+      db_->Put(WriteOptions(), ToString(table * 100 + i), "val");
+    }
+    db_->Flush(FlushOptions());
+  }
+
+  // 1. Read table properties directly from file
+  Reopen(options);
+  VerifyTableProperties(db_, 10 + 11 + 12 + 13);
+
+  // 2. Put two tables to table cache and
+  Reopen(options);
+  // fetch key from 1st and 2nd table, which will internally place that table to
+  // the table cache.
+  for (int i = 0; i < 2; ++i) {
+    Get(ToString(i * 100 + 0));
+  }
+
+  VerifyTableProperties(db_, 10 + 11 + 12 + 13);
+
+  // 3. Put all tables to table cache
+  Reopen(options);
+  // fetch key from 1st and 2nd table, which will internally place that table to
+  // the table cache.
+  for (int i = 0; i < 4; ++i) {
+    Get(ToString(i * 100 + 0));
+  }
+  VerifyTableProperties(db_, 10 + 11 + 12 + 13);
+}
+
+TablePropertiesCollection
+DBTablePropertiesTest::TestGetPropertiesOfTablesInRange(
+    std::vector<Range> ranges, std::size_t* num_properties,
+    std::size_t* num_files) {
+  // run the query
+  TablePropertiesCollection props;
+  EXPECT_OK(db_->GetPropertiesOfTablesInRange(
+      db_->DefaultColumnFamily(), &ranges[0], ranges.size(), &props));
+
+  // Make sure that we've received properties for those and for those files
+  // only which fall within requested ranges
+  std::vector<LiveFileMetaData> vmd;
+  db_->GetLiveFilesMetaData(&vmd);
+  for (auto md : vmd) {
+    std::string fn = md.db_path + md.name;
+    bool in_range = false;
+    for (auto r : ranges) {
+      // smallestkey < limit && largestkey >= start
+      if (r.limit.compare(md.smallestkey) >= 0 &&
+          r.start.compare(md.largestkey) <= 0) {
+        in_range = true;
+        EXPECT_GT(props.count(fn), 0);
+      }
+    }
+    if (!in_range) {
+      EXPECT_EQ(props.count(fn), 0);
+    }
+  }
+
+  if (num_properties) {
+    *num_properties = props.size();
+  }
+
+  if (num_files) {
+    *num_files = vmd.size();
+  }
+  return props;
+}
+
+TEST_F(DBTablePropertiesTest, GetPropertiesOfTablesInRange) {
+  // Fixed random sead
+  Random rnd(301);
+
+  Options options;
+  options.create_if_missing = true;
+  options.write_buffer_size = 4096;
+  options.max_write_buffer_number = 8;
+  options.level0_file_num_compaction_trigger = 2;
+  options.level0_slowdown_writes_trigger = 2;
+  options.level0_stop_writes_trigger = 4;
+  options.target_file_size_base = 2048;
+  options.max_bytes_for_level_base = 10240;
+  options.max_bytes_for_level_multiplier = 4;
+  options.soft_rate_limit = 1.1;
+  options.num_levels = 8;
+
+  DestroyAndReopen(options);
+
+  // build a decent LSM
+  for (int i = 0; i < 10000; i++) {
+    ASSERT_OK(Put(test::RandomKey(&rnd, 5), RandomString(&rnd, 102)));
+  }
+  Flush();
+  db_->PauseBackgroundWork();
+
+  // Ensure that we have at least L0, L1 and L2
+  ASSERT_GT(NumTableFilesAtLevel(0), 0);
+  ASSERT_GT(NumTableFilesAtLevel(1), 0);
+  ASSERT_GT(NumTableFilesAtLevel(2), 0);
+
+  // Query the largest range
+  std::size_t num_properties, num_files;
+  TestGetPropertiesOfTablesInRange(
+      {Range(test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST),
+             test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
+      &num_properties, &num_files);
+  ASSERT_EQ(num_properties, num_files);
+
+  // Query the empty range
+  TestGetPropertiesOfTablesInRange(
+      {Range(test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST),
+             test::RandomKey(&rnd, 5, test::RandomKeyType::SMALLEST))},
+      &num_properties, &num_files);
+  ASSERT_GT(num_files, 0);
+  ASSERT_EQ(num_properties, 0);
+
+  // Query the middle rangee
+  TestGetPropertiesOfTablesInRange(
+      {Range(test::RandomKey(&rnd, 5, test::RandomKeyType::MIDDLE),
+             test::RandomKey(&rnd, 5, test::RandomKeyType::LARGEST))},
+      &num_properties, &num_files);
+  ASSERT_GT(num_files, 0);
+  ASSERT_GT(num_files, num_properties);
+  ASSERT_GT(num_properties, 0);
+
+  // Query a bunch of random ranges
+  for (int j = 0; j < 100; j++) {
+    // create a bunch of ranges
+    std::vector<std::string> random_keys;
+    auto n = 2 * rnd.Uniform(50);
+    for (uint32_t i = 0; i < n; ++i) {
+      random_keys.push_back(test::RandomKey(&rnd, 5));
+    }
+
+    std::vector<Range> ranges;
+    auto it = random_keys.begin();
+    while (it != random_keys.end()) {
+      ranges.push_back(Range(*it, *(it + 1)));
+      it += 2;
+    }
+
+    TestGetPropertiesOfTablesInRange(std::move(ranges));
+  }
+}
+}  // namespace rocksdb
+
+#endif  // ROCKSDB_LITE
+
+int main(int argc, char** argv) {
+  rocksdb::port::InstallStackTraceHandler();
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

db/db_tailing_iter_test.cc

@@ -10,11 +10,11 @@
 // Introduction of SyncPoint effectively disabled building and running this test
 // in Release build.
 // which is a pity, it is a good test
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
 
+#include "db/db_test_util.h"
 #include "db/forward_iterator.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
 
 namespace rocksdb {
 
@@ -646,10 +646,10 @@ TEST_F(DBTestTailingIterator, ManagedTailingIteratorSeekToSame) {
 
 }  // namespace rocksdb
 
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif  // !defined(ROCKSDB_LITE)
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();

db/db_test_util.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/db_test_util.h"
+#include "db/db_test_util.h"
 
 namespace rocksdb {
 
@@ -38,11 +38,10 @@ SpecialEnv::SpecialEnv(Env* base)
   table_write_callback_ = nullptr;
 }
 
-
-DBTestBase::DBTestBase(const std::string path) : option_config_(kDefault),
-           mem_env_(!getenv("MEM_ENV") ? nullptr :
-                                         new MockEnv(Env::Default())),
-           env_(new SpecialEnv(mem_env_ ? mem_env_ : Env::Default())) {
+DBTestBase::DBTestBase(const std::string path)
+    : option_config_(kDefault),
+      mem_env_(!getenv("MEM_ENV") ? nullptr : new MockEnv(Env::Default())),
+      env_(new SpecialEnv(mem_env_ ? mem_env_ : Env::Default())) {
   env_->SetBackgroundThreads(1, Env::LOW);
   env_->SetBackgroundThreads(1, Env::HIGH);
   dbname_ = test::TmpDir(env_) + path;
@@ -59,12 +58,9 @@ DBTestBase::DBTestBase(const std::string path) : option_config_(kDefault),
 }
 
 DBTestBase::~DBTestBase() {
-// SyncPoint is not supported in Released Windows Mode.
-#if !(defined NDEBUG) || !defined(OS_WIN)
   rocksdb::SyncPoint::GetInstance()->DisableProcessing();
   rocksdb::SyncPoint::GetInstance()->LoadDependency({});
   rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
   Close();
   Options options;
   options.db_paths.emplace_back(dbname_, 0);
@@ -79,6 +75,23 @@ DBTestBase::~DBTestBase() {
 // test.  Return false if there are no more configurations to test.
 bool DBTestBase::ChangeOptions(int skip_mask) {
   for (option_config_++; option_config_ < kEnd; option_config_++) {
+#ifdef ROCKSDB_LITE
+    // These options are not supported in ROCKSDB_LITE
+    if (option_config_ == kHashSkipList ||
+        option_config_ == kPlainTableFirstBytePrefix ||
+        option_config_ == kPlainTableCappedPrefix ||
+        option_config_ == kPlainTableCappedPrefixNonMmap ||
+        option_config_ == kPlainTableAllBytesPrefix ||
+        option_config_ == kVectorRep || option_config_ == kHashLinkList ||
+        option_config_ == kHashCuckoo ||
+        option_config_ == kUniversalCompaction ||
+        option_config_ == kUniversalCompactionMultiLevel ||
+        option_config_ == kUniversalSubcompactions ||
+        option_config_ == kFIFOCompaction) {
+      continue;
+    }
+#endif
+
     if ((skip_mask & kSkipDeletesFilterFirst) &&
         option_config_ == kDeletesFilterFirst) {
       continue;
@@ -92,8 +105,7 @@ bool DBTestBase::ChangeOptions(int skip_mask) {
       continue;
     }
     if ((skip_mask & kSkipNoSeekToLast) &&
-        (option_config_ == kHashLinkList ||
-         option_config_ == kHashSkipList)) {;
+        (option_config_ == kHashLinkList || option_config_ == kHashSkipList)) {
       continue;
     }
     if ((skip_mask & kSkipPlainTable) &&
@@ -115,8 +127,7 @@ bool DBTestBase::ChangeOptions(int skip_mask) {
         option_config_ == kFIFOCompaction) {
       continue;
     }
-    if ((skip_mask & kSkipMmapReads) &&
-        option_config_ == kWalDirAndMmapReads) {
+    if ((skip_mask & kSkipMmapReads) && option_config_ == kWalDirAndMmapReads) {
       continue;
     }
     break;
@@ -205,10 +216,10 @@ Options DBTestBase::CurrentOptions(
   BlockBasedTableOptions table_options;
   bool set_block_based_table_factory = true;
   switch (option_config_) {
+#ifndef ROCKSDB_LITE
     case kHashSkipList:
       options.prefix_extractor.reset(NewFixedPrefixTransform(1));
-      options.memtable_factory.reset(
-          NewHashSkipListRepFactory(16));
+      options.memtable_factory.reset(NewHashSkipListRepFactory(16));
       break;
     case kPlainTableFirstBytePrefix:
       options.table_factory.reset(new PlainTableFactory());
@@ -238,6 +249,19 @@ Options DBTestBase::CurrentOptions(
       options.max_sequential_skip_in_iterations = 999999;
       set_block_based_table_factory = false;
       break;
+    case kVectorRep:
+      options.memtable_factory.reset(new VectorRepFactory(100));
+      break;
+    case kHashLinkList:
+      options.prefix_extractor.reset(NewFixedPrefixTransform(1));
+      options.memtable_factory.reset(
+          NewHashLinkListRepFactory(4, 0, 3, true, 4));
+      break;
+    case kHashCuckoo:
+      options.memtable_factory.reset(
+          NewHashCuckooRepFactory(options.write_buffer_size));
+      break;
+#endif  // ROCKSDB_LITE
     case kMergePut:
       options.merge_operator = MergeOperators::CreatePutOperator();
       break;
@@ -274,18 +298,6 @@ Options DBTestBase::CurrentOptions(
     case kDeletesFilterFirst:
       options.filter_deletes = true;
       break;
-    case kVectorRep:
-      options.memtable_factory.reset(new VectorRepFactory(100));
-      break;
-    case kHashLinkList:
-      options.prefix_extractor.reset(NewFixedPrefixTransform(1));
-      options.memtable_factory.reset(
-          NewHashLinkListRepFactory(4, 0, 3, true, 4));
-      break;
-    case kHashCuckoo:
-      options.memtable_factory.reset(
-          NewHashCuckooRepFactory(options.write_buffer_size));
-      break;
     case kUniversalCompaction:
       options.compaction_style = kCompactionStyleUniversal;
       options.num_levels = 1;
@@ -296,7 +308,7 @@ Options DBTestBase::CurrentOptions(
       break;
     case kCompressedBlockCache:
       options.allow_mmap_writes = true;
-      table_options.block_cache_compressed = NewLRUCache(8*1024*1024);
+      table_options.block_cache_compressed = NewLRUCache(8 * 1024 * 1024);
       break;
     case kInfiniteMaxOpenFiles:
       options.max_open_files = -1;
@@ -328,6 +340,10 @@ Options DBTestBase::CurrentOptions(
       options.row_cache = NewLRUCache(1024 * 1024);
       break;
     }
+    case kRecycleLogFiles: {
+      options.recycle_log_file_num = 2;
+      break;
+    }
     case kLevelSubcompactions: {
       options.max_subcompactions = 4;
       break;
@@ -355,7 +371,7 @@ Options DBTestBase::CurrentOptions(
 }
 
 void DBTestBase::CreateColumnFamilies(const std::vector<std::string>& cfs,
-                          const Options& options) {
+                                      const Options& options) {
   ColumnFamilyOptions cf_opts(options);
   size_t cfi = handles_.size();
   handles_.resize(cfi + cfs.size());
@@ -365,7 +381,7 @@ void DBTestBase::CreateColumnFamilies(const std::vector<std::string>& cfs,
 }
 
 void DBTestBase::CreateAndReopenWithCF(const std::vector<std::string>& cfs,
-                           const Options& options) {
+                                       const Options& options) {
   CreateColumnFamilies(cfs, options);
   std::vector<std::string> cfs_plus_default = cfs;
   cfs_plus_default.insert(cfs_plus_default.begin(), kDefaultColumnFamilyName);
@@ -373,18 +389,17 @@ void DBTestBase::CreateAndReopenWithCF(const std::vector<std::string>& cfs,
 }
 
 void DBTestBase::ReopenWithColumnFamilies(const std::vector<std::string>& cfs,
-                              const std::vector<Options>& options) {
+                                          const std::vector<Options>& options) {
   ASSERT_OK(TryReopenWithColumnFamilies(cfs, options));
 }
 
 void DBTestBase::ReopenWithColumnFamilies(const std::vector<std::string>& cfs,
-                              const Options& options) {
+                                          const Options& options) {
   ASSERT_OK(TryReopenWithColumnFamilies(cfs, options));
 }
 
 Status DBTestBase::TryReopenWithColumnFamilies(
-    const std::vector<std::string>& cfs,
-    const std::vector<Options>& options) {
+    const std::vector<std::string>& cfs, const std::vector<Options>& options) {
   Close();
   EXPECT_EQ(cfs.size(), options.size());
   std::vector<ColumnFamilyDescriptor> column_families;
@@ -396,8 +411,7 @@ Status DBTestBase::TryReopenWithColumnFamilies(
 }
 
 Status DBTestBase::TryReopenWithColumnFamilies(
-    const std::vector<std::string>& cfs,
-    const Options& options) {
+    const std::vector<std::string>& cfs, const Options& options) {
   Close();
   std::vector<Options> v_opts(cfs.size(), options);
   return TryReopenWithColumnFamilies(cfs, v_opts);
@@ -454,7 +468,7 @@ Status DBTestBase::Put(const Slice& k, const Slice& v, WriteOptions wo) {
 }
 
 Status DBTestBase::Put(int cf, const Slice& k, const Slice& v,
-           WriteOptions wo) {
+                       WriteOptions wo) {
   if (kMergePut == option_config_) {
     return db_->Merge(wo, handles_[cf], k, v);
   } else {
@@ -493,7 +507,7 @@ std::string DBTestBase::Get(const std::string& k, const Snapshot* snapshot) {
 }
 
 std::string DBTestBase::Get(int cf, const std::string& k,
-                const Snapshot* snapshot) {
+                            const Snapshot* snapshot) {
   ReadOptions options;
   options.verify_checksums = true;
   options.snapshot = snapshot;
@@ -552,9 +566,9 @@ std::string DBTestBase::AllEntriesFor(const Slice& user_key, int cf) {
   Arena arena;
   ScopedArenaIterator iter;
   if (cf == 0) {
-    iter.set(dbfull()->TEST_NewInternalIterator(&arena));
+    iter.set(dbfull()->NewInternalIterator(&arena));
   } else {
-    iter.set(dbfull()->TEST_NewInternalIterator(&arena, handles_[cf]));
+    iter.set(dbfull()->NewInternalIterator(&arena, handles_[cf]));
   }
   InternalKey target(user_key, kMaxSequenceNumber, kTypeValue);
   iter->Seek(target.Encode());
@@ -605,6 +619,7 @@ std::string DBTestBase::AllEntriesFor(const Slice& user_key, int cf) {
   return result;
 }
 
+#ifndef ROCKSDB_LITE
 int DBTestBase::NumSortedRuns(int cf) {
   ColumnFamilyMetaData cf_meta;
   if (cf == 0) {
@@ -631,20 +646,6 @@ uint64_t DBTestBase::TotalSize(int cf) {
   return cf_meta.size;
 }
 
-int DBTestBase::NumTableFilesAtLevel(int level, int cf) {
-  std::string property;
-  if (cf == 0) {
-    // default cfd
-    EXPECT_TRUE(db_->GetProperty(
-        "rocksdb.num-files-at-level" + NumberToString(level), &property));
-  } else {
-    EXPECT_TRUE(db_->GetProperty(
-        handles_[cf], "rocksdb.num-files-at-level" + NumberToString(level),
-        &property));
-  }
-  return atoi(property.c_str());
-}
-
 uint64_t DBTestBase::SizeAtLevel(int level) {
   std::vector<LiveFileMetaData> metadata;
   db_->GetLiveFilesMetaData(&metadata);
@@ -671,6 +672,27 @@ int DBTestBase::TotalLiveFiles(int cf) {
   return num_files;
 }
 
+size_t DBTestBase::CountLiveFiles() {
+  std::vector<LiveFileMetaData> metadata;
+  db_->GetLiveFilesMetaData(&metadata);
+  return metadata.size();
+}
+#endif  // ROCKSDB_LITE
+
+int DBTestBase::NumTableFilesAtLevel(int level, int cf) {
+  std::string property;
+  if (cf == 0) {
+    // default cfd
+    EXPECT_TRUE(db_->GetProperty(
+        "rocksdb.num-files-at-level" + NumberToString(level), &property));
+  } else {
+    EXPECT_TRUE(db_->GetProperty(
+        handles_[cf], "rocksdb.num-files-at-level" + NumberToString(level),
+        &property));
+  }
+  return atoi(property.c_str());
+}
+
 int DBTestBase::TotalTableFiles(int cf, int levels) {
   if (levels == -1) {
     levels = CurrentOptions().num_levels;
@@ -713,12 +735,6 @@ size_t DBTestBase::CountFiles() {
   return files.size() + logfiles.size();
 }
 
-size_t DBTestBase::CountLiveFiles() {
-  std::vector<LiveFileMetaData> metadata;
-  db_->GetLiveFilesMetaData(&metadata);
-  return metadata.size();
-}
-
 uint64_t DBTestBase::Size(const Slice& start, const Slice& limit, int cf) {
   Range r(start, limit);
   uint64_t size;
@@ -731,7 +747,7 @@ uint64_t DBTestBase::Size(const Slice& start, const Slice& limit, int cf) {
 }
 
 void DBTestBase::Compact(int cf, const Slice& start, const Slice& limit,
-             uint32_t target_path_id) {
+                         uint32_t target_path_id) {
   CompactRangeOptions compact_options;
   compact_options.target_path_id = target_path_id;
   ASSERT_OK(db_->CompactRange(compact_options, handles_[cf], &start, &limit));
@@ -748,9 +764,8 @@ void DBTestBase::Compact(const Slice& start, const Slice& limit) {
 
 // Do n memtable compactions, each of which produces an sstable
 // covering the range [small,large].
-void DBTestBase::MakeTables(
-    int n, const std::string& small,
-    const std::string& large, int cf) {
+void DBTestBase::MakeTables(int n, const std::string& small,
+                            const std::string& large, int cf) {
   for (int i = 0; i < n; i++) {
     ASSERT_OK(Put(cf, small, "begin"));
     ASSERT_OK(Put(cf, large, "end"));
@@ -761,8 +776,8 @@ void DBTestBase::MakeTables(
 
 // Prevent pushing of new sstables into deeper levels by adding
 // tables that cover a specified range to all levels.
-void DBTestBase::FillLevels(
-    const std::string& smallest, const std::string& largest, int cf) {
+void DBTestBase::FillLevels(const std::string& smallest,
+                            const std::string& largest, int cf) {
   MakeTables(db_->NumberLevels(handles_[cf]), smallest, largest, cf);
 }
 
@@ -779,7 +794,7 @@ void DBTestBase::MoveFilesToLevel(int level, int cf) {
 void DBTestBase::DumpFileCounts(const char* label) {
   fprintf(stderr, "---\n%s:\n", label);
   fprintf(stderr, "maxoverlap: %" PRIu64 "\n",
-      dbfull()->TEST_MaxNextLevelOverlappingBytes());
+          dbfull()->TEST_MaxNextLevelOverlappingBytes());
   for (int level = 0; level < db_->NumberLevels(); level++) {
     int num = NumTableFilesAtLevel(level);
     if (num > 0) {
@@ -888,9 +903,10 @@ void DBTestBase::VerifyIterLast(std::string expected_key, int cf) {
 //   sets newValue with delta
 // If previous value is not empty,
 //   updates previous value with 'b' string of previous value size - 1.
-UpdateStatus DBTestBase::updateInPlaceSmallerSize(
-    char* prevValue, uint32_t* prevSize,
-    Slice delta, std::string* newValue) {
+UpdateStatus DBTestBase::updateInPlaceSmallerSize(char* prevValue,
+                                                  uint32_t* prevSize,
+                                                  Slice delta,
+                                                  std::string* newValue) {
   if (prevValue == nullptr) {
     *newValue = std::string(delta.size(), 'c');
     return UpdateStatus::UPDATED;
@@ -902,9 +918,10 @@ UpdateStatus DBTestBase::updateInPlaceSmallerSize(
   }
 }
 
-UpdateStatus DBTestBase::updateInPlaceSmallerVarintSize(
-    char* prevValue, uint32_t* prevSize,
-    Slice delta, std::string* newValue) {
+UpdateStatus DBTestBase::updateInPlaceSmallerVarintSize(char* prevValue,
+                                                        uint32_t* prevSize,
+                                                        Slice delta,
+                                                        std::string* newValue) {
   if (prevValue == nullptr) {
     *newValue = std::string(delta.size(), 'c');
     return UpdateStatus::UPDATED;
@@ -916,16 +933,17 @@ UpdateStatus DBTestBase::updateInPlaceSmallerVarintSize(
   }
 }
 
-UpdateStatus DBTestBase::updateInPlaceLargerSize(
-    char* prevValue, uint32_t* prevSize,
-    Slice delta, std::string* newValue) {
+UpdateStatus DBTestBase::updateInPlaceLargerSize(char* prevValue,
+                                                 uint32_t* prevSize,
+                                                 Slice delta,
+                                                 std::string* newValue) {
   *newValue = std::string(delta.size(), 'c');
   return UpdateStatus::UPDATED;
 }
 
-UpdateStatus DBTestBase::updateInPlaceNoAction(
-    char* prevValue, uint32_t* prevSize,
-    Slice delta, std::string* newValue) {
+UpdateStatus DBTestBase::updateInPlaceNoAction(char* prevValue,
+                                               uint32_t* prevSize, Slice delta,
+                                               std::string* newValue) {
   return UpdateStatus::UPDATE_FAILED;
 }
 
@@ -934,9 +952,9 @@ void DBTestBase::validateNumberOfEntries(int numValues, int cf) {
   ScopedArenaIterator iter;
   Arena arena;
   if (cf != 0) {
-    iter.set(dbfull()->TEST_NewInternalIterator(&arena, handles_[cf]));
+    iter.set(dbfull()->NewInternalIterator(&arena, handles_[cf]));
   } else {
-    iter.set(dbfull()->TEST_NewInternalIterator(&arena));
+    iter.set(dbfull()->NewInternalIterator(&arena));
   }
   iter->SeekToFirst();
   ASSERT_EQ(iter->status().ok(), true);
@@ -953,9 +971,8 @@ void DBTestBase::validateNumberOfEntries(int numValues, int cf) {
   ASSERT_EQ(0, seq);
 }
 
-void DBTestBase::CopyFile(
-    const std::string& source, const std::string& destination,
-    uint64_t size) {
+void DBTestBase::CopyFile(const std::string& source,
+                          const std::string& destination, uint64_t size) {
   const EnvOptions soptions;
   unique_ptr<SequentialFile> srcfile;
   ASSERT_OK(env_->NewSequentialFile(source, &srcfile, soptions));

db/db_test_util.h

@@ -27,8 +27,10 @@
 #include <vector>
 
 #include "db/db_impl.h"
+#include "db/db_test_util.h"
 #include "db/dbformat.h"
 #include "db/filename.h"
+#include "memtable/hash_linklist_rep.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/convenience.h"
@@ -42,12 +44,11 @@
 #include "table/block_based_table_factory.h"
 #include "table/mock_table.h"
 #include "table/plain_table_factory.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/compression.h"
-#include "util/db_test_util.h"
-#include "util/hash_linklist_rep.h"
 #include "util/mock_env.h"
 #include "util/mutexlock.h"
-#include "util/scoped_arena_iterator.h"
+
 #include "util/string_util.h"
 // SyncPoint is not supported in Released Windows Mode.
 #if !(defined NDEBUG) || !defined(OS_WIN)
@@ -131,9 +132,7 @@ class SpecialEnv : public EnvWrapper {
 
      public:
       SSTableFile(SpecialEnv* env, unique_ptr<WritableFile>&& base)
-          : env_(env),
-            base_(std::move(base)) {
-      }
+          : env_(env), base_(std::move(base)) {}
       Status Append(const Slice& data) override {
         if (env_->table_write_callback_) {
           (*env_->table_write_callback_)();
@@ -148,9 +147,7 @@ class SpecialEnv : public EnvWrapper {
           return base_->Append(data);
         }
       }
-      Status Truncate(uint64_t size) override {
-        return base_->Truncate(size);
-      }
+      Status Truncate(uint64_t size) override { return base_->Truncate(size); }
       Status Close() override {
 // SyncPoint is not supported in Released Windows Mode.
 #if !(defined NDEBUG) || !defined(OS_WIN)
@@ -180,7 +177,7 @@ class SpecialEnv : public EnvWrapper {
     class ManifestFile : public WritableFile {
      public:
       ManifestFile(SpecialEnv* env, unique_ptr<WritableFile>&& b)
-          : env_(env), base_(std::move(b)) { }
+          : env_(env), base_(std::move(b)) {}
       Status Append(const Slice& data) override {
         if (env_->manifest_write_error_.load(std::memory_order_acquire)) {
           return Status::IOError("simulated writer error");
@@ -283,8 +280,7 @@ class SpecialEnv : public EnvWrapper {
      public:
       CountingFile(unique_ptr<RandomAccessFile>&& target,
                    anon::AtomicCounter* counter)
-          : target_(std::move(target)), counter_(counter) {
-      }
+          : target_(std::move(target)), counter_(counter) {}
       virtual Status Read(uint64_t offset, size_t n, Slice* result,
                           char* scratch) const override {
         counter_->Increment();
@@ -329,7 +325,6 @@ class SpecialEnv : public EnvWrapper {
     return s;
   }
 
-
   virtual void SleepForMicroseconds(int micros) override {
     sleep_counter_.Increment();
     if (no_sleep_) {
@@ -406,7 +401,7 @@ class SpecialEnv : public EnvWrapper {
   std::atomic<int64_t> addon_time_;
   bool no_sleep_;
 
-  std::atomic<bool> is_wal_sync_thread_safe_ {true};
+  std::atomic<bool> is_wal_sync_thread_safe_{true};
 };
 
 class DBTestBase : public testing::Test {
@@ -442,9 +437,10 @@ class DBTestBase : public testing::Test {
     kFIFOCompaction = 26,
     kOptimizeFiltersForHits = 27,
     kRowCache = 28,
-    kLevelSubcompactions = 29,
-    kUniversalSubcompactions = 30,
-    kEnd = 29
+    kRecycleLogFiles = 29,
+    kLevelSubcompactions = 30,
+    kUniversalSubcompactions = 31,
+    kEnd = 30
   };
   int option_config_;
 
@@ -509,9 +505,7 @@ class DBTestBase : public testing::Test {
       const Options& defaultOptions,
       const anon::OptionsOverride& options_override = anon::OptionsOverride());
 
-  DBImpl* dbfull() {
-    return reinterpret_cast<DBImpl*>(db_);
-  }
+  DBImpl* dbfull() { return reinterpret_cast<DBImpl*>(db_); }
 
   void CreateColumnFamilies(const std::vector<std::string>& cfs,
                             const Options& options);
@@ -525,9 +519,8 @@ class DBTestBase : public testing::Test {
   void ReopenWithColumnFamilies(const std::vector<std::string>& cfs,
                                 const Options& options);
 
-  Status TryReopenWithColumnFamilies(
-      const std::vector<std::string>& cfs,
-      const std::vector<Options>& options);
+  Status TryReopenWithColumnFamilies(const std::vector<std::string>& cfs,
+                                     const std::vector<Options>& options);
 
   Status TryReopenWithColumnFamilies(const std::vector<std::string>& cfs,
                                      const Options& options);
@@ -574,16 +567,20 @@ class DBTestBase : public testing::Test {
 
   std::string AllEntriesFor(const Slice& user_key, int cf = 0);
 
+#ifndef ROCKSDB_LITE
   int NumSortedRuns(int cf = 0);
 
   uint64_t TotalSize(int cf = 0);
 
-  int NumTableFilesAtLevel(int level, int cf = 0);
-
   uint64_t SizeAtLevel(int level);
 
   int TotalLiveFiles(int cf = 0);
 
+  size_t CountLiveFiles();
+#endif  // ROCKSDB_LITE
+
+  int NumTableFilesAtLevel(int level, int cf = 0);
+
   int TotalTableFiles(int cf = 0, int levels = -1);
 
   // Return spread of files per level
@@ -591,8 +588,6 @@ class DBTestBase : public testing::Test {
 
   size_t CountFiles();
 
-  size_t CountLiveFiles();
-
   uint64_t Size(const Slice& start, const Slice& limit, int cf = 0);
 
   void Compact(int cf, const Slice& start, const Slice& limit,
@@ -643,21 +638,21 @@ class DBTestBase : public testing::Test {
   //   sets newValue with delta
   // If previous value is not empty,
   //   updates previous value with 'b' string of previous value size - 1.
-  static UpdateStatus updateInPlaceSmallerSize(
-      char* prevValue, uint32_t* prevSize,
-      Slice delta, std::string* newValue);
+  static UpdateStatus updateInPlaceSmallerSize(char* prevValue,
+                                               uint32_t* prevSize, Slice delta,
+                                               std::string* newValue);
 
-  static UpdateStatus updateInPlaceSmallerVarintSize(
-      char* prevValue, uint32_t* prevSize,
-      Slice delta, std::string* newValue);
+  static UpdateStatus updateInPlaceSmallerVarintSize(char* prevValue,
+                                                     uint32_t* prevSize,
+                                                     Slice delta,
+                                                     std::string* newValue);
 
-  static UpdateStatus updateInPlaceLargerSize(
-      char* prevValue, uint32_t* prevSize,
-      Slice delta, std::string* newValue);
+  static UpdateStatus updateInPlaceLargerSize(char* prevValue,
+                                              uint32_t* prevSize, Slice delta,
+                                              std::string* newValue);
 
-  static UpdateStatus updateInPlaceNoAction(
-      char* prevValue, uint32_t* prevSize,
-      Slice delta, std::string* newValue);
+  static UpdateStatus updateInPlaceNoAction(char* prevValue, uint32_t* prevSize,
+                                            Slice delta, std::string* newValue);
 
   // Utility method to test InplaceUpdate
   void validateNumberOfEntries(int numValues, int cf = 0);

db/db_universal_compaction_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 "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
 #include "util/sync_point.h"
 
 namespace rocksdb {
@@ -1210,10 +1210,10 @@ INSTANTIATE_TEST_CASE_P(DBTestUniversalManualCompactionOutputPathId,
 
 }  // namespace rocksdb
 
-#endif  // !(defined NDEBUG) || !defined(OS_WIN)
+#endif  // !defined(ROCKSDB_LITE)
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
+#if !defined(ROCKSDB_LITE)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();

db/db_wal_test.cc

@@ -7,11 +7,9 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
+#include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "util/db_test_util.h"
-#if !(defined NDEBUG) || !defined(OS_WIN)
 #include "util/sync_point.h"
-#endif
 
 namespace rocksdb {
 class DBWALTest : public DBTestBase {
@@ -70,7 +68,6 @@ TEST_F(DBWALTest, RollLog) {
   } while (ChangeOptions());
 }
 
-#if !(defined NDEBUG) || !defined(OS_WIN)
 TEST_F(DBWALTest, SyncWALNotBlockWrite) {
   Options options = CurrentOptions();
   options.max_write_buffer_number = 4;
@@ -130,15 +127,10 @@ TEST_F(DBWALTest, SyncWALNotWaitWrite) {
   ASSERT_EQ(Get("foo2"), "bar2");
   rocksdb::SyncPoint::GetInstance()->DisableProcessing();
 }
-#endif
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
   rocksdb::port::InstallStackTraceHandler();
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();
-#else
-  return 0;
-#endif
 }

db/event_helpers.cc

@@ -53,7 +53,7 @@ void EventHelpers::LogAndNotifyTableFileCreation(
                 info.table_properties.filter_policy_name;
 
     // user collected properties
-    for (const auto& prop : info.table_properties.user_collected_properties) {
+    for (const auto& prop : info.table_properties.readable_properties) {
       jwriter << prop.first << prop.second;
     }
     jwriter.EndObject();

db/fault_injection_test.cc

@@ -11,8 +11,6 @@
 // the last "sync". It then checks for data loss errors by purposely dropping
 // file data (or entire files) not protected by a "sync".
 
-#if !(defined NDEBUG) || !defined(OS_WIN)
-
 #include <map>
 #include <set>
 #include "db/db_impl.h"
@@ -902,13 +900,7 @@ INSTANTIATE_TEST_CASE_P(FaultTest, FaultInjectionTest, ::testing::Bool());
 
 }  // namespace rocksdb
 
-#endif // #if !(defined NDEBUG) || !defined(OS_WIN)
-
 int main(int argc, char** argv) {
-#if !(defined NDEBUG) || !defined(OS_WIN)
   ::testing::InitGoogleTest(&argc, argv);
   return RUN_ALL_TESTS();
-#else
-  return 0;
-#endif
 }

db/filename.cc

@@ -21,6 +21,7 @@
 #include "util/file_reader_writer.h"
 #include "util/logging.h"
 #include "util/stop_watch.h"
+#include "util/sync_point.h"
 
 namespace rocksdb {
 
@@ -322,7 +323,9 @@ Status SetCurrentFile(Env* env, const std::string& dbname,
   std::string tmp = TempFileName(dbname, descriptor_number);
   Status s = WriteStringToFile(env, contents.ToString() + "\n", tmp, true);
   if (s.ok()) {
+    TEST_KILL_RANDOM("SetCurrentFile:0", rocksdb_kill_odds * REDUCE_ODDS2);
     s = env->RenameFile(tmp, CurrentFileName(dbname));
+    TEST_KILL_RANDOM("SetCurrentFile:1", rocksdb_kill_odds * REDUCE_ODDS2);
   }
   if (s.ok()) {
     if (directory_to_fsync != nullptr) {
@@ -351,6 +354,7 @@ Status SetIdentityFile(Env* env, const std::string& dbname) {
 
 Status SyncManifest(Env* env, const DBOptions* db_options,
                     WritableFileWriter* file) {
+  TEST_KILL_RANDOM("SyncManifest:0", rocksdb_kill_odds * REDUCE_ODDS2);
   if (db_options->disableDataSync) {
     return Status::OK();
   } else {

db/flush_job.cc

@@ -197,7 +197,7 @@ Status FlushJob::WriteLevel0Table(const autovector<MemTable*>& mems,
     if (log_buffer_) {
       log_buffer_->FlushBufferToLog();
     }
-    std::vector<Iterator*> memtables;
+    std::vector<InternalIterator*> memtables;
     ReadOptions ro;
     ro.total_order_seek = true;
     Arena arena;
@@ -231,13 +231,15 @@ Status FlushJob::WriteLevel0Table(const autovector<MemTable*>& mems,
 
       TEST_SYNC_POINT_CALLBACK("FlushJob::WriteLevel0Table:output_compression",
                                &output_compression_);
-      s = BuildTable(
-          dbname_, db_options_.env, *cfd_->ioptions(), env_options_,
-          cfd_->table_cache(), iter.get(), meta, cfd_->internal_comparator(),
-          cfd_->int_tbl_prop_collector_factories(), existing_snapshots_,
-          output_compression_, cfd_->ioptions()->compression_opts,
-          mutable_cf_options_.paranoid_file_checks, cfd_->internal_stats(),
-          Env::IO_HIGH, &info.table_properties);
+      s = BuildTable(dbname_, db_options_.env, *cfd_->ioptions(), env_options_,
+                     cfd_->table_cache(), iter.get(), meta,
+                     cfd_->internal_comparator(),
+                     cfd_->int_tbl_prop_collector_factories(), cfd_->GetID(),
+                     existing_snapshots_, output_compression_,
+                     cfd_->ioptions()->compression_opts,
+                     mutable_cf_options_.paranoid_file_checks,
+                     cfd_->internal_stats(), Env::IO_HIGH, &table_properties_);
+      info.table_properties = table_properties_;
       LogFlush(db_options_.info_log);
     }
     Log(InfoLogLevel::INFO_LEVEL, db_options_.info_log,

db/flush_job.h

@@ -27,12 +27,12 @@
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/transaction_log.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/autovector.h"
 #include "util/event_logger.h"
 #include "util/instrumented_mutex.h"
 #include "util/stop_watch.h"
 #include "util/thread_local.h"
-#include "util/scoped_arena_iterator.h"
 #include "db/internal_stats.h"
 #include "db/write_controller.h"
 #include "db/flush_scheduler.h"
@@ -66,6 +66,7 @@ class FlushJob {
   ~FlushJob();
 
   Status Run(FileMetaData* file_meta = nullptr);
+  TableProperties GetTableProperties() const { return table_properties_; }
 
  private:
   void ReportStartedFlush();
@@ -89,6 +90,7 @@ class FlushJob {
   CompressionType output_compression_;
   Statistics* stats_;
   EventLogger* event_logger_;
+  TableProperties table_properties_;
 };
 
 }  // namespace rocksdb

db/flush_job_test.cc

@@ -61,7 +61,7 @@ class FlushJobTest : public testing::Test {
     unique_ptr<WritableFileWriter> file_writer(
         new WritableFileWriter(std::move(file), EnvOptions()));
     {
-      log::Writer log(std::move(file_writer));
+      log::Writer log(std::move(file_writer), 0, false);
       std::string record;
       new_db.EncodeTo(&record);
       s = log.AddRecord(record);

db/forward_iterator.cc

@@ -28,7 +28,7 @@ namespace rocksdb {
 //     iter.SetFileIndex(file_index);
 //     iter.Seek(target);
 //     iter.Next()
-class LevelIterator : public Iterator {
+class LevelIterator : public InternalIterator {
  public:
   LevelIterator(const ColumnFamilyData* const cfd,
       const ReadOptions& read_options,
@@ -113,7 +113,7 @@ class LevelIterator : public Iterator {
   bool valid_;
   uint32_t file_index_;
   Status status_;
-  std::unique_ptr<Iterator> file_iter_;
+  std::unique_ptr<InternalIterator> file_iter_;
 };
 
 ForwardIterator::ForwardIterator(DBImpl* db, const ReadOptions& read_options,
@@ -146,10 +146,10 @@ ForwardIterator::~ForwardIterator() {
 
 void ForwardIterator::Cleanup(bool release_sv) {
   if (mutable_iter_ != nullptr) {
-    mutable_iter_->~Iterator();
+    mutable_iter_->~InternalIterator();
   }
   for (auto* m : imm_iters_) {
-    m->~Iterator();
+    m->~InternalIterator();
   }
   imm_iters_.clear();
   for (auto* f : l0_iters_) {

db/forward_iterator.h

@@ -14,6 +14,7 @@
 #include "rocksdb/iterator.h"
 #include "rocksdb/options.h"
 #include "db/dbformat.h"
+#include "table/internal_iterator.h"
 #include "util/arena.h"
 
 namespace rocksdb {
@@ -30,16 +31,15 @@ class MinIterComparator {
   explicit MinIterComparator(const Comparator* comparator) :
     comparator_(comparator) {}
 
-  bool operator()(Iterator* a, Iterator* b) {
+  bool operator()(InternalIterator* a, InternalIterator* b) {
     return comparator_->Compare(a->key(), b->key()) > 0;
   }
  private:
   const Comparator* comparator_;
 };
 
-typedef std::priority_queue<Iterator*,
-          std::vector<Iterator*>,
-          MinIterComparator> MinIterHeap;
+typedef std::priority_queue<InternalIterator*, std::vector<InternalIterator*>,
+                            MinIterComparator> MinIterHeap;
 
 /**
  * ForwardIterator is a special type of iterator that only supports Seek()
@@ -48,7 +48,7 @@ typedef std::priority_queue<Iterator*,
  * the iterator. At the current implementation, snapshot is taken at the
  * time Seek() is called. The Next() followed do not see new values after.
  */
-class ForwardIterator : public Iterator {
+class ForwardIterator : public InternalIterator {
  public:
   ForwardIterator(DBImpl* db, const ReadOptions& read_options,
                   ColumnFamilyData* cfd, SuperVersion* current_sv = nullptr);
@@ -94,11 +94,11 @@ class ForwardIterator : public Iterator {
   MinIterHeap immutable_min_heap_;
 
   SuperVersion* sv_;
-  Iterator* mutable_iter_;
-  std::vector<Iterator*> imm_iters_;
-  std::vector<Iterator*> l0_iters_;
+  InternalIterator* mutable_iter_;
+  std::vector<InternalIterator*> imm_iters_;
+  std::vector<InternalIterator*> l0_iters_;
   std::vector<LevelIterator*> level_iters_;
-  Iterator* current_;
+  InternalIterator* current_;
   bool valid_;
 
   // Internal iterator status; set only by one of the unsupported methods.

db/internal_stats.cc

@@ -130,6 +130,8 @@ static const std::string aggregated_table_properties =
     "aggregated-table-properties";
 static const std::string aggregated_table_properties_at_level =
     aggregated_table_properties + "-at-level";
+static const std::string num_running_compactions = "num-running-compactions";
+static const std::string num_running_flushes = "num-running-flushes";
 
 const std::string DB::Properties::kNumFilesAtLevelPrefix =
                       rocksdb_prefix + num_files_at_level_prefix;
@@ -143,6 +145,10 @@ const std::string DB::Properties::kMemTableFlushPending =
                       rocksdb_prefix + mem_table_flush_pending;
 const std::string DB::Properties::kCompactionPending =
                       rocksdb_prefix + compaction_pending;
+const std::string DB::Properties::kNumRunningCompactions =
+    rocksdb_prefix + num_running_compactions;
+const std::string DB::Properties::kNumRunningFlushes =
+    rocksdb_prefix + num_running_flushes;
 const std::string DB::Properties::kBackgroundErrors =
                       rocksdb_prefix + background_errors;
 const std::string DB::Properties::kCurSizeActiveMemTable =
@@ -260,6 +266,10 @@ DBPropertyType GetPropertyType(const Slice& property, bool* is_int_property,
     return kTotalSstFilesSize;
   } else if (in == estimate_pending_comp_bytes) {
     return kEstimatePendingCompactionBytes;
+  } else if (in == num_running_flushes) {
+    return kNumRunningFlushes;
+  } else if (in == num_running_compactions) {
+    return kNumRunningCompactions;
   }
   return kUnknown;
 }
@@ -388,11 +398,17 @@ bool InternalStats::GetIntProperty(DBPropertyType property_type,
       // Return number of mem tables that are ready to flush (made immutable)
       *value = (cfd_->imm()->IsFlushPending() ? 1 : 0);
       return true;
+    case kNumRunningFlushes:
+      *value = db->num_running_flushes();
+      return true;
     case kCompactionPending:
       // 1 if the system already determines at least one compaction is needed.
       // 0 otherwise,
       *value = (cfd_->compaction_picker()->NeedsCompaction(vstorage) ? 1 : 0);
       return true;
+    case kNumRunningCompactions:
+      *value = db->num_running_compactions_;
+      return true;
     case kBackgroundErrors:
       // Accumulated number of  errors in background flushes or compactions.
       *value = GetBackgroundErrorCount();

db/internal_stats.h

@@ -38,7 +38,10 @@ enum DBPropertyType : uint32_t {
                                  // in memory that have already been flushed
   kMemtableFlushPending,         // Return 1 if mem table flushing is pending,
                                  // otherwise 0.
+  kNumRunningFlushes,      // Return the number of currently running flushes.
   kCompactionPending,      // Return 1 if a compaction is pending. Otherwise 0.
+  kNumRunningCompactions,  // Return the number of currently running
+                           // compactions.
   kBackgroundErrors,       // Return accumulated background errors encountered.
   kCurSizeActiveMemTable,  // Return current size of the active memtable
   kCurSizeAllMemTables,    // Return current size of unflushed
@@ -49,10 +52,10 @@ enum DBPropertyType : uint32_t {
                                    // memtable.
   kNumEntriesInImmutableMemtable,  // Return sum of number of entries in all
                                    // the immutable mem tables.
-  kNumDeletesInMutableMemtable,    // Return number of entries in the mutable
-                                   // memtable.
-  kNumDeletesInImmutableMemtable,  // Return sum of number of deletes in all
-                                   // the immutable mem tables.
+  kNumDeletesInMutableMemtable,    // Return number of deletion entries in the
+                                   // mutable memtable.
+  kNumDeletesInImmutableMemtable,  // Return the total number of deletion
+                                   // entries in all the immutable mem tables.
   kEstimatedNumKeys,  // Estimated total number of keys in the database.
   kEstimatedUsageByTableReaders,  // Estimated memory by table readers.
   kIsFileDeletionEnabled,         // Equals disable_delete_obsolete_files_,

db/listener_test.cc

@@ -8,6 +8,7 @@
 #include "db/filename.h"
 #include "db/version_set.h"
 #include "db/write_batch_internal.h"
+#include "memtable/hash_linklist_rep.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/db.h"
@@ -22,7 +23,6 @@
 #include "table/block_based_table_factory.h"
 #include "table/plain_table_factory.h"
 #include "util/hash.h"
-#include "util/hash_linklist_rep.h"
 #include "util/logging.h"
 #include "util/mutexlock.h"
 #include "util/rate_limiter.h"
@@ -152,6 +152,40 @@ class EventListenerTest : public testing::Test {
   std::vector<ColumnFamilyHandle*> handles_;
 };
 
+struct TestPropertiesCollector : public rocksdb::TablePropertiesCollector {
+  virtual rocksdb::Status AddUserKey(const rocksdb::Slice& key,
+                                     const rocksdb::Slice& value,
+                                     rocksdb::EntryType type,
+                                     rocksdb::SequenceNumber seq,
+                                     uint64_t file_size) override {
+    return Status::OK();
+  }
+  virtual rocksdb::Status Finish(
+      rocksdb::UserCollectedProperties* properties) override {
+    properties->insert({"0", "1"});
+    return Status::OK();
+  }
+
+  virtual const char* Name() const override {
+    return "TestTablePropertiesCollector";
+  }
+
+  rocksdb::UserCollectedProperties GetReadableProperties() const override {
+    rocksdb::UserCollectedProperties ret;
+    ret["2"] = "3";
+    return ret;
+  }
+};
+
+class TestPropertiesCollectorFactory : public TablePropertiesCollectorFactory {
+ public:
+  virtual TablePropertiesCollector* CreateTablePropertiesCollector(
+      TablePropertiesCollectorFactory::Context context) override {
+    return new TestPropertiesCollector;
+  }
+  const char* Name() const override { return "TestTablePropertiesCollector"; }
+};
+
 class TestCompactionListener : public EventListener {
  public:
   void OnCompactionCompleted(DB *db, const CompactionJobInfo& ci) override {
@@ -161,6 +195,16 @@ class TestCompactionListener : public EventListener {
     ASSERT_GT(ci.output_files.size(), 0U);
     ASSERT_EQ(db->GetEnv()->GetThreadID(), ci.thread_id);
     ASSERT_GT(ci.thread_id, 0U);
+
+    for (auto fl : {ci.input_files, ci.output_files}) {
+      for (auto fn : fl) {
+        auto it = ci.table_properties.find(fn);
+        ASSERT_NE(it, ci.table_properties.end());
+        auto tp = it->second;
+        ASSERT_TRUE(tp != nullptr);
+        ASSERT_EQ(tp->user_collected_properties.find("0")->second, "1");
+      }
+    }
   }
 
   std::vector<DB*> compacted_dbs_;
@@ -186,6 +230,8 @@ TEST_F(EventListenerTest, OnSingleDBCompactionTest) {
   options.enable_thread_tracking = true;
 #endif  // ROCKSDB_USING_THREAD_STATUS
   options.level0_file_num_compaction_trigger = kNumL0Files;
+  options.table_properties_collector_factories.push_back(
+      std::make_shared<TestPropertiesCollectorFactory>());
 
   TestCompactionListener* listener = new TestCompactionListener();
   options.listeners.emplace_back(listener);
@@ -274,6 +320,8 @@ class TestFlushListener : public EventListener {
     ASSERT_EQ(prev_fc_info_.file_path, info.file_path);
     ASSERT_EQ(db->GetEnv()->GetThreadID(), info.thread_id);
     ASSERT_GT(info.thread_id, 0U);
+    ASSERT_EQ(info.table_properties.user_collected_properties.find("0")->second,
+              "1");
   }
 
   std::vector<std::string> flushed_column_family_names_;
@@ -299,6 +347,8 @@ TEST_F(EventListenerTest, OnSingleDBFlushTest) {
   std::vector<std::string> cf_names = {
       "pikachu", "ilya", "muromec", "dobrynia",
       "nikitich", "alyosha", "popovich"};
+  options.table_properties_collector_factories.push_back(
+      std::make_shared<TestPropertiesCollectorFactory>());
   CreateAndReopenWithCF(cf_names, &options);
 
   ASSERT_OK(Put(1, "pikachu", std::string(90000, 'p')));
@@ -330,6 +380,8 @@ TEST_F(EventListenerTest, MultiCF) {
 #endif  // ROCKSDB_USING_THREAD_STATUS
   TestFlushListener* listener = new TestFlushListener(options.env);
   options.listeners.emplace_back(listener);
+  options.table_properties_collector_factories.push_back(
+      std::make_shared<TestPropertiesCollectorFactory>());
   std::vector<std::string> cf_names = {
       "pikachu", "ilya", "muromec", "dobrynia",
       "nikitich", "alyosha", "popovich"};
@@ -360,6 +412,8 @@ TEST_F(EventListenerTest, MultiDBMultiListeners) {
 #if ROCKSDB_USING_THREAD_STATUS
   options.enable_thread_tracking = true;
 #endif  // ROCKSDB_USING_THREAD_STATUS
+  options.table_properties_collector_factories.push_back(
+      std::make_shared<TestPropertiesCollectorFactory>());
   std::vector<TestFlushListener*> listeners;
   const int kNumDBs = 5;
   const int kNumListeners = 10;
@@ -454,6 +508,8 @@ TEST_F(EventListenerTest, DisableBGCompaction) {
   options.compaction_style = kCompactionStyleNone;
   options.compression = kNoCompression;
   options.write_buffer_size = 100000;  // Small write buffer
+  options.table_properties_collector_factories.push_back(
+      std::make_shared<TestPropertiesCollectorFactory>());
 
   CreateAndReopenWithCF({"pikachu"}, &options);
   ColumnFamilyMetaData cf_meta;

db/log_format.h

@@ -22,14 +22,24 @@ enum RecordType {
   // For fragments
   kFirstType = 2,
   kMiddleType = 3,
-  kLastType = 4
+  kLastType = 4,
+
+  // For recycled log files
+  kRecyclableFullType = 5,
+  kRecyclableFirstType = 6,
+  kRecyclableMiddleType = 7,
+  kRecyclableLastType = 8,
 };
-static const int kMaxRecordType = kLastType;
+static const int kMaxRecordType = kRecyclableLastType;
 
 static const unsigned int kBlockSize = 32768;
 
 // Header is checksum (4 bytes), type (1 byte), length (2 bytes).
 static const int kHeaderSize = 4 + 1 + 2;
 
+// Recyclable header is checksum (4 bytes), type (1 byte), log number
+// (4 bytes), length (2 bytes).
+static const int kRecyclableHeaderSize = 4 + 1 + 4 + 2;
+
 }  // namespace log
 }  // namespace rocksdb

db/log_reader.cc

@@ -21,9 +21,12 @@ namespace log {
 Reader::Reporter::~Reporter() {
 }
 
-Reader::Reader(unique_ptr<SequentialFileReader>&& _file, Reporter* reporter,
-               bool checksum, uint64_t initial_offset)
-    : file_(std::move(_file)),
+Reader::Reader(std::shared_ptr<Logger> info_log,
+	       unique_ptr<SequentialFileReader>&& _file,
+               Reporter* reporter, bool checksum, uint64_t initial_offset,
+               uint64_t log_num)
+    : info_log_(info_log),
+      file_(std::move(_file)),
       reporter_(reporter),
       checksum_(checksum),
       backing_store_(new char[kBlockSize]),
@@ -33,7 +36,8 @@ Reader::Reader(unique_ptr<SequentialFileReader>&& _file, Reporter* reporter,
       eof_offset_(0),
       last_record_offset_(0),
       end_of_buffer_offset_(0),
-      initial_offset_(initial_offset) {}
+      initial_offset_(initial_offset),
+      log_number_(log_num) {}
 
 Reader::~Reader() {
   delete[] backing_store_;
@@ -62,8 +66,15 @@ bool Reader::SkipToInitialBlock() {
   return true;
 }
 
+// For kAbsoluteConsistency, on clean shutdown we don't expect any error
+// in the log files.  For other modes, we can ignore only incomplete records
+// in the last log file, which are presumably due to a write in progress
+// during restart (or from log recycling).
+//
+// TODO krad: Evaluate if we need to move to a more strict mode where we
+// restrict the inconsistency to only the last log
 bool Reader::ReadRecord(Slice* record, std::string* scratch,
-                        const bool report_eof_inconsistency) {
+                        WALRecoveryMode wal_recovery_mode) {
   if (last_record_offset_ < initial_offset_) {
     if (!SkipToInitialBlock()) {
       return false;
@@ -80,10 +91,11 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
   Slice fragment;
   while (true) {
     uint64_t physical_record_offset = end_of_buffer_offset_ - buffer_.size();
-    const unsigned int record_type =
-        ReadPhysicalRecord(&fragment, report_eof_inconsistency);
+    size_t drop_size;
+    const unsigned int record_type = ReadPhysicalRecord(&fragment, &drop_size);
     switch (record_type) {
       case kFullType:
+      case kRecyclableFullType:
         if (in_fragmented_record && !scratch->empty()) {
           // Handle bug in earlier versions of log::Writer where
           // it could emit an empty kFirstType record at the tail end
@@ -98,6 +110,7 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
         return true;
 
       case kFirstType:
+      case kRecyclableFirstType:
         if (in_fragmented_record && !scratch->empty()) {
           // Handle bug in earlier versions of log::Writer where
           // it could emit an empty kFirstType record at the tail end
@@ -111,6 +124,7 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
         break;
 
       case kMiddleType:
+      case kRecyclableMiddleType:
         if (!in_fragmented_record) {
           ReportCorruption(fragment.size(),
                            "missing start of fragmented record(1)");
@@ -120,6 +134,7 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
         break;
 
       case kLastType:
+      case kRecyclableLastType:
         if (!in_fragmented_record) {
           ReportCorruption(fragment.size(),
                            "missing start of fragmented record(2)");
@@ -131,9 +146,17 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
         }
         break;
 
+      case kBadHeader:
+        if (wal_recovery_mode == WALRecoveryMode::kAbsoluteConsistency) {
+          // in clean shutdown we don't expect any error in the log files
+          ReportCorruption(drop_size, "truncated header");
+        }
+      // fall-thru
+
       case kEof:
         if (in_fragmented_record) {
-          if (report_eof_inconsistency) {
+          if (wal_recovery_mode == WALRecoveryMode::kAbsoluteConsistency) {
+            // in clean shutdown we don't expect any error in the log files
             ReportCorruption(scratch->size(), "error reading trailing data");
           }
           // This can be caused by the writer dying immediately after
@@ -143,6 +166,23 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
         }
         return false;
 
+      case kOldRecord:
+        if (wal_recovery_mode != WALRecoveryMode::kSkipAnyCorruptedRecords) {
+          // Treat a record from a previous instance of the log as EOF.
+          if (in_fragmented_record) {
+            if (wal_recovery_mode == WALRecoveryMode::kAbsoluteConsistency) {
+              // in clean shutdown we don't expect any error in the log files
+              ReportCorruption(scratch->size(), "error reading trailing data");
+            }
+            // This can be caused by the writer dying immediately after
+            //  writing a physical record but before completing the next; don't
+            //  treat it as a corruption, just ignore the entire logical record.
+            scratch->clear();
+          }
+          return false;
+        }
+      // fall-thru
+
       case kBadRecord:
         if (in_fragmented_record) {
           ReportCorruption(scratch->size(), "error in middle of record");
@@ -244,36 +284,49 @@ void Reader::ReportDrop(size_t bytes, const Status& reason) {
   }
 }
 
-unsigned int Reader::ReadPhysicalRecord(Slice* result,
-                                        const bool report_eof_inconsistency) {
+bool Reader::ReadMore(size_t* drop_size, int *error) {
+  if (!eof_ && !read_error_) {
+    // Last read was a full read, so this is a trailer to skip
+    buffer_.clear();
+    Status status = file_->Read(kBlockSize, &buffer_, backing_store_);
+    end_of_buffer_offset_ += buffer_.size();
+    if (!status.ok()) {
+      buffer_.clear();
+      ReportDrop(kBlockSize, status);
+      read_error_ = true;
+      *error = kEof;
+      return false;
+    } else if (buffer_.size() < (size_t)kBlockSize) {
+      eof_ = true;
+      eof_offset_ = buffer_.size();
+    }
+    return true;
+  } else {
+    // Note that if buffer_ is non-empty, we have a truncated header at the
+    //  end of the file, which can be caused by the writer crashing in the
+    //  middle of writing the header. Unless explicitly requested we don't
+    //  considering this an error, just report EOF.
+    if (buffer_.size()) {
+      *drop_size = buffer_.size();
+      buffer_.clear();
+      *error = kBadHeader;
+      return false;
+    }
+    buffer_.clear();
+    *error = kEof;
+    return false;
+  }
+}
+
+unsigned int Reader::ReadPhysicalRecord(Slice* result, size_t* drop_size) {
   while (true) {
+    // We need at least the minimum header size
     if (buffer_.size() < (size_t)kHeaderSize) {
-      if (!eof_ && !read_error_) {
-        // Last read was a full read, so this is a trailer to skip
-        buffer_.clear();
-        Status status = file_->Read(kBlockSize, &buffer_, backing_store_);
-        end_of_buffer_offset_ += buffer_.size();
-        if (!status.ok()) {
-          buffer_.clear();
-          ReportDrop(kBlockSize, status);
-          read_error_ = true;
-          return kEof;
-        } else if (buffer_.size() < (size_t)kBlockSize) {
-          eof_ = true;
-          eof_offset_ = buffer_.size();
-        }
-        continue;
-      } else {
-        // Note that if buffer_ is non-empty, we have a truncated header at the
-        //  end of the file, which can be caused by the writer crashing in the
-        //  middle of writing the header. Unless explicitly requested we don't
-        //  considering this an error, just report EOF.
-        if (buffer_.size() && report_eof_inconsistency) {
-          ReportCorruption(buffer_.size(), "truncated header");
-        }
-        buffer_.clear();
-        return kEof;
+      int r;
+      if (!ReadMore(drop_size, &r)) {
+	return r;
       }
+      continue;
     }
 
     // Parse the header
@@ -282,18 +335,34 @@ unsigned int Reader::ReadPhysicalRecord(Slice* result,
     const uint32_t b = static_cast<uint32_t>(header[5]) & 0xff;
     const unsigned int type = header[6];
     const uint32_t length = a | (b << 8);
-    if (kHeaderSize + length > buffer_.size()) {
-      size_t drop_size = buffer_.size();
+    int header_size = kHeaderSize;
+    if (type >= kRecyclableFullType && type <= kRecyclableLastType) {
+      header_size = kRecyclableHeaderSize;
+      // We need enough for the larger header
+      if (buffer_.size() < (size_t)kRecyclableHeaderSize) {
+	int r;
+	if (!ReadMore(drop_size, &r)) {
+          return r;
+        }
+	continue;
+      }
+      const uint32_t log_num = DecodeFixed32(header + 7);
+      if (log_num != log_number_) {
+        return kOldRecord;
+      }
+    }
+    if (header_size + length > buffer_.size()) {
+      *drop_size = buffer_.size();
       buffer_.clear();
       if (!eof_) {
-        ReportCorruption(drop_size, "bad record length");
+        ReportCorruption(*drop_size, "bad record length");
         return kBadRecord;
       }
       // If the end of the file has been reached without reading |length| bytes
       // of payload, assume the writer died in the middle of writing the record.
       // Don't report a corruption unless requested.
-      if (drop_size && report_eof_inconsistency) {
-        ReportCorruption(drop_size, "truncated header");
+      if (*drop_size) {
+        return kBadHeader;
       }
       return kEof;
     }
@@ -311,29 +380,29 @@ unsigned int Reader::ReadPhysicalRecord(Slice* result,
     // Check crc
     if (checksum_) {
       uint32_t expected_crc = crc32c::Unmask(DecodeFixed32(header));
-      uint32_t actual_crc = crc32c::Value(header + 6, 1 + length);
+      uint32_t actual_crc = crc32c::Value(header + 6, length + header_size - 6);
       if (actual_crc != expected_crc) {
         // Drop the rest of the buffer since "length" itself may have
         // been corrupted and if we trust it, we could find some
         // fragment of a real log record that just happens to look
         // like a valid log record.
-        size_t drop_size = buffer_.size();
+        *drop_size = buffer_.size();
         buffer_.clear();
-        ReportCorruption(drop_size, "checksum mismatch");
+        ReportCorruption(*drop_size, "checksum mismatch");
         return kBadRecord;
       }
     }
 
-    buffer_.remove_prefix(kHeaderSize + length);
+    buffer_.remove_prefix(header_size + length);
 
     // Skip physical record that started before initial_offset_
-    if (end_of_buffer_offset_ - buffer_.size() - kHeaderSize - length <
+    if (end_of_buffer_offset_ - buffer_.size() - header_size - length <
         initial_offset_) {
       result->clear();
       return kBadRecord;
     }
 
-    *result = Slice(header + kHeaderSize, length);
+    *result = Slice(header + header_size, length);
     return type;
   }
 }

db/log_reader.h

@@ -14,10 +14,12 @@
 #include "db/log_format.h"
 #include "rocksdb/slice.h"
 #include "rocksdb/status.h"
+#include "rocksdb/options.h"
 
 namespace rocksdb {
 
 class SequentialFileReader;
+class Logger;
 using std::unique_ptr;
 
 namespace log {
@@ -51,8 +53,10 @@ class Reader {
   //
   // The Reader will start reading at the first record located at physical
   // position >= initial_offset within the file.
-  Reader(unique_ptr<SequentialFileReader>&& file, Reporter* reporter,
-         bool checksum, uint64_t initial_offset);
+  Reader(std::shared_ptr<Logger> info_log,
+	 unique_ptr<SequentialFileReader>&& file,
+         Reporter* reporter, bool checksum, uint64_t initial_offset,
+         uint64_t log_num);
 
   ~Reader();
 
@@ -62,7 +66,8 @@ class Reader {
   // will only be valid until the next mutating operation on this
   // reader or the next mutation to *scratch.
   bool ReadRecord(Slice* record, std::string* scratch,
-                  bool report_eof_inconsistency = false);
+                  WALRecoveryMode wal_recovery_mode =
+                      WALRecoveryMode::kTolerateCorruptedTailRecords);
 
   // Returns the physical offset of the last record returned by ReadRecord.
   //
@@ -84,6 +89,7 @@ class Reader {
   SequentialFileReader* file() { return file_.get(); }
 
  private:
+  std::shared_ptr<Logger> info_log_;
   const unique_ptr<SequentialFileReader> file_;
   Reporter* const reporter_;
   bool const checksum_;
@@ -104,6 +110,9 @@ class Reader {
   // Offset at which to start looking for the first record to return
   uint64_t const initial_offset_;
 
+  // which log number this is
+  uint64_t const log_number_;
+
   // Extend record types with the following special values
   enum {
     kEof = kMaxRecordType + 1,
@@ -112,7 +121,11 @@ class Reader {
     // * The record has an invalid CRC (ReadPhysicalRecord reports a drop)
     // * The record is a 0-length record (No drop is reported)
     // * The record is below constructor's initial_offset (No drop is reported)
-    kBadRecord = kMaxRecordType + 2
+    kBadRecord = kMaxRecordType + 2,
+    // Returned when we fail to read a valid header.
+    kBadHeader = kMaxRecordType + 3,
+    // Returned when we read an old record from a previous user of the log.
+    kOldRecord = kMaxRecordType + 4,
   };
 
   // Skips all blocks that are completely before "initial_offset_".
@@ -121,8 +134,10 @@ class Reader {
   bool SkipToInitialBlock();
 
   // Return type, or one of the preceding special values
-  unsigned int ReadPhysicalRecord(Slice* result,
-                                  bool report_eof_inconsistency = false);
+  unsigned int ReadPhysicalRecord(Slice* result, size_t* drop_size);
+
+  // Read some more
+  bool ReadMore(size_t* drop_size, int *error);
 
   // Reports dropped bytes to the reporter.
   // buffer_ must be updated to remove the dropped bytes prior to invocation.

db/log_test.cc

@@ -43,7 +43,7 @@ static std::string RandomSkewedString(int i, Random* rnd) {
   return BigString(NumberString(i), rnd->Skewed(17));
 }
 
-class LogTest : public testing::Test {
+class LogTest : public ::testing::TestWithParam<int> {
  private:
   class StringSource : public SequentialFile {
    public:
@@ -153,19 +153,26 @@ class LogTest : public testing::Test {
 
   // Record metadata for testing initial offset functionality
   static size_t initial_offset_record_sizes_[];
-  static uint64_t initial_offset_last_record_offsets_[];
+  uint64_t initial_offset_last_record_offsets_[4];
 
  public:
   LogTest()
       : reader_contents_(),
-        dest_holder_(
-            test::GetWritableFileWriter(
-              new test::StringSink(&reader_contents_))),
+        dest_holder_(test::GetWritableFileWriter(
+            new test::StringSink(&reader_contents_))),
         source_holder_(
             test::GetSequentialFileReader(new StringSource(reader_contents_))),
-        writer_(std::move(dest_holder_)),
-        reader_(std::move(source_holder_), &report_, true /*checksum*/,
-                0 /*initial_offset*/) {}
+        writer_(std::move(dest_holder_), 123, GetParam()),
+        reader_(NULL, std::move(source_holder_), &report_, true /*checksum*/,
+                0 /*initial_offset*/, 123) {
+    int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+    initial_offset_last_record_offsets_[0] = 0;
+    initial_offset_last_record_offsets_[1] = header_size + 10000;
+    initial_offset_last_record_offsets_[2] = 2 * (header_size + 10000);
+    initial_offset_last_record_offsets_[3] = 2 * (header_size + 10000) +
+                                             (2 * log::kBlockSize - 1000) +
+                                             3 * header_size;
+  }
 
   void Write(const std::string& msg) {
     writer_.AddRecord(Slice(msg));
@@ -175,10 +182,11 @@ class LogTest : public testing::Test {
     return dest_contents().size();
   }
 
-  std::string Read(const bool report_eof_inconsistency = false) {
+  std::string Read(const WALRecoveryMode wal_recovery_mode =
+                       WALRecoveryMode::kTolerateCorruptedTailRecords) {
     std::string scratch;
     Slice record;
-    if (reader_.ReadRecord(&record, &scratch, report_eof_inconsistency)) {
+    if (reader_.ReadRecord(&record, &scratch, wal_recovery_mode)) {
       return record.ToString();
     } else {
       return "EOF";
@@ -200,9 +208,11 @@ class LogTest : public testing::Test {
     dest->Drop(bytes);
   }
 
-  void FixChecksum(int header_offset, int len) {
+  void FixChecksum(int header_offset, int len, bool recyclable) {
     // Compute crc of type/len/data
-    uint32_t crc = crc32c::Value(&dest_contents()[header_offset+6], 1 + len);
+    int header_size = recyclable ? kRecyclableHeaderSize : kHeaderSize;
+    uint32_t crc = crc32c::Value(&dest_contents()[header_offset + 6],
+                                 header_size - 6 + len);
     crc = crc32c::Mask(crc);
     EncodeFixed32(&dest_contents()[header_offset], crc);
   }
@@ -259,8 +269,8 @@ class LogTest : public testing::Test {
     unique_ptr<SequentialFileReader> file_reader(
         test::GetSequentialFileReader(new StringSource(reader_contents_)));
     unique_ptr<Reader> offset_reader(
-        new Reader(std::move(file_reader), &report_, true /*checksum*/,
-                   WrittenBytes() + offset_past_end));
+        new Reader(NULL, std::move(file_reader), &report_,
+                   true /*checksum*/, WrittenBytes() + offset_past_end, 123));
     Slice record;
     std::string scratch;
     ASSERT_TRUE(!offset_reader->ReadRecord(&record, &scratch));
@@ -271,8 +281,9 @@ class LogTest : public testing::Test {
     WriteInitialOffsetLog();
     unique_ptr<SequentialFileReader> file_reader(
         test::GetSequentialFileReader(new StringSource(reader_contents_)));
-    unique_ptr<Reader> offset_reader(new Reader(
-        std::move(file_reader), &report_, true /*checksum*/, initial_offset));
+    unique_ptr<Reader> offset_reader(
+        new Reader(NULL, std::move(file_reader), &report_,
+                   true /*checksum*/, initial_offset, 123));
     Slice record;
     std::string scratch;
     ASSERT_TRUE(offset_reader->ReadRecord(&record, &scratch));
@@ -291,16 +302,9 @@ size_t LogTest::initial_offset_record_sizes_[] =
      2 * log::kBlockSize - 1000,  // Span three blocks
      1};
 
-uint64_t LogTest::initial_offset_last_record_offsets_[] =
-    {0,
-     kHeaderSize + 10000,
-     2 * (kHeaderSize + 10000),
-     2 * (kHeaderSize + 10000) +
-         (2 * log::kBlockSize - 1000) + 3 * kHeaderSize};
+TEST_P(LogTest, Empty) { ASSERT_EQ("EOF", Read()); }
 
-TEST_F(LogTest, Empty) { ASSERT_EQ("EOF", Read()); }
-
-TEST_F(LogTest, ReadWrite) {
+TEST_P(LogTest, ReadWrite) {
   Write("foo");
   Write("bar");
   Write("");
@@ -313,7 +317,7 @@ TEST_F(LogTest, ReadWrite) {
   ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
 }
 
-TEST_F(LogTest, ManyBlocks) {
+TEST_P(LogTest, ManyBlocks) {
   for (int i = 0; i < 100000; i++) {
     Write(NumberString(i));
   }
@@ -323,7 +327,7 @@ TEST_F(LogTest, ManyBlocks) {
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, Fragmentation) {
+TEST_P(LogTest, Fragmentation) {
   Write("small");
   Write(BigString("medium", 50000));
   Write(BigString("large", 100000));
@@ -333,11 +337,12 @@ TEST_F(LogTest, Fragmentation) {
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, MarginalTrailer) {
+TEST_P(LogTest, MarginalTrailer) {
   // Make a trailer that is exactly the same length as an empty record.
-  const int n = kBlockSize - 2*kHeaderSize;
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  const int n = kBlockSize - 2 * header_size;
   Write(BigString("foo", n));
-  ASSERT_EQ((unsigned int)(kBlockSize - kHeaderSize), WrittenBytes());
+  ASSERT_EQ((unsigned int)(kBlockSize - header_size), WrittenBytes());
   Write("");
   Write("bar");
   ASSERT_EQ(BigString("foo", n), Read());
@@ -346,11 +351,12 @@ TEST_F(LogTest, MarginalTrailer) {
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, MarginalTrailer2) {
+TEST_P(LogTest, MarginalTrailer2) {
   // Make a trailer that is exactly the same length as an empty record.
-  const int n = kBlockSize - 2*kHeaderSize;
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  const int n = kBlockSize - 2 * header_size;
   Write(BigString("foo", n));
-  ASSERT_EQ((unsigned int)(kBlockSize - kHeaderSize), WrittenBytes());
+  ASSERT_EQ((unsigned int)(kBlockSize - header_size), WrittenBytes());
   Write("bar");
   ASSERT_EQ(BigString("foo", n), Read());
   ASSERT_EQ("bar", Read());
@@ -359,10 +365,11 @@ TEST_F(LogTest, MarginalTrailer2) {
   ASSERT_EQ("", ReportMessage());
 }
 
-TEST_F(LogTest, ShortTrailer) {
-  const int n = kBlockSize - 2*kHeaderSize + 4;
+TEST_P(LogTest, ShortTrailer) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  const int n = kBlockSize - 2 * header_size + 4;
   Write(BigString("foo", n));
-  ASSERT_EQ((unsigned int)(kBlockSize - kHeaderSize + 4), WrittenBytes());
+  ASSERT_EQ((unsigned int)(kBlockSize - header_size + 4), WrittenBytes());
   Write("");
   Write("bar");
   ASSERT_EQ(BigString("foo", n), Read());
@@ -371,15 +378,16 @@ TEST_F(LogTest, ShortTrailer) {
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, AlignedEof) {
-  const int n = kBlockSize - 2*kHeaderSize + 4;
+TEST_P(LogTest, AlignedEof) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  const int n = kBlockSize - 2 * header_size + 4;
   Write(BigString("foo", n));
-  ASSERT_EQ((unsigned int)(kBlockSize - kHeaderSize + 4), WrittenBytes());
+  ASSERT_EQ((unsigned int)(kBlockSize - header_size + 4), WrittenBytes());
   ASSERT_EQ(BigString("foo", n), Read());
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, RandomRead) {
+TEST_P(LogTest, RandomRead) {
   const int N = 500;
   Random write_rnd(301);
   for (int i = 0; i < N; i++) {
@@ -394,7 +402,7 @@ TEST_F(LogTest, RandomRead) {
 
 // Tests of all the error paths in log_reader.cc follow:
 
-TEST_F(LogTest, ReadError) {
+TEST_P(LogTest, ReadError) {
   Write("foo");
   ForceError();
   ASSERT_EQ("EOF", Read());
@@ -402,17 +410,17 @@ TEST_F(LogTest, ReadError) {
   ASSERT_EQ("OK", MatchError("read error"));
 }
 
-TEST_F(LogTest, BadRecordType) {
+TEST_P(LogTest, BadRecordType) {
   Write("foo");
   // Type is stored in header[6]
   IncrementByte(6, 100);
-  FixChecksum(0, 3);
+  FixChecksum(0, 3, false);
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ(3U, DroppedBytes());
   ASSERT_EQ("OK", MatchError("unknown record type"));
 }
 
-TEST_F(LogTest, TruncatedTrailingRecordIsIgnored) {
+TEST_P(LogTest, TruncatedTrailingRecordIsIgnored) {
   Write("foo");
   ShrinkSize(4);   // Drop all payload as well as a header byte
   ASSERT_EQ("EOF", Read());
@@ -421,17 +429,18 @@ TEST_F(LogTest, TruncatedTrailingRecordIsIgnored) {
   ASSERT_EQ("", ReportMessage());
 }
 
-TEST_F(LogTest, TruncatedTrailingRecordIsNotIgnored) {
+TEST_P(LogTest, TruncatedTrailingRecordIsNotIgnored) {
   Write("foo");
   ShrinkSize(4);  // Drop all payload as well as a header byte
-  ASSERT_EQ("EOF", Read(/*report_eof_inconsistency*/ true));
+  ASSERT_EQ("EOF", Read(WALRecoveryMode::kAbsoluteConsistency));
   // Truncated last record is ignored, not treated as an error
   ASSERT_GT(DroppedBytes(), 0U);
   ASSERT_EQ("OK", MatchError("Corruption: truncated header"));
 }
 
-TEST_F(LogTest, BadLength) {
-  const int kPayloadSize = kBlockSize - kHeaderSize;
+TEST_P(LogTest, BadLength) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  const int kPayloadSize = kBlockSize - header_size;
   Write(BigString("bar", kPayloadSize));
   Write("foo");
   // Least significant size byte is stored in header[4].
@@ -441,7 +450,7 @@ TEST_F(LogTest, BadLength) {
   ASSERT_EQ("OK", MatchError("bad record length"));
 }
 
-TEST_F(LogTest, BadLengthAtEndIsIgnored) {
+TEST_P(LogTest, BadLengthAtEndIsIgnored) {
   Write("foo");
   ShrinkSize(1);
   ASSERT_EQ("EOF", Read());
@@ -449,63 +458,63 @@ TEST_F(LogTest, BadLengthAtEndIsIgnored) {
   ASSERT_EQ("", ReportMessage());
 }
 
-TEST_F(LogTest, BadLengthAtEndIsNotIgnored) {
+TEST_P(LogTest, BadLengthAtEndIsNotIgnored) {
   Write("foo");
   ShrinkSize(1);
-  ASSERT_EQ("EOF", Read(/*report_eof_inconsistency=*/true));
+  ASSERT_EQ("EOF", Read(WALRecoveryMode::kAbsoluteConsistency));
   ASSERT_GT(DroppedBytes(), 0U);
   ASSERT_EQ("OK", MatchError("Corruption: truncated header"));
 }
 
-TEST_F(LogTest, ChecksumMismatch) {
-  Write("foo");
-  IncrementByte(0, 10);
+TEST_P(LogTest, ChecksumMismatch) {
+  Write("foooooo");
+  IncrementByte(0, 14);
   ASSERT_EQ("EOF", Read());
-  ASSERT_EQ(10U, DroppedBytes());
+  ASSERT_EQ(14U + 4 * !!GetParam(), DroppedBytes());
   ASSERT_EQ("OK", MatchError("checksum mismatch"));
 }
 
-TEST_F(LogTest, UnexpectedMiddleType) {
+TEST_P(LogTest, UnexpectedMiddleType) {
   Write("foo");
-  SetByte(6, kMiddleType);
-  FixChecksum(0, 3);
+  SetByte(6, GetParam() ? kRecyclableMiddleType : kMiddleType);
+  FixChecksum(0, 3, !!GetParam());
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ(3U, DroppedBytes());
   ASSERT_EQ("OK", MatchError("missing start"));
 }
 
-TEST_F(LogTest, UnexpectedLastType) {
+TEST_P(LogTest, UnexpectedLastType) {
   Write("foo");
-  SetByte(6, kLastType);
-  FixChecksum(0, 3);
+  SetByte(6, GetParam() ? kRecyclableLastType : kLastType);
+  FixChecksum(0, 3, !!GetParam());
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ(3U, DroppedBytes());
   ASSERT_EQ("OK", MatchError("missing start"));
 }
 
-TEST_F(LogTest, UnexpectedFullType) {
+TEST_P(LogTest, UnexpectedFullType) {
   Write("foo");
   Write("bar");
-  SetByte(6, kFirstType);
-  FixChecksum(0, 3);
+  SetByte(6, GetParam() ? kRecyclableFirstType : kFirstType);
+  FixChecksum(0, 3, !!GetParam());
   ASSERT_EQ("bar", Read());
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ(3U, DroppedBytes());
   ASSERT_EQ("OK", MatchError("partial record without end"));
 }
 
-TEST_F(LogTest, UnexpectedFirstType) {
+TEST_P(LogTest, UnexpectedFirstType) {
   Write("foo");
   Write(BigString("bar", 100000));
-  SetByte(6, kFirstType);
-  FixChecksum(0, 3);
+  SetByte(6, GetParam() ? kRecyclableFirstType : kFirstType);
+  FixChecksum(0, 3, !!GetParam());
   ASSERT_EQ(BigString("bar", 100000), Read());
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ(3U, DroppedBytes());
   ASSERT_EQ("OK", MatchError("partial record without end"));
 }
 
-TEST_F(LogTest, MissingLastIsIgnored) {
+TEST_P(LogTest, MissingLastIsIgnored) {
   Write(BigString("bar", kBlockSize));
   // Remove the LAST block, including header.
   ShrinkSize(14);
@@ -514,16 +523,16 @@ TEST_F(LogTest, MissingLastIsIgnored) {
   ASSERT_EQ(0U, DroppedBytes());
 }
 
-TEST_F(LogTest, MissingLastIsNotIgnored) {
+TEST_P(LogTest, MissingLastIsNotIgnored) {
   Write(BigString("bar", kBlockSize));
   // Remove the LAST block, including header.
   ShrinkSize(14);
-  ASSERT_EQ("EOF", Read(/*report_eof_inconsistency=*/true));
+  ASSERT_EQ("EOF", Read(WALRecoveryMode::kAbsoluteConsistency));
   ASSERT_GT(DroppedBytes(), 0U);
   ASSERT_EQ("OK", MatchError("Corruption: error reading trailing data"));
 }
 
-TEST_F(LogTest, PartialLastIsIgnored) {
+TEST_P(LogTest, PartialLastIsIgnored) {
   Write(BigString("bar", kBlockSize));
   // Cause a bad record length in the LAST block.
   ShrinkSize(1);
@@ -532,18 +541,18 @@ TEST_F(LogTest, PartialLastIsIgnored) {
   ASSERT_EQ(0U, DroppedBytes());
 }
 
-TEST_F(LogTest, PartialLastIsNotIgnored) {
+TEST_P(LogTest, PartialLastIsNotIgnored) {
   Write(BigString("bar", kBlockSize));
   // Cause a bad record length in the LAST block.
   ShrinkSize(1);
-  ASSERT_EQ("EOF", Read(/*report_eof_inconsistency=*/true));
+  ASSERT_EQ("EOF", Read(WALRecoveryMode::kAbsoluteConsistency));
   ASSERT_GT(DroppedBytes(), 0U);
   ASSERT_EQ("OK", MatchError(
                       "Corruption: truncated headerCorruption: "
                       "error reading trailing data"));
 }
 
-TEST_F(LogTest, ErrorJoinsRecords) {
+TEST_P(LogTest, ErrorJoinsRecords) {
   // Consider two fragmented records:
   //    first(R1) last(R1) first(R2) last(R2)
   // where the middle two fragments disappear.  We do not want
@@ -566,46 +575,60 @@ TEST_F(LogTest, ErrorJoinsRecords) {
   ASSERT_GE(dropped, 2 * kBlockSize);
 }
 
-TEST_F(LogTest, ReadStart) { CheckInitialOffsetRecord(0, 0); }
+TEST_P(LogTest, ReadStart) { CheckInitialOffsetRecord(0, 0); }
 
-TEST_F(LogTest, ReadSecondOneOff) { CheckInitialOffsetRecord(1, 1); }
+TEST_P(LogTest, ReadSecondOneOff) { CheckInitialOffsetRecord(1, 1); }
 
-TEST_F(LogTest, ReadSecondTenThousand) { CheckInitialOffsetRecord(10000, 1); }
+TEST_P(LogTest, ReadSecondTenThousand) { CheckInitialOffsetRecord(10000, 1); }
 
-TEST_F(LogTest, ReadSecondStart) { CheckInitialOffsetRecord(10007, 1); }
+TEST_P(LogTest, ReadSecondStart) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  CheckInitialOffsetRecord(10000 + header_size, 1);
+}
 
-TEST_F(LogTest, ReadThirdOneOff) { CheckInitialOffsetRecord(10008, 2); }
+TEST_P(LogTest, ReadThirdOneOff) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  CheckInitialOffsetRecord(10000 + header_size + 1, 2);
+}
 
-TEST_F(LogTest, ReadThirdStart) { CheckInitialOffsetRecord(20014, 2); }
+TEST_P(LogTest, ReadThirdStart) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  CheckInitialOffsetRecord(20000 + 2 * header_size, 2);
+}
 
-TEST_F(LogTest, ReadFourthOneOff) { CheckInitialOffsetRecord(20015, 3); }
+TEST_P(LogTest, ReadFourthOneOff) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  CheckInitialOffsetRecord(20000 + 2 * header_size + 1, 3);
+}
 
-TEST_F(LogTest, ReadFourthFirstBlockTrailer) {
+TEST_P(LogTest, ReadFourthFirstBlockTrailer) {
   CheckInitialOffsetRecord(log::kBlockSize - 4, 3);
 }
 
-TEST_F(LogTest, ReadFourthMiddleBlock) {
+TEST_P(LogTest, ReadFourthMiddleBlock) {
   CheckInitialOffsetRecord(log::kBlockSize + 1, 3);
 }
 
-TEST_F(LogTest, ReadFourthLastBlock) {
+TEST_P(LogTest, ReadFourthLastBlock) {
   CheckInitialOffsetRecord(2 * log::kBlockSize + 1, 3);
 }
 
-TEST_F(LogTest, ReadFourthStart) {
+TEST_P(LogTest, ReadFourthStart) {
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
   CheckInitialOffsetRecord(
-      2 * (kHeaderSize + 1000) + (2 * log::kBlockSize - 1000) + 3 * kHeaderSize,
+      2 * (header_size + 1000) + (2 * log::kBlockSize - 1000) + 3 * header_size,
       3);
 }
 
-TEST_F(LogTest, ReadEnd) { CheckOffsetPastEndReturnsNoRecords(0); }
+TEST_P(LogTest, ReadEnd) { CheckOffsetPastEndReturnsNoRecords(0); }
 
-TEST_F(LogTest, ReadPastEnd) { CheckOffsetPastEndReturnsNoRecords(5); }
+TEST_P(LogTest, ReadPastEnd) { CheckOffsetPastEndReturnsNoRecords(5); }
 
-TEST_F(LogTest, ClearEofSingleBlock) {
+TEST_P(LogTest, ClearEofSingleBlock) {
   Write("foo");
   Write("bar");
-  ForceEOF(3 + kHeaderSize + 2);
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  ForceEOF(3 + header_size + 2);
   ASSERT_EQ("foo", Read());
   UnmarkEOF();
   ASSERT_EQ("bar", Read());
@@ -617,12 +640,13 @@ TEST_F(LogTest, ClearEofSingleBlock) {
   ASSERT_TRUE(IsEOF());
 }
 
-TEST_F(LogTest, ClearEofMultiBlock) {
+TEST_P(LogTest, ClearEofMultiBlock) {
   size_t num_full_blocks = 5;
-  size_t n = (kBlockSize - kHeaderSize) * num_full_blocks + 25;
+  int header_size = GetParam() ? kRecyclableHeaderSize : kHeaderSize;
+  size_t n = (kBlockSize - header_size) * num_full_blocks + 25;
   Write(BigString("foo", n));
   Write(BigString("bar", n));
-  ForceEOF(n + num_full_blocks * kHeaderSize + 10);
+  ForceEOF(n + num_full_blocks * header_size + header_size + 3);
   ASSERT_EQ(BigString("foo", n), Read());
   ASSERT_TRUE(IsEOF());
   UnmarkEOF();
@@ -634,7 +658,7 @@ TEST_F(LogTest, ClearEofMultiBlock) {
   ASSERT_TRUE(IsEOF());
 }
 
-TEST_F(LogTest, ClearEofError) {
+TEST_P(LogTest, ClearEofError) {
   // If an error occurs during Read() in UnmarkEOF(), the records contained
   // in the buffer should be returned on subsequent calls of ReadRecord()
   // until no more full records are left, whereafter ReadRecord() should return
@@ -652,7 +676,7 @@ TEST_F(LogTest, ClearEofError) {
   ASSERT_EQ("EOF", Read());
 }
 
-TEST_F(LogTest, ClearEofError2) {
+TEST_P(LogTest, ClearEofError2) {
   Write("foo");
   Write("bar");
   UnmarkEOF();
@@ -666,6 +690,8 @@ TEST_F(LogTest, ClearEofError2) {
   ASSERT_EQ("OK", MatchError("read error"));
 }
 
+INSTANTIATE_TEST_CASE_P(bool, LogTest, ::testing::Values(0, 2));
+
 }  // namespace log
 }  // namespace rocksdb
 

db/log_writer.cc

@@ -18,8 +18,12 @@
 namespace rocksdb {
 namespace log {
 
-Writer::Writer(unique_ptr<WritableFileWriter>&& dest)
-    : dest_(std::move(dest)), block_offset_(0) {
+Writer::Writer(unique_ptr<WritableFileWriter>&& dest,
+               uint64_t log_number, bool recycle_log_files)
+    : dest_(std::move(dest)),
+      block_offset_(0),
+      log_number_(log_number),
+      recycle_log_files_(recycle_log_files) {
   for (int i = 0; i <= kMaxRecordType; i++) {
     char t = static_cast<char>(i);
     type_crc_[i] = crc32c::Value(&t, 1);
@@ -33,6 +37,10 @@ Status Writer::AddRecord(const Slice& slice) {
   const char* ptr = slice.data();
   size_t left = slice.size();
 
+  // Header size varies depending on whether we are recycling or not.
+  const int header_size =
+      recycle_log_files_ ? kRecyclableHeaderSize : kHeaderSize;
+
   // Fragment the record if necessary and emit it.  Note that if slice
   // is empty, we still want to iterate once to emit a single
   // zero-length record
@@ -41,32 +49,34 @@ Status Writer::AddRecord(const Slice& slice) {
   do {
     const int leftover = kBlockSize - block_offset_;
     assert(leftover >= 0);
-    if (leftover < kHeaderSize) {
+    if (leftover < header_size) {
       // Switch to a new block
       if (leftover > 0) {
-        // Fill the trailer (literal below relies on kHeaderSize being 7)
-        assert(kHeaderSize == 7);
-        dest_->Append(Slice("\x00\x00\x00\x00\x00\x00", leftover));
+        // Fill the trailer (literal below relies on kHeaderSize and
+        // kRecyclableHeaderSize being <= 11)
+        assert(header_size <= 11);
+        dest_->Append(
+            Slice("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00", leftover));
       }
       block_offset_ = 0;
     }
 
-    // Invariant: we never leave < kHeaderSize bytes in a block.
-    assert(static_cast<int>(kBlockSize) - block_offset_ >= kHeaderSize);
+    // Invariant: we never leave < header_size bytes in a block.
+    assert(static_cast<int>(kBlockSize) - block_offset_ >= header_size);
 
-    const size_t avail = kBlockSize - block_offset_ - kHeaderSize;
+    const size_t avail = kBlockSize - block_offset_ - header_size;
     const size_t fragment_length = (left < avail) ? left : avail;
 
     RecordType type;
     const bool end = (left == fragment_length);
     if (begin && end) {
-      type = kFullType;
+      type = recycle_log_files_ ? kRecyclableFullType : kFullType;
     } else if (begin) {
-      type = kFirstType;
+      type = recycle_log_files_ ? kRecyclableFirstType : kFirstType;
     } else if (end) {
-      type = kLastType;
+      type = recycle_log_files_ ? kRecyclableLastType : kLastType;
     } else {
-      type = kMiddleType;
+      type = recycle_log_files_ ? kRecyclableMiddleType : kMiddleType;
     }
 
     s = EmitPhysicalRecord(type, ptr, fragment_length);
@@ -79,28 +89,48 @@ Status Writer::AddRecord(const Slice& slice) {
 
 Status Writer::EmitPhysicalRecord(RecordType t, const char* ptr, size_t n) {
   assert(n <= 0xffff);  // Must fit in two bytes
-  assert(block_offset_ + kHeaderSize + n <= kBlockSize);
+
+  size_t header_size;
+  char buf[kRecyclableHeaderSize];
 
   // Format the header
-  char buf[kHeaderSize];
   buf[4] = static_cast<char>(n & 0xff);
   buf[5] = static_cast<char>(n >> 8);
   buf[6] = static_cast<char>(t);
 
+  uint32_t crc = type_crc_[t];
+  if (t < kRecyclableFullType) {
+    // Legacy record format
+    assert(block_offset_ + kHeaderSize + n <= kBlockSize);
+    header_size = kHeaderSize;
+  } else {
+    // Recyclable record format
+    assert(block_offset_ + kRecyclableHeaderSize + n <= kBlockSize);
+    header_size = kRecyclableHeaderSize;
+
+    // Only encode low 32-bits of the 64-bit log number.  This means
+    // we will fail to detect an old record if we recycled a log from
+    // ~4 billion logs ago, but that is effectively impossible, and
+    // even if it were we'dbe far more likely to see a false positive
+    // on the 32-bit CRC.
+    EncodeFixed32(buf + 7, static_cast<uint32_t>(log_number_));
+    crc = crc32c::Extend(crc, buf + 7, 4);
+  }
+
   // Compute the crc of the record type and the payload.
-  uint32_t crc = crc32c::Extend(type_crc_[t], ptr, n);
-  crc = crc32c::Mask(crc);                 // Adjust for storage
+  crc = crc32c::Extend(crc, ptr, n);
+  crc = crc32c::Mask(crc);  // Adjust for storage
   EncodeFixed32(buf, crc);
 
   // Write the header and the payload
-  Status s = dest_->Append(Slice(buf, kHeaderSize));
+  Status s = dest_->Append(Slice(buf, header_size));
   if (s.ok()) {
     s = dest_->Append(Slice(ptr, n));
     if (s.ok()) {
       s = dest_->Flush();
     }
   }
-  block_offset_ += kHeaderSize + n;
+  block_offset_ += header_size + n;
   return s;
 }
 

db/log_writer.h

@@ -43,7 +43,7 @@ namespace log {
  * Data is written out in kBlockSize chunks. If next record does not fit
  * into the space left, the leftover space will be padded with \0.
  *
- * Record format:
+ * Legacy record format:
  *
  * +---------+-----------+-----------+--- ... ---+
  * |CRC (4B) | Size (2B) | Type (1B) | Payload   |
@@ -57,13 +57,23 @@ namespace log {
  *        blocks that are larger than kBlockSize
  * Payload = Byte stream as long as specified by the payload size
  *
+ * Recyclable record format:
+ *
+ * +---------+-----------+-----------+----------------+--- ... ---+
+ * |CRC (4B) | Size (2B) | Type (1B) | Log number (4B)| Payload   |
+ * +---------+-----------+-----------+----------------+--- ... ---+
+ *
+ * Same as above, with the addition of
+ * Log number = 32bit log file number, so that we can distinguish between
+ * records written by the most recent log writer vs a previous one.
  */
 class Writer {
  public:
   // Create a writer that will append data to "*dest".
   // "*dest" must be initially empty.
   // "*dest" must remain live while this Writer is in use.
-  explicit Writer(unique_ptr<WritableFileWriter>&& dest);
+  explicit Writer(unique_ptr<WritableFileWriter>&& dest,
+                  uint64_t log_number, bool recycle_log_files);
   ~Writer();
 
   Status AddRecord(const Slice& slice);
@@ -74,6 +84,8 @@ class Writer {
  private:
   unique_ptr<WritableFileWriter> dest_;
   int block_offset_;       // Current offset in block
+  uint64_t log_number_;
+  bool recycle_log_files_;
 
   // crc32c values for all supported record types.  These are
   // pre-computed to reduce the overhead of computing the crc of the

db/memtable.cc

@@ -21,6 +21,7 @@
 #include "rocksdb/iterator.h"
 #include "rocksdb/merge_operator.h"
 #include "rocksdb/slice_transform.h"
+#include "table/internal_iterator.h"
 #include "table/merger.h"
 #include "util/arena.h"
 #include "util/coding.h"
@@ -202,7 +203,7 @@ const char* EncodeKey(std::string* scratch, const Slice& target) {
   return scratch->data();
 }
 
-class MemTableIterator: public Iterator {
+class MemTableIterator : public InternalIterator {
  public:
   MemTableIterator(
       const MemTable& mem, const ReadOptions& read_options, Arena* arena)
@@ -285,7 +286,8 @@ class MemTableIterator: public Iterator {
   void operator=(const MemTableIterator&);
 };
 
-Iterator* MemTable::NewIterator(const ReadOptions& read_options, Arena* arena) {
+InternalIterator* MemTable::NewIterator(const ReadOptions& read_options,
+                                        Arena* arena) {
   assert(arena != nullptr);
   auto mem = arena->AllocateAligned(sizeof(MemTableIterator));
   return new (mem) MemTableIterator(*this, read_options, arena);

db/memtable.h

@@ -31,6 +31,7 @@ class Mutex;
 class MemTableIterator;
 class MergeContext;
 class WriteBuffer;
+class InternalIterator;
 
 struct MemTableOptions {
   explicit MemTableOptions(
@@ -140,7 +141,7 @@ class MemTable {
   // arena: If not null, the arena needs to be used to allocate the Iterator.
   //        Calling ~Iterator of the iterator will destroy all the states but
   //        those allocated in arena.
-  Iterator* NewIterator(const ReadOptions& read_options, Arena* arena);
+  InternalIterator* NewIterator(const ReadOptions& read_options, Arena* arena);
 
   // Add an entry into memtable that maps key to value at the
   // specified sequence number and with the specified type.

db/memtable_list.cc

@@ -138,9 +138,9 @@ bool MemTableListVersion::GetFromList(std::list<MemTable*>* list,
   return false;
 }
 
-void MemTableListVersion::AddIterators(const ReadOptions& options,
-                                       std::vector<Iterator*>* iterator_list,
-                                       Arena* arena) {
+void MemTableListVersion::AddIterators(
+    const ReadOptions& options, std::vector<InternalIterator*>* iterator_list,
+    Arena* arena) {
   for (auto& m : memlist_) {
     iterator_list->push_back(m->NewIterator(options, arena));
   }

db/memtable_list.h

@@ -74,7 +74,8 @@ class MemTableListVersion {
   }
 
   void AddIterators(const ReadOptions& options,
-                    std::vector<Iterator*>* iterator_list, Arena* arena);
+                    std::vector<InternalIterator*>* iterator_list,
+                    Arena* arena);
 
   void AddIterators(const ReadOptions& options,
                     MergeIteratorBuilder* merge_iter_builder);

db/memtablerep_bench.cc

@@ -590,6 +590,7 @@ int main(int argc, char** argv) {
   std::unique_ptr<rocksdb::MemTableRepFactory> factory;
   if (FLAGS_memtablerep == "skiplist") {
     factory.reset(new rocksdb::SkipListFactory);
+#ifndef ROCKSDB_LITE
   } else if (FLAGS_memtablerep == "vector") {
     factory.reset(new rocksdb::VectorRepFactory);
   } else if (FLAGS_memtablerep == "hashskiplist") {
@@ -611,6 +612,7 @@ int main(int argc, char** argv) {
         static_cast<uint32_t>(FLAGS_hash_function_count)));
     options.prefix_extractor.reset(
         rocksdb::NewFixedPrefixTransform(FLAGS_prefix_length));
+#endif  // ROCKSDB_LITE
   } else {
     fprintf(stdout, "Unknown memtablerep: %s\n", FLAGS_memtablerep.c_str());
     exit(1);

db/merge_helper.cc

@@ -12,6 +12,7 @@
 #include "rocksdb/comparator.h"
 #include "rocksdb/db.h"
 #include "rocksdb/merge_operator.h"
+#include "table/internal_iterator.h"
 #include "util/perf_context_imp.h"
 #include "util/statistics.h"
 
@@ -56,7 +57,8 @@ Status MergeHelper::TimedFullMerge(const Slice& key, const Slice* value,
 //       keys_ stores the list of keys encountered while merging.
 //       operands_ stores the list of merge operands encountered while merging.
 //       keys_[i] corresponds to operands_[i] for each i.
-Status MergeHelper::MergeUntil(Iterator* iter, const SequenceNumber stop_before,
+Status MergeHelper::MergeUntil(InternalIterator* iter,
+                               const SequenceNumber stop_before,
                                const bool at_bottom) {
   // Get a copy of the internal key, before it's invalidated by iter->Next()
   // Also maintain the list of merge operands seen.

db/merge_helper.h

@@ -22,6 +22,7 @@ class Iterator;
 class Logger;
 class MergeOperator;
 class Statistics;
+class InternalIterator;
 
 class MergeHelper {
  public:
@@ -82,7 +83,8 @@ class MergeHelper {
   //   with asserts removed).
   //
   // REQUIRED: The first key in the input is not corrupted.
-  Status MergeUntil(Iterator* iter, const SequenceNumber stop_before = 0,
+  Status MergeUntil(InternalIterator* iter,
+                    const SequenceNumber stop_before = 0,
                     const bool at_bottom = false);
 
   // Filters a merge operand using the compaction filter specified

db/perf_context_test.cc

@@ -5,6 +5,7 @@
 //
 #include <algorithm>
 #include <iostream>
+#include <thread>
 #include <vector>
 
 #include "rocksdb/db.h"
@@ -444,6 +445,7 @@ void ProfileQueries(bool enabled_time = false) {
   }
 }
 
+#ifndef ROCKSDB_LITE
 TEST_F(PerfContextTest, KeyComparisonCount) {
   SetPerfLevel(kEnableCount);
   ProfileQueries();
@@ -454,6 +456,7 @@ TEST_F(PerfContextTest, KeyComparisonCount) {
   SetPerfLevel(kEnableTime);
   ProfileQueries(true);
 }
+#endif  // ROCKSDB_LITE
 
 // make perf_context_test
 // export ROCKSDB_TESTS=PerfContextTest.SeekKeyComparison
@@ -539,6 +542,49 @@ TEST_F(PerfContextTest, SeekKeyComparison) {
   }
 }
 
+TEST_F(PerfContextTest, DBMutexLockCounter) {
+  SetPerfLevel(kEnableTime);
+  int stats_code[] = {0, static_cast<int>(DB_MUTEX_WAIT_MICROS)};
+  for (int c = 0; c < 2; ++c) {
+    InstrumentedMutex mutex(nullptr, Env::Default(), stats_code[c]);
+    mutex.Lock();
+    std::thread child_thread([&] {
+      SetPerfLevel(kEnableTime);
+      perf_context.Reset();
+      ASSERT_EQ(perf_context.db_mutex_lock_nanos, 0);
+      mutex.Lock();
+      mutex.Unlock();
+      if (stats_code[c] == DB_MUTEX_WAIT_MICROS) {
+        // increment the counter only when it's a DB Mutex
+        ASSERT_GT(perf_context.db_mutex_lock_nanos, 0);
+      } else {
+        ASSERT_EQ(perf_context.db_mutex_lock_nanos, 0);
+      }
+    });
+    Env::Default()->SleepForMicroseconds(100);
+    mutex.Unlock();
+    child_thread.join();
+  }
+}
+
+TEST_F(PerfContextTest, FalseDBMutexWait) {
+  SetPerfLevel(kEnableTime);
+  int stats_code[] = {0, static_cast<int>(DB_MUTEX_WAIT_MICROS)};
+  for (int c = 0; c < 2; ++c) {
+    InstrumentedMutex mutex(nullptr, Env::Default(), stats_code[c]);
+    InstrumentedCondVar lock(&mutex);
+    perf_context.Reset();
+    mutex.Lock();
+    lock.TimedWait(100);
+    mutex.Unlock();
+    if (stats_code[c] == static_cast<int>(DB_MUTEX_WAIT_MICROS)) {
+      // increment the counter only when it's a DB Mutex
+      ASSERT_GT(perf_context.db_condition_wait_nanos, 0);
+    } else {
+      ASSERT_EQ(perf_context.db_condition_wait_nanos, 0);
+    }
+  }
+}
 }
 
 int main(int argc, char** argv) {

db/prefix_test.cc

@@ -20,9 +20,11 @@ int main() {
 #include <gflags/gflags.h>
 #include "rocksdb/comparator.h"
 #include "rocksdb/db.h"
+#include "rocksdb/filter_policy.h"
 #include "rocksdb/perf_context.h"
 #include "rocksdb/slice_transform.h"
 #include "rocksdb/memtablerep.h"
+#include "rocksdb/table.h"
 #include "util/histogram.h"
 #include "util/stop_watch.h"
 #include "util/string_util.h"
@@ -163,6 +165,12 @@ class PrefixTest : public testing::Test {
     options.memtable_prefix_bloom_huge_page_tlb_size =
         FLAGS_memtable_prefix_bloom_huge_page_tlb_size;
 
+    options.prefix_extractor.reset(NewFixedPrefixTransform(8));
+    BlockBasedTableOptions bbto;
+    bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
+    bbto.whole_key_filtering = false;
+    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
+
     Status s = DB::Open(options, kDbName,  &db);
     EXPECT_OK(s);
     return std::shared_ptr<DB>(db);
@@ -393,6 +401,55 @@ TEST_F(PrefixTest, TestResult) {
   }
 }
 
+// Show results in prefix
+TEST_F(PrefixTest, PrefixValid) {
+  for (int num_buckets = 1; num_buckets <= 2; num_buckets++) {
+    FirstOption();
+    while (NextOptions(num_buckets)) {
+      std::cout << "*** Mem table: " << options.memtable_factory->Name()
+                << " number of buckets: " << num_buckets << std::endl;
+      DestroyDB(kDbName, Options());
+      auto db = OpenDb();
+      WriteOptions write_options;
+      ReadOptions read_options;
+
+      // Insert keys with common prefix and one key with different
+      Slice v16("v16");
+      Slice v17("v17");
+      Slice v18("v18");
+      Slice v19("v19");
+      PutKey(db.get(), write_options, 12345, 6, v16);
+      PutKey(db.get(), write_options, 12345, 7, v17);
+      PutKey(db.get(), write_options, 12345, 8, v18);
+      PutKey(db.get(), write_options, 12345, 9, v19);
+      PutKey(db.get(), write_options, 12346, 8, v16);
+      db->Flush(FlushOptions());
+      db->Delete(write_options, TestKeyToSlice(TestKey(12346, 8)));
+      db->Flush(FlushOptions());
+      read_options.prefix_same_as_start = true;
+      std::unique_ptr<Iterator> iter(db->NewIterator(read_options));
+      SeekIterator(iter.get(), 12345, 6);
+      ASSERT_TRUE(iter->Valid());
+      ASSERT_TRUE(v16 == iter->value());
+
+      iter->Next();
+      ASSERT_TRUE(iter->Valid());
+      ASSERT_TRUE(v17 == iter->value());
+
+      iter->Next();
+      ASSERT_TRUE(iter->Valid());
+      ASSERT_TRUE(v18 == iter->value());
+
+      iter->Next();
+      ASSERT_TRUE(iter->Valid());
+      ASSERT_TRUE(v19 == iter->value());
+      iter->Next();
+      ASSERT_FALSE(iter->Valid());
+      ASSERT_EQ(kNotFoundResult, Get(db.get(), read_options, 12346, 8));
+    }
+  }
+}
+
 TEST_F(PrefixTest, DynamicPrefixIterator) {
   while (NextOptions(FLAGS_bucket_count)) {
     std::cout << "*** Mem table: " << options.memtable_factory->Name()

db/repair.cc

@@ -81,8 +81,8 @@
 #include "rocksdb/env.h"
 #include "rocksdb/options.h"
 #include "rocksdb/immutable_options.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/file_reader_writer.h"
-#include "util/scoped_arena_iterator.h"
 
 namespace rocksdb {
 
@@ -249,8 +249,8 @@ class Repairer {
     // corruptions cause entire commits to be skipped instead of
     // propagating bad information (like overly large sequence
     // numbers).
-    log::Reader reader(std::move(lfile_reader), &reporter,
-                       true /*enable checksum*/, 0 /*initial_offset*/);
+    log::Reader reader(options_.info_log, std::move(lfile_reader), &reporter,
+                       true /*enable checksum*/, 0 /*initial_offset*/, log);
 
     // Read all the records and add to a memtable
     std::string scratch;
@@ -290,10 +290,11 @@ class Repairer {
       ro.total_order_seek = true;
       Arena arena;
       ScopedArenaIterator iter(mem->NewIterator(ro, &arena));
-      status = BuildTable(dbname_, env_, ioptions_, env_options_, table_cache_,
-                          iter.get(), &meta, icmp_,
-                          &int_tbl_prop_collector_factories_, {},
-                          kNoCompression, CompressionOptions(), false, nullptr);
+      status = BuildTable(
+          dbname_, env_, ioptions_, env_options_, table_cache_, iter.get(),
+          &meta, icmp_, &int_tbl_prop_collector_factories_,
+          TablePropertiesCollectorFactory::Context::kUnknownColumnFamily, {},
+          kNoCompression, CompressionOptions(), false, nullptr);
     }
     delete mem->Unref();
     delete cf_mems_default;
@@ -339,7 +340,7 @@ class Repairer {
     t->meta.fd = FileDescriptor(t->meta.fd.GetNumber(), t->meta.fd.GetPathId(),
                                 file_size);
     if (status.ok()) {
-      Iterator* iter = table_cache_->NewIterator(
+      InternalIterator* iter = table_cache_->NewIterator(
           ReadOptions(), env_options_, icmp_, t->meta.fd);
       bool empty = true;
       ParsedInternalKey parsed;
@@ -412,7 +413,7 @@ class Repairer {
     {
       unique_ptr<WritableFileWriter> file_writer(
           new WritableFileWriter(std::move(file), env_options));
-      log::Writer log(std::move(file_writer));
+      log::Writer log(std::move(file_writer), 0, false);
       std::string record;
       edit_->EncodeTo(&record);
       status = log.AddRecord(record);

db/table_cache.cc

@@ -14,6 +14,7 @@
 #include "db/version_edit.h"
 
 #include "rocksdb/statistics.h"
+#include "table/internal_iterator.h"
 #include "table/iterator_wrapper.h"
 #include "table/table_builder.h"
 #include "table/table_reader.h"
@@ -148,13 +149,11 @@ Status TableCache::FindTable(const EnvOptions& env_options,
   return s;
 }
 
-Iterator* TableCache::NewIterator(const ReadOptions& options,
-                                  const EnvOptions& env_options,
-                                  const InternalKeyComparator& icomparator,
-                                  const FileDescriptor& fd,
-                                  TableReader** table_reader_ptr,
-                                  HistogramImpl* file_read_hist,
-                                  bool for_compaction, Arena* arena) {
+InternalIterator* TableCache::NewIterator(
+    const ReadOptions& options, const EnvOptions& env_options,
+    const InternalKeyComparator& icomparator, const FileDescriptor& fd,
+    TableReader** table_reader_ptr, HistogramImpl* file_read_hist,
+    bool for_compaction, Arena* arena) {
   PERF_TIMER_GUARD(new_table_iterator_nanos);
 
   if (table_reader_ptr != nullptr) {
@@ -171,7 +170,7 @@ Iterator* TableCache::NewIterator(const ReadOptions& options,
         env_options, icomparator, fd, /* sequential mode */ true,
         /* record stats */ false, nullptr, &table_reader_unique_ptr);
     if (!s.ok()) {
-      return NewErrorIterator(s, arena);
+      return NewErrorInternalIterator(s, arena);
     }
     table_reader = table_reader_unique_ptr.release();
   } else {
@@ -182,13 +181,13 @@ Iterator* TableCache::NewIterator(const ReadOptions& options,
                     options.read_tier == kBlockCacheTier /* no_io */,
                     !for_compaction /* record read_stats */, file_read_hist);
       if (!s.ok()) {
-        return NewErrorIterator(s, arena);
+        return NewErrorInternalIterator(s, arena);
       }
       table_reader = GetTableReaderFromHandle(handle);
     }
   }
 
-  Iterator* result = table_reader->NewIterator(options, arena);
+  InternalIterator* result = table_reader->NewIterator(options, arena);
 
   if (create_new_table_reader) {
     assert(handle == nullptr);

db/table_cache.h

@@ -29,6 +29,7 @@ class Arena;
 struct FileDescriptor;
 class GetContext;
 class HistogramImpl;
+class InternalIterator;
 
 class TableCache {
  public:
@@ -43,12 +44,12 @@ class TableCache {
   // the returned iterator.  The returned "*tableptr" object is owned by
   // the cache and should not be deleted, and is valid for as long as the
   // returned iterator is live.
-  Iterator* NewIterator(const ReadOptions& options, const EnvOptions& toptions,
-                        const InternalKeyComparator& internal_comparator,
-                        const FileDescriptor& file_fd,
-                        TableReader** table_reader_ptr = nullptr,
-                        HistogramImpl* file_read_hist = nullptr,
-                        bool for_compaction = false, Arena* arena = nullptr);
+  InternalIterator* NewIterator(
+      const ReadOptions& options, const EnvOptions& toptions,
+      const InternalKeyComparator& internal_comparator,
+      const FileDescriptor& file_fd, TableReader** table_reader_ptr = nullptr,
+      HistogramImpl* file_read_hist = nullptr, bool for_compaction = false,
+      Arena* arena = nullptr);
 
   // If a seek to internal key "k" in specified file finds an entry,
   // call (*handle_result)(arg, found_key, found_value) repeatedly until

db/table_properties_collector.h

@@ -41,7 +41,8 @@ class IntTblPropCollectorFactory {
  public:
   virtual ~IntTblPropCollectorFactory() {}
   // has to be thread-safe
-  virtual IntTblPropCollector* CreateIntTblPropCollector() = 0;
+  virtual IntTblPropCollector* CreateIntTblPropCollector(
+      uint32_t column_family_id) = 0;
 
   // The name of the properties collector can be used for debugging purpose.
   virtual const char* Name() const = 0;
@@ -69,7 +70,8 @@ class InternalKeyPropertiesCollector : public IntTblPropCollector {
 class InternalKeyPropertiesCollectorFactory
     : public IntTblPropCollectorFactory {
  public:
-  virtual IntTblPropCollector* CreateIntTblPropCollector() override {
+  virtual IntTblPropCollector* CreateIntTblPropCollector(
+      uint32_t column_family_id) override {
     return new InternalKeyPropertiesCollector();
   }
 
@@ -114,9 +116,12 @@ class UserKeyTablePropertiesCollectorFactory
   explicit UserKeyTablePropertiesCollectorFactory(
       std::shared_ptr<TablePropertiesCollectorFactory> user_collector_factory)
       : user_collector_factory_(user_collector_factory) {}
-  virtual IntTblPropCollector* CreateIntTblPropCollector() override {
+  virtual IntTblPropCollector* CreateIntTblPropCollector(
+      uint32_t column_family_id) override {
+    TablePropertiesCollectorFactory::Context context;
+    context.column_family_id = column_family_id;
     return new UserKeyTablePropertiesCollector(
-        user_collector_factory_->CreateTablePropertiesCollector());
+        user_collector_factory_->CreateTablePropertiesCollector(context));
   }
 
   virtual const char* Name() const override {

db/table_properties_collector_test.cc

@@ -35,6 +35,8 @@ class TablePropertiesTest : public testing::Test,
 
 // Utilities test functions
 namespace {
+static const uint32_t kTestColumnFamilyId = 66;
+
 void MakeBuilder(const Options& options, const ImmutableCFOptions& ioptions,
                  const InternalKeyComparator& internal_comparator,
                  const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
@@ -46,7 +48,8 @@ void MakeBuilder(const Options& options, const ImmutableCFOptions& ioptions,
 
   builder->reset(NewTableBuilder(
       ioptions, internal_comparator, int_tbl_prop_collector_factories,
-      writable->get(), options.compression, options.compression_opts));
+      kTestColumnFamilyId /* column_family_id */, writable->get(),
+      options.compression, options.compression_opts));
 }
 }  // namespace
 
@@ -178,14 +181,17 @@ class RegularKeysStartWithAFactory : public IntTblPropCollectorFactory,
  public:
   explicit RegularKeysStartWithAFactory(bool backward_mode)
       : backward_mode_(backward_mode) {}
-  virtual TablePropertiesCollector* CreateTablePropertiesCollector() override {
+  virtual TablePropertiesCollector* CreateTablePropertiesCollector(
+      TablePropertiesCollectorFactory::Context context) override {
+    EXPECT_EQ(kTestColumnFamilyId, context.column_family_id);
     if (!backward_mode_) {
       return new RegularKeysStartWithA();
     } else {
       return new RegularKeysStartWithABackwardCompatible();
     }
   }
-  virtual IntTblPropCollector* CreateIntTblPropCollector() override {
+  virtual IntTblPropCollector* CreateIntTblPropCollector(
+      uint32_t column_family_id) override {
     return new RegularKeysStartWithAInternal();
   }
   const char* Name() const override { return "RegularKeysStartWithA"; }

db/transaction_log_impl.cc

@@ -262,8 +262,10 @@ Status TransactionLogIteratorImpl::OpenLogReader(const LogFile* logFile) {
     return s;
   }
   assert(file);
-  currentLogReader_.reset(new log::Reader(std::move(file), &reporter_,
-                                          read_options_.verify_checksums_, 0));
+  currentLogReader_.reset(new log::Reader(options_->info_log,
+					  std::move(file), &reporter_,
+                                          read_options_.verify_checksums_, 0,
+                                          logFile->LogNumber()));
   return Status::OK();
 }
 }  //  namespace rocksdb

db/version_builder.cc

@@ -82,6 +82,7 @@ class VersionBuilder::Rep {
   };
 
   const EnvOptions& env_options_;
+  Logger* info_log_;
   TableCache* table_cache_;
   VersionStorageInfo* base_vstorage_;
   LevelState* levels_;
@@ -89,9 +90,10 @@ class VersionBuilder::Rep {
   FileComparator level_nonzero_cmp_;
 
  public:
-  Rep(const EnvOptions& env_options, TableCache* table_cache,
+  Rep(const EnvOptions& env_options, Logger* info_log, TableCache* table_cache,
       VersionStorageInfo* base_vstorage)
       : env_options_(env_options),
+        info_log_(info_log),
         table_cache_(table_cache),
         base_vstorage_(base_vstorage) {
     levels_ = new LevelState[base_vstorage_->num_levels()];
@@ -134,7 +136,10 @@ class VersionBuilder::Rep {
         auto f2 = level_files[i];
         if (level == 0) {
           assert(level_zero_cmp_(f1, f2));
-          assert(f1->largest_seqno > f2->largest_seqno);
+          assert(f1->largest_seqno > f2->largest_seqno ||
+                 // We can have multiple files with seqno = 0 as a result of
+                 // using DB::AddFile()
+                 (f1->largest_seqno == 0 && f2->largest_seqno == 0));
         } else {
           assert(level_nonzero_cmp_(f1, f2));
 
@@ -335,15 +340,16 @@ class VersionBuilder::Rep {
     if (levels_[level].deleted_files.count(f->fd.GetNumber()) > 0) {
       // File is deleted: do nothing
     } else {
-      vstorage->AddFile(level, f);
+      vstorage->AddFile(level, f, info_log_);
     }
   }
 };
 
 VersionBuilder::VersionBuilder(const EnvOptions& env_options,
                                TableCache* table_cache,
-                               VersionStorageInfo* base_vstorage)
-    : rep_(new Rep(env_options, table_cache, base_vstorage)) {}
+                               VersionStorageInfo* base_vstorage,
+                               Logger* info_log)
+    : rep_(new Rep(env_options, info_log, table_cache, base_vstorage)) {}
 VersionBuilder::~VersionBuilder() { delete rep_; }
 void VersionBuilder::CheckConsistency(VersionStorageInfo* vstorage) {
   rep_->CheckConsistency(vstorage);

db/version_builder.h

@@ -24,7 +24,7 @@ class InternalStats;
 class VersionBuilder {
  public:
   VersionBuilder(const EnvOptions& env_options, TableCache* table_cache,
-                 VersionStorageInfo* base_vstorage);
+                 VersionStorageInfo* base_vstorage, Logger* info_log = nullptr);
   ~VersionBuilder();
   void CheckConsistency(VersionStorageInfo* vstorage);
   void CheckConsistencyForDeletes(VersionEdit* edit, uint64_t number,

db/version_edit.cc

@@ -12,6 +12,7 @@
 #include "db/version_set.h"
 #include "util/coding.h"
 #include "util/event_logger.h"
+#include "util/sync_point.h"
 #include "rocksdb/slice.h"
 
 namespace rocksdb {
@@ -32,12 +33,22 @@ enum Tag {
   // these are new formats divergent from open source leveldb
   kNewFile2 = 100,
   kNewFile3 = 102,
+  kNewFile4 = 103,      // 4th (the latest) format version of adding files
   kColumnFamily = 200,  // specify column family for version edit
   kColumnFamilyAdd = 201,
   kColumnFamilyDrop = 202,
   kMaxColumnFamily = 203,
 };
 
+enum CustomTag {
+  kTerminate = 1,  // The end of customized fields
+  kNeedCompaction = 2,
+  kPathId = 65,
+};
+// If this bit for the custom tag is set, opening DB should fail if
+// we don't know this field.
+uint32_t kCustomTagNonSafeIgnoreMask = 1 << 6;
+
 uint64_t PackFileNumberAndPathId(uint64_t number, uint64_t path_id) {
   assert(number <= kFileNumberMask);
   return number | (path_id * (kFileNumberMask + 1));
@@ -102,7 +113,11 @@ bool VersionEdit::EncodeTo(std::string* dst) const {
     if (!f.smallest.Valid() || !f.largest.Valid()) {
       return false;
     }
-    if (f.fd.GetPathId() == 0) {
+    bool has_customized_fields = false;
+    if (f.marked_for_compaction) {
+      PutVarint32(dst, kNewFile4);
+      has_customized_fields = true;
+    } else if (f.fd.GetPathId() == 0) {
       // Use older format to make sure user can roll back the build if they
       // don't config multiple DB paths.
       PutVarint32(dst, kNewFile2);
@@ -111,7 +126,8 @@ bool VersionEdit::EncodeTo(std::string* dst) const {
     }
     PutVarint32(dst, new_files_[i].first);  // level
     PutVarint64(dst, f.fd.GetNumber());
-    if (f.fd.GetPathId() != 0) {
+    if (f.fd.GetPathId() != 0 && !has_customized_fields) {
+      // kNewFile3
       PutVarint32(dst, f.fd.GetPathId());
     }
     PutVarint64(dst, f.fd.GetFileSize());
@@ -119,6 +135,48 @@ bool VersionEdit::EncodeTo(std::string* dst) const {
     PutLengthPrefixedSlice(dst, f.largest.Encode());
     PutVarint64(dst, f.smallest_seqno);
     PutVarint64(dst, f.largest_seqno);
+    if (has_customized_fields) {
+      // Customized fields' format:
+      // +-----------------------------+
+      // | 1st field's tag (varint32)  |
+      // +-----------------------------+
+      // | 1st field's size (varint32) |
+      // +-----------------------------+
+      // |    bytes for 1st field      |
+      // |  (based on size decoded)    |
+      // +-----------------------------+
+      // |                             |
+      // |          ......             |
+      // |                             |
+      // +-----------------------------+
+      // | last field's size (varint32)|
+      // +-----------------------------+
+      // |    bytes for last field     |
+      // |  (based on size decoded)    |
+      // +-----------------------------+
+      // | terminating tag (varint32)  |
+      // +-----------------------------+
+      //
+      // Customized encoding for fields:
+      //   tag kPathId: 1 byte as path_id
+      //   tag kNeedCompaction:
+      //        now only can take one char value 1 indicating need-compaction
+      //
+      if (f.fd.GetPathId() != 0) {
+        PutVarint32(dst, CustomTag::kPathId);
+        char p = static_cast<char>(f.fd.GetPathId());
+        PutLengthPrefixedSlice(dst, Slice(&p, 1));
+      }
+      if (f.marked_for_compaction) {
+        PutVarint32(dst, CustomTag::kNeedCompaction);
+        char p = static_cast<char>(1);
+        PutLengthPrefixedSlice(dst, Slice(&p, 1));
+      }
+      TEST_SYNC_POINT_CALLBACK("VersionEdit::EncodeTo:NewFile4:CustomizeFields",
+                               dst);
+
+      PutVarint32(dst, CustomTag::kTerminate);
+    }
   }
 
   // 0 is default and does not need to be explicitly written
@@ -161,6 +219,63 @@ bool VersionEdit::GetLevel(Slice* input, int* level, const char** msg) {
   }
 }
 
+const char* VersionEdit::DecodeNewFile4From(Slice* input) {
+  const char* msg = nullptr;
+  int level;
+  FileMetaData f;
+  uint64_t number;
+  uint32_t path_id = 0;
+  uint64_t file_size;
+  if (GetLevel(input, &level, &msg) && GetVarint64(input, &number) &&
+      GetVarint64(input, &file_size) && GetInternalKey(input, &f.smallest) &&
+      GetInternalKey(input, &f.largest) &&
+      GetVarint64(input, &f.smallest_seqno) &&
+      GetVarint64(input, &f.largest_seqno)) {
+    // See comments in VersionEdit::EncodeTo() for format of customized fields
+    while (true) {
+      uint32_t custom_tag;
+      Slice field;
+      if (!GetVarint32(input, &custom_tag)) {
+        return "new-file4 custom field";
+      }
+      if (custom_tag == kTerminate) {
+        break;
+      }
+      if (!GetLengthPrefixedSlice(input, &field)) {
+        return "new-file4 custom field lenth prefixed slice error";
+      }
+      switch (custom_tag) {
+        case kPathId:
+          if (field.size() != 1) {
+            return "path_id field wrong size";
+          }
+          path_id = field[0];
+          if (path_id > 3) {
+            return "path_id wrong vaue";
+          }
+          break;
+        case kNeedCompaction:
+          if (field.size() != 1) {
+            return "need_compaction field wrong size";
+          }
+          f.marked_for_compaction = (field[0] == 1);
+          break;
+        default:
+          if ((custom_tag & kCustomTagNonSafeIgnoreMask) != 0) {
+            // Should not proceed if cannot understand it
+            return "new-file4 custom field not supported";
+          }
+          break;
+      }
+    }
+  } else {
+    return "new-file4 entry";
+  }
+  f.fd = FileDescriptor(number, path_id, file_size);
+  new_files_.push_back(std::make_pair(level, f));
+  return nullptr;
+}
+
 Status VersionEdit::DecodeFrom(const Slice& src) {
   Clear();
   Slice input = src;
@@ -304,6 +419,11 @@ Status VersionEdit::DecodeFrom(const Slice& src) {
         break;
       }
 
+      case kNewFile4: {
+        msg = DecodeNewFile4From(&input);
+        break;
+      }
+
       case kColumnFamily:
         if (!GetVarint32(&input, &column_family_)) {
           if (!msg) {

db/version_edit.h

@@ -237,6 +237,8 @@ class VersionEdit {
   bool EncodeTo(std::string* dst) const;
   Status DecodeFrom(const Slice& src);
 
+  const char* DecodeNewFile4From(Slice* input);
+
   typedef std::set<std::pair<int, uint64_t>> DeletedFileSet;
 
   const DeletedFileSet& GetDeletedFiles() { return deleted_files_; }

db/version_edit_test.cc

@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
 #include "db/version_edit.h"
+#include "util/sync_point.h"
 #include "util/testharness.h"
 
 namespace rocksdb {
@@ -45,6 +46,121 @@ TEST_F(VersionEditTest, EncodeDecode) {
   TestEncodeDecode(edit);
 }
 
+TEST_F(VersionEditTest, EncodeDecodeNewFile4) {
+  static const uint64_t kBig = 1ull << 50;
+
+  VersionEdit edit;
+  edit.AddFile(3, 300, 3, 100, InternalKey("foo", kBig + 500, kTypeValue),
+               InternalKey("zoo", kBig + 600, kTypeDeletion), kBig + 500,
+               kBig + 600, true);
+  edit.AddFile(4, 301, 3, 100, InternalKey("foo", kBig + 501, kTypeValue),
+               InternalKey("zoo", kBig + 601, kTypeDeletion), kBig + 501,
+               kBig + 601, false);
+  edit.AddFile(5, 302, 0, 100, InternalKey("foo", kBig + 502, kTypeValue),
+               InternalKey("zoo", kBig + 602, kTypeDeletion), kBig + 502,
+               kBig + 602, true);
+
+  edit.DeleteFile(4, 700);
+
+  edit.SetComparatorName("foo");
+  edit.SetLogNumber(kBig + 100);
+  edit.SetNextFile(kBig + 200);
+  edit.SetLastSequence(kBig + 1000);
+  TestEncodeDecode(edit);
+
+  std::string encoded, encoded2;
+  edit.EncodeTo(&encoded);
+  VersionEdit parsed;
+  Status s = parsed.DecodeFrom(encoded);
+  ASSERT_TRUE(s.ok()) << s.ToString();
+  auto& new_files = parsed.GetNewFiles();
+  ASSERT_TRUE(new_files[0].second.marked_for_compaction);
+  ASSERT_TRUE(!new_files[1].second.marked_for_compaction);
+  ASSERT_TRUE(new_files[2].second.marked_for_compaction);
+  ASSERT_EQ(3, new_files[0].second.fd.GetPathId());
+  ASSERT_EQ(3, new_files[1].second.fd.GetPathId());
+  ASSERT_EQ(0, new_files[2].second.fd.GetPathId());
+}
+
+TEST_F(VersionEditTest, ForwardCompatibleNewFile4) {
+  static const uint64_t kBig = 1ull << 50;
+  VersionEdit edit;
+  edit.AddFile(3, 300, 3, 100, InternalKey("foo", kBig + 500, kTypeValue),
+               InternalKey("zoo", kBig + 600, kTypeDeletion), kBig + 500,
+               kBig + 600, true);
+  edit.AddFile(4, 301, 3, 100, InternalKey("foo", kBig + 501, kTypeValue),
+               InternalKey("zoo", kBig + 601, kTypeDeletion), kBig + 501,
+               kBig + 601, false);
+  edit.DeleteFile(4, 700);
+
+  edit.SetComparatorName("foo");
+  edit.SetLogNumber(kBig + 100);
+  edit.SetNextFile(kBig + 200);
+  edit.SetLastSequence(kBig + 1000);
+
+  std::string encoded;
+
+  // Call back function to add extra customized builds.
+  bool first = true;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "VersionEdit::EncodeTo:NewFile4:CustomizeFields", [&](void* arg) {
+        std::string* str = reinterpret_cast<std::string*>(arg);
+        PutVarint32(str, 33);
+        const std::string str1 = "random_string";
+        PutLengthPrefixedSlice(str, str1);
+        if (first) {
+          first = false;
+          PutVarint32(str, 22);
+          const std::string str2 = "s";
+          PutLengthPrefixedSlice(str, str2);
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  edit.EncodeTo(&encoded);
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  VersionEdit parsed;
+  Status s = parsed.DecodeFrom(encoded);
+  ASSERT_TRUE(s.ok()) << s.ToString();
+  ASSERT_TRUE(!first);
+  auto& new_files = parsed.GetNewFiles();
+  ASSERT_TRUE(new_files[0].second.marked_for_compaction);
+  ASSERT_TRUE(!new_files[1].second.marked_for_compaction);
+  ASSERT_EQ(3, new_files[0].second.fd.GetPathId());
+  ASSERT_EQ(3, new_files[1].second.fd.GetPathId());
+  ASSERT_EQ(1u, parsed.GetDeletedFiles().size());
+}
+
+TEST_F(VersionEditTest, NewFile4NotSupportedField) {
+  static const uint64_t kBig = 1ull << 50;
+  VersionEdit edit;
+  edit.AddFile(3, 300, 3, 100, InternalKey("foo", kBig + 500, kTypeValue),
+               InternalKey("zoo", kBig + 600, kTypeDeletion), kBig + 500,
+               kBig + 600, true);
+
+  edit.SetComparatorName("foo");
+  edit.SetLogNumber(kBig + 100);
+  edit.SetNextFile(kBig + 200);
+  edit.SetLastSequence(kBig + 1000);
+
+  std::string encoded;
+
+  // Call back function to add extra customized builds.
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "VersionEdit::EncodeTo:NewFile4:CustomizeFields", [&](void* arg) {
+        std::string* str = reinterpret_cast<std::string*>(arg);
+        const std::string str1 = "s";
+        PutLengthPrefixedSlice(str, str1);
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  edit.EncodeTo(&encoded);
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+
+  VersionEdit parsed;
+  Status s = parsed.DecodeFrom(encoded);
+  ASSERT_NOK(s);
+}
+
 TEST_F(VersionEditTest, EncodeEmptyFile) {
   VersionEdit edit;
   edit.AddFile(0, 0, 0, 0, InternalKey(), InternalKey(), 0, 0, false);

db/version_set.cc

@@ -35,6 +35,7 @@
 #include "db/writebuffer.h"
 #include "rocksdb/env.h"
 #include "rocksdb/merge_operator.h"
+#include "table/internal_iterator.h"
 #include "table/table_reader.h"
 #include "table/merger.h"
 #include "table/two_level_iterator.h"
@@ -420,7 +421,7 @@ namespace {
 // is the largest key that occurs in the file, and value() is an
 // 16-byte value containing the file number and file size, both
 // encoded using EncodeFixed64.
-class LevelFileNumIterator : public Iterator {
+class LevelFileNumIterator : public InternalIterator {
  public:
   LevelFileNumIterator(const InternalKeyComparator& icmp,
                        const LevelFilesBrief* flevel)
@@ -488,9 +489,9 @@ class LevelFileIteratorState : public TwoLevelIteratorState {
         file_read_hist_(file_read_hist),
         for_compaction_(for_compaction) {}
 
-  Iterator* NewSecondaryIterator(const Slice& meta_handle) override {
+  InternalIterator* NewSecondaryIterator(const Slice& meta_handle) override {
     if (meta_handle.size() != sizeof(FileDescriptor)) {
-      return NewErrorIterator(
+      return NewErrorInternalIterator(
           Status::Corruption("FileReader invoked with unexpected value"));
     } else {
       const FileDescriptor* fd =
@@ -523,7 +524,7 @@ class BaseReferencedVersionBuilder {
   explicit BaseReferencedVersionBuilder(ColumnFamilyData* cfd)
       : version_builder_(new VersionBuilder(
             cfd->current()->version_set()->env_options(), cfd->table_cache(),
-            cfd->current()->storage_info())),
+            cfd->current()->storage_info(), cfd->ioptions()->info_log)),
         version_(cfd->current()) {
     version_->Ref();
   }
@@ -541,7 +542,7 @@ class BaseReferencedVersionBuilder {
 
 Status Version::GetTableProperties(std::shared_ptr<const TableProperties>* tp,
                                    const FileMetaData* file_meta,
-                                   const std::string* fname) {
+                                   const std::string* fname) const {
   auto table_cache = cfd_->table_cache();
   auto ioptions = cfd_->ioptions();
   Status s = table_cache->GetTableProperties(
@@ -623,6 +624,38 @@ Status Version::GetPropertiesOfAllTables(TablePropertiesCollection* props,
   return Status::OK();
 }
 
+Status Version::GetPropertiesOfTablesInRange(
+    const Range* range, std::size_t n, TablePropertiesCollection* props) const {
+  for (int level = 0; level < storage_info_.num_non_empty_levels(); level++) {
+    for (decltype(n) i = 0; i < n; i++) {
+      // Convert user_key into a corresponding internal key.
+      InternalKey k1(range[i].start, kMaxSequenceNumber, kValueTypeForSeek);
+      InternalKey k2(range[i].limit, kMaxSequenceNumber, kValueTypeForSeek);
+      std::vector<FileMetaData*> files;
+      storage_info_.GetOverlappingInputs(level, &k1, &k2, &files, -1, nullptr,
+                                         false);
+      for (const auto& file_meta : files) {
+        auto fname =
+            TableFileName(vset_->db_options_->db_paths,
+                          file_meta->fd.GetNumber(), file_meta->fd.GetPathId());
+        if (props->count(fname) == 0) {
+          // 1. If the table is already present in table cache, load table
+          // properties from there.
+          std::shared_ptr<const TableProperties> table_properties;
+          Status s = GetTableProperties(&table_properties, file_meta, &fname);
+          if (s.ok()) {
+            props->insert({fname, table_properties});
+          } else {
+            return s;
+          }
+        }
+      }
+    }
+  }
+
+  return Status::OK();
+}
+
 Status Version::GetAggregatedTableProperties(
     std::shared_ptr<const TableProperties>* tp, int level) {
   TablePropertiesCollection props;
@@ -1064,7 +1097,7 @@ void VersionStorageInfo::EstimateCompactionBytesNeeded(
   // We keep doing it to Level 2, 3, etc, until the last level and return the
   // accumulated bytes.
 
-  size_t bytes_compact_to_next_level = 0;
+  uint64_t bytes_compact_to_next_level = 0;
   // Level 0
   bool level0_compact_triggered = false;
   if (static_cast<int>(files_[0].size()) >
@@ -1080,7 +1113,7 @@ void VersionStorageInfo::EstimateCompactionBytesNeeded(
 
   // Level 1 and up.
   for (int level = base_level(); level <= MaxInputLevel(); level++) {
-    size_t level_size = 0;
+    uint64_t level_size = 0;
     for (auto* f : files_[level]) {
       level_size += f->fd.GetFileSize();
     }
@@ -1091,7 +1124,7 @@ void VersionStorageInfo::EstimateCompactionBytesNeeded(
     // Add size added by previous compaction
     level_size += bytes_compact_to_next_level;
     bytes_compact_to_next_level = 0;
-    size_t level_target = MaxBytesForLevel(level);
+    uint64_t level_target = MaxBytesForLevel(level);
     if (level_size > level_target) {
       bytes_compact_to_next_level = level_size - level_target;
       // Simplify to assume the actual compaction fan-out ratio is always
@@ -1229,13 +1262,27 @@ bool CompareCompensatedSizeDescending(const Fsize& first, const Fsize& second) {
 }
 } // anonymous namespace
 
-void VersionStorageInfo::AddFile(int level, FileMetaData* f) {
+void VersionStorageInfo::AddFile(int level, FileMetaData* f, Logger* info_log) {
   auto* level_files = &files_[level];
   // Must not overlap
-  assert(level <= 0 || level_files->empty() ||
-         internal_comparator_->Compare(
-             (*level_files)[level_files->size() - 1]->largest, f->smallest) <
-             0);
+#ifndef NDEBUG
+  if (level > 0 && !level_files->empty() &&
+      internal_comparator_->Compare(
+          (*level_files)[level_files->size() - 1]->largest, f->smallest) >= 0) {
+    auto* f2 = (*level_files)[level_files->size() - 1];
+    if (info_log != nullptr) {
+      Error(info_log, "Adding new file %" PRIu64
+                      " range (%s, %s) to level %d but overlapping "
+                      "with existing file %" PRIu64 " %s %s",
+            f->fd.GetNumber(), f->smallest.DebugString(true).c_str(),
+            f->largest.DebugString(true).c_str(), level, f2->fd.GetNumber(),
+            f2->smallest.DebugString(true).c_str(),
+            f2->largest.DebugString(true).c_str());
+      LogFlush(info_log);
+    }
+    assert(false);
+  }
+#endif
   f->refs++;
   level_files->push_back(f);
 }
@@ -1405,7 +1452,8 @@ bool VersionStorageInfo::OverlapInLevel(int level,
 // The file_index returns a pointer to any file in an overlapping range.
 void VersionStorageInfo::GetOverlappingInputs(
     int level, const InternalKey* begin, const InternalKey* end,
-    std::vector<FileMetaData*>* inputs, int hint_index, int* file_index) {
+    std::vector<FileMetaData*>* inputs, int hint_index, int* file_index,
+    bool expand_range) const {
   if (level >= num_non_empty_levels_) {
     // this level is empty, no overlapping inputs
     return;
@@ -1438,7 +1486,7 @@ void VersionStorageInfo::GetOverlappingInputs(
       // "f" is completely after specified range; skip it
     } else {
       inputs->push_back(files_[level][i-1]);
-      if (level == 0) {
+      if (level == 0 && expand_range) {
         // Level-0 files may overlap each other.  So check if the newly
         // added file has expanded the range.  If so, restart search.
         if (begin != nullptr && user_cmp->Compare(file_start, user_begin) < 0) {
@@ -1464,7 +1512,7 @@ void VersionStorageInfo::GetOverlappingInputs(
 // forwards to find all overlapping files.
 void VersionStorageInfo::GetOverlappingInputsBinarySearch(
     int level, const Slice& user_begin, const Slice& user_end,
-    std::vector<FileMetaData*>* inputs, int hint_index, int* file_index) {
+    std::vector<FileMetaData*>* inputs, int hint_index, int* file_index) const {
   assert(level > 0);
   int min = 0;
   int mid = 0;
@@ -1512,8 +1560,7 @@ void VersionStorageInfo::GetOverlappingInputsBinarySearch(
 // Use FileLevel in searching, make it faster
 void VersionStorageInfo::ExtendOverlappingInputs(
     int level, const Slice& user_begin, const Slice& user_end,
-    std::vector<FileMetaData*>* inputs, unsigned int midIndex) {
-
+    std::vector<FileMetaData*>* inputs, unsigned int midIndex) const {
   const Comparator* user_cmp = user_comparator_;
   const FdWithKeyRange* files = level_files_brief_[level].files;
 #ifndef NDEBUG
@@ -2070,8 +2117,8 @@ Status VersionSet::LogAndApply(ColumnFamilyData* column_family_data,
           "Creating manifest %" PRIu64 "\n", pending_manifest_file_number_);
       unique_ptr<WritableFile> descriptor_file;
       EnvOptions opt_env_opts = env_->OptimizeForManifestWrite(env_options_);
-      s = env_->NewWritableFile(
-          DescriptorFileName(dbname_, pending_manifest_file_number_),
+      s = NewWritableFile(
+          env_, DescriptorFileName(dbname_, pending_manifest_file_number_),
           &descriptor_file, opt_env_opts);
       if (s.ok()) {
         descriptor_file->SetPreallocationBlockSize(
@@ -2079,7 +2126,7 @@ Status VersionSet::LogAndApply(ColumnFamilyData* column_family_data,
 
         unique_ptr<WritableFileWriter> file_writer(
             new WritableFileWriter(std::move(descriptor_file), opt_env_opts));
-        descriptor_log_.reset(new log::Writer(std::move(file_writer)));
+        descriptor_log_.reset(new log::Writer(std::move(file_writer), 0, false));
         s = WriteSnapshot(descriptor_log_.get());
       }
     }
@@ -2098,6 +2145,8 @@ Status VersionSet::LogAndApply(ColumnFamilyData* column_family_data,
               "Unable to Encode VersionEdit:" + e->DebugString(true));
           break;
         }
+        TEST_KILL_RANDOM("VersionSet::LogAndApply:BeforeAddRecord",
+                         rocksdb_kill_odds * REDUCE_ODDS2);
         s = descriptor_log_->AddRecord(record);
         if (!s.ok()) {
           break;
@@ -2341,8 +2390,8 @@ Status VersionSet::Recover(
   {
     VersionSet::LogReporter reporter;
     reporter.status = &s;
-    log::Reader reader(std::move(manifest_file_reader), &reporter,
-                       true /*checksum*/, 0 /*initial_offset*/);
+    log::Reader reader(NULL, std::move(manifest_file_reader), &reporter,
+                       true /*checksum*/, 0 /*initial_offset*/, 0);
     Slice record;
     std::string scratch;
     while (reader.ReadRecord(&record, &scratch) && s.ok()) {
@@ -2594,8 +2643,8 @@ Status VersionSet::ListColumnFamilies(std::vector<std::string>* column_families,
   column_family_names.insert({0, kDefaultColumnFamilyName});
   VersionSet::LogReporter reporter;
   reporter.status = &s;
-  log::Reader reader(std::move(file_reader), &reporter, true /*checksum*/,
-                     0 /*initial_offset*/);
+  log::Reader reader(NULL, std::move(file_reader), &reporter, true /*checksum*/,
+                     0 /*initial_offset*/, 0);
   Slice record;
   std::string scratch;
   while (reader.ReadRecord(&record, &scratch) && s.ok()) {
@@ -2752,8 +2801,8 @@ Status VersionSet::DumpManifest(Options& options, std::string& dscname,
   {
     VersionSet::LogReporter reporter;
     reporter.status = &s;
-    log::Reader reader(std::move(file_reader), &reporter, true /*checksum*/,
-                       0 /*initial_offset*/);
+    log::Reader reader(NULL, std::move(file_reader), &reporter,
+                       true /*checksum*/, 0 /*initial_offset*/, 0);
     Slice record;
     std::string scratch;
     while (reader.ReadRecord(&record, &scratch) && s.ok()) {
@@ -3005,7 +3054,8 @@ bool VersionSet::ManifestContains(uint64_t manifest_file_num,
     }
     file_reader.reset(new SequentialFileReader(std::move(file)));
   }
-  log::Reader reader(std::move(file_reader), nullptr, true /*checksum*/, 0);
+  log::Reader reader(NULL, std::move(file_reader), nullptr,
+                     true /*checksum*/, 0, 0);
   Slice r;
   std::string scratch;
   bool result = false;
@@ -3119,7 +3169,7 @@ uint64_t VersionSet::ApproximateSize(Version* v, const FdWithKeyRange& f,
     // "key" falls in the range for this table.  Add the
     // approximate offset of "key" within the table.
     TableReader* table_reader_ptr;
-    Iterator* iter = v->cfd_->table_cache()->NewIterator(
+    InternalIterator* iter = v->cfd_->table_cache()->NewIterator(
         ReadOptions(), env_options_, v->cfd_->internal_comparator(), f.fd,
         &table_reader_ptr);
     if (table_reader_ptr != nullptr) {
@@ -3166,7 +3216,7 @@ void VersionSet::AddLiveFiles(std::vector<FileDescriptor>* live_list) {
   }
 }
 
-Iterator* VersionSet::MakeInputIterator(Compaction* c) {
+InternalIterator* VersionSet::MakeInputIterator(Compaction* c) {
   auto cfd = c->column_family_data();
   ReadOptions read_options;
   read_options.verify_checksums =
@@ -3182,7 +3232,7 @@ Iterator* VersionSet::MakeInputIterator(Compaction* c) {
   const size_t space = (c->level() == 0 ? c->input_levels(0)->num_files +
                                               c->num_input_levels() - 1
                                         : c->num_input_levels());
-  Iterator** list = new Iterator* [space];
+  InternalIterator** list = new InternalIterator* [space];
   size_t num = 0;
   for (size_t which = 0; which < c->num_input_levels(); which++) {
     if (c->input_levels(which)->num_files != 0) {
@@ -3209,7 +3259,7 @@ Iterator* VersionSet::MakeInputIterator(Compaction* c) {
     }
   }
   assert(num <= space);
-  Iterator* result =
+  InternalIterator* result =
       NewMergingIterator(&c->column_family_data()->internal_comparator(), list,
                          static_cast<int>(num));
   delete[] list;

db/version_set.h

@@ -48,7 +48,7 @@ class Writer;
 }
 
 class Compaction;
-class Iterator;
+class InternalIterator;
 class LogBuffer;
 class LookupKey;
 class MemTable;
@@ -97,7 +97,7 @@ class VersionStorageInfo {
 
   void Reserve(int level, size_t size) { files_[level].reserve(size); }
 
-  void AddFile(int level, FileMetaData* f);
+  void AddFile(int level, FileMetaData* f, Logger* info_log = nullptr);
 
   void SetFinalized();
 
@@ -159,23 +159,26 @@ class VersionStorageInfo {
       int level, const InternalKey* begin,  // nullptr means before all keys
       const InternalKey* end,               // nullptr means after all keys
       std::vector<FileMetaData*>* inputs,
-      int hint_index = -1,         // index of overlap file
-      int* file_index = nullptr);  // return index of overlap file
+      int hint_index = -1,        // index of overlap file
+      int* file_index = nullptr,  // return index of overlap file
+      bool expand_range = true)   // if set, returns files which overlap the
+      const;                      // range and overlap each other. If false,
+                                  // then just files intersecting the range
 
   void GetOverlappingInputsBinarySearch(
       int level,
       const Slice& begin,  // nullptr means before all keys
       const Slice& end,    // nullptr means after all keys
       std::vector<FileMetaData*>* inputs,
-      int hint_index,    // index of overlap file
-      int* file_index);  // return index of overlap file
+      int hint_index,          // index of overlap file
+      int* file_index) const;  // return index of overlap file
 
   void ExtendOverlappingInputs(
       int level,
       const Slice& begin,  // nullptr means before all keys
       const Slice& end,    // nullptr means after all keys
       std::vector<FileMetaData*>* inputs,
-      unsigned int index);  // start extending from this index
+      unsigned int index) const;  // start extending from this index
 
   // Returns true iff some file in the specified level overlaps
   // some part of [*smallest_user_key,*largest_user_key].
@@ -456,15 +459,16 @@ class Version {
   // file-name conversion.
   Status GetTableProperties(std::shared_ptr<const TableProperties>* tp,
                             const FileMetaData* file_meta,
-                            const std::string* fname = nullptr);
+                            const std::string* fname = nullptr) const;
 
   // REQUIRES: lock is held
   // On success, *props will be populated with all SSTables' table properties.
   // The keys of `props` are the sst file name, the values of `props` are the
   // tables' propertis, represented as shared_ptr.
   Status GetPropertiesOfAllTables(TablePropertiesCollection* props);
-
   Status GetPropertiesOfAllTables(TablePropertiesCollection* props, int level);
+  Status GetPropertiesOfTablesInRange(const Range* range, std::size_t n,
+                                      TablePropertiesCollection* props) const;
 
   // REQUIRES: lock is held
   // On success, "tp" will contains the aggregated table property amoug
@@ -502,7 +506,8 @@ class Version {
     return storage_info_.user_comparator_;
   }
 
-  bool PrefixMayMatch(const ReadOptions& read_options, Iterator* level_iter,
+  bool PrefixMayMatch(const ReadOptions& read_options,
+                      InternalIterator* level_iter,
                       const Slice& internal_prefix) const;
 
   // The helper function of UpdateAccumulatedStats, which may fill the missing
@@ -643,7 +648,7 @@ class VersionSet {
 
   // Create an iterator that reads over the compaction inputs for "*c".
   // The caller should delete the iterator when no longer needed.
-  Iterator* MakeInputIterator(Compaction* c);
+  InternalIterator* MakeInputIterator(Compaction* c);
 
   // Add all files listed in any live version to *live.
   void AddLiveFiles(std::vector<FileDescriptor>* live_list);

db/wal_manager.cc

@@ -448,8 +448,8 @@ Status WalManager::ReadFirstLine(const std::string& fname,
   reporter.fname = fname.c_str();
   reporter.status = &status;
   reporter.ignore_error = !db_options_.paranoid_checks;
-  log::Reader reader(std::move(file_reader), &reporter, true /*checksum*/,
-                     0 /*initial_offset*/);
+  log::Reader reader(db_options_.info_log, std::move(file_reader), &reporter,
+                     true /*checksum*/, 0 /*initial_offset*/, *sequence);
   std::string scratch;
   Slice record;
 

db/wal_manager_test.cc

@@ -77,7 +77,7 @@ class WalManagerTest : public testing::Test {
     ASSERT_OK(env_->NewWritableFile(fname, &file, env_options_));
     unique_ptr<WritableFileWriter> file_writer(
         new WritableFileWriter(std::move(file), env_options_));
-    current_log_writer_.reset(new log::Writer(std::move(file_writer)));
+    current_log_writer_.reset(new log::Writer(std::move(file_writer), 0, false));
   }
 
   void CreateArchiveLogs(int num_logs, int entries_per_log) {
@@ -127,7 +127,8 @@ TEST_F(WalManagerTest, ReadFirstRecordCache) {
 
   unique_ptr<WritableFileWriter> file_writer(
       new WritableFileWriter(std::move(file), EnvOptions()));
-  log::Writer writer(std::move(file_writer));
+  log::Writer writer(std::move(file_writer), 1,
+		     db_options_.recycle_log_file_num > 0);
   WriteBatch batch;
   batch.Put("foo", "bar");
   WriteBatchInternal::SetSequence(&batch, 10);

db/write_batch_test.cc

@@ -17,10 +17,10 @@
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/utilities/write_batch_with_index.h"
+#include "table/scoped_arena_iterator.h"
 #include "util/logging.h"
 #include "util/string_util.h"
 #include "util/testharness.h"
-#include "util/scoped_arena_iterator.h"
 
 namespace rocksdb {
 

db/write_thread.cc

@@ -120,12 +120,13 @@ size_t WriteThread::EnterAsBatchGroupLeader(
       break;
     }
 
-    size += WriteBatchInternal::ByteSize(w->batch);
-    if (size > max_size) {
+    auto batch_size = WriteBatchInternal::ByteSize(w->batch);
+    if (size + batch_size > max_size) {
       // Do not make batch too big
       break;
     }
 
+    size += batch_size;
     write_batch_group->push_back(w->batch);
     w->in_batch_group = true;
     *last_writer = w;

examples/rocksdb_option_file_example.ini

@@ -29,25 +29,167 @@
 #
 # Below is an example of a RocksDB options file:
 [Version]
-  # The Version section stores the version information about rocksdb
-  # and option file.  This is used for handling potential format
-  # change in the future.
-  rocksdb_version=4.0.0  # We support "#" style comment.
-  options_file_version=1.0
+  rocksdb_version=4.0.0
+  options_file_version=1.1
+
 [DBOptions]
-  # Followed by the Version section is the DBOptions section.
-  # The value of an options can be assigned using a statement.
-  # Note that for those options that is not set in the options file,
-  # we will use the default value.
-  max_open_files=12345
-  max_background_flushes=301
+  stats_dump_period_sec=600
+  max_manifest_file_size=18446744073709551615
+  bytes_per_sync=0
+  delayed_write_rate=1048576
+  WAL_ttl_seconds=0
+  WAL_size_limit_MB=0
+  max_subcompactions=1
+  wal_dir=
+  wal_bytes_per_sync=0
+  db_write_buffer_size=0
+  max_total_wal_size=0
+  skip_stats_update_on_db_open=false
+  max_open_files=5000
+  max_file_opening_threads=1
+  use_fsync=false
+  max_background_compactions=1
+  manifest_preallocation_size=4194304
+  max_background_flushes=1
+  is_fd_close_on_exec=true
+  create_if_missing=false
+  use_adaptive_mutex=false
+  enable_thread_tracking=false
+  disableDataSync=false
+  max_log_file_size=0
+  advise_random_on_open=true
+  create_missing_column_families=false
+  keep_log_file_num=1000
+  table_cache_numshardbits=4
+  error_if_exists=false
+  skip_log_error_on_recovery=false
+  allow_os_buffer=true
+  allow_mmap_reads=false
+  paranoid_checks=true
+  delete_obsolete_files_period_micros=21600000000
+  disable_data_sync=false
+  log_file_time_to_roll=0
+  compaction_readahead_size=0
+  db_log_dir=
+  new_table_reader_for_compaction_inputs=false
+  allow_mmap_writes=false
+
 [CFOptions "default"]
-  # ColumnFamilyOptions section must follow the format of
-  # [CFOptions "cf name"].  If a rocksdb instance
-  # has multiple column families, then its CFOptions must be
-  # specified in the same order as column family creation order.
-[CFOptions "the second column family"]
-  # Each column family must have one section in the RocksDB option
-  # file even all the options of this column family are set to
-  # default value.
-[CFOptions "the third column family"]
+  compaction_style=kCompactionStyleLevel
+  compaction_filter=nullptr
+  num_levels=7
+  table_factory=BlockBasedTable
+  comparator=leveldb.BytewiseComparator
+  max_sequential_skip_in_iterations=8
+  soft_rate_limit=0.000000
+  max_bytes_for_level_base=536870912
+  memtable_prefix_bloom_probes=6
+  memtable_prefix_bloom_bits=0
+  memtable_prefix_bloom_huge_page_tlb_size=0
+  max_successive_merges=0
+  arena_block_size=0
+  min_write_buffer_number_to_merge=2
+  target_file_size_multiplier=1
+  source_compaction_factor=1
+  max_bytes_for_level_multiplier=10
+  compaction_filter_factory=nullptr
+  max_write_buffer_number=6
+  level0_stop_writes_trigger=24
+  compression=kSnappyCompression
+  level0_file_num_compaction_trigger=2
+  purge_redundant_kvs_while_flush=true
+  max_write_buffer_number_to_maintain=0
+  memtable_factory=SkipListFactory
+  max_grandparent_overlap_factor=10
+  expanded_compaction_factor=25
+  hard_pending_compaction_bytes_limit=0
+  inplace_update_num_locks=10000
+  level_compaction_dynamic_level_bytes=false
+  level0_slowdown_writes_trigger=20
+  filter_deletes=false
+  verify_checksums_in_compaction=true
+  min_partial_merge_operands=2
+  paranoid_file_checks=false
+  target_file_size_base=67108864
+  optimize_filters_for_hits=false
+  merge_operator=nullptr
+  compression_per_level=kNoCompression:kNoCompression:kSnappyCompression:kSnappyCompression:kSnappyCompression:kSnappyCompression:kSnappyCompression
+  compaction_measure_io_stats=false
+  prefix_extractor=nullptr
+  bloom_locality=0
+  write_buffer_size=134217728
+  disable_auto_compactions=false
+  inplace_update_support=false
+[TableOptions/BlockBasedTable "default"]
+  format_version=0
+  whole_key_filtering=true
+  block_size_deviation=10
+  block_size=4096
+  block_restart_interval=16
+  filter_policy=nullptr
+  no_block_cache=false
+  checksum=kCRC32c
+  cache_index_and_filter_blocks=false
+  index_type=kBinarySearch
+  hash_index_allow_collision=true
+  flush_block_policy_factory=FlushBlockBySizePolicyFactory
+
+[CFOptions "universal"]
+  compaction_style=kCompactionStyleUniversal
+  compaction_filter=nullptr
+  num_levels=7
+  table_factory=BlockBasedTable
+  comparator=leveldb.BytewiseComparator
+  max_sequential_skip_in_iterations=8
+  soft_rate_limit=0.000000
+  max_bytes_for_level_base=10485760
+  memtable_prefix_bloom_probes=6
+  memtable_prefix_bloom_bits=0
+  memtable_prefix_bloom_huge_page_tlb_size=0
+  max_successive_merges=0
+  arena_block_size=0
+  min_write_buffer_number_to_merge=2
+  target_file_size_multiplier=1
+  source_compaction_factor=1
+  max_bytes_for_level_multiplier=10
+  compaction_filter_factory=nullptr
+  max_write_buffer_number=6
+  level0_stop_writes_trigger=24
+  compression=kSnappyCompression
+  level0_file_num_compaction_trigger=4
+  purge_redundant_kvs_while_flush=true
+  max_write_buffer_number_to_maintain=0
+  memtable_factory=SkipListFactory
+  max_grandparent_overlap_factor=10
+  expanded_compaction_factor=25
+  hard_pending_compaction_bytes_limit=0
+  inplace_update_num_locks=10000
+  level_compaction_dynamic_level_bytes=false
+  level0_slowdown_writes_trigger=20
+  filter_deletes=false
+  verify_checksums_in_compaction=true
+  min_partial_merge_operands=2
+  paranoid_file_checks=false
+  target_file_size_base=2097152
+  optimize_filters_for_hits=false
+  merge_operator=nullptr
+  compression_per_level=
+  compaction_measure_io_stats=false
+  prefix_extractor=nullptr
+  bloom_locality=0
+  write_buffer_size=134217728
+  disable_auto_compactions=false
+  inplace_update_support=false
+[TableOptions/BlockBasedTable "universal"]
+  format_version=0
+  whole_key_filtering=true
+  block_size_deviation=10
+  block_size=4096
+  block_restart_interval=16
+  filter_policy=nullptr
+  no_block_cache=false
+  checksum=kCRC32c
+  cache_index_and_filter_blocks=false
+  index_type=kBinarySearch
+  hash_index_allow_collision=true
+  flush_block_policy_factory=FlushBlockBySizePolicyFactory

include/rocksdb/c.h

@@ -450,6 +450,9 @@ rocksdb_block_based_options_set_hash_index_allow_collision(
 extern ROCKSDB_LIBRARY_API void
 rocksdb_block_based_options_set_cache_index_and_filter_blocks(
     rocksdb_block_based_table_options_t*, unsigned char);
+extern ROCKSDB_LIBRARY_API void
+rocksdb_block_based_options_set_skip_table_builder_flush(
+    rocksdb_block_based_table_options_t* options, unsigned char);
 extern ROCKSDB_LIBRARY_API void rocksdb_options_set_block_based_table_factory(
     rocksdb_options_t* opt, rocksdb_block_based_table_options_t* table_options);
 
@@ -572,6 +575,8 @@ extern ROCKSDB_LIBRARY_API void rocksdb_options_set_log_file_time_to_roll(
     rocksdb_options_t*, size_t);
 extern ROCKSDB_LIBRARY_API void rocksdb_options_set_keep_log_file_num(
     rocksdb_options_t*, size_t);
+extern ROCKSDB_LIBRARY_API void rocksdb_options_set_recycle_log_file_num(
+    rocksdb_options_t*, size_t);
 extern ROCKSDB_LIBRARY_API void rocksdb_options_set_soft_rate_limit(
     rocksdb_options_t*, double);
 extern ROCKSDB_LIBRARY_API void rocksdb_options_set_hard_rate_limit(

include/rocksdb/compaction_filter.h

@@ -39,6 +39,8 @@ class CompactionFilter {
     // Is this compaction requested by the client (true),
     // or is it occurring as an automatic compaction process
     bool is_manual_compaction;
+    // Which column family this compaction is for.
+    uint32_t column_family_id;
   };
 
   virtual ~CompactionFilter() {}
@@ -114,20 +116,6 @@ class CompactionFilterFactory {
   virtual const char* Name() const = 0;
 };
 
-// Default implementation of CompactionFilterFactory which does not
-// return any filter
-class DefaultCompactionFilterFactory : public CompactionFilterFactory {
- public:
-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
-      const CompactionFilter::Context& context) override {
-    return std::unique_ptr<CompactionFilter>(nullptr);
-  }
-
-  virtual const char* Name() const override {
-    return "DefaultCompactionFilterFactory";
-  }
-};
-
 }  // namespace rocksdb
 
 #endif  // STORAGE_ROCKSDB_INCLUDE_COMPACTION_FILTER_H_

include/rocksdb/convenience.h

@@ -40,7 +40,14 @@ Status GetDBOptionsFromMap(
 Status GetBlockBasedTableOptionsFromMap(
     const BlockBasedTableOptions& table_options,
     const std::unordered_map<std::string, std::string>& opts_map,
-    BlockBasedTableOptions* new_table_options);
+    BlockBasedTableOptions* new_table_options,
+    bool input_strings_escaped = false);
+
+Status GetPlainTableOptionsFromMap(
+    const PlainTableOptions& table_options,
+    const std::unordered_map<std::string, std::string>& opts_map,
+    PlainTableOptions* new_table_options,
+    bool input_strings_escaped = false);
 
 // Take a string representation of option names and  values, apply them into the
 // base_options, and return the new options as a result. The string has the
@@ -73,6 +80,11 @@ Status GetBlockBasedTableOptionsFromString(
     const std::string& opts_str,
     BlockBasedTableOptions* new_table_options);
 
+Status GetPlainTableOptionsFromString(
+    const PlainTableOptions& table_options,
+    const std::string& opts_str,
+    PlainTableOptions* new_table_options);
+
 Status GetOptionsFromString(const Options& base_options,
                             const std::string& opts_str, Options* new_options);
 

include/rocksdb/db.h

@@ -15,15 +15,16 @@
 #include <vector>
 #include <string>
 #include <unordered_map>
-#include "rocksdb/metadata.h"
-#include "rocksdb/version.h"
+#include "rocksdb/immutable_options.h"
 #include "rocksdb/iterator.h"
-#include "rocksdb/options.h"
-#include "rocksdb/types.h"
-#include "rocksdb/transaction_log.h"
 #include "rocksdb/listener.h"
+#include "rocksdb/metadata.h"
+#include "rocksdb/options.h"
 #include "rocksdb/snapshot.h"
 #include "rocksdb/thread_status.h"
+#include "rocksdb/transaction_log.h"
+#include "rocksdb/types.h"
+#include "rocksdb/version.h"
 
 #ifdef _WIN32
 // Windows API macro interference
@@ -188,6 +189,18 @@ class DB {
   // Remove the database entry for "key". Requires that the key exists
   // and was not overwritten. Returns OK on success, and a non-OK status
   // on error.  It is not an error if "key" did not exist in the database.
+  //
+  // If a key is overwritten (by calling Put() multiple times), then the result
+  // of calling SingleDelete() on this key is undefined.  SingleDelete() only
+  // behaves correctly if there has been only one Put() for this key since the
+  // previous call to SingleDelete() for this key.
+  //
+  // This feature is currently an experimental performance optimization
+  // for a very specific workload.  It is up to the caller to ensure that
+  // SingleDelete is only used for a key that is not deleted using Delete() or
+  // written using Merge().  Mixing SingleDelete operations with Deletes and
+  // Merges can result in undefined behavior.
+  //
   // Note: consider setting options.sync = true.
   virtual Status SingleDelete(const WriteOptions& options,
                               ColumnFamilyHandle* column_family,
@@ -258,9 +271,10 @@ class DB {
   // This check is potentially lighter-weight than invoking DB::Get(). One way
   // to make this lighter weight is to avoid doing any IOs.
   // Default implementation here returns true and sets 'value_found' to false
-  virtual bool KeyMayExist(const ReadOptions& options,
-                           ColumnFamilyHandle* column_family, const Slice& key,
-                           std::string* value, bool* value_found = nullptr) {
+  virtual bool KeyMayExist(const ReadOptions& /*options*/,
+                           ColumnFamilyHandle* /*column_family*/,
+                           const Slice& /*key*/, std::string* /*value*/,
+                           bool* value_found = nullptr) {
     if (value_found != nullptr) {
       *value_found = false;
     }
@@ -324,34 +338,37 @@ class DB {
   //  "rocksdb.compaction-pending" - 1 if at least one compaction is pending
   //  "rocksdb.background-errors" - accumulated number of background errors
   //  "rocksdb.cur-size-active-mem-table"
-//  "rocksdb.size-all-mem-tables"
-//  "rocksdb.num-entries-active-mem-table"
-//  "rocksdb.num-entries-imm-mem-tables"
-//  "rocksdb.num-deletes-active-mem-table"
-//  "rocksdb.num-deletes-imm-mem-tables"
-//  "rocksdb.estimate-num-keys" - estimated keys in the column family
-//  "rocksdb.estimate-table-readers-mem" - estimated memory used for reding
-//      SST tables, that is not counted as a part of block cache.
-//  "rocksdb.is-file-deletions-enabled"
-//  "rocksdb.num-snapshots"
-//  "rocksdb.oldest-snapshot-time"
-//  "rocksdb.num-live-versions" - `version` is an internal data structure.
-//      See version_set.h for details. More live versions often mean more SST
-//      files are held from being deleted, by iterators or unfinished
-//      compactions.
-//  "rocksdb.estimate-live-data-size"
-//  "rocksdb.total-sst-files-size" - total size of all used sst files, this may
-//      slow down online queries if there are too many files.
-//  "rocksdb.base-level"
-//  "rocksdb.estimate-pending-compaction-bytes" - estimated total number of
-//      bytes compaction needs to rewrite the data to get all levels down
-//      to under target size. Not valid for other compactions than level-based.
-//  "rocksdb.aggregated-table-properties" - returns a string representation of
-//      the aggregated table properties of the target column family.
-//  "rocksdb.aggregated-table-properties-at-level<N>", same as the previous
-//      one but only returns the aggregated table properties of the specified
-//      level "N" at the target column family.
-//  replaced by the target level.
+  //  "rocksdb.size-all-mem-tables"
+  //  "rocksdb.num-entries-active-mem-table"
+  //  "rocksdb.num-entries-imm-mem-tables"
+  //  "rocksdb.num-deletes-active-mem-table"
+  //  "rocksdb.num-deletes-imm-mem-tables"
+  //  "rocksdb.estimate-num-keys" - estimated keys in the column family
+  //  "rocksdb.estimate-table-readers-mem" - estimated memory used for reding
+  //      SST tables, that is not counted as a part of block cache.
+  //  "rocksdb.is-file-deletions-enabled"
+  //  "rocksdb.num-snapshots"
+  //  "rocksdb.oldest-snapshot-time"
+  //  "rocksdb.num-live-versions" - `version` is an internal data structure.
+  //      See version_set.h for details. More live versions often mean more SST
+  //      files are held from being deleted, by iterators or unfinished
+  //      compactions.
+  //  "rocksdb.estimate-live-data-size"
+  //  "rocksdb.total-sst-files-size" - total size of all used sst files, this
+  //      may slow down online queries if there are too many files.
+  //  "rocksdb.base-level"
+  //  "rocksdb.estimate-pending-compaction-bytes" - estimated total number of
+  //      bytes compaction needs to rewrite the data to get all levels down
+  //      to under target size. Not valid for other compactions than
+  //      level-based.
+  //  "rocksdb.aggregated-table-properties" - returns a string representation
+  //      of the aggregated table properties of the target column family.
+  //  "rocksdb.aggregated-table-properties-at-level<N>", same as the previous
+  //      one but only returns the aggregated table properties of the specified
+  //      level "N" at the target column family.
+  //  "rocksdb.num-running-compactions" - the number of currently running
+  //      compacitons.
+  //  "rocksdb.num-running-flushes" - the number of currently running flushes.
 #ifndef ROCKSDB_LITE
   struct Properties {
     static const std::string kNumFilesAtLevelPrefix;
@@ -361,7 +378,9 @@ class DB {
     static const std::string kDBStats;
     static const std::string kNumImmutableMemTable;
     static const std::string kMemTableFlushPending;
+    static const std::string kNumRunningFlushes;
     static const std::string kCompactionPending;
+    static const std::string kNumRunningCompactions;
     static const std::string kBackgroundErrors;
     static const std::string kCurSizeActiveMemTable;
     static const std::string kCurSizeAllMemTables;
@@ -414,12 +433,19 @@ class DB {
   //  "rocksdb.total-sst-files-size"
   //  "rocksdb.base-level"
   //  "rocksdb.estimate-pending-compaction-bytes"
+  //  "rocksdb.num-running-compactions"
+  //  "rocksdb.num-running-flushes"
   virtual bool GetIntProperty(ColumnFamilyHandle* column_family,
                               const Slice& property, uint64_t* value) = 0;
   virtual bool GetIntProperty(const Slice& property, uint64_t* value) {
     return GetIntProperty(DefaultColumnFamily(), property, value);
   }
 
+  // Same as GetIntProperty(), but this one returns the aggregated int
+  // property from all column families.
+  virtual bool GetAggregatedIntProperty(const Slice& property,
+                                        uint64_t* value) = 0;
+
   // For each i in [0,n-1], store in "sizes[i]", the approximate
   // file system space used by keys in "[range[i].start .. range[i].limit)".
   //
@@ -495,8 +521,9 @@ class DB {
     return CompactRange(options, DefaultColumnFamily(), begin, end);
   }
 
-  virtual Status SetOptions(ColumnFamilyHandle* column_family,
-      const std::unordered_map<std::string, std::string>& new_options) {
+  virtual Status SetOptions(
+      ColumnFamilyHandle* /*column_family*/,
+      const std::unordered_map<std::string, std::string>& /*new_options*/) {
     return Status::NotSupported("Not implemented");
   }
   virtual Status SetOptions(
@@ -646,7 +673,8 @@ class DB {
 
   // Returns a list of all table files with their level, start key
   // and end key
-  virtual void GetLiveFilesMetaData(std::vector<LiveFileMetaData>* metadata) {}
+  virtual void GetLiveFilesMetaData(
+      std::vector<LiveFileMetaData>* /*metadata*/) {}
 
   // Obtains the meta data of the specified column family of the DB.
   // Status::NotFound() will be returned if the current DB does not have
@@ -654,9 +682,8 @@ class DB {
   //
   // If cf_name is not specified, then the metadata of the default
   // column family will be returned.
-  virtual void GetColumnFamilyMetaData(
-      ColumnFamilyHandle* column_family,
-      ColumnFamilyMetaData* metadata) {}
+  virtual void GetColumnFamilyMetaData(ColumnFamilyHandle* /*column_family*/,
+                                       ColumnFamilyMetaData* /*metadata*/) {}
 
   // Get the metadata of the default column family.
   void GetColumnFamilyMetaData(
@@ -670,13 +697,11 @@ class DB {
   // move the file instead of copying it.
   //
   // Current Requirements:
-  // (1) Memtable is empty.
-  // (2) All existing files (if any) have sequence number = 0.
-  // (3) Key range in loaded table file don't overlap with existing
-  //     files key ranges.
-  // (4) No other writes happen during AddFile call, otherwise
+  // (1) Key range in loaded table file don't overlap with
+  //     existing keys or tombstones in DB.
+  // (2) No other writes happen during AddFile call, otherwise
   //     DB may get corrupted.
-  // (5) Database have at least 2 levels.
+  // (3) No snapshots are held.
   virtual Status AddFile(ColumnFamilyHandle* column_family,
                          const std::string& file_path,
                          bool move_file = false) = 0;
@@ -709,6 +734,9 @@ class DB {
   virtual Status GetPropertiesOfAllTables(TablePropertiesCollection* props) {
     return GetPropertiesOfAllTables(DefaultColumnFamily(), props);
   }
+  virtual Status GetPropertiesOfTablesInRange(
+      ColumnFamilyHandle* column_family, const Range* range, std::size_t n,
+      TablePropertiesCollection* props) = 0;
 #endif  // ROCKSDB_LITE
 
   // Needed for StackableDB

include/rocksdb/env.h

@@ -88,6 +88,15 @@ struct EnvOptions {
   // WAL writes
   bool fallocate_with_keep_size = true;
 
+  // See DBOPtions doc
+  size_t compaction_readahead_size;
+
+  // See DBOPtions doc
+  size_t random_access_max_buffer_size;
+
+  // See DBOptions doc
+  size_t writable_file_max_buffer_size = 1024 * 1024;
+
   // If not nullptr, write rate limiting is enabled for flush and compaction
   RateLimiter* rate_limiter = nullptr;
 };
@@ -139,6 +148,12 @@ class Env {
                                  unique_ptr<WritableFile>* result,
                                  const EnvOptions& options) = 0;
 
+  // Reuse an existing file by renaming it and opening it as writable.
+  virtual Status ReuseWritableFile(const std::string& fname,
+                                   const std::string& old_fname,
+                                   unique_ptr<WritableFile>* result,
+                                   const EnvOptions& options);
+
   // Create an object that represents a directory. Will fail if directory
   // doesn't exist. If the directory exists, it will open the directory
   // and create a new Directory object.
@@ -402,6 +417,10 @@ class RandomAccessFile {
     return false;
   }
 
+  // For cases when read-ahead is implemented in the platform dependent
+  // layer
+  virtual void EnableReadAhead() {}
+
   // Tries to get an unique ID for this file that will be the same each time
   // the file is opened (and will stay the same while the file is open).
   // Furthermore, it tries to make this ID at most "max_size" bytes. If such an

include/rocksdb/iterator.h

@@ -24,10 +24,32 @@
 
 namespace rocksdb {
 
-class Iterator {
+class Cleanable {
  public:
-  Iterator();
-  virtual ~Iterator();
+  Cleanable();
+  ~Cleanable();
+  // Clients are allowed to register function/arg1/arg2 triples that
+  // will be invoked when this iterator is destroyed.
+  //
+  // Note that unlike all of the preceding methods, this method is
+  // not abstract and therefore clients should not override it.
+  typedef void (*CleanupFunction)(void* arg1, void* arg2);
+  void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
+
+ protected:
+  struct Cleanup {
+    CleanupFunction function;
+    void* arg1;
+    void* arg2;
+    Cleanup* next;
+  };
+  Cleanup cleanup_;
+};
+
+class Iterator : public Cleanable {
+ public:
+  Iterator() {}
+  virtual ~Iterator() {}
 
   // An iterator is either positioned at a key/value pair, or
   // not valid.  This method returns true iff the iterator is valid.
@@ -73,23 +95,7 @@ class Iterator {
   // satisfied without doing some IO, then this returns Status::Incomplete().
   virtual Status status() const = 0;
 
-  // Clients are allowed to register function/arg1/arg2 triples that
-  // will be invoked when this iterator is destroyed.
-  //
-  // Note that unlike all of the preceding methods, this method is
-  // not abstract and therefore clients should not override it.
-  typedef void (*CleanupFunction)(void* arg1, void* arg2);
-  void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
-
  private:
-  struct Cleanup {
-    CleanupFunction function;
-    void* arg1;
-    void* arg2;
-    Cleanup* next;
-  };
-  Cleanup cleanup_;
-
   // No copying allowed
   Iterator(const Iterator&);
   void operator=(const Iterator&);

include/rocksdb/listener.h

@@ -4,7 +4,9 @@
 
 #pragma once
 
+#include <memory>
 #include <string>
+#include <unordered_map>
 #include <vector>
 #include "rocksdb/compaction_job_stats.h"
 #include "rocksdb/status.h"
@@ -12,6 +14,9 @@
 
 namespace rocksdb {
 
+typedef std::unordered_map<std::string, std::shared_ptr<const TableProperties>>
+    TablePropertiesCollection;
+
 class DB;
 class Status;
 struct CompactionJobStats;
@@ -72,6 +77,8 @@ struct FlushJobInfo {
   SequenceNumber smallest_seqno;
   // The largest sequence number in the newly created file
   SequenceNumber largest_seqno;
+  // Table properties of the table being flushed
+  TableProperties table_properties;
 };
 
 struct CompactionJobInfo {
@@ -93,8 +100,13 @@ struct CompactionJobInfo {
   int output_level;
   // the names of the compaction input files.
   std::vector<std::string> input_files;
+
   // the names of the compaction output files.
   std::vector<std::string> output_files;
+  // Table properties for input and output tables.
+  // The map is keyed by values from input_files and output_files.
+  TablePropertiesCollection table_properties;
+
   // If non-null, this variable stores detailed information
   // about this compaction.
   CompactionJobStats stats;
@@ -138,8 +150,8 @@ class EventListener {
   // Note that the this function must be implemented in a way such that
   // it should not run for an extended period of time before the function
   // returns.  Otherwise, RocksDB may be blocked.
-  virtual void OnFlushCompleted(
-      DB* db, const FlushJobInfo& flush_job_info) {}
+  virtual void OnFlushCompleted(DB* /*db*/,
+                                const FlushJobInfo& /*flush_job_info*/) {}
 
   // A call-back function for RocksDB which will be called whenever
   // a SST file is deleted.  Different from OnCompactionCompleted and
@@ -152,8 +164,7 @@ class EventListener {
   // Note that if applications would like to use the passed reference
   // outside this function call, they should make copies from the
   // returned value.
-  virtual void OnTableFileDeleted(
-      const TableFileDeletionInfo& info) {}
+  virtual void OnTableFileDeleted(const TableFileDeletionInfo& /*info*/) {}
 
   // A call-back function for RocksDB which will be called whenever
   // a registered RocksDB compacts a file. The default implementation
@@ -168,7 +179,8 @@ class EventListener {
   // @param ci a reference to a CompactionJobInfo struct. 'ci' is released
   //  after this function is returned, and must be copied if it is needed
   //  outside of this function.
-  virtual void OnCompactionCompleted(DB *db, const CompactionJobInfo& ci) {}
+  virtual void OnCompactionCompleted(DB* /*db*/,
+                                     const CompactionJobInfo& /*ci*/) {}
 
   // A call-back function for RocksDB which will be called whenever
   // a SST file is created.  Different from OnCompactionCompleted and
@@ -181,8 +193,7 @@ class EventListener {
   // Note that if applications would like to use the passed reference
   // outside this function call, they should make copies from these
   // returned value.
-  virtual void OnTableFileCreated(
-      const TableFileCreationInfo& info) {}
+  virtual void OnTableFileCreated(const TableFileCreationInfo& /*info*/) {}
 
   virtual ~EventListener() {}
 };

include/rocksdb/options.h

@@ -46,6 +46,7 @@ class Slice;
 class SliceTransform;
 class Statistics;
 class InternalKeyComparator;
+class WalFilter;
 
 // DB contents are stored in a set of blocks, each of which holds a
 // sequence of key,value pairs.  Each block may be compressed before
@@ -957,6 +958,16 @@ struct DBOptions {
   // Default: 1000
   size_t keep_log_file_num;
 
+  // Recycle log files.
+  // If non-zero, we will reuse previously written log files for new
+  // logs, overwriting the old data.  The value indicates how many
+  // such files we will keep around at any point in time for later
+  // use.  This is more efficient because the blocks are already
+  // allocated and fdatasync does not need to update the inode after
+  // each write.
+  // Default: 0
+  size_t recycle_log_file_num;
+
   // manifest file is rolled over on reaching this limit.
   // The older manifest file be deleted.
   // The default value is MAX_INT so that roll-over does not take place.
@@ -1064,6 +1075,28 @@ struct DBOptions {
   // Default: 0
   size_t compaction_readahead_size;
 
+  // This is a maximum buffer size that is used by WinMmapReadableFile in
+  // unbuffered disk I/O mode. We need to maintain an aligned buffer for
+  // reads. We allow the buffer to grow until the specified value and then
+  // for bigger requests allocate one shot buffers. In unbuffered mode we
+  // always bypass read-ahead buffer at ReadaheadRandomAccessFile
+  // When read-ahead is required we then make use of compaction_readahead_size
+  // value and always try to read ahead. With read-ahead we always
+  // pre-allocate buffer to the size instead of growing it up to a limit.
+  //
+  // This option is currently honored only on Windows
+  //
+  // Default: 1 Mb
+  size_t random_access_max_buffer_size;
+
+  // This is the maximum buffer size that is used by WritableFileWriter.
+  // On Windows, we need to maintain an aligned buffer for writes. 
+  // We allow the buffer to grow until it's size hits the limit. 
+  //
+  // Default: 1024 * 1024 (1 MB)
+  size_t writable_file_max_buffer_size;
+
+
   // Use adaptive mutex, which spins in the user space before resorting
   // to kernel. This could reduce context switch when the mutex is not
   // heavily contended. However, if the mutex is hot, we could end up
@@ -1130,6 +1163,15 @@ struct DBOptions {
   // Default: nullptr (disabled)
   // Not supported in ROCKSDB_LITE mode!
   std::shared_ptr<Cache> row_cache;
+
+#ifndef ROCKSDB_LITE
+  // A filter object supplied to be invoked while processing write-ahead-logs
+  // (WALs) during recovery. The filter provides a way to inspect log
+  // records, ignoring a particular record or skipping replay.
+  // The filter is invoked at startup and is invoked from a single-thread
+  // currently.
+  const WalFilter* wal_filter;
+#endif  // ROCKSDB_LITE
 };
 
 // Options to control the behavior of a database (passed to DB::Open)
@@ -1248,6 +1290,14 @@ struct ReadOptions {
   // this option.
   bool total_order_seek;
 
+  // Enforce that the iterator only iterates over the same prefix as the seek.
+  // This option is effective only for prefix seeks, i.e. prefix_extractor is
+  // non-null for the column family and total_order_seek is false.  Unlike
+  // iterate_upper_bound, prefix_same_as_start only works within a prefix
+  // but in both directions.
+  // Default: false
+  bool prefix_same_as_start;
+
   ReadOptions();
   ReadOptions(bool cksum, bool cache);
 };
@@ -1312,6 +1362,11 @@ extern Options GetOptions(size_t total_write_buffer_limit,
                           int write_amplification_threshold = 32,
                           uint64_t target_db_size = 68719476736 /* 64GB */);
 
+// Create a Logger from provided DBOptions
+extern Status CreateLoggerFromOptions(const std::string& dbname,
+                                      const DBOptions& options,
+                                      std::shared_ptr<Logger>* logger);
+
 // CompactionOptions are used in CompactFiles() call.
 struct CompactionOptions {
   // Compaction output compression type

include/rocksdb/perf_context.h

@@ -27,42 +27,42 @@ struct PerfContext {
   uint64_t block_cache_hit_count;     // total number of block cache hits
   uint64_t block_read_count;          // total number of block reads (with IO)
   uint64_t block_read_byte;           // total number of bytes from block reads
-  uint64_t block_read_time;           // total time spent on block reads
-  uint64_t block_checksum_time;       // total time spent on block checksum
-  uint64_t block_decompress_time;     // total time spent on block decompression
+  uint64_t block_read_time;           // total nanos spent on block reads
+  uint64_t block_checksum_time;       // total nanos spent on block checksum
+  uint64_t block_decompress_time;  // total nanos spent on block decompression
   // total number of internal keys skipped over during iteration (overwritten or
   // deleted, to be more specific, hidden by a put or delete of the same key)
   uint64_t internal_key_skipped_count;
   // total number of deletes and single deletes skipped over during iteration
   uint64_t internal_delete_skipped_count;
 
-  uint64_t get_snapshot_time;          // total time spent on getting snapshot
-  uint64_t get_from_memtable_time;     // total time spent on querying memtables
+  uint64_t get_snapshot_time;       // total nanos spent on getting snapshot
+  uint64_t get_from_memtable_time;  // total nanos spent on querying memtables
   uint64_t get_from_memtable_count;    // number of mem tables queried
-  // total time spent after Get() finds a key
+  // total nanos spent after Get() finds a key
   uint64_t get_post_process_time;
-  uint64_t get_from_output_files_time; // total time reading from output files
-  // total time spent on seeking memtable
+  uint64_t get_from_output_files_time;  // total nanos reading from output files
+  // total nanos spent on seeking memtable
   uint64_t seek_on_memtable_time;
   // number of seeks issued on memtable
   uint64_t seek_on_memtable_count;
-  // total time spent on seeking child iters
+  // total nanos spent on seeking child iters
   uint64_t seek_child_seek_time;
   // number of seek issued in child iterators
   uint64_t seek_child_seek_count;
-  uint64_t seek_min_heap_time;         // total time spent on the merge heap
-  // total time spent on seeking the internal entries
+  uint64_t seek_min_heap_time;  // total nanos spent on the merge heap
+  // total nanos spent on seeking the internal entries
   uint64_t seek_internal_seek_time;
-  // total time spent on iterating internal entries to find the next user entry
+  // total nanos spent on iterating internal entries to find the next user entry
   uint64_t find_next_user_entry_time;
 
-  // total time spent on writing to WAL
+  // total nanos spent on writing to WAL
   uint64_t write_wal_time;
-  // total time spent on writing to mem tables
+  // total nanos spent on writing to mem tables
   uint64_t write_memtable_time;
-  // total time spent on delaying write
+  // total nanos spent on delaying write
   uint64_t write_delay_time;
-  // total time spent on writing a record, excluding the above three times
+  // total nanos spent on writing a record, excluding the above three times
   uint64_t write_pre_and_post_process_time;
 
   uint64_t db_mutex_lock_nanos;      // time spent on acquiring DB mutex.