Making platform 007 (gcc 7) default in build_detect_platform.sh (#5947 )

Summary: Making platform 007 (gcc 7) default in build_detect_platform.sh. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5947 Differential Revision: D18038837 Pulled By: vjnadimpalli fbshipit-source-id: 9ac2ddaa93bf328a416faec028970e039886378e
Bump version to 6.2.4
2026-11-06 02:52:35 -05:00 · 2019-10-30 10:29:05 -07:00 · 2019-09-18 14:42:43 -07:00 · 2019-09-18 14:39:45 -07:00 · 2019-09-18 14:37:22 -07:00 · 2019-09-18 14:37:16 -07:00
736 changed files with 47902 additions and 33366 deletions
@@ -92,9 +92,15 @@ else()
  endif()

  if(WITH_ZLIB)
-    find_package(zlib REQUIRED)
+    find_package(ZLIB REQUIRED)
    add_definitions(-DZLIB)
-    include_directories(${ZLIB_INCLUDE_DIR})
+    if(ZLIB_INCLUDE_DIRS)
+      # CMake 3
+      include_directories(${ZLIB_INCLUDE_DIRS})
+    else()
+      # CMake 2
+      include_directories(${ZLIB_INCLUDE_DIR})
+    endif()
    list(APPEND THIRDPARTY_LIBS ${ZLIB_LIBRARIES})
  endif()

@@ -311,6 +317,10 @@ if(DISABLE_STALL_NOTIF)
  add_definitions(-DROCKSDB_DISABLE_STALL_NOTIFICATION)
 endif()

+option(WITH_DYNAMIC_EXTENSION "build with dynamic extension support" OFF)
+if(NOT WITH_DYNAMIC_EXTENSION)
+  add_definitions(-DROCKSDB_NO_DYNAMIC_EXTENSION)
+endif()

 if(DEFINED USE_RTTI)
  if(USE_RTTI)
@@ -410,7 +420,7 @@ if(WITH_FALLOCATE)
 #include <linux/falloc.h>
 int main() {
 int fd = open(\"/dev/null\", 0);
- fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE, 0, 1024);
+ fallocate(fd, FALLOC_FL_KEEP_SIZE, 0, 1024);
 }
 " HAVE_FALLOCATE)
  if(HAVE_FALLOCATE)
@@ -455,8 +465,6 @@ include_directories(${PROJECT_SOURCE_DIR}/include)
 include_directories(SYSTEM ${PROJECT_SOURCE_DIR}/third-party/gtest-1.7.0/fused-src)
 find_package(Threads REQUIRED)

-add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
-
 # Main library source code

 set(SOURCES
@@ -483,6 +491,7 @@ set(SOURCES
        db/db_impl_debug.cc
        db/db_impl_experimental.cc
        db/db_impl_readonly.cc
+        db/db_impl_secondary.cc
        db/db_info_dumper.cc
        db/db_iter.cc
        db/dbformat.cc
@@ -495,6 +504,7 @@ set(SOURCES
        db/flush_scheduler.cc
        db/forward_iterator.cc
        db/internal_stats.cc
+        db/in_memory_stats_history.cc
        db/logs_with_prep_tracker.cc
        db/log_reader.cc
        db/log_writer.cc
@@ -504,7 +514,6 @@ set(SOURCES
        db/merge_helper.cc
        db/merge_operator.cc
        db/range_del_aggregator.cc
-        db/range_del_aggregator_v2.cc
        db/range_tombstone_fragmenter.cc
        db/repair.cc
        db/snapshot_impl.cc
@@ -525,7 +534,6 @@ set(SOURCES
        env/env_hdfs.cc
        env/mock_env.cc
        memtable/alloc_tracker.cc
-        memtable/hash_cuckoo_rep.cc
        memtable/hash_linklist_rep.cc
        memtable/hash_skiplist_rep.cc
        memtable/skiplistrep.cc
@@ -597,6 +605,7 @@ set(SOURCES
        util/comparator.cc
        util/compression_context_cache.cc
        util/concurrent_arena.cc
+        util/concurrent_task_limiter_impl.cc
        util/crc32c.cc
        util/delete_scheduler.cc
        util/dynamic_bloom.cc
@@ -637,20 +646,11 @@ set(SOURCES
        utilities/cassandra/format.cc
        utilities/cassandra/merge_operator.cc
        utilities/checkpoint/checkpoint_impl.cc
-        utilities/col_buf_decoder.cc
-        utilities/col_buf_encoder.cc
-        utilities/column_aware_encoding_util.cc
        utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc
-        utilities/date_tiered/date_tiered_db_impl.cc
        utilities/debug.cc
-        utilities/document/document_db.cc
-        utilities/document/json_document.cc
-        utilities/document/json_document_builder.cc
        utilities/env_mirror.cc
        utilities/env_timed.cc
-        utilities/geodb/geodb_impl.cc
        utilities/leveldb_options/leveldb_options.cc
-        utilities/lua/rocks_lua_compaction_filter.cc
        utilities/memory/memory_util.cc
        utilities/merge_operators/bytesxor.cc
        utilities/merge_operators/max.cc
@@ -665,9 +665,7 @@ set(SOURCES
        utilities/persistent_cache/block_cache_tier_metadata.cc
        utilities/persistent_cache/persistent_cache_tier.cc
        utilities/persistent_cache/volatile_tier_impl.cc
-        utilities/redis/redis_lists.cc
        utilities/simulator_cache/sim_cache.cc
-        utilities/spatialdb/spatial_db.cc
        utilities/table_properties_collectors/compact_on_deletion_collector.cc
        utilities/trace/file_trace_reader_writer.cc
        utilities/transactions/optimistic_transaction_db_impl.cc
@@ -823,6 +821,7 @@ if(NOT WIN32 OR ROCKSDB_INSTALL_ON_WINDOWS)
    TARGETS ${ROCKSDB_SHARED_LIB}
    EXPORT RocksDBTargets
    COMPONENT runtime
+    ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}"
    RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}"
    LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}"
    INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
@@ -846,6 +845,7 @@ endif()

 option(WITH_TESTS "build with tests" ON)
 if(WITH_TESTS)
+  add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
  set(TESTS
        cache/cache_test.cc
        cache/lru_cache_test.cc
@@ -877,6 +877,7 @@ if(WITH_TESTS)
        db/db_options_test.cc
        db/db_properties_test.cc
        db/db_range_del_test.cc
+        db/db_secondary_test.cc
        db/db_sst_test.cc
        db/db_statistics_test.cc
        db/db_table_properties_test.cc
@@ -907,7 +908,6 @@ if(WITH_TESTS)
        db/plain_table_db_test.cc
        db/prefix_test.cc
        db/range_del_aggregator_test.cc
-        db/range_del_aggregator_v2_test.cc
        db/range_tombstone_fragmenter_test.cc
        db/repair_test.cc
        db/table_properties_collector_test.cc
@@ -969,12 +969,6 @@ if(WITH_TESTS)
        utilities/cassandra/cassandra_row_merge_test.cc
        utilities/cassandra/cassandra_serialize_test.cc
        utilities/checkpoint/checkpoint_test.cc
-        utilities/column_aware_encoding_test.cc
-        utilities/date_tiered/date_tiered_test.cc
-        utilities/document/document_db_test.cc
-        utilities/document/json_document_test.cc
-        utilities/geodb/geodb_test.cc
-        utilities/lua/rocks_lua_test.cc
        utilities/memory/memory_test.cc
        utilities/merge_operators/string_append/stringappend_test.cc
        utilities/object_registry_test.cc
@@ -982,8 +976,6 @@ if(WITH_TESTS)
        utilities/options/options_util_test.cc
        utilities/persistent_cache/hash_table_test.cc
        utilities/persistent_cache/persistent_cache_test.cc
-        utilities/redis/redis_lists_test.cc
-        utilities/spatialdb/spatial_db_test.cc
        utilities/simulator_cache/sim_cache_test.cc
        utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
        utilities/transactions/optimistic_transaction_test.cc
@@ -1003,7 +995,6 @@ if(WITH_TESTS)
    db/range_del_aggregator_bench.cc
    tools/db_bench.cc
    table/table_reader_bench.cc
-    utilities/column_aware_encoding_exp.cc
    utilities/persistent_cache/hash_table_bench.cc)
  add_library(testharness OBJECT util/testharness.cc)
  foreach(sourcefile ${BENCHMARKS})
@@ -1,11 +1,111 @@
 # Rocksdb Change Log
-## Unreleased
+## 6.2.4 (9/18/2019)
+### Bug Fixes
+* Disable snap_refresh_nanos by default. The feature is to be deprecated in the next release.
+
+## 6.2.3 (9/3/2019)
+### Bug Fixes
+* Fix a bug in file ingestion caused by incorrect file number allocation when the number of column families involved in the ingestion exceeds 2.
+
+## 6.2.2 (6/7/2019)
+### Bug Fixes
+* Disable dynamic extension support by default for CMake.
+
+## 6.2.1 (6/4/2019)
+### Bug Fixes
+* Fix flush's/compaction's merge processing logic which allowed `Put`s covered by range tombstones to reappear. Note `Put`s may exist even if the user only ever called `Merge()` due to an internal conversion during compaction to the bottommost level.
+
+## 6.2.0 (4/30/2019)
 ### New Features
+* Add an option `strict_bytes_per_sync` that causes a file-writing thread to block rather than exceed the limit on bytes pending writeback specified by `bytes_per_sync` or `wal_bytes_per_sync`.
+* Improve range scan performance by avoiding per-key upper bound check in BlockBasedTableIterator.
+* Introduce Periodic Compaction for Level style compaction. Files are re-compacted periodically and put in the same level.
+* Block-based table index now contains exact highest key in the file, rather than an upper bound. This may improve Get() and iterator Seek() performance in some situations, especially when direct IO is enabled and block cache is disabled. A setting BlockBasedTableOptions::index_shortening is introduced to control this behavior. Set it to kShortenSeparatorsAndSuccessor to get the old behavior.
+* When reading from option file/string/map, customized envs can be filled according to object registry.
+* Add an option `snap_refresh_nanos` (default to 0.5s) to periodically refresh the snapshot list in compaction jobs. Assign to 0 to disable the feature.
+* Improve range scan performance when using explicit user readahead by not creating new table readers for every iterator.
+
+### Public API Change
+* Change the behavior of OptimizeForPointLookup(): move away from hash-based block-based-table index, and use whole key memtable filtering.
+* Change the behavior of OptimizeForSmallDb(): use a 16MB block cache, put index and filter blocks into it, and cost the memtable size to it. DBOptions.OptimizeForSmallDb() and ColumnFamilyOptions.OptimizeForSmallDb() start to take an optional cache object.
+* Added BottommostLevelCompaction::kForceOptimized to avoid double compacting newly compacted files in the bottommost level compaction of manual compaction. Note this option may prohibit the manual compaction to produce a single file in the bottommost level.
+
+### Bug Fixes
+* Adjust WriteBufferManager's dummy entry size to block cache from 1MB to 256KB.
+* Fix a race condition between WritePrepared::Get and ::Put with duplicate keys.
+* Fix crash when memtable prefix bloom is enabled and read/write a key out of domain of prefix extractor.
+* Close a WAL file before another thread deletes it.
+* Fix an assertion failure `IsFlushPending() == true` caused by one bg thread releasing the db mutex in ~ColumnFamilyData and another thread clearing `flush_requested_` flag.
+
+## 6.1.1 (4/9/2019)
+### New Features
+* When reading from option file/string/map, customized comparators and/or merge operators can be filled according to object registry.

 ### Public API Change

+### Bug Fixes
+* Fix a bug in 2PC where a sequence of txn prepare, memtable flush, and crash could result in losing the prepared transaction.
+* Fix a bug in Encryption Env which could cause encrypted files to be read beyond file boundaries.
+
+## 6.1.0 (3/27/2019)
+### New Features
+* Introduce two more stats levels, kExceptHistogramOrTimers and kExceptTimers.
+* Added a feature to perform data-block sampling for compressibility, and report stats to user.
+* Add support for trace filtering.
+* Add DBOptions.avoid_unnecessary_blocking_io. If true, we avoid file deletion when destorying ColumnFamilyHandle and Iterator. Instead, a job is scheduled to delete the files in background.
+
+### Public API Change
+* Remove bundled fbson library.
+* statistics.stats_level_ becomes atomic. It is preferred to use statistics.set_stats_level() and statistics.get_stats_level() to access it.
+* Introduce a new IOError subcode, PathNotFound, to indicate trying to open a nonexistent file or directory for read.
+* Add initial support for multiple db instances sharing the same data in single-writer, multi-reader mode.
+* Removed some "using std::xxx" from public headers.
+
+### Bug Fixes
+* Fix JEMALLOC_CXX_THROW macro missing from older Jemalloc versions, causing build failures on some platforms.
+* Fix SstFileReader not able to open file ingested with write_glbal_seqno=true.
+
+## 6.0.0 (2/19/2019)
+### New Features
+* Enabled checkpoint on readonly db (DBImplReadOnly).
+* Make DB ignore dropped column families while committing results of atomic flush.
+* RocksDB may choose to preopen some files even if options.max_open_files != -1. This may make DB open slightly longer.
+* For users of dictionary compression with ZSTD v0.7.0+, we now reuse the same digested dictionary when compressing each of an SST file's data blocks for faster compression speeds.
+* For all users of dictionary compression who set `cache_index_and_filter_blocks == true`, we now store dictionary data used for decompression in the block cache for better control over memory usage. For users of ZSTD v1.1.4+ who compile with -DZSTD_STATIC_LINKING_ONLY, this includes a digested dictionary, which is used to increase decompression speed.
+* Add support for block checksums verification for external SST files before ingestion.
+* Introduce stats history which periodically saves Statistics snapshots and added `GetStatsHistory` API to retrieve these snapshots.
+* Add a place holder in manifest which indicate a record from future that can be safely ignored.
+* Add support for trace sampling.
+* Enable properties block checksum verification for block-based tables.
+* For all users of dictionary compression, we now generate a separate dictionary for compressing each bottom-level SST file. Previously we reused a single dictionary for a whole compaction to bottom level. The new approach achieves better compression ratios; however, it uses more memory and CPU for buffering/sampling data blocks and training dictionaries.
+* Add whole key bloom filter support in memtable.
+* Files written by `SstFileWriter` will now use dictionary compression if it is configured in the file writer's `CompressionOptions`.
+
+### Public API Change
+* Disallow CompactionFilter::IgnoreSnapshots() = false, because it is not very useful and the behavior is confusing. The filter will filter everything if there is no snapshot declared by the time the compaction starts. However, users can define a snapshot after the compaction starts and before it finishes and this new snapshot won't be repeatable, because after the compaction finishes, some keys may be dropped.
+* CompactionPri = kMinOverlappingRatio also uses compensated file size, which boosts file with lots of tombstones to be compacted first.
+* Transaction::GetForUpdate is extended with a do_validate parameter with default value of true. If false it skips validating the snapshot before doing the read. Similarly ::Merge, ::Put, ::Delete, and ::SingleDelete are extended with assume_tracked with default value of false. If true it indicates that call is assumed to be after a ::GetForUpdate.
+* `TableProperties::num_entries` and `TableProperties::num_deletions` now also account for number of range tombstones.
+* Remove geodb, spatial_db, document_db, json_document, date_tiered_db, and redis_lists.
+* With "ldb ----try_load_options", when wal_dir specified by the option file doesn't exist, ignore it.
+* Change time resolution in FileOperationInfo.
+* Deleting Blob files also go through SStFileManager.
+* Remove CuckooHash memtable.
+* The counter stat `number.block.not_compressed` now also counts blocks not compressed due to poor compression ratio.
+* Remove ttl option from `CompactionOptionsFIFO`. The option has been deprecated and ttl in `ColumnFamilyOptions` is used instead.
+* Support SST file ingestion across multiple column families via DB::IngestExternalFiles. See the function's comment about atomicity.
+* Remove Lua compaction filter.
+
 ### Bug Fixes
 * Fix a deadlock caused by compaction and file ingestion waiting for each other in the event of write stalls.
+* Fix a memory leak when files with range tombstones are read in mmap mode and block cache is enabled
+* Fix handling of corrupt range tombstone blocks such that corruptions cannot cause deleted keys to reappear
+* Lock free MultiGet
+* Fix incorrect `NotFound` point lookup result when querying the endpoint of a file that has been extended by a range tombstone.
+* Fix with pipelined write, write leaders's callback failure lead to the whole write group fail.
+
+### Change Default Options
+* Change options.compaction_pri's default to kMinOverlappingRatio

 ## 5.18.0 (11/30/2018)
 ### New Features
@@ -43,6 +43,8 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
      command line flags processing. You can compile rocksdb library even
      if you don't have gflags installed.

+* If you wish to build the RocksJava static target, then cmake is required for building Snappy.
+
 ## Supported platforms

 * **Linux - Ubuntu**
@@ -1,7 +1,9 @@
 This is the list of all known third-party language bindings for RocksDB. If something is missing, please open a pull request to add it.

 * Java - https://github.com/facebook/rocksdb/tree/master/java
-* Python - http://pyrocksdb.readthedocs.org/en/latest/
+* Python
+    * http://python-rocksdb.readthedocs.io/en/latest/
+    * http://pyrocksdb.readthedocs.org/en/latest/ (unmaintained)
 * Perl - https://metacpan.org/pod/RocksDB
 * Node.js - https://npmjs.org/package/rocksdb
 * Go - https://github.com/tecbot/gorocksdb
@@ -137,6 +137,12 @@ CFLAGS +=  -DHAVE_POWER8
 HAVE_POWER8=1
 endif

+ifeq (,$(shell $(CXX) -fsyntax-only -march=armv8-a+crc -xc /dev/null 2>&1))
+CXXFLAGS += -march=armv8-a+crc
+CFLAGS += -march=armv8-a+crc
+ARMCRC_SOURCE=1
+endif
+
 # if we're compiling for release, compile without debug code (-DNDEBUG)
 ifeq ($(DEBUG_LEVEL),0)
 OPT += -DNDEBUG
@@ -198,7 +204,7 @@ include make_config.mk
 CLEAN_FILES += make_config.mk

 missing_make_config_paths := $(shell				\
-	grep "\/\S*" -o $(CURDIR)/make_config.mk | 		\
+	grep "\./\S*\|/\S*" -o $(CURDIR)/make_config.mk | 	\
 	while read path;					\
 		do [ -e $$path ] || echo $$path; 		\
 	done | sort | uniq)
@@ -403,7 +409,7 @@ BENCHTOOLOBJECTS = $(BENCH_LIB_SOURCES:.cc=.o) $(LIBOBJECTS) $(TESTUTIL)

 ANALYZETOOLOBJECTS = $(ANALYZER_LIB_SOURCES:.cc=.o)

-EXPOBJECTS = $(EXP_LIB_SOURCES:.cc=.o) $(LIBOBJECTS) $(TESTUTIL)
+EXPOBJECTS = $(LIBOBJECTS) $(TESTUTIL)

 TESTS = \
 	db_basic_test \
@@ -443,6 +449,7 @@ TESTS = \
 	db_merge_operator_test \
 	db_options_test \
 	db_range_del_test \
+	db_secondary_test \
 	db_sst_test \
 	db_tailing_iter_test \
 	db_io_failure_test \
@@ -482,7 +489,6 @@ TESTS = \
 	merger_test \
 	util_merge_operators_test \
 	options_file_test \
-	redis_test \
 	reduce_levels_test \
 	plain_table_db_test \
 	comparator_db_test \
@@ -496,12 +502,8 @@ TESTS = \
 	cassandra_row_merge_test \
 	cassandra_serialize_test \
 	ttl_test \
-	date_tiered_test \
 	backupable_db_test \
-	document_db_test \
-	json_document_test \
 	sim_cache_test \
-	spatial_db_test \
 	version_edit_test \
 	version_set_test \
 	compaction_picker_test \
@@ -513,7 +515,6 @@ TESTS = \
 	deletefile_test \
 	obsolete_files_test \
 	table_test \
-	geodb_test \
 	delete_scheduler_test \
 	options_test \
 	options_settable_test \
@@ -530,7 +531,6 @@ TESTS = \
 	compaction_job_test \
 	thread_list_test \
 	sst_dump_test \
-	column_aware_encoding_test \
 	compact_files_test \
 	optimistic_transaction_test \
 	write_callback_test \
@@ -542,8 +542,6 @@ TESTS = \
 	ldb_cmd_test \
 	persistent_cache_test \
 	statistics_test \
-	lua_test \
-	range_del_aggregator_test \
 	lru_cache_test \
 	object_registry_test \
 	repair_test \
@@ -554,8 +552,9 @@ TESTS = \
 	trace_analyzer_test \
 	repeatable_thread_test \
 	range_tombstone_fragmenter_test \
-	range_del_aggregator_v2_test \
+	range_del_aggregator_test \
 	sst_file_reader_test \
+	db_secondary_test \

 PARALLEL_TEST = \
 	backupable_db_test \
@@ -605,7 +604,7 @@ TEST_LIBS = \
 	librocksdb_env_basic_test.a

 # TODO: add back forward_iterator_bench, after making it build in all environemnts.
-BENCHMARKS = db_bench table_reader_bench cache_bench memtablerep_bench column_aware_encoding_exp persistent_cache_bench range_del_aggregator_bench
+BENCHMARKS = db_bench table_reader_bench cache_bench memtablerep_bench persistent_cache_bench range_del_aggregator_bench

 # if user didn't config LIBNAME, set the default
 ifeq ($(LIBNAME),)
@@ -695,7 +694,8 @@ endif  # PLATFORM_SHARED_EXT
 .PHONY: blackbox_crash_test check clean coverage crash_test ldb_tests package \
 	release tags tags0 valgrind_check whitebox_crash_test format static_lib shared_lib all \
 	dbg rocksdbjavastatic rocksdbjava install install-static install-shared uninstall \
-	analyze tools tools_lib
+	analyze tools tools_lib \
+	blackbox_crash_test_with_atomic_flush whitebox_crash_test_with_atomic_flush


 all: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(TESTS)
@@ -825,7 +825,6 @@ J ?= 100%
 # Use this regexp to select the subset of tests whose names match.
 tests-regexp = .

-t_run = $(wildcard t/run-*)
 .PHONY: check_0
 check_0:
 	$(AM_V_GEN)export TEST_TMPDIR=$(TMPD); \
@@ -835,7 +834,7 @@ check_0:
 	test -t 1 && eta=--eta || eta=; \
 	{ \
 		printf './%s\n' $(filter-out $(PARALLEL_TEST),$(TESTS)); \
-		printf '%s\n' $(t_run); \
+		find t -name 'run-*' -print; \
 	} \
 	  | $(prioritize_long_running_tests)				\
 	  | grep -E '$(tests-regexp)'					\
@@ -852,7 +851,7 @@ valgrind_check_0:
 	test -t 1 && eta=--eta || eta=;					\
 	{								\
 	  printf './%s\n' $(filter-out $(PARALLEL_TEST) %skiplist_test options_settable_test, $(TESTS));		\
-	  printf '%s\n' $(t_run);					\
+	  find t -name 'run-*' -print; \
 	}								\
 	  | $(prioritize_long_running_tests)				\
 	  | grep -E '$(tests-regexp)'					\
@@ -904,23 +903,29 @@ ldb_tests: ldb

 crash_test: whitebox_crash_test blackbox_crash_test

+crash_test_with_atomic_flush: whitebox_crash_test_with_atomic_flush blackbox_crash_test_with_atomic_flush
+
 blackbox_crash_test: db_stress
 	python -u tools/db_crashtest.py --simple blackbox $(CRASH_TEST_EXT_ARGS)
-	python -u tools/db_crashtest.py --enable_atomic_flush blackbox $(CRASH_TEST_EXT_ARGS)
 	python -u tools/db_crashtest.py blackbox $(CRASH_TEST_EXT_ARGS)

+blackbox_crash_test_with_atomic_flush: db_stress
+	python -u tools/db_crashtest.py --enable_atomic_flush blackbox $(CRASH_TEST_EXT_ARGS)
+
 ifeq ($(CRASH_TEST_KILL_ODD),)
  CRASH_TEST_KILL_ODD=888887
 endif

 whitebox_crash_test: db_stress
 	python -u tools/db_crashtest.py --simple whitebox --random_kill_odd \
-      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)
-	python -u tools/db_crashtest.py --enable_atomic_flush whitebox  --random_kill_odd \
      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)
 	python -u tools/db_crashtest.py whitebox  --random_kill_odd \
      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)

+whitebox_crash_test_with_atomic_flush: db_stress
+	python -u tools/db_crashtest.py --enable_atomic_flush whitebox  --random_kill_odd \
+      $(CRASH_TEST_KILL_ODD) $(CRASH_TEST_EXT_ARGS)
+
 asan_check:
 	$(MAKE) clean
 	COMPILE_WITH_ASAN=1 $(MAKE) check -j32
@@ -931,6 +936,11 @@ asan_crash_test:
 	COMPILE_WITH_ASAN=1 $(MAKE) crash_test
 	$(MAKE) clean

+asan_crash_test_with_atomic_flush:
+	$(MAKE) clean
+	COMPILE_WITH_ASAN=1 $(MAKE) crash_test_with_atomic_flush
+	$(MAKE) clean
+
 ubsan_check:
 	$(MAKE) clean
 	COMPILE_WITH_UBSAN=1 $(MAKE) check -j32
@@ -941,6 +951,11 @@ ubsan_crash_test:
 	COMPILE_WITH_UBSAN=1 $(MAKE) crash_test
 	$(MAKE) clean

+ubsan_crash_test_with_atomic_flush:
+	$(MAKE) clean
+	COMPILE_WITH_UBSAN=1 $(MAKE) crash_test_with_atomic_flush
+	$(MAKE) clean
+
 valgrind_test:
 	ROCKSDB_VALGRIND_RUN=1 DISABLE_JEMALLOC=1 $(MAKE) valgrind_check

@@ -1154,9 +1169,6 @@ cassandra_row_merge_test: utilities/cassandra/cassandra_row_merge_test.o utiliti
 cassandra_serialize_test: utilities/cassandra/cassandra_serialize_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-redis_test: utilities/redis/redis_lists_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 hash_table_test: utilities/persistent_cache/hash_table_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1295,18 +1307,9 @@ backupable_db_test: utilities/backupable/backupable_db_test.o $(LIBOBJECTS) $(TE
 checkpoint_test: utilities/checkpoint/checkpoint_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-document_db_test: utilities/document/document_db_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
-json_document_test: utilities/document/json_document_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 sim_cache_test: utilities/simulator_cache/sim_cache_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-spatial_db_test: utilities/spatialdb/spatial_db_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 env_mirror_test: utilities/env_mirror_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1324,9 +1327,6 @@ object_registry_test: utilities/object_registry_test.o $(LIBOBJECTS) $(TESTHARNE
 ttl_test: utilities/ttl/ttl_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-date_tiered_test: utilities/date_tiered/date_tiered_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 write_batch_with_index_test: utilities/write_batch_with_index/write_batch_with_index_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1453,9 +1453,6 @@ deletefile_test: db/deletefile_test.o $(LIBOBJECTS) $(TESTHARNESS)
 obsolete_files_test: db/obsolete_files_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-geodb_test: utilities/geodb/geodb_test.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 rocksdb_dump: tools/dump/rocksdb_dump.o $(LIBOBJECTS)
 	$(AM_LINK)

@@ -1504,9 +1501,6 @@ timer_queue_test: util/timer_queue_test.o $(LIBOBJECTS) $(TESTHARNESS)
 sst_dump_test: tools/sst_dump_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-column_aware_encoding_test: utilities/column_aware_encoding_test.o $(TESTHARNESS) $(EXPOBJECTS)
-	$(AM_LINK)
-
 optimistic_transaction_test: utilities/transactions/optimistic_transaction_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1546,9 +1540,6 @@ sst_dump: tools/sst_dump.o $(LIBOBJECTS)
 blob_dump: tools/blob_dump.o $(LIBOBJECTS)
 	$(AM_LINK)

-column_aware_encoding_exp: utilities/column_aware_encoding_exp.o $(EXPOBJECTS)
-	$(AM_LINK)
-
 repair_test: db/repair_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1570,9 +1561,6 @@ statistics_test: monitoring/statistics_test.o $(LIBOBJECTS) $(TESTHARNESS)
 lru_cache_test: cache/lru_cache_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-lua_test: utilities/lua/rocks_lua_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
-	$(AM_LINK)
-
 range_del_aggregator_test: db/range_del_aggregator_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

@@ -1588,10 +1576,10 @@ repeatable_thread_test: util/repeatable_thread_test.o $(LIBOBJECTS) $(TESTHARNES
 range_tombstone_fragmenter_test: db/range_tombstone_fragmenter_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-range_del_aggregator_v2_test: db/range_del_aggregator_v2_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
+sst_file_reader_test: table/sst_file_reader_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

-sst_file_reader_test: table/sst_file_reader_test.o $(LIBOBJECTS) $(TESTHARNESS)
+db_secondary_test: db/db_secondary_test.o db/db_test_util.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(AM_LINK)

 #-------------------------------------------------
@@ -1665,14 +1653,14 @@ ZLIB_DOWNLOAD_BASE ?= http://zlib.net
 BZIP2_VER ?= 1.0.6
 BZIP2_SHA256 ?= a2848f34fcd5d6cf47def00461fcb528a0484d8edef8208d6d2e2909dc61d9cd
 BZIP2_DOWNLOAD_BASE ?= https://web.archive.org/web/20180624184835/http://www.bzip.org
-SNAPPY_VER ?= 1.1.4
-SNAPPY_SHA256 ?= 134bfe122fd25599bb807bb8130e7ba6d9bdb851e0b16efcb83ac4f5d0b70057
-SNAPPY_DOWNLOAD_BASE ?= https://github.com/google/snappy/releases/download
-LZ4_VER ?= 1.8.0
-LZ4_SHA256 ?= 2ca482ea7a9bb103603108b5a7510b7592b90158c151ff50a28f1ca8389fccf6
+SNAPPY_VER ?= 1.1.7
+SNAPPY_SHA256 ?= 3dfa02e873ff51a11ee02b9ca391807f0c8ea0529a4924afa645fbf97163f9d4
+SNAPPY_DOWNLOAD_BASE ?= https://github.com/google/snappy/archive
+LZ4_VER ?= 1.8.3
+LZ4_SHA256 ?= 33af5936ac06536805f9745e0b6d61da606a1f8b4cc5c04dd3cbaca3b9b4fc43
 LZ4_DOWNLOAD_BASE ?= https://github.com/lz4/lz4/archive
-ZSTD_VER ?= 1.3.3
-ZSTD_SHA256 ?= a77c47153ee7de02626c5b2a097005786b71688be61e9fb81806a011f90b297b
+ZSTD_VER ?= 1.3.7
+ZSTD_SHA256 ?= 5dd1e90eb16c25425880c8a91327f63de22891ffed082fcc17e5ae84fce0d5fb
 ZSTD_DOWNLOAD_BASE ?= https://github.com/facebook/zstd/archive
 CURL_SSL_OPTS ?= --tlsv1

@@ -1711,7 +1699,9 @@ endif

 libz.a:
 	-rm -rf zlib-$(ZLIB_VER)
-	curl -O -L ${ZLIB_DOWNLOAD_BASE}/zlib-$(ZLIB_VER).tar.gz
+ifeq (,$(wildcard ./zlib-$(ZLIB_VER).tar.gz))
+	curl --output zlib-$(ZLIB_VER).tar.gz -L ${ZLIB_DOWNLOAD_BASE}/zlib-$(ZLIB_VER).tar.gz
+endif
 	ZLIB_SHA256_ACTUAL=`$(SHA256_CMD) zlib-$(ZLIB_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(ZLIB_SHA256)" != "$$ZLIB_SHA256_ACTUAL" ]; then \
 		echo zlib-$(ZLIB_VER).tar.gz checksum mismatch, expected=\"$(ZLIB_SHA256)\" actual=\"$$ZLIB_SHA256_ACTUAL\"; \
@@ -1723,7 +1713,9 @@ libz.a:

 libbz2.a:
 	-rm -rf bzip2-$(BZIP2_VER)
-	curl -O -L ${BZIP2_DOWNLOAD_BASE}/$(BZIP2_VER)/bzip2-$(BZIP2_VER).tar.gz
+ifeq (,$(wildcard ./bzip2-$(BZIP2_VER).tar.gz))
+	curl --output bzip2-$(BZIP2_VER).tar.gz -L ${BZIP2_DOWNLOAD_BASE}/$(BZIP2_VER)/bzip2-$(BZIP2_VER).tar.gz
+endif
 	BZIP2_SHA256_ACTUAL=`$(SHA256_CMD) bzip2-$(BZIP2_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(BZIP2_SHA256)" != "$$BZIP2_SHA256_ACTUAL" ]; then \
 		echo bzip2-$(BZIP2_VER).tar.gz checksum mismatch, expected=\"$(BZIP2_SHA256)\" actual=\"$$BZIP2_SHA256_ACTUAL\"; \
@@ -1735,21 +1727,24 @@ libbz2.a:

 libsnappy.a:
 	-rm -rf snappy-$(SNAPPY_VER)
-	curl -O -L ${CURL_SSL_OPTS} ${SNAPPY_DOWNLOAD_BASE}/$(SNAPPY_VER)/snappy-$(SNAPPY_VER).tar.gz
+ifeq (,$(wildcard ./snappy-$(SNAPPY_VER).tar.gz))
+	curl --output snappy-$(SNAPPY_VER).tar.gz -L ${CURL_SSL_OPTS} ${SNAPPY_DOWNLOAD_BASE}/$(SNAPPY_VER).tar.gz
+endif
 	SNAPPY_SHA256_ACTUAL=`$(SHA256_CMD) snappy-$(SNAPPY_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(SNAPPY_SHA256)" != "$$SNAPPY_SHA256_ACTUAL" ]; then \
 		echo snappy-$(SNAPPY_VER).tar.gz checksum mismatch, expected=\"$(SNAPPY_SHA256)\" actual=\"$$SNAPPY_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
 	tar xvzf snappy-$(SNAPPY_VER).tar.gz
-	cd snappy-$(SNAPPY_VER) && CFLAGS='${EXTRA_CFLAGS}' CXXFLAGS='${EXTRA_CXXFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --with-pic --enable-static --disable-shared
-	cd snappy-$(SNAPPY_VER) && $(MAKE) ${SNAPPY_MAKE_TARGET}
-	cp snappy-$(SNAPPY_VER)/.libs/libsnappy.a .
+	mkdir snappy-$(SNAPPY_VER)/build
+	cd snappy-$(SNAPPY_VER)/build && CFLAGS='${EXTRA_CFLAGS}' CXXFLAGS='${EXTRA_CXXFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' cmake -DCMAKE_POSITION_INDEPENDENT_CODE=ON .. && $(MAKE) ${SNAPPY_MAKE_TARGET}
+	cp snappy-$(SNAPPY_VER)/build/libsnappy.a .

 liblz4.a:
 	-rm -rf lz4-$(LZ4_VER)
-	curl -O -L ${CURL_SSL_OPTS} ${LZ4_DOWNLOAD_BASE}/v$(LZ4_VER).tar.gz
-	mv v$(LZ4_VER).tar.gz lz4-$(LZ4_VER).tar.gz
+ifeq (,$(wildcard ./lz4-$(LZ4_VER).tar.gz))
+	curl --output lz4-$(LZ4_VER).tar.gz -L ${CURL_SSL_OPTS} ${LZ4_DOWNLOAD_BASE}/v$(LZ4_VER).tar.gz
+endif
 	LZ4_SHA256_ACTUAL=`$(SHA256_CMD) lz4-$(LZ4_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(LZ4_SHA256)" != "$$LZ4_SHA256_ACTUAL" ]; then \
 		echo lz4-$(LZ4_VER).tar.gz checksum mismatch, expected=\"$(LZ4_SHA256)\" actual=\"$$LZ4_SHA256_ACTUAL\"; \
@@ -1761,8 +1756,9 @@ liblz4.a:

 libzstd.a:
 	-rm -rf zstd-$(ZSTD_VER)
-	curl -O -L ${CURL_SSL_OPTS} ${ZSTD_DOWNLOAD_BASE}/v$(ZSTD_VER).tar.gz
-	mv v$(ZSTD_VER).tar.gz zstd-$(ZSTD_VER).tar.gz
+ifeq (,$(wildcard ./zstd-$(ZSTD_VER).tar.gz))
+	curl --output zstd-$(ZSTD_VER).tar.gz -L ${CURL_SSL_OPTS} ${ZSTD_DOWNLOAD_BASE}/v$(ZSTD_VER).tar.gz
+endif
 	ZSTD_SHA256_ACTUAL=`$(SHA256_CMD) zstd-$(ZSTD_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(ZSTD_SHA256)" != "$$ZSTD_SHA256_ACTUAL" ]; then \
 		echo zstd-$(ZSTD_VER).tar.gz checksum mismatch, expected=\"$(ZSTD_SHA256)\" actual=\"$$ZSTD_SHA256_ACTUAL\"; \
@@ -1931,7 +1927,8 @@ commit_prereq: build_tools/rocksdb-lego-determinator \
 ifeq ($(PLATFORM), IOS)
 # For iOS, create universal object files to be used on both the simulator and
 # a device.
-PLATFORMSROOT=/Applications/Xcode.app/Contents/Developer/Platforms
+XCODEROOT=$(shell xcode-select -print-path)
+PLATFORMSROOT=$(XCODEROOT)/Platforms
 SIMULATORROOT=$(PLATFORMSROOT)/iPhoneSimulator.platform/Developer
 DEVICEROOT=$(PLATFORMSROOT)/iPhoneOS.platform/Developer
 IOSVERSION=$(shell defaults read $(PLATFORMSROOT)/iPhoneOS.platform/version CFBundleShortVersionString)
@@ -1968,7 +1965,7 @@ endif
 #  	Source files dependencies detection
 # ---------------------------------------------------------------------------

-all_sources = $(LIB_SOURCES) $(MAIN_SOURCES) $(MOCK_LIB_SOURCES) $(TOOL_LIB_SOURCES) $(BENCH_LIB_SOURCES) $(TEST_LIB_SOURCES) $(EXP_LIB_SOURCES) $(ANALYZER_LIB_SOURCES)
+all_sources = $(LIB_SOURCES) $(MAIN_SOURCES) $(MOCK_LIB_SOURCES) $(TOOL_LIB_SOURCES) $(BENCH_LIB_SOURCES) $(TEST_LIB_SOURCES) $(ANALYZER_LIB_SOURCES)
 DEPFILES = $(all_sources:.cc=.cc.d)

 # Add proper dependency support so changing a .h file forces a .cc file to
@@ -5,7 +5,7 @@
 [![PPC64le Build Status](http://140.211.168.68:8080/buildStatus/icon?job=Rocksdb)](http://140.211.168.68:8080/job/Rocksdb)

 RocksDB is developed and maintained by Facebook Database Engineering Team.
-It is built on earlier work on LevelDB by Sanjay Ghemawat (sanjay@google.com)
+It is built on earlier work on [LevelDB](https://github.com/google/leveldb) by Sanjay Ghemawat (sanjay@google.com)
 and Jeff Dean (jeff@google.com)

 This code is a library that forms the core building block for a fast
@@ -1,12 +1,10 @@
 load("@fbcode_macros//build_defs:auto_headers.bzl", "AutoHeaders")
+load("@fbcode_macros//build_defs:cpp_library.bzl", "cpp_library")
+load(":defs.bzl", "test_binary")

 REPO_PATH = package_name() + "/"

-BUCK_BINS = "buck-out/gen/" + REPO_PATH
-
-TEST_RUNNER = REPO_PATH + "buckifier/rocks_test_runner.sh"
-
-rocksdb_compiler_flags = [
+ROCKSDB_COMPILER_FLAGS = [
    "-fno-builtin-memcmp",
    "-DROCKSDB_PLATFORM_POSIX",
    "-DROCKSDB_LIB_IO_POSIX",
@@ -22,6 +20,7 @@ rocksdb_compiler_flags = [
    "-DBZIP2",
    "-DLZ4",
    "-DZSTD",
+    "-DZSTD_STATIC_LINKING_ONLY",
    "-DGFLAGS=gflags",
    "-DNUMA",
    "-DTBB",
@@ -33,7 +32,7 @@ rocksdb_compiler_flags = [
    "-Wnarrowing",
 ]

-rocksdb_external_deps = [
+ROCKSDB_EXTERNAL_DEPS = [
    ("bzip2", None, "bz2"),
    ("snappy", None, "snappy"),
    ("zlib", None, "z"),
@@ -45,13 +44,13 @@ rocksdb_external_deps = [
    ("googletest", None, "gtest"),
 ]

-rocksdb_preprocessor_flags = [
+ROCKSDB_PREPROCESSOR_FLAGS = [
    # Directories with files for #include
    "-I" + REPO_PATH + "include/",
    "-I" + REPO_PATH,
 ]

-rocksdb_arch_preprocessor_flags = {
+ROCKSDB_ARCH_PREPROCESSOR_FLAGS = {
    "x86_64": [
        "-DHAVE_SSE42",
        "-DHAVE_PCLMUL",
@@ -64,18 +63,15 @@ is_opt_mode = build_mode.startswith("opt")

 # -DNDEBUG is added by default in opt mode in fbcode. But adding it twice
 # doesn't harm and avoid forgetting to add it.
-if is_opt_mode:
-    rocksdb_compiler_flags.append("-DNDEBUG")
-
-default_allocator = read_config("fbcode", "default_allocator")
+ROCKSDB_COMPILER_FLAGS += (["-DNDEBUG"] if is_opt_mode else [])

 sanitizer = read_config("fbcode", "sanitizer")

-# Let RocksDB aware of jemalloc existence.
-# Do not enable it if sanitizer presents.
-if is_opt_mode and default_allocator.startswith("jemalloc") and sanitizer == "":
-    rocksdb_compiler_flags.append("-DROCKSDB_JEMALLOC")
-    rocksdb_external_deps.append(("jemalloc", None, "headers"))
+# Do not enable jemalloc if sanitizer presents. RocksDB will further detect
+# whether the binary is linked with jemalloc at runtime.
+ROCKSDB_COMPILER_FLAGS += (["-DROCKSDB_JEMALLOC"] if sanitizer == "" else [])
+
+ROCKSDB_EXTERNAL_DEPS += ([("jemalloc", None, "headers")] if sanitizer == "" else [])

 cpp_library(
    name = "rocksdb_lib",
@@ -102,6 +98,7 @@ cpp_library(
        "db/db_impl_files.cc",
        "db/db_impl_open.cc",
        "db/db_impl_readonly.cc",
+        "db/db_impl_secondary.cc",
        "db/db_impl_write.cc",
        "db/db_info_dumper.cc",
        "db/db_iter.cc",
@@ -114,6 +111,7 @@ cpp_library(
        "db/flush_job.cc",
        "db/flush_scheduler.cc",
        "db/forward_iterator.cc",
+        "db/in_memory_stats_history.cc",
        "db/internal_stats.cc",
        "db/log_reader.cc",
        "db/log_writer.cc",
@@ -124,7 +122,6 @@ cpp_library(
        "db/merge_helper.cc",
        "db/merge_operator.cc",
        "db/range_del_aggregator.cc",
-        "db/range_del_aggregator_v2.cc",
        "db/range_tombstone_fragmenter.cc",
        "db/repair.cc",
        "db/snapshot_impl.cc",
@@ -147,7 +144,6 @@ cpp_library(
        "env/io_posix.cc",
        "env/mock_env.cc",
        "memtable/alloc_tracker.cc",
-        "memtable/hash_cuckoo_rep.cc",
        "memtable/hash_linklist_rep.cc",
        "memtable/hash_skiplist_rep.cc",
        "memtable/skiplistrep.cc",
@@ -220,6 +216,7 @@ cpp_library(
        "util/comparator.cc",
        "util/compression_context_cache.cc",
        "util/concurrent_arena.cc",
+        "util/concurrent_task_limiter_impl.cc",
        "util/crc32c.cc",
        "util/delete_scheduler.cc",
        "util/dynamic_bloom.cc",
@@ -261,16 +258,10 @@ cpp_library(
        "utilities/checkpoint/checkpoint_impl.cc",
        "utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc",
        "utilities/convenience/info_log_finder.cc",
-        "utilities/date_tiered/date_tiered_db_impl.cc",
        "utilities/debug.cc",
-        "utilities/document/document_db.cc",
-        "utilities/document/json_document.cc",
-        "utilities/document/json_document_builder.cc",
        "utilities/env_mirror.cc",
        "utilities/env_timed.cc",
-        "utilities/geodb/geodb_impl.cc",
        "utilities/leveldb_options/leveldb_options.cc",
-        "utilities/lua/rocks_lua_compaction_filter.cc",
        "utilities/memory/memory_util.cc",
        "utilities/merge_operators/bytesxor.cc",
        "utilities/merge_operators/max.cc",
@@ -285,9 +276,7 @@ cpp_library(
        "utilities/persistent_cache/block_cache_tier_metadata.cc",
        "utilities/persistent_cache/persistent_cache_tier.cc",
        "utilities/persistent_cache/volatile_tier_impl.cc",
-        "utilities/redis/redis_lists.cc",
        "utilities/simulator_cache/sim_cache.cc",
-        "utilities/spatialdb/spatial_db.cc",
        "utilities/table_properties_collectors/compact_on_deletion_collector.cc",
        "utilities/trace/file_trace_reader_writer.cc",
        "utilities/transactions/optimistic_transaction.cc",
@@ -308,11 +297,11 @@ cpp_library(
        "utilities/write_batch_with_index/write_batch_with_index_internal.cc",
    ],
    auto_headers = AutoHeaders.RECURSIVE_GLOB,
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )

 cpp_library(
@@ -325,16 +314,13 @@ cpp_library(
        "util/testharness.cc",
        "util/testutil.cc",
        "utilities/cassandra/test_utils.cc",
-        "utilities/col_buf_decoder.cc",
-        "utilities/col_buf_encoder.cc",
-        "utilities/column_aware_encoding_util.cc",
    ],
    auto_headers = AutoHeaders.RECURSIVE_GLOB,
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [":rocksdb_lib"],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )

 cpp_library(
@@ -345,22 +331,22 @@ cpp_library(
        "util/testutil.cc",
    ],
    auto_headers = AutoHeaders.RECURSIVE_GLOB,
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [":rocksdb_lib"],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )

 cpp_library(
    name = "env_basic_test_lib",
    srcs = ["env/env_basic_test.cc"],
    auto_headers = AutoHeaders.RECURSIVE_GLOB,
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [":rocksdb_test_lib"],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )

 # [test_name, test_src, test_type]
@@ -450,11 +436,6 @@ ROCKS_TESTS = [
        "util/coding_test.cc",
        "serial",
    ],
-    [
-        "column_aware_encoding_test",
-        "utilities/column_aware_encoding_test.cc",
-        "serial",
-    ],
    [
        "column_family_test",
        "db/column_family_test.cc",
@@ -525,11 +506,6 @@ ROCKS_TESTS = [
        "table/data_block_hash_index_test.cc",
        "serial",
    ],
-    [
-        "date_tiered_test",
-        "utilities/date_tiered/date_tiered_test.cc",
-        "serial",
-    ],
    [
        "db_basic_test",
        "db/db_basic_test.cc",
@@ -630,6 +606,11 @@ ROCKS_TESTS = [
        "db/db_range_del_test.cc",
        "serial",
    ],
+    [
+        "db_secondary_test",
+        "db/db_secondary_test.cc",
+        "serial",
+    ],
    [
        "db_sst_test",
        "db/db_sst_test.cc",
@@ -690,11 +671,6 @@ ROCKS_TESTS = [
        "db/deletefile_test.cc",
        "serial",
    ],
-    [
-        "document_db_test",
-        "utilities/document/document_db_test.cc",
-        "serial",
-    ],
    [
        "dynamic_bloom_test",
        "util/dynamic_bloom_test.cc",
@@ -770,11 +746,6 @@ ROCKS_TESTS = [
        "table/full_filter_block_test.cc",
        "serial",
    ],
-    [
-        "geodb_test",
-        "utilities/geodb/geodb_test.cc",
-        "serial",
-    ],
    [
        "hash_table_test",
        "utilities/persistent_cache/hash_table_test.cc",
@@ -805,11 +776,6 @@ ROCKS_TESTS = [
        "monitoring/iostats_context_test.cc",
        "serial",
    ],
-    [
-        "json_document_test",
-        "utilities/document/json_document_test.cc",
-        "serial",
-    ],
    [
        "ldb_cmd_test",
        "tools/ldb_cmd_test.cc",
@@ -935,11 +901,6 @@ ROCKS_TESTS = [
        "db/range_del_aggregator_test.cc",
        "serial",
    ],
-    [
-        "range_del_aggregator_v2_test",
-        "db/range_del_aggregator_v2_test.cc",
-        "serial",
-    ],
    [
        "range_tombstone_fragmenter_test",
        "db/range_tombstone_fragmenter_test.cc",
@@ -980,11 +941,6 @@ ROCKS_TESTS = [
        "util/slice_transform_test.cc",
        "serial",
    ],
-    [
-        "spatial_db_test",
-        "utilities/spatialdb/spatial_db_test.cc",
-        "serial",
-    ],
    [
        "sst_dump_test",
        "tools/sst_dump_test.cc",
@@ -1110,26 +1066,16 @@ ROCKS_TESTS = [
 # Generate a test rule for each entry in ROCKS_TESTS
 # Do not build the tests in opt mode, since SyncPoint and other test code
 # will not be included.
-if not is_opt_mode:
-    for test_cfg in ROCKS_TESTS:
-        test_name = test_cfg[0]
-        test_cc = test_cfg[1]
-        ttype = "gtest" if test_cfg[2] == "parallel" else "simple"
-        test_bin = test_name + "_bin"
-
-        cpp_binary(
-            name = test_bin,
-            srcs = [test_cc],
-            arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-            compiler_flags = rocksdb_compiler_flags,
-            preprocessor_flags = rocksdb_preprocessor_flags,
-            deps = [":rocksdb_test_lib"],
-            external_deps = rocksdb_external_deps,
-        )
-
-        custom_unittest(
-            name = test_name,
-            command = [TEST_RUNNER, BUCK_BINS + test_bin],
-            type = ttype,
-            deps = [":" + test_bin],
-        )
+[
+    test_binary(
+        parallelism = parallelism,
+        rocksdb_arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+        rocksdb_compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        rocksdb_external_deps = ROCKSDB_EXTERNAL_DEPS,
+        rocksdb_preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+        test_cc = test_cc,
+        test_name = test_name,
+    )
+    for test_name, test_cc, parallelism in ROCKS_TESTS
+    if not is_opt_mode
+]
@@ -43,7 +43,7 @@ We plan to use this port for our business purposes here at Bing and this provide

 * Certain headers that are not present and not necessary on Windows were simply `#ifndef OS_WIN` in a few places (`unistd.h`)
 * All posix specific headers were replaced to port/port.h which worked well
-* Replaced `dirent.h` for `port/dirent.h` (very few places) with the implementation of the relevant interfaces within `rocksdb::port` namespace
+* Replaced `dirent.h` for `port/port_dirent.h` (very few places) with the implementation of the relevant interfaces within `rocksdb::port` namespace
 * Replaced `sys/time.h` to `port/sys_time.h` (few places) implemented equivalents within `rocksdb::port`
 * `printf %z` specification is not supported on Windows. To imitate existing standards we came up with a string macro `ROCKSDB_PRIszt` which expands to `zu` on posix systems and to `Iu` on windows.
 * in class member initialization were moved to a __ctors in some cases
@@ -3,7 +3,7 @@ image: Visual Studio 2017
 before_build:
 - md %APPVEYOR_BUILD_FOLDER%\build
 - cd %APPVEYOR_BUILD_FOLDER%\build
- cmake -G "Visual Studio 15 Win64" -DOPTDBG=1 -DWITH_XPRESS=1 -DPORTABLE=1 ..
+- cmake -G "Visual Studio 15 Win64" -DOPTDBG=1 -DWITH_XPRESS=1 -DPORTABLE=1 -DJNI=1 ..
 - cd ..
 build:
  project: build\rocksdb.sln
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
@@ -81,7 +82,7 @@ def get_tests(repo_path):
            else:
                # we consumed all the parallel tests
                break
-    
+
    return tests


@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 # Create a tmp directory for the test to use
 TEST_DIR=$(mktemp -d /dev/shm/fbcode_rocksdb_XXXXXXX)
 # shellcheck disable=SC2068
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
@@ -1,16 +1,15 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 from __future__ import unicode_literals
 rocksdb_target_header = """load("@fbcode_macros//build_defs:auto_headers.bzl", "AutoHeaders")
+load("@fbcode_macros//build_defs:cpp_library.bzl", "cpp_library")
+load(":defs.bzl", "test_binary")

 REPO_PATH = package_name() + "/"

-BUCK_BINS = "buck-out/gen/" + REPO_PATH
-
-TEST_RUNNER = REPO_PATH + "buckifier/rocks_test_runner.sh"
-
-rocksdb_compiler_flags = [
+ROCKSDB_COMPILER_FLAGS = [
    "-fno-builtin-memcmp",
    "-DROCKSDB_PLATFORM_POSIX",
    "-DROCKSDB_LIB_IO_POSIX",
@@ -26,6 +25,7 @@ rocksdb_compiler_flags = [
    "-DBZIP2",
    "-DLZ4",
    "-DZSTD",
+    "-DZSTD_STATIC_LINKING_ONLY",
    "-DGFLAGS=gflags",
    "-DNUMA",
    "-DTBB",
@@ -37,7 +37,7 @@ rocksdb_compiler_flags = [
    "-Wnarrowing",
 ]

-rocksdb_external_deps = [
+ROCKSDB_EXTERNAL_DEPS = [
    ("bzip2", None, "bz2"),
    ("snappy", None, "snappy"),
    ("zlib", None, "z"),
@@ -49,13 +49,13 @@ rocksdb_external_deps = [
    ("googletest", None, "gtest"),
 ]

-rocksdb_preprocessor_flags = [
+ROCKSDB_PREPROCESSOR_FLAGS = [
    # Directories with files for #include
    "-I" + REPO_PATH + "include/",
    "-I" + REPO_PATH,
 ]

-rocksdb_arch_preprocessor_flags = {
+ROCKSDB_ARCH_PREPROCESSOR_FLAGS = {
    "x86_64": [
        "-DHAVE_SSE42",
        "-DHAVE_PCLMUL",
@@ -68,18 +68,15 @@ is_opt_mode = build_mode.startswith("opt")

 # -DNDEBUG is added by default in opt mode in fbcode. But adding it twice
 # doesn't harm and avoid forgetting to add it.
-if is_opt_mode:
-    rocksdb_compiler_flags.append("-DNDEBUG")
-
-default_allocator = read_config("fbcode", "default_allocator")
+ROCKSDB_COMPILER_FLAGS += (["-DNDEBUG"] if is_opt_mode else [])

 sanitizer = read_config("fbcode", "sanitizer")

-# Let RocksDB aware of jemalloc existence.
-# Do not enable it if sanitizer presents.
-if is_opt_mode and default_allocator.startswith("jemalloc") and sanitizer == "":
-    rocksdb_compiler_flags.append("-DROCKSDB_JEMALLOC")
-    rocksdb_external_deps.append(("jemalloc", None, "headers"))
+# Do not enable jemalloc if sanitizer presents. RocksDB will further detect
+# whether the binary is linked with jemalloc at runtime.
+ROCKSDB_COMPILER_FLAGS += (["-DROCKSDB_JEMALLOC"] if sanitizer == "" else [])
+
+ROCKSDB_EXTERNAL_DEPS += ([("jemalloc", None, "headers")] if sanitizer == "" else [])
 """


@@ -88,11 +85,11 @@ cpp_library(
    name = "{name}",
    srcs = [{srcs}],
    {headers_attr_prefix}headers = {headers},
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [{deps}],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 """

@@ -100,11 +97,11 @@ binary_template = """
 cpp_binary(
    name = "%s",
    srcs = [%s],
-    arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    preprocessor_flags = rocksdb_preprocessor_flags,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
    deps = [%s],
-    external_deps = rocksdb_external_deps,
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 """

@@ -123,27 +120,17 @@ ROCKS_TESTS = [
 # Generate a test rule for each entry in ROCKS_TESTS
 # Do not build the tests in opt mode, since SyncPoint and other test code
 # will not be included.
-if not is_opt_mode:
-    for test_cfg in ROCKS_TESTS:
-        test_name = test_cfg[0]
-        test_cc = test_cfg[1]
-        ttype = "gtest" if test_cfg[2] == "parallel" else "simple"
-        test_bin = test_name + "_bin"
-
-        cpp_binary(
-            name = test_bin,
-            srcs = [test_cc],
-            arch_preprocessor_flags = rocksdb_arch_preprocessor_flags,
-            compiler_flags = rocksdb_compiler_flags,
-            preprocessor_flags = rocksdb_preprocessor_flags,
-            deps = [":rocksdb_test_lib"],
-            external_deps = rocksdb_external_deps,
-        )
-
-        custom_unittest(
-            name = test_name,
-            command = [TEST_RUNNER, BUCK_BINS + test_bin],
-            type = ttype,
-            deps = [":" + test_bin],
-        )
+[
+    test_binary(
+        parallelism = parallelism,
+        rocksdb_arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+        rocksdb_compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        rocksdb_external_deps = ROCKSDB_EXTERNAL_DEPS,
+        rocksdb_preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+        test_cc = test_cc,
+        test_name = test_name,
+    )
+    for test_name, test_cc, parallelism in ROCKS_TESTS
+    if not is_opt_mode
+]
 """
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 """
 This module keeps commonly used components.
 """
@@ -97,7 +97,7 @@ function getSteps($applyDiff, $diffID, $username, $test) {
  }

  // fbcode is a sub-repo. We cannot patch until we add it to ignore otherwise
-  // Git thinks it is an uncommited change.
+  // Git thinks it is an uncommitted change.
  $fix_git_ignore = array(
    "name" => "Fix git ignore",
    "shell" => "echo fbcode >> .git/info/exclude",
@@ -1,4 +1,5 @@
 #!/usr/bin/python
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 # amalgamate.py creates an amalgamation from a unity build.
 # It can be run with either Python 2 or 3.
@@ -53,11 +53,13 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
    FBCODE_BUILD="true"
    # If we're compiling with TSAN we need pic build
    PIC_BUILD=$COMPILE_WITH_TSAN
-    if [ -z "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
-      source "$PWD/build_tools/fbcode_config.sh"
-    else
+    if [ -n "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
      # we need this to build with MySQL. Don't use for other purposes.
      source "$PWD/build_tools/fbcode_config4.8.1.sh"
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_5xx" ]; then
+      source "$PWD/build_tools/fbcode_config.sh"
+    else
+      source "$PWD/build_tools/fbcode_config_platform007.sh"
    fi
 fi

@@ -232,7 +234,7 @@ else
          #include <linux/falloc.h>
          int main() {
      int fd = open("/dev/null", 0);
-      fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE, 0, 1024);
+      fallocate(fd, FALLOC_FL_KEEP_SIZE, 0, 1024);
          }
 EOF
        if [ "$?" = 0 ]; then
@@ -368,6 +370,16 @@ EOF
            # JEMALLOC can be enabled either using the flag (like here) or by
            # providing direct link to the jemalloc library
            WITH_JEMALLOC_FLAG=1
+            # check for JEMALLOC installed with HomeBrew
+            if [ "$PLATFORM" == "OS_MACOSX" ]; then
+                if hash brew 2>/dev/null && brew ls --versions jemalloc > /dev/null; then
+                    JEMALLOC_VER=$(brew ls --versions jemalloc | tail -n 1 | cut -f 2 -d ' ')
+                    JEMALLOC_INCLUDE="-I/usr/local/Cellar/jemalloc/${JEMALLOC_VER}/include"
+                    JEMALLOC_LIB="/usr/local/Cellar/jemalloc/${JEMALLOC_VER}/lib/libjemalloc_pic.a"
+                    PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS $JEMALLOC_LIB"
+                    JAVA_STATIC_LDFLAGS="$JAVA_STATIC_LDFLAGS $JEMALLOC_LIB"
+                fi
+            fi
        fi
    fi
    if ! test $JEMALLOC && ! test $ROCKSDB_DISABLE_TCMALLOC; then
@@ -506,8 +518,8 @@ if test "$USE_HDFS"; then
    echo "JAVA_HOME has to be set for HDFS usage."
    exit 1
  fi
-  HDFS_CCFLAGS="$HDFS_CCFLAGS -I$JAVA_HOME/include -I$JAVA_HOME/include/linux -DUSE_HDFS"
-  HDFS_LDFLAGS="$HDFS_LDFLAGS -lhdfs -L$JAVA_HOME/jre/lib/amd64"
+  HDFS_CCFLAGS="$HDFS_CCFLAGS -I$JAVA_HOME/include -I$JAVA_HOME/include/linux -DUSE_HDFS -I$HADOOP_HOME/include"
+  HDFS_LDFLAGS="$HDFS_LDFLAGS -lhdfs -L$JAVA_HOME/jre/lib/amd64 -L$HADOOP_HOME/lib/native"
  HDFS_LDFLAGS="$HDFS_LDFLAGS -L$JAVA_HOME/jre/lib/amd64/server -L$GLIBC_RUNTIME_PATH/lib"
  HDFS_LDFLAGS="$HDFS_LDFLAGS -ldl -lverify -ljava -ljvm"
  COMMON_FLAGS="$COMMON_FLAGS $HDFS_CCFLAGS"
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 GCC_BASE=/mnt/gvfs/third-party2/gcc/112ec378fec7002ad3e09afde022e656049f7191/5.x/centos7-native/c447969
 CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/04999bdb3ce81a11073535dcb00b5e13dc1cbaf5/stable/centos7-native/c9f9104
 LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/92b0c8e5c8eecc71eb042594ce1ab3413799b385/5.x/gcc-5-glibc-2.23/339d858
@@ -1,4 +1,5 @@
 # shellcheck disable=SC2148
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 GCC_BASE=/mnt/gvfs/third-party2/gcc/cf7d14c625ce30bae1a4661c2319c5a283e4dd22/4.8.1/centos6-native/cc6c9dc
 CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/8598c375b0e94e1448182eb3df034704144a838d/stable/centos6-native/3f16ddd
 LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/d6e0a7da6faba45f5e5b1638f9edd7afc2f34e7d/4.8.1/gcc-4.8.1-glibc-2.17/8aac7fc
@@ -0,0 +1,19 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+GCC_BASE=/mnt/gvfs/third-party2/gcc/6e8e715624fd15256a7970073387793dfcf79b46/7.x/centos7-native/b2ef2b6
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/ef37e1faa1c29782abfac1ae65a291b9b7966f6d/stable/centos7-native/c9f9104
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/c67031f0f739ac61575a061518d6ef5038f99f90/7.x/platform007/5620abc
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/60d6f124a78798b73944f5ba87c2306ae3460153/2.26/platform007/f259413
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/7f9bdaada18f59bc27ec2b0871eb8a6144343aef/1.1.3/platform007/ca4da3d
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/22c2d65676fb7c23cfa797c4f6937f38b026f3cf/1.2.8/platform007/ca4da3d
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/dc49a21c5fceec6456a7a28a94dcd16690af1337/1.0.6/platform007/ca4da3d
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/907b498203d297947f3bb70b9466f47e100f1873/r131/platform007/ca4da3d
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/3ee276cbacfad3074e3f07bf826ac47f06970f4e/1.3.5/platform007/15a3614
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/0b9929d2588991c65a57168bf88aff2db87c5d48/2.2.0/platform007/ca4da3d
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/9c910d36d6235cc40e8ff559358f1833452300ca/master/platform007/5b0f53e
+NUMA_BASE=/mnt/gvfs/third-party2/numa/9cbf2460284c669ed19c3ccb200a71f7dd7e53c7/2.0.11/platform007/ca4da3d
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/bf3d7497fe4e6d007354f0adffa16ce3003f8338/1.3/platform007/6f3e0a9
+TBB_BASE=/mnt/gvfs/third-party2/tbb/ff4e0b093534704d8abab678a4fd7f5ea7b094c7/2018_U5/platform007/ca4da3d
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/b5c4a61a5c483ba24722005ae07895971a2ac707/fb/platform007/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/92ff90349e2f43ea0a8246d8b1cf17b6869013e3/2.29.1/centos7-native/da39a3e
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/f3f697a28122e6bcd513273dd9c1ff23852fc59f/3.13.0/platform007/ca4da3d
+LUA_BASE=/mnt/gvfs/third-party2/lua/f0cd714433206d5139df61659eb7b28b1dea6683/5.3.4/platform007/5007832
@@ -1,2 +1,3 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 docker run -v $PWD:/rocks -w /rocks buildpack-deps make
@@ -132,11 +132,14 @@ _TEST_NAME_TO_PARSERS = {
    'lite': [CompilerErrorParser],
    'lite_test': [CompilerErrorParser, GTestErrorParser],
    'stress_crash': [CompilerErrorParser, DbCrashErrorParser],
+    'stress_crash_with_atomic_flush': [CompilerErrorParser, DbCrashErrorParser],
    'write_stress': [CompilerErrorParser, WriteStressErrorParser],
    'asan': [CompilerErrorParser, GTestErrorParser, AsanErrorParser],
    'asan_crash': [CompilerErrorParser, AsanErrorParser, DbCrashErrorParser],
+    'asan_crash_with_atomic_flush': [CompilerErrorParser, AsanErrorParser, DbCrashErrorParser],
    'ubsan': [CompilerErrorParser, GTestErrorParser, UbsanErrorParser],
    'ubsan_crash': [CompilerErrorParser, UbsanErrorParser, DbCrashErrorParser],
+    'ubsan_crash_with_atomic_flush': [CompilerErrorParser, UbsanErrorParser, DbCrashErrorParser],
    'valgrind': [CompilerErrorParser, GTestErrorParser, ValgrindErrorParser],
    'tsan': [CompilerErrorParser, GTestErrorParser, TsanErrorParser],
    'format_compatible': [CompilerErrorParser, CompatErrorParser],
@@ -1,5 +1,5 @@
 #!/bin/sh
-
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 # fail early
 set -e

@@ -1,4 +1,5 @@
 #!/bin/sh
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 #
 # Set environment variables so that we can compile rocksdb using
 # fbcode settings.  It uses the latest g++ and clang compilers and also
@@ -51,7 +52,7 @@ if test -z $PIC_BUILD; then
 else
  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
 fi
-CFLAGS+=" -DZSTD"
+CFLAGS+=" -DZSTD -DZSTD_STATIC_LINKING_ONLY"

 # location of gflags headers and libraries
 GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
@@ -1,4 +1,5 @@
 #!/bin/sh
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 #
 # Set environment variables so that we can compile rocksdb using
 # fbcode settings.  It uses the latest g++ compiler and also
@@ -0,0 +1,158 @@
+#!/bin/sh
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+#
+# Set environment variables so that we can compile rocksdb using
+# fbcode settings.  It uses the latest g++ and clang compilers and also
+# uses jemalloc
+# Environment variables that change the behavior of this script:
+# PIC_BUILD -- if true, it will only take pic versions of libraries from fbcode. libraries that don't have pic variant will not be included
+
+
+BASEDIR=`dirname $BASH_SOURCE`
+source "$BASEDIR/dependencies_platform007.sh"
+
+CFLAGS=""
+
+# libgcc
+LIBGCC_INCLUDE="$LIBGCC_BASE/include/c++/7.3.0"
+LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
+
+# glibc
+GLIBC_INCLUDE="$GLIBC_BASE/include"
+GLIBC_LIBS=" -L $GLIBC_BASE/lib"
+
+# snappy
+SNAPPY_INCLUDE=" -I $SNAPPY_BASE/include/"
+if test -z $PIC_BUILD; then
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy.a"
+else
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy_pic.a"
+fi
+CFLAGS+=" -DSNAPPY"
+
+if test -z $PIC_BUILD; then
+  # location of zlib headers and libraries
+  ZLIB_INCLUDE=" -I $ZLIB_BASE/include/"
+  ZLIB_LIBS=" $ZLIB_BASE/lib/libz.a"
+  CFLAGS+=" -DZLIB"
+
+  # location of bzip headers and libraries
+  BZIP_INCLUDE=" -I $BZIP2_BASE/include/"
+  BZIP_LIBS=" $BZIP2_BASE/lib/libbz2.a"
+  CFLAGS+=" -DBZIP2"
+
+  LZ4_INCLUDE=" -I $LZ4_BASE/include/"
+  LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
+  CFLAGS+=" -DLZ4"
+fi
+
+ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+if test -z $PIC_BUILD; then
+  ZSTD_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
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags.a"
+else
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags_pic.a"
+fi
+CFLAGS+=" -DGFLAGS=gflags"
+
+# location of jemalloc
+JEMALLOC_INCLUDE=" -I $JEMALLOC_BASE/include/"
+JEMALLOC_LIB=" $JEMALLOC_BASE/lib/libjemalloc.a"
+
+if test -z $PIC_BUILD; then
+  # location of numa
+  NUMA_INCLUDE=" -I $NUMA_BASE/include/"
+  NUMA_LIB=" $NUMA_BASE/lib/libnuma.a"
+  CFLAGS+=" -DNUMA"
+
+  # location of libunwind
+  LIBUNWIND="$LIBUNWIND_BASE/lib/libunwind.a"
+fi
+
+# location of TBB
+TBB_INCLUDE=" -isystem $TBB_BASE/include/"
+if test -z $PIC_BUILD; then
+  TBB_LIBS="$TBB_BASE/lib/libtbb.a"
+else
+  TBB_LIBS="$TBB_BASE/lib/libtbb_pic.a"
+fi
+CFLAGS+=" -DTBB"
+
+# use Intel SSE support for checksum calculations
+export USE_SSE=1
+export PORTABLE=1
+
+BINUTILS="$BINUTILS_BASE/bin"
+AR="$BINUTILS/ar"
+
+DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE"
+
+STDLIBS="-L $GCC_BASE/lib64"
+
+CLANG_BIN="$CLANG_BASE/bin"
+CLANG_LIB="$CLANG_BASE/lib"
+CLANG_SRC="$CLANG_BASE/../../src"
+
+CLANG_ANALYZER="$CLANG_BIN/clang++"
+CLANG_SCAN_BUILD="$CLANG_SRC/llvm/tools/clang/tools/scan-build/bin/scan-build"
+
+if [ -z "$USE_CLANG" ]; then
+  # gcc
+  CC="$GCC_BASE/bin/gcc"
+  CXX="$GCC_BASE/bin/g++"
+
+  CFLAGS+=" -B$BINUTILS/gold"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  JEMALLOC=1
+else
+  # clang
+  CLANG_INCLUDE="$CLANG_LIB/clang/stable/include"
+  CC="$CLANG_BIN/clang"
+  CXX="$CLANG_BIN/clang++"
+
+  KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
+
+  CFLAGS+=" -B$BINUTILS/gold -nostdinc -nostdlib"
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/7.x/x86_64-facebook-linux "
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $CLANG_INCLUDE"
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE/linux "
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE "
+  CFLAGS+=" -Wno-expansion-to-defined "
+  CXXFLAGS="-nostdinc++"
+fi
+
+CFLAGS+=" $DEPS_INCLUDE"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_SUPPORT_THREAD_LOCAL -DHAVE_SSE42"
+CXXFLAGS+=" $CFLAGS"
+
+EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
+EXEC_LDFLAGS+=" -B$BINUTILS/gold"
+EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/platform007/lib/ld.so"
+EXEC_LDFLAGS+=" $LIBUNWIND"
+EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/platform007/lib"
+# required by libtbb
+EXEC_LDFLAGS+=" -ldl"
+
+PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
+
+EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS"
+
+VALGRIND_VER="$VALGRIND_BASE/bin/"
+
+# lua not supported because it's on track for deprecation, I think
+LUA_PATH=
+LUA_LIB=
+
+export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD LUA_PATH LUA_LIB
@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 # If clang_format_diff.py command is not specfied, we assume we are able to
 # access directly without any path.
 if [ -z $CLANG_FORMAT_DIFF ]
@@ -1,5 +1,6 @@
 # shellcheck disable=SC1113
 #/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 set -e

@@ -1,4 +1,5 @@
 #!/usr/local/fbcode/gcc-4.9-glibc-2.20-fb/bin/python2.7
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 from __future__ import absolute_import
 from __future__ import division
@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 set -e

@@ -85,7 +85,9 @@ NON_SHM="TMPD=/tmp/rocksdb_test_tmp"
 GCC_481="ROCKSDB_FBCODE_BUILD_WITH_481=1"
 ASAN="COMPILE_WITH_ASAN=1"
 CLANG="USE_CLANG=1"
-TSAN="COMPILE_WITH_TSAN=1"
+# in gcc-5 there are known problems with TSAN like https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71090.
+# using platform007 gives us gcc-8 or higher which has that bug fixed.
+TSAN="ROCKSDB_FBCODE_BUILD_WITH_PLATFORM007=1 COMPILE_WITH_TSAN=1"
 UBSAN="COMPILE_WITH_UBSAN=1"
 TSAN_CRASH='CRASH_TEST_EXT_ARGS="--compression_type=zstd --log2_keys_per_lock=22"'
 NON_TSAN_CRASH="CRASH_TEST_EXT_ARGS=--compression_type=zstd"
@@ -398,6 +400,35 @@ STRESS_CRASH_TEST_COMMANDS="[
    }
 ]"

+#
+# RocksDB stress/crash test with atomic flush
+#
+STRESS_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS="[
+    {
+        'name':'Rocksdb Stress/Crash Test (atomic flush)',
+        'oncall':'$ONCALL',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug stress tests',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            {
+                'name':'Build and run RocksDB debug crash tests with atomic flush',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 crash_test_with_atomic_flush || $CONTRUN_NAME=crash_test_with_atomic_flush $TASK_CREATION_TOOL',
+                '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
@@ -462,6 +493,28 @@ ASAN_CRASH_TEST_COMMANDS="[
    }
 ]"

+#
+# RocksDB crash testing with atomic flush under address sanitizer
+#
+ASAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS="[
+    {
+        'name':'Rocksdb crash test (atomic flush) under ASAN',
+        'oncall':'$ONCALL',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug asan_crash_test_with_atomic_flush',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 asan_crash_test_with_atomic_flush || $CONTRUN_NAME=asan_crash_test_with_atomic_flush $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB test under undefined behavior sanitizer
 #
@@ -504,6 +557,28 @@ UBSAN_CRASH_TEST_COMMANDS="[
    }
 ]"

+#
+# RocksDB crash testing with atomic flush under undefined behavior sanitizer
+#
+UBSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS="[
+    {
+        'name':'Rocksdb crash test (atomic flush) under UBSAN',
+        'oncall':'$ONCALL',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug ubsan_crash_test_with_atomic_flush',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 ubsan_crash_test_with_atomic_flush || $CONTRUN_NAME=ubsan_crash_test_with_atomic_flush $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB unit test under valgrind
 #
@@ -570,6 +645,28 @@ TSAN_CRASH_TEST_COMMANDS="[
    }
 ]"

+#
+# RocksDB crash test with atomic flush under TSAN
+#
+TSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS="[
+    {
+        'name':'Rocksdb Crash Test with atomic flush under TSAN',
+        'oncall':'$ONCALL',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Compile and run',
+                'timeout': 86400,
+                'shell':'set -o pipefail && $SHM $DEBUG $TSAN $TSAN_CRASH CRASH_TEST_KILL_ODD=1887 make J=1 crash_test_with_atomic_flush || $CONTRUN_NAME=tsan_crash_test_with_atomic_flush $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB format compatible
 #
@@ -609,7 +706,7 @@ run_no_compression()
  rm -rf /dev/shm/rocksdb
  mkdir /dev/shm/rocksdb
  make clean
-  cat build_tools/fbcode_config.sh | grep -iv dzlib | grep -iv dlz4 | grep -iv dsnappy | grep -iv dbzip2 > .tmp.fbcode_config.sh
+  cat build_tools/fbcode_config.sh | grep -iv dzstd | grep -iv dzlib | grep -iv dlz4 | grep -iv dsnappy | grep -iv dbzip2 > .tmp.fbcode_config.sh
  mv .tmp.fbcode_config.sh build_tools/fbcode_config.sh
  cat Makefile | grep -v tools/ldb_test.py > .tmp.Makefile
  mv .tmp.Makefile Makefile
@@ -751,6 +848,9 @@ case $1 in
  stress_crash)
    echo $STRESS_CRASH_TEST_COMMANDS
    ;;
+  stress_crash_with_atomic_flush)
+    echo $STRESS_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
+    ;;
  write_stress)
    echo $WRITE_STRESS_COMMANDS
    ;;
@@ -760,12 +860,18 @@ case $1 in
  asan_crash)
    echo $ASAN_CRASH_TEST_COMMANDS
    ;;
+  asan_crash_with_atomic_flush)
+    echo $ASAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
+    ;;
  ubsan)
    echo $UBSAN_TEST_COMMANDS
    ;;
  ubsan_crash)
    echo $UBSAN_CRASH_TEST_COMMANDS
    ;;
+  ubsan_crash_with_atomic_flush)
+    echo $UBSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
+    ;;
  valgrind)
    echo $VALGRIND_TEST_COMMANDS
    ;;
@@ -775,6 +881,9 @@ case $1 in
  tsan_crash)
    echo $TSAN_CRASH_TEST_COMMANDS
    ;;
+  tsan_crash_with_atomic_flush)
+    echo $TSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
+    ;;
  format_compatible)
    echo $FORMAT_COMPATIBLE_COMMANDS
    ;;
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 # This script enables you running RocksDB tests by running
 # All the tests concurrently and utilizing all the cores
 Param(
@@ -1,4 +1,5 @@
 #!/bin/bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 set -e

 ROCKSDB_VERSION="5.10.3"
@@ -1,4 +1,5 @@
 #!/bin/sh
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 #
 # Update dependencies.sh file with the latest avaliable versions

@@ -53,6 +54,45 @@ function get_lib_base()
  log_variable $__res_var
 }

+###########################################################
+#                platform007 dependencies                 #
+###########################################################
+
+OUTPUT="$BASEDIR/dependencies_platform007.sh"
+
+rm -f "$OUTPUT"
+touch "$OUTPUT"
+
+echo "Writing dependencies to $OUTPUT"
+
+# Compilers locations
+GCC_BASE=`readlink -f $TP2_LATEST/gcc/7.x/centos7-native/*/`
+CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos7-native/*/`
+
+log_variable GCC_BASE
+log_variable CLANG_BASE
+
+# Libraries locations
+get_lib_base libgcc     7.x     platform007
+get_lib_base glibc      2.26    platform007
+get_lib_base snappy     LATEST  platform007
+get_lib_base zlib       LATEST  platform007
+get_lib_base bzip2      LATEST  platform007
+get_lib_base lz4        LATEST  platform007
+get_lib_base zstd       LATEST  platform007
+get_lib_base gflags     LATEST  platform007
+get_lib_base jemalloc   LATEST  platform007
+get_lib_base numa       LATEST  platform007
+get_lib_base libunwind  LATEST  platform007
+get_lib_base tbb        LATEST  platform007
+
+get_lib_base kernel-headers fb platform007
+get_lib_base binutils   LATEST centos7-native
+get_lib_base valgrind   LATEST platform007
+get_lib_base lua        5.3.4  platform007
+
+git diff $OUTPUT
+
 ###########################################################
 #                   5.x dependencies                      #
 ###########################################################
@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 if [ "$#" = "0" ]; then
  echo "Usage: $0 major|minor|patch|full"
  exit 1
@@ -73,8 +73,7 @@ class CacheTest : public testing::TestWithParam<std::string> {
    current_ = this;
  }

-  ~CacheTest() {
-  }
+  ~CacheTest() override {}

  std::shared_ptr<Cache> NewCache(size_t capacity) {
    auto type = GetParam();
@@ -99,7 +98,7 @@ class CacheTest : public testing::TestWithParam<std::string> {
    return nullptr;
  }

-  int Lookup(shared_ptr<Cache> cache, int key) {
+  int Lookup(std::shared_ptr<Cache> cache, int key) {
    Cache::Handle* handle = cache->Lookup(EncodeKey(key));
    const int r = (handle == nullptr) ? -1 : DecodeValue(cache->Value(handle));
    if (handle != nullptr) {
@@ -108,16 +107,16 @@ class CacheTest : public testing::TestWithParam<std::string> {
    return r;
  }

-  void Insert(shared_ptr<Cache> cache, int key, int value, int charge = 1) {
+  void Insert(std::shared_ptr<Cache> cache, int key, int value,
+              int charge = 1) {
    cache->Insert(EncodeKey(key), EncodeValue(value), charge,
                  &CacheTest::Deleter);
  }

-  void Erase(shared_ptr<Cache> cache, int key) {
+  void Erase(std::shared_ptr<Cache> cache, int key) {
    cache->Erase(EncodeKey(key));
  }

-
  int Lookup(int key) {
    return Lookup(cache_, key);
  }
@@ -307,7 +306,7 @@ TEST_P(CacheTest, EvictionPolicy) {
  Insert(200, 201);

  // Frequently used entry must be kept around
-  for (int i = 0; i < kCacheSize + 100; i++) {
+  for (int i = 0; i < kCacheSize + 200; i++) {
    Insert(1000+i, 2000+i);
    ASSERT_EQ(101, Lookup(100));
  }
@@ -360,7 +359,7 @@ TEST_P(CacheTest, EvictionPolicyRef) {
  Insert(303, 104);

  // Insert entries much more than Cache capacity
-  for (int i = 0; i < kCacheSize + 100; i++) {
+  for (int i = 0; i < kCacheSize + 200; i++) {
    Insert(1000 + i, 2000 + i);
  }

@@ -688,7 +687,8 @@ TEST_P(CacheTest, DefaultShardBits) {
 }

 #ifdef SUPPORT_CLOCK_CACHE
-shared_ptr<Cache> (*new_clock_cache_func)(size_t, int, bool) = NewClockCache;
+std::shared_ptr<Cache> (*new_clock_cache_func)(size_t, int,
+                                               bool) = NewClockCache;
 INSTANTIATE_TEST_CASE_P(CacheTestInstance, CacheTest,
                        testing::Values(kLRU, kClock));
 #else
@@ -234,38 +234,35 @@ struct CleanupContext {
 };

 // A cache shard which maintains its own CLOCK cache.
-class ClockCacheShard : public CacheShard {
+class ClockCacheShard final : public CacheShard {
 public:
  // Hash map type.
  typedef tbb::concurrent_hash_map<CacheKey, CacheHandle*, CacheKey> HashTable;

  ClockCacheShard();
-  ~ClockCacheShard();
+  ~ClockCacheShard() override;

  // Interfaces
-  virtual void SetCapacity(size_t capacity) override;
-  virtual void SetStrictCapacityLimit(bool strict_capacity_limit) override;
-  virtual Status Insert(const Slice& key, uint32_t hash, void* value,
-                        size_t charge,
-                        void (*deleter)(const Slice& key, void* value),
-                        Cache::Handle** handle,
-                        Cache::Priority priority) override;
-  virtual Cache::Handle* Lookup(const Slice& key, uint32_t hash) override;
+  void SetCapacity(size_t capacity) override;
+  void SetStrictCapacityLimit(bool strict_capacity_limit) override;
+  Status Insert(const Slice& key, uint32_t hash, void* value, size_t charge,
+                void (*deleter)(const Slice& key, void* value),
+                Cache::Handle** handle, Cache::Priority priority) override;
+  Cache::Handle* Lookup(const Slice& key, uint32_t hash) override;
  // If the entry in in cache, increase reference count and return true.
  // Return false otherwise.
  //
  // Not necessary to hold mutex_ before being called.
-  virtual bool Ref(Cache::Handle* handle) override;
-  virtual bool Release(Cache::Handle* handle,
-                       bool force_erase = false) override;
-  virtual void Erase(const Slice& key, uint32_t hash) override;
+  bool Ref(Cache::Handle* handle) override;
+  bool Release(Cache::Handle* handle, bool force_erase = false) override;
+  void Erase(const Slice& key, uint32_t hash) override;
  bool EraseAndConfirm(const Slice& key, uint32_t hash,
                       CleanupContext* context);
-  virtual size_t GetUsage() const override;
-  virtual size_t GetPinnedUsage() const override;
-  virtual void EraseUnRefEntries() override;
-  virtual void ApplyToAllCacheEntries(void (*callback)(void*, size_t),
-                                      bool thread_safe) override;
+  size_t GetUsage() const override;
+  size_t GetPinnedUsage() const override;
+  void EraseUnRefEntries() override;
+  void ApplyToAllCacheEntries(void (*callback)(void*, size_t),
+                              bool thread_safe) override;

 private:
  static const uint32_t kInCacheBit = 1;
@@ -675,7 +672,7 @@ void ClockCacheShard::EraseUnRefEntries() {
  Cleanup(context);
 }

-class ClockCache : public ShardedCache {
+class ClockCache final : public ShardedCache {
 public:
  ClockCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit)
      : ShardedCache(capacity, num_shard_bits, strict_capacity_limit) {
@@ -685,31 +682,31 @@ class ClockCache : public ShardedCache {
    SetStrictCapacityLimit(strict_capacity_limit);
  }

-  virtual ~ClockCache() { delete[] shards_; }
+  ~ClockCache() override { delete[] shards_; }

-  virtual const char* Name() const override { return "ClockCache"; }
+  const char* Name() const override { return "ClockCache"; }

-  virtual CacheShard* GetShard(int shard) override {
+  CacheShard* GetShard(int shard) override {
    return reinterpret_cast<CacheShard*>(&shards_[shard]);
  }

-  virtual const CacheShard* GetShard(int shard) const override {
+  const CacheShard* GetShard(int shard) const override {
    return reinterpret_cast<CacheShard*>(&shards_[shard]);
  }

-  virtual void* Value(Handle* handle) override {
+  void* Value(Handle* handle) override {
    return reinterpret_cast<const CacheHandle*>(handle)->value;
  }

-  virtual size_t GetCharge(Handle* handle) const override {
+  size_t GetCharge(Handle* handle) const override {
    return reinterpret_cast<const CacheHandle*>(handle)->charge;
  }

-  virtual uint32_t GetHash(Handle* handle) const override {
+  uint32_t GetHash(Handle* handle) const override {
    return reinterpret_cast<const CacheHandle*>(handle)->hash;
  }

-  virtual void DisownData() override { shards_ = nullptr; }
+  void DisownData() override { shards_ = nullptr; }

 private:
  ClockCacheShard* shards_;
@@ -100,14 +100,16 @@ void LRUHandleTable::Resize() {
 }

 LRUCacheShard::LRUCacheShard(size_t capacity, bool strict_capacity_limit,
-                             double high_pri_pool_ratio)
+                             double high_pri_pool_ratio,
+                             bool use_adaptive_mutex)
    : capacity_(0),
      high_pri_pool_usage_(0),
      strict_capacity_limit_(strict_capacity_limit),
      high_pri_pool_ratio_(high_pri_pool_ratio),
      high_pri_pool_capacity_(0),
      usage_(0),
-      lru_usage_(0) {
+      lru_usage_(0),
+      mutex_(use_adaptive_mutex) {
  // Make empty circular linked list
  lru_.next = &lru_;
  lru_.prev = &lru_;
@@ -462,7 +464,8 @@ std::string LRUCacheShard::GetPrintableOptions() const {

 LRUCache::LRUCache(size_t capacity, int num_shard_bits,
                   bool strict_capacity_limit, double high_pri_pool_ratio,
-                   std::shared_ptr<MemoryAllocator> allocator)
+                   std::shared_ptr<MemoryAllocator> allocator,
+                   bool use_adaptive_mutex)
    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit,
                   std::move(allocator)) {
  num_shards_ = 1 << num_shard_bits;
@@ -471,7 +474,8 @@ LRUCache::LRUCache(size_t capacity, int num_shard_bits,
  size_t per_shard = (capacity + (num_shards_ - 1)) / num_shards_;
  for (int i = 0; i < num_shards_; i++) {
    new (&shards_[i])
-        LRUCacheShard(per_shard, strict_capacity_limit, high_pri_pool_ratio);
+        LRUCacheShard(per_shard, strict_capacity_limit, high_pri_pool_ratio,
+            use_adaptive_mutex);
  }
 }

@@ -540,13 +544,15 @@ std::shared_ptr<Cache> NewLRUCache(const LRUCacheOptions& cache_opts) {
  return NewLRUCache(cache_opts.capacity, cache_opts.num_shard_bits,
                     cache_opts.strict_capacity_limit,
                     cache_opts.high_pri_pool_ratio,
-                     cache_opts.memory_allocator);
+                     cache_opts.memory_allocator,
+                     cache_opts.use_adaptive_mutex);
 }

 std::shared_ptr<Cache> NewLRUCache(
    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
    double high_pri_pool_ratio,
-    std::shared_ptr<MemoryAllocator> memory_allocator) {
+    std::shared_ptr<MemoryAllocator> memory_allocator,
+    bool use_adaptive_mutex) {
  if (num_shard_bits >= 20) {
    return nullptr;  // the cache cannot be sharded into too many fine pieces
  }
@@ -559,7 +565,8 @@ std::shared_ptr<Cache> NewLRUCache(
  }
  return std::make_shared<LRUCache>(capacity, num_shard_bits,
                                    strict_capacity_limit, high_pri_pool_ratio,
-                                    std::move(memory_allocator));
+                                    std::move(memory_allocator),
+                                    use_adaptive_mutex);
 }

 }  // namespace rocksdb
@@ -55,10 +55,18 @@ struct LRUHandle {
                     // cache itself is counted as 1

  // Include the following flags:
-  //   in_cache:    whether this entry is referenced by the hash table.
-  //   is_high_pri: whether this entry is high priority entry.
-  //   in_high_pri_pool: whether this entry is in high-pri pool.
-  char flags;
+  //   IN_CACHE:         whether this entry is referenced by the hash table.
+  //   IS_HIGH_PRI:      whether this entry is high priority entry.
+  //   IN_HIGH_PRI_POOL: whether this entry is in high-pri pool.
+  //   HAS_HIT:          whether this entry has had any lookups (hits).
+  enum Flags : uint8_t {
+    IN_CACHE = (1 << 0),
+    IS_HIGH_PRI = (1 << 1),
+    IN_HIGH_PRI_POOL = (1 << 2),
+    HAS_HIT = (1 << 3),
+  };
+
+  uint8_t flags;

  uint32_t hash;     // Hash of key(); used for fast sharding and comparisons

@@ -74,36 +82,36 @@ struct LRUHandle {
    }
  }

-  bool InCache() { return flags & 1; }
-  bool IsHighPri() { return flags & 2; }
-  bool InHighPriPool() { return flags & 4; }
-  bool HasHit() { return flags & 8; }
+  bool InCache() const { return flags & IN_CACHE; }
+  bool IsHighPri() const { return flags & IS_HIGH_PRI; }
+  bool InHighPriPool() const { return flags & IN_HIGH_PRI_POOL; }
+  bool HasHit() const { return flags & HAS_HIT; }

  void SetInCache(bool in_cache) {
    if (in_cache) {
-      flags |= 1;
+      flags |= IN_CACHE;
    } else {
-      flags &= ~1;
+      flags &= ~IN_CACHE;
    }
  }

  void SetPriority(Cache::Priority priority) {
    if (priority == Cache::Priority::HIGH) {
-      flags |= 2;
+      flags |= IS_HIGH_PRI;
    } else {
-      flags &= ~2;
+      flags &= ~IS_HIGH_PRI;
    }
  }

  void SetInHighPriPool(bool in_high_pri_pool) {
    if (in_high_pri_pool) {
-      flags |= 4;
+      flags |= IN_HIGH_PRI_POOL;
    } else {
-      flags &= ~4;
+      flags &= ~IN_HIGH_PRI_POOL;
    }
  }

-  void SetHit() { flags |= 8; }
+  void SetHit() { flags |= HAS_HIT; }

  void Free() {
    assert((refs == 1 && InCache()) || (refs == 0 && !InCache()));
@@ -157,10 +165,10 @@ class LRUHandleTable {
 };

 // A single shard of sharded cache.
-class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard : public CacheShard {
+class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShard {
 public:
  LRUCacheShard(size_t capacity, bool strict_capacity_limit,
-                double high_pri_pool_ratio);
+                double high_pri_pool_ratio, bool use_adaptive_mutex);
  virtual ~LRUCacheShard();

  // Separate from constructor so caller can easily make an array of LRUCache
@@ -276,11 +284,16 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard : public CacheShard {
  mutable port::Mutex mutex_;
 };

-class LRUCache : public ShardedCache {
+class LRUCache
+#ifdef NDEBUG
+    final
+#endif
+    : public ShardedCache {
 public:
  LRUCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit,
           double high_pri_pool_ratio,
-           std::shared_ptr<MemoryAllocator> memory_allocator = nullptr);
+           std::shared_ptr<MemoryAllocator> memory_allocator = nullptr,
+           bool use_adaptive_mutex = kDefaultToAdaptiveMutex);
  virtual ~LRUCache();
  virtual const char* Name() const override { return "LRUCache"; }
  virtual CacheShard* GetShard(int shard) override;
@@ -15,7 +15,7 @@ namespace rocksdb {
 class LRUCacheTest : public testing::Test {
 public:
  LRUCacheTest() {}
-  ~LRUCacheTest() { DeleteCache(); }
+  ~LRUCacheTest() override { DeleteCache(); }

  void DeleteCache() {
    if (cache_ != nullptr) {
@@ -25,12 +25,13 @@ class LRUCacheTest : public testing::Test {
    }
  }

-  void NewCache(size_t capacity, double high_pri_pool_ratio = 0.0) {
+  void NewCache(size_t capacity, double high_pri_pool_ratio = 0.0,
+                bool use_adaptive_mutex = kDefaultToAdaptiveMutex) {
    DeleteCache();
    cache_ = reinterpret_cast<LRUCacheShard*>(
        port::cacheline_aligned_alloc(sizeof(LRUCacheShard)));
    new (cache_) LRUCacheShard(capacity, false /*strict_capcity_limit*/,
-                               high_pri_pool_ratio);
+                               high_pri_pool_ratio, use_adaptive_mutex);
  }

  void Insert(const std::string& key,
@@ -83,7 +83,7 @@ class ShardedCache : public Cache {

 private:
  static inline uint32_t HashSlice(const Slice& s) {
-    return Hash(s.data(), s.size(), 0);
+    return static_cast<uint32_t>(GetSliceNPHash64(s));
  }

  uint32_t Shard(uint32_t hash) {
@@ -1,21 +0,0 @@
-# - Find zlib
-# Find the zlib compression library and includes
-#
-# ZLIB_INCLUDE_DIR - where to find zlib.h, etc.
-# ZLIB_LIBRARIES - List of libraries when using zlib.
-# ZLIB_FOUND - True if zlib found.
-
-find_path(ZLIB_INCLUDE_DIR
-  NAMES zlib.h
-  HINTS ${ZLIB_ROOT_DIR}/include)
-
-find_library(ZLIB_LIBRARIES
-  NAMES z
-  HINTS ${ZLIB_ROOT_DIR}/lib)
-
-include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(zlib DEFAULT_MSG ZLIB_LIBRARIES ZLIB_INCLUDE_DIR)
-
-mark_as_advanced(
-  ZLIB_LIBRARIES
-  ZLIB_INCLUDE_DIR)
@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 # Exit on error.
 set -e
@@ -1,3 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 import re
 import sys

@@ -18,6 +18,7 @@
 #include "db/event_helpers.h"
 #include "db/internal_stats.h"
 #include "db/merge_helper.h"
+#include "db/range_del_aggregator.h"
 #include "db/table_cache.h"
 #include "db/version_edit.h"
 #include "monitoring/iostats_context_imp.h"
@@ -46,18 +47,20 @@ TableBuilder* NewTableBuilder(
        int_tbl_prop_collector_factories,
    uint32_t column_family_id, const std::string& column_family_name,
    WritableFileWriter* file, const CompressionType compression_type,
-    const CompressionOptions& compression_opts, int level,
-    const std::string* compression_dict, const bool skip_filters,
-    const uint64_t creation_time, const uint64_t oldest_key_time) {
+    uint64_t sample_for_compression, const CompressionOptions& compression_opts,
+    int level, const bool skip_filters, const uint64_t creation_time,
+    const uint64_t oldest_key_time, const uint64_t target_file_size,
+    const uint64_t file_creation_time) {
  assert((column_family_id ==
          TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
         column_family_name.empty());
  return ioptions.table_factory->NewTableBuilder(
      TableBuilderOptions(ioptions, moptions, internal_comparator,
                          int_tbl_prop_collector_factories, compression_type,
-                          compression_opts, compression_dict, skip_filters,
-                          column_family_name, level, creation_time,
-                          oldest_key_time),
+                          sample_for_compression, compression_opts,
+                          skip_filters, column_family_name, level,
+                          creation_time, oldest_key_time, target_file_size,
+                          file_creation_time),
      column_family_id, file);
 }

@@ -65,19 +68,21 @@ Status BuildTable(
    const std::string& dbname, Env* env, const ImmutableCFOptions& ioptions,
    const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
    TableCache* table_cache, InternalIterator* iter,
-    std::unique_ptr<InternalIterator> range_del_iter, FileMetaData* meta,
-    const InternalKeyComparator& internal_comparator,
+    std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
+        range_del_iters,
+    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
    const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
        int_tbl_prop_collector_factories,
    uint32_t column_family_id, const std::string& column_family_name,
    std::vector<SequenceNumber> snapshots,
    SequenceNumber earliest_write_conflict_snapshot,
    SnapshotChecker* snapshot_checker, const CompressionType compression,
-    const CompressionOptions& compression_opts, bool paranoid_file_checks,
-    InternalStats* internal_stats, TableFileCreationReason reason,
-    EventLogger* event_logger, int job_id, const Env::IOPriority io_priority,
-    TableProperties* table_properties, int level, const uint64_t creation_time,
-    const uint64_t oldest_key_time, Env::WriteLifeTimeHint write_hint) {
+    uint64_t sample_for_compression, const CompressionOptions& compression_opts,
+    bool paranoid_file_checks, InternalStats* internal_stats,
+    TableFileCreationReason reason, EventLogger* event_logger, int job_id,
+    const Env::IOPriority io_priority, TableProperties* table_properties,
+    int level, const uint64_t creation_time, const uint64_t oldest_key_time,
+    Env::WriteLifeTimeHint write_hint, const uint64_t file_creation_time) {
  assert((column_family_id ==
          TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
         column_family_name.empty());
@@ -86,12 +91,10 @@ Status BuildTable(
  Status s;
  meta->fd.file_size = 0;
  iter->SeekToFirst();
-  std::unique_ptr<RangeDelAggregator> range_del_agg(
-      new RangeDelAggregator(internal_comparator, snapshots));
-  s = range_del_agg->AddTombstones(std::move(range_del_iter));
-  if (!s.ok()) {
-    // may be non-ok if a range tombstone key is unparsable
-    return s;
+  std::unique_ptr<CompactionRangeDelAggregator> range_del_agg(
+      new CompactionRangeDelAggregator(&internal_comparator, snapshots));
+  for (auto& range_del_iter : range_del_iters) {
+    range_del_agg->AddTombstones(std::move(range_del_iter));
  }

  std::string fname = TableFileName(ioptions.cf_paths, meta->fd.GetNumber(),
@@ -105,6 +108,11 @@ Status BuildTable(
  if (iter->Valid() || !range_del_agg->IsEmpty()) {
    TableBuilder* builder;
    std::unique_ptr<WritableFileWriter> file_writer;
+    // Currently we only enable dictionary compression during compaction to the
+    // bottommost level.
+    CompressionOptions compression_opts_for_flush(compression_opts);
+    compression_opts_for_flush.max_dict_bytes = 0;
+    compression_opts_for_flush.zstd_max_train_bytes = 0;
    {
      std::unique_ptr<WritableFile> file;
 #ifndef NDEBUG
@@ -121,15 +129,16 @@ Status BuildTable(
      file->SetIOPriority(io_priority);
      file->SetWriteLifeTimeHint(write_hint);

-      file_writer.reset(new WritableFileWriter(std::move(file), fname,
-                                               env_options, ioptions.statistics,
-                                               ioptions.listeners));
+      file_writer.reset(
+          new WritableFileWriter(std::move(file), fname, env_options, env,
+                                 ioptions.statistics, ioptions.listeners));
      builder = NewTableBuilder(
          ioptions, mutable_cf_options, internal_comparator,
          int_tbl_prop_collector_factories, column_family_id,
-          column_family_name, file_writer.get(), compression, compression_opts,
-          level, nullptr /* compression_dict */, false /* skip_filters */,
-          creation_time, oldest_key_time);
+          column_family_name, file_writer.get(), compression,
+          sample_for_compression, compression_opts_for_flush, level,
+          false /* skip_filters */, creation_time, oldest_key_time,
+          0 /*target_file_size*/, file_creation_time);
    }

    MergeHelper merge(env, internal_comparator.user_comparator(),
@@ -158,8 +167,10 @@ Status BuildTable(
      }
    }

-    for (auto it = range_del_agg->NewIterator(); it->Valid(); it->Next()) {
-      auto tombstone = it->Tombstone();
+    auto range_del_it = range_del_agg->NewIterator();
+    for (range_del_it->SeekToFirst(); range_del_it->Valid();
+         range_del_it->Next()) {
+      auto tombstone = range_del_it->Tombstone();
      auto kv = tombstone.Serialize();
      builder->Add(kv.first.Encode(), kv.second);
      meta->UpdateBoundariesForRange(kv.first, tombstone.SerializeEndKey(),
@@ -9,6 +9,7 @@
 #include <string>
 #include <utility>
 #include <vector>
+#include "db/range_tombstone_fragmenter.h"
 #include "db/table_properties_collector.h"
 #include "options/cf_options.h"
 #include "rocksdb/comparator.h"
@@ -39,8 +40,6 @@ class InternalStats;
 // @param column_family_name Name of the column family that is also identified
 //    by column_family_id, or empty string if unknown. It must outlive the
 //    TableBuilder returned by this function.
-// @param compression_dict Data for presetting the compression library's
-//    dictionary, or nullptr.
 TableBuilder* NewTableBuilder(
    const ImmutableCFOptions& options, const MutableCFOptions& moptions,
    const InternalKeyComparator& internal_comparator,
@@ -48,10 +47,11 @@ TableBuilder* NewTableBuilder(
        int_tbl_prop_collector_factories,
    uint32_t column_family_id, const std::string& column_family_name,
    WritableFileWriter* file, const CompressionType compression_type,
+    const uint64_t sample_for_compression,
    const CompressionOptions& compression_opts, int level,
-    const std::string* compression_dict = nullptr,
    const bool skip_filters = false, const uint64_t creation_time = 0,
-    const uint64_t oldest_key_time = 0);
+    const uint64_t oldest_key_time = 0, const uint64_t target_file_size = 0,
+    const uint64_t file_creation_time = 0);

 // Build a Table file from the contents of *iter.  The generated file
 // will be named according to number specified in meta. On success, the rest of
@@ -65,20 +65,23 @@ extern Status BuildTable(
    const std::string& dbname, Env* env, const ImmutableCFOptions& options,
    const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
    TableCache* table_cache, InternalIterator* iter,
-    std::unique_ptr<InternalIterator> range_del_iter, FileMetaData* meta,
-    const InternalKeyComparator& internal_comparator,
+    std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
+        range_del_iters,
+    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
    const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
        int_tbl_prop_collector_factories,
    uint32_t column_family_id, const std::string& column_family_name,
    std::vector<SequenceNumber> snapshots,
    SequenceNumber earliest_write_conflict_snapshot,
    SnapshotChecker* snapshot_checker, const CompressionType compression,
+    const uint64_t sample_for_compression,
    const CompressionOptions& compression_opts, bool paranoid_file_checks,
    InternalStats* internal_stats, TableFileCreationReason reason,
    EventLogger* event_logger = nullptr, int job_id = 0,
    const Env::IOPriority io_priority = Env::IO_HIGH,
    TableProperties* table_properties = nullptr, int level = -1,
    const uint64_t creation_time = 0, const uint64_t oldest_key_time = 0,
-    Env::WriteLifeTimeHint write_hint = Env::WLTH_NOT_SET);
+    Env::WriteLifeTimeHint write_hint = Env::WLTH_NOT_SET,
+    const uint64_t file_creation_time = 0);

 }  // namespace rocksdb
@@ -208,13 +208,10 @@ struct rocksdb_compactionfilter_t : public CompactionFilter {
  const char* (*name_)(void*);
  unsigned char ignore_snapshots_;

-  virtual ~rocksdb_compactionfilter_t() {
-    (*destructor_)(state_);
-  }
+  ~rocksdb_compactionfilter_t() override { (*destructor_)(state_); }

-  virtual bool Filter(int level, const Slice& key, const Slice& existing_value,
-                      std::string* new_value,
-                      bool* value_changed) const override {
+  bool Filter(int level, const Slice& key, const Slice& existing_value,
+              std::string* new_value, bool* value_changed) const override {
    char* c_new_value = nullptr;
    size_t new_value_length = 0;
    unsigned char c_value_changed = 0;
@@ -231,9 +228,9 @@ struct rocksdb_compactionfilter_t : public CompactionFilter {
    return result;
  }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }

-  virtual bool IgnoreSnapshots() const override { return ignore_snapshots_; }
+  bool IgnoreSnapshots() const override { return ignore_snapshots_; }
 };

 struct rocksdb_compactionfilterfactory_t : public CompactionFilterFactory {
@@ -243,9 +240,9 @@ struct rocksdb_compactionfilterfactory_t : public CompactionFilterFactory {
      void*, rocksdb_compactionfiltercontext_t* context);
  const char* (*name_)(void*);

-  virtual ~rocksdb_compactionfilterfactory_t() { (*destructor_)(state_); }
+  ~rocksdb_compactionfilterfactory_t() override { (*destructor_)(state_); }

-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& context) override {
    rocksdb_compactionfiltercontext_t ccontext;
    ccontext.rep = context;
@@ -253,7 +250,7 @@ struct rocksdb_compactionfilterfactory_t : public CompactionFilterFactory {
    return std::unique_ptr<CompactionFilter>(cf);
  }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }
 };

 struct rocksdb_comparator_t : public Comparator {
@@ -265,20 +262,17 @@ struct rocksdb_comparator_t : public Comparator {
      const char* b, size_t blen);
  const char* (*name_)(void*);

-  virtual ~rocksdb_comparator_t() {
-    (*destructor_)(state_);
-  }
+  ~rocksdb_comparator_t() override { (*destructor_)(state_); }

-  virtual int Compare(const Slice& a, const Slice& b) const override {
+  int Compare(const Slice& a, const Slice& b) const override {
    return (*compare_)(state_, a.data(), a.size(), b.data(), b.size());
  }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }

  // No-ops since the C binding does not support key shortening methods.
-  virtual void FindShortestSeparator(std::string*,
-                                     const Slice&) const override {}
-  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
+  void FindShortestSeparator(std::string*, const Slice&) const override {}
+  void FindShortSuccessor(std::string* /*key*/) const override {}
 };

 struct rocksdb_filterpolicy_t : public FilterPolicy {
@@ -298,14 +292,11 @@ struct rocksdb_filterpolicy_t : public FilterPolicy {
      void*,
      const char* filter, size_t filter_length);

-  virtual ~rocksdb_filterpolicy_t() {
-    (*destructor_)(state_);
-  }
+  ~rocksdb_filterpolicy_t() override { (*destructor_)(state_); }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }

-  virtual void CreateFilter(const Slice* keys, int n,
-                            std::string* dst) const override {
+  void CreateFilter(const Slice* keys, int n, std::string* dst) const override {
    std::vector<const char*> key_pointers(n);
    std::vector<size_t> key_sizes(n);
    for (int i = 0; i < n; i++) {
@@ -323,8 +314,7 @@ struct rocksdb_filterpolicy_t : public FilterPolicy {
    }
  }

-  virtual bool KeyMayMatch(const Slice& key,
-                           const Slice& filter) const override {
+  bool KeyMayMatch(const Slice& key, const Slice& filter) const override {
    return (*key_match_)(state_, key.data(), key.size(),
                         filter.data(), filter.size());
  }
@@ -349,14 +339,12 @@ struct rocksdb_mergeoperator_t : public MergeOperator {
      void*,
      const char* value, size_t value_length);

-  virtual ~rocksdb_mergeoperator_t() {
-    (*destructor_)(state_);
-  }
+  ~rocksdb_mergeoperator_t() override { (*destructor_)(state_); }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }

-  virtual bool FullMergeV2(const MergeOperationInput& merge_in,
-                           MergeOperationOutput* merge_out) const override {
+  bool FullMergeV2(const MergeOperationInput& merge_in,
+                   MergeOperationOutput* merge_out) const override {
    size_t n = merge_in.operand_list.size();
    std::vector<const char*> operand_pointers(n);
    std::vector<size_t> operand_sizes(n);
@@ -390,10 +378,10 @@ struct rocksdb_mergeoperator_t : public MergeOperator {
    return success;
  }

-  virtual bool PartialMergeMulti(const Slice& key,
-                                 const std::deque<Slice>& operand_list,
-                                 std::string* new_value,
-                                 Logger* /*logger*/) const override {
+  bool PartialMergeMulti(const Slice& key,
+                         const std::deque<Slice>& operand_list,
+                         std::string* new_value,
+                         Logger* /*logger*/) const override {
    size_t operand_count = operand_list.size();
    std::vector<const char*> operand_pointers(operand_count);
    std::vector<size_t> operand_sizes(operand_count);
@@ -444,23 +432,21 @@ struct rocksdb_slicetransform_t : public SliceTransform {
      void*,
      const char* key, size_t length);

-  virtual ~rocksdb_slicetransform_t() {
-    (*destructor_)(state_);
-  }
+  ~rocksdb_slicetransform_t() override { (*destructor_)(state_); }

-  virtual const char* Name() const override { return (*name_)(state_); }
+  const char* Name() const override { return (*name_)(state_); }

-  virtual Slice Transform(const Slice& src) const override {
+  Slice Transform(const Slice& src) const override {
    size_t len;
    char* dst = (*transform_)(state_, src.data(), src.size(), &len);
    return Slice(dst, len);
  }

-  virtual bool InDomain(const Slice& src) const override {
+  bool InDomain(const Slice& src) const override {
    return (*in_domain_)(state_, src.data(), src.size());
  }

-  virtual bool InRange(const Slice& src) const override {
+  bool InRange(const Slice& src) const override {
    return (*in_range_)(state_, src.data(), src.size());
  }
 };
@@ -1084,6 +1070,18 @@ int rocksdb_property_int(
  }
 }

+int rocksdb_property_int_cf(
+    rocksdb_t* db,
+    rocksdb_column_family_handle_t* column_family,
+    const char* propname,
+    uint64_t *out_val) {
+  if (db->rep->GetIntProperty(column_family->rep, Slice(propname), out_val)) {
+    return 0;
+  } else {
+    return -1;
+  }
+}
+
 char* rocksdb_property_value_cf(
    rocksdb_t* db,
    rocksdb_column_family_handle_t* column_family,
@@ -1197,6 +1195,14 @@ void rocksdb_flush(
  SaveError(errptr, db->rep->Flush(options->rep));
 }

+void rocksdb_flush_cf(
+    rocksdb_t* db,
+    const rocksdb_flushoptions_t* options,
+    rocksdb_column_family_handle_t* column_family,
+    char** errptr) {
+  SaveError(errptr, db->rep->Flush(options->rep, column_family->rep));
+}
+
 void rocksdb_disable_file_deletions(
    rocksdb_t* db,
    char** errptr) {
@@ -1495,10 +1501,10 @@ class H : public WriteBatch::Handler {
  void* state_;
  void (*put_)(void*, const char* k, size_t klen, const char* v, size_t vlen);
  void (*deleted_)(void*, const char* k, size_t klen);
-  virtual void Put(const Slice& key, const Slice& value) override {
+  void Put(const Slice& key, const Slice& value) override {
    (*put_)(state_, key.data(), key.size(), value.data(), value.size());
  }
-  virtual void Delete(const Slice& key) override {
+  void Delete(const Slice& key) override {
    (*deleted_)(state_, key.data(), key.size());
  }
 };
@@ -1784,11 +1790,11 @@ rocksdb_iterator_t* rocksdb_writebatch_wi_create_iterator_with_base(
 }

 rocksdb_iterator_t* rocksdb_writebatch_wi_create_iterator_with_base_cf(
-    rocksdb_writebatch_wi_t* wbwi,
-    rocksdb_iterator_t* base_iterator,
+    rocksdb_writebatch_wi_t* wbwi, rocksdb_iterator_t* base_iterator,
    rocksdb_column_family_handle_t* column_family) {
  rocksdb_iterator_t* result = new rocksdb_iterator_t;
-  result->rep = wbwi->rep->NewIteratorWithBase(column_family->rep, base_iterator->rep);
+  result->rep =
+      wbwi->rep->NewIteratorWithBase(column_family->rep, base_iterator->rep);
  delete base_iterator;
  return result;
 }
@@ -2067,6 +2073,15 @@ void rocksdb_set_options(
            db->rep->SetOptions(options_map));
    }

+void rocksdb_set_options_cf(
+    rocksdb_t* db, rocksdb_column_family_handle_t* handle, int count, const char* const keys[], const char* const values[], char** errptr) {
+        std::unordered_map<std::string, std::string> options_map;
+        for (int i=0; i<count; i++)
+            options_map[keys[i]] = values[i];
+        SaveError(errptr,
+            db->rep->SetOptions(handle->rep, options_map));
+    }
+
 rocksdb_options_t* rocksdb_options_create() {
  return new rocksdb_options_t;
 }
@@ -2211,6 +2226,11 @@ void rocksdb_options_set_max_bytes_for_level_base(
  opt->rep.max_bytes_for_level_base = n;
 }

+void rocksdb_options_set_snap_refresh_nanos(rocksdb_options_t* opt,
+                                            uint64_t n) {
+  opt->rep.snap_refresh_nanos = n;
+}
+
 void rocksdb_options_set_level_compaction_dynamic_level_bytes(
    rocksdb_options_t* opt, unsigned char v) {
  opt->rep.level_compaction_dynamic_level_bytes = v;
@@ -2857,7 +2877,7 @@ rocksdb_compactionfilter_t* rocksdb_compactionfilter_create(
  result->state_ = state;
  result->destructor_ = destructor;
  result->filter_ = filter;
-  result->ignore_snapshots_ = false;
+  result->ignore_snapshots_ = true;
  result->name_ = name;
  return result;
 }
@@ -2957,7 +2977,7 @@ rocksdb_filterpolicy_t* rocksdb_filterpolicy_create_bloom_format(int bits_per_ke
  // supplied C functions.
  struct Wrapper : public rocksdb_filterpolicy_t {
    const FilterPolicy* rep_;
-    ~Wrapper() { delete rep_; }
+    ~Wrapper() override { delete rep_; }
    const char* Name() const override { return rep_->Name(); }
    void CreateFilter(const Slice* keys, int n,
                      std::string* dst) const override {
@@ -3412,7 +3432,7 @@ void rocksdb_slicetransform_destroy(rocksdb_slicetransform_t* st) {

 struct Wrapper : public rocksdb_slicetransform_t {
  const SliceTransform* rep_;
-  ~Wrapper() { delete rep_; }
+  ~Wrapper() override { delete rep_; }
  const char* Name() const override { return rep_->Name(); }
  Slice Transform(const Slice& src) const override {
    return rep_->Transform(src);
@@ -3707,6 +3727,38 @@ rocksdb_transactiondb_t* rocksdb_transactiondb_open(
  return result;
 }

+rocksdb_transactiondb_t* rocksdb_transactiondb_open_column_families(
+    const rocksdb_options_t* options,
+    const rocksdb_transactiondb_options_t* txn_db_options, const char* name,
+    int num_column_families, const char** column_family_names,
+    const rocksdb_options_t** column_family_options,
+    rocksdb_column_family_handle_t** column_family_handles, char** errptr) {
+  std::vector<ColumnFamilyDescriptor> column_families;
+  for (int i = 0; i < num_column_families; i++) {
+    column_families.push_back(ColumnFamilyDescriptor(
+        std::string(column_family_names[i]),
+        ColumnFamilyOptions(column_family_options[i]->rep)));
+  }
+
+  TransactionDB* txn_db;
+  std::vector<ColumnFamilyHandle*> handles;
+  if (SaveError(errptr, TransactionDB::Open(options->rep, txn_db_options->rep,
+                                            std::string(name), column_families,
+                                            &handles, &txn_db))) {
+    return nullptr;
+  }
+
+  for (size_t i = 0; i < handles.size(); i++) {
+    rocksdb_column_family_handle_t* c_handle =
+        new rocksdb_column_family_handle_t;
+    c_handle->rep = handles[i];
+    column_family_handles[i] = c_handle;
+  }
+  rocksdb_transactiondb_t* result = new rocksdb_transactiondb_t;
+  result->rep = txn_db;
+  return result;
+}
+
 const rocksdb_snapshot_t* rocksdb_transactiondb_create_snapshot(
    rocksdb_transactiondb_t* txn_db) {
  rocksdb_snapshot_t* result = new rocksdb_snapshot_t;
@@ -3827,6 +3879,26 @@ char* rocksdb_transaction_get_for_update(rocksdb_transaction_t* txn,
  return result;
 }

+char* rocksdb_transaction_get_for_update_cf(
+    rocksdb_transaction_t* txn, const rocksdb_readoptions_t* options,
+    rocksdb_column_family_handle_t* column_family, const char* key, size_t klen,
+    size_t* vlen, unsigned char exclusive, char** errptr) {
+  char* result = nullptr;
+  std::string tmp;
+  Status s = txn->rep->GetForUpdate(options->rep, column_family->rep,
+                                    Slice(key, klen), &tmp, exclusive);
+  if (s.ok()) {
+    *vlen = tmp.size();
+    result = CopyString(tmp);
+  } else {
+    *vlen = 0;
+    if (!s.IsNotFound()) {
+      SaveError(errptr, s);
+    }
+  }
+  return result;
+}
+
 // Read a key outside a transaction
 char* rocksdb_transactiondb_get(
    rocksdb_transactiondb_t* txn_db,
@@ -3919,6 +3991,14 @@ void rocksdb_transaction_merge(rocksdb_transaction_t* txn, const char* key,
  SaveError(errptr, txn->rep->Merge(Slice(key, klen), Slice(val, vlen)));
 }

+void rocksdb_transaction_merge_cf(rocksdb_transaction_t* txn,
+                                  rocksdb_column_family_handle_t* column_family,
+                                  const char* key, size_t klen, const char* val,
+                                  size_t vlen, char** errptr) {
+  SaveError(errptr, txn->rep->Merge(column_family->rep, Slice(key, klen),
+                                    Slice(val, vlen)));
+}
+
 // Merge a key outside a transaction
 void rocksdb_transactiondb_merge(rocksdb_transactiondb_t* txn_db,
                                 const rocksdb_writeoptions_t* options,
@@ -3928,6 +4008,14 @@ void rocksdb_transactiondb_merge(rocksdb_transactiondb_t* txn_db,
                                       Slice(val, vlen)));
 }

+void rocksdb_transactiondb_merge_cf(
+    rocksdb_transactiondb_t* txn_db, const rocksdb_writeoptions_t* options,
+    rocksdb_column_family_handle_t* column_family, const char* key, size_t klen,
+    const char* val, size_t vlen, char** errptr) {
+  SaveError(errptr, txn_db->rep->Merge(options->rep, column_family->rep,
+                                       Slice(key, klen), Slice(val, vlen)));
+}
+
 // Delete a key inside a transaction
 void rocksdb_transaction_delete(rocksdb_transaction_t* txn, const char* key,
                                size_t klen, char** errptr) {
@@ -3980,6 +4068,14 @@ rocksdb_iterator_t* rocksdb_transactiondb_create_iterator(
  return result;
 }

+rocksdb_iterator_t* rocksdb_transactiondb_create_iterator_cf(
+    rocksdb_transactiondb_t* txn_db, const rocksdb_readoptions_t* options,
+    rocksdb_column_family_handle_t* column_family) {
+  rocksdb_iterator_t* result = new rocksdb_iterator_t;
+  result->rep = txn_db->rep->NewIterator(options->rep, column_family->rep);
+  return result;
+}
+
 void rocksdb_transactiondb_close(rocksdb_transactiondb_t* txn_db) {
  delete txn_db->rep;
  delete txn_db;
@@ -1,6 +1,7 @@
 /* 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. */
+// Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.

 #include <stdio.h>

@@ -917,7 +918,8 @@ int main(int argc, char** argv) {
    rocksdb_writebatch_wi_t* wbi = rocksdb_writebatch_wi_create(0, 1);
    rocksdb_writebatch_wi_put(wbi, "bar", 3, "b", 1);
    rocksdb_writebatch_wi_delete(wbi, "foo", 3);
-    rocksdb_iterator_t* iter = rocksdb_writebatch_wi_create_iterator_with_base(wbi, base_iter);
+    rocksdb_iterator_t* iter =
+        rocksdb_writebatch_wi_create_iterator_with_base(wbi, base_iter);
    CheckCondition(!rocksdb_iter_valid(iter));
    rocksdb_iter_seek_to_first(iter);
    CheckCondition(rocksdb_iter_valid(iter));
@@ -1186,6 +1188,12 @@ int main(int argc, char** argv) {
    rocksdb_put_cf(db, woptions, handles[1], "foo", 3, "hello", 5, &err);
    CheckNoError(err);

+    rocksdb_flushoptions_t *flush_options = rocksdb_flushoptions_create();
+    rocksdb_flushoptions_set_wait(flush_options, 1);
+    rocksdb_flush_cf(db, flush_options, handles[1], &err);
+    CheckNoError(err)
+    rocksdb_flushoptions_destroy(flush_options);
+
    CheckGetCF(db, roptions, handles[1], "foo", "hello");
    CheckPinGetCF(db, roptions, handles[1], "foo", "hello");

@@ -1527,7 +1535,7 @@ int main(int argc, char** argv) {
    const rocksdb_snapshot_t* snapshot;
    snapshot = rocksdb_transactiondb_create_snapshot(txn_db);
    rocksdb_readoptions_set_snapshot(roptions, snapshot);
-  
+
    rocksdb_transactiondb_put(txn_db, woptions, "foo", 3, "hey", 3,  &err);
    CheckNoError(err);

@@ -25,7 +25,7 @@
 #include "db/db_impl.h"
 #include "db/internal_stats.h"
 #include "db/job_context.h"
-#include "db/range_del_aggregator_v2.h"
+#include "db/range_del_aggregator.h"
 #include "db/table_properties_collector.h"
 #include "db/version_set.h"
 #include "db/write_controller.h"
@@ -63,12 +63,24 @@ ColumnFamilyHandleImpl::~ColumnFamilyHandleImpl() {
    JobContext job_context(0);
    mutex_->Lock();
    if (cfd_->Unref()) {
+      bool dropped = cfd_->IsDropped();
+
      delete cfd_;
+
+      if (dropped) {
+        db_->FindObsoleteFiles(&job_context, false, true);
+      }
    }
-    db_->FindObsoleteFiles(&job_context, false, true);
    mutex_->Unlock();
    if (job_context.HaveSomethingToDelete()) {
-      db_->PurgeObsoleteFiles(job_context);
+      bool defer_purge =
+          db_->immutable_db_options().avoid_unnecessary_blocking_io;
+      db_->PurgeObsoleteFiles(job_context, defer_purge);
+      if (defer_purge) {
+        mutex_->Lock();
+        db_->SchedulePurge();
+        mutex_->Unlock();
+      }
    }
    job_context.Clean();
  }
@@ -129,14 +141,10 @@ Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
    }
  }
  if (cf_options.compression_opts.zstd_max_train_bytes > 0) {
-    if (!CompressionTypeSupported(CompressionType::kZSTD)) {
-      // Dictionary trainer is available since v0.6.1, but ZSTD was marked
-      // stable only since v0.8.0. For now we enable the feature in stable
-      // versions only.
+    if (!ZSTD_TrainDictionarySupported()) {
      return Status::InvalidArgument(
-          "zstd dictionary trainer cannot be used because " +
-          CompressionTypeToString(CompressionType::kZSTD) +
-          " is not linked with the binary.");
+          "zstd dictionary trainer cannot be used because ZSTD 1.1.3+ "
+          "is not linked with the binary.");
    }
    if (cf_options.compression_opts.max_dict_bytes == 0) {
      return Status::InvalidArgument(
@@ -945,7 +953,7 @@ Status ColumnFamilyData::RangesOverlapWithMemtables(
  ScopedArenaIterator memtable_iter(merge_iter_builder.Finish());

  auto read_seq = super_version->current->version_set()->LastSequence();
-  RangeDelAggregatorV2 range_del_agg(&internal_comparator_, read_seq);
+  ReadRangeDelAggregator range_del_agg(&internal_comparator_, read_seq);
  auto* active_range_del_iter =
      super_version->mem->NewRangeTombstoneIterator(read_opts, read_seq);
  range_del_agg.AddTombstones(
@@ -986,13 +994,14 @@ const int ColumnFamilyData::kCompactToBaseLevel = -2;

 Compaction* ColumnFamilyData::CompactRange(
    const MutableCFOptions& mutable_cf_options, int input_level,
-    int output_level, uint32_t output_path_id, uint32_t max_subcompactions,
+    int output_level, const CompactRangeOptions& compact_range_options,
    const InternalKey* begin, const InternalKey* end,
-    InternalKey** compaction_end, bool* conflict) {
+    InternalKey** compaction_end, bool* conflict,
+    uint64_t max_file_num_to_ignore) {
  auto* result = compaction_picker_->CompactRange(
      GetName(), mutable_cf_options, current_->storage_info(), input_level,
-      output_level, output_path_id, max_subcompactions, begin, end,
-      compaction_end, conflict);
+      output_level, compact_range_options, begin, end, compaction_end, conflict,
+      max_file_num_to_ignore);
  if (result != nullptr) {
    result->SetInputVersion(current_);
  }
@@ -296,9 +296,10 @@ class ColumnFamilyData {
  // REQUIRES: DB mutex held
  Compaction* CompactRange(const MutableCFOptions& mutable_cf_options,
                           int input_level, int output_level,
-                           uint32_t output_path_id, uint32_t max_subcompactions,
+                           const CompactRangeOptions& compact_range_options,
                           const InternalKey* begin, const InternalKey* end,
-                           InternalKey** compaction_end, bool* manual_conflict);
+                           InternalKey** compaction_end, bool* manual_conflict,
+                           uint64_t max_file_num_to_ignore);

  CompactionPicker* compaction_picker() { return compaction_picker_.get(); }
  // thread-safe
@@ -383,6 +384,8 @@ class ColumnFamilyData {

  Directory* GetDataDir(size_t path_id) const;

+  ThreadLocalPtr* TEST_GetLocalSV() { return local_sv_.get(); }
+
 private:
  friend class ColumnFamilySet;
  ColumnFamilyData(uint32_t id, const std::string& name,
@@ -14,6 +14,7 @@

 #include "db/db_impl.h"
 #include "db/db_test_util.h"
+#include "memtable/hash_skiplist_rep.h"
 #include "options/options_parser.h"
 #include "port/port.h"
 #include "rocksdb/db.h"
@@ -68,7 +69,7 @@ class ColumnFamilyTestBase : public testing::Test {
    DestroyDB(dbname_, Options(db_options_, column_family_options_));
  }

-  virtual ~ColumnFamilyTestBase() {
+  ~ColumnFamilyTestBase() override {
    std::vector<ColumnFamilyDescriptor> column_families;
    for (auto h : handles_) {
      ColumnFamilyDescriptor cfdescriptor;
@@ -1206,29 +1207,32 @@ TEST_P(ColumnFamilyTest, DifferentWriteBufferSizes) {
 }
 #endif  // !ROCKSDB_LITE

-#ifndef ROCKSDB_LITE  // Cuckoo is not supported in lite
-TEST_P(ColumnFamilyTest, MemtableNotSupportSnapshot) {
-  db_options_.allow_concurrent_memtable_write = false;
-  Open();
-  auto* s1 = dbfull()->GetSnapshot();
-  ASSERT_TRUE(s1 != nullptr);
-  dbfull()->ReleaseSnapshot(s1);
+// The test is commented out because we want to test that snapshot is
+// not created for memtables not supported it, but There isn't a memtable
+// that doesn't support snapshot right now. If we have one later, we can
+// re-enable the test.
+//
+// #ifndef ROCKSDB_LITE  // Cuckoo is not supported in lite
+//   TEST_P(ColumnFamilyTest, MemtableNotSupportSnapshot) {
+//   db_options_.allow_concurrent_memtable_write = false;
+//   Open();
+//   auto* s1 = dbfull()->GetSnapshot();
+//   ASSERT_TRUE(s1 != nullptr);
+//   dbfull()->ReleaseSnapshot(s1);

-  // Add a column family that doesn't support snapshot
-  ColumnFamilyOptions first;
-  first.memtable_factory.reset(NewHashCuckooRepFactory(1024 * 1024));
-  CreateColumnFamilies({"first"}, {first});
-  auto* s2 = dbfull()->GetSnapshot();
-  ASSERT_TRUE(s2 == nullptr);
+//   // Add a column family that doesn't support snapshot
+//   ColumnFamilyOptions first;
+//   first.memtable_factory.reset(new DummyMemtableNotSupportingSnapshot());
+//   CreateColumnFamilies({"first"}, {first});
+//   auto* s2 = dbfull()->GetSnapshot();
+//   ASSERT_TRUE(s2 == nullptr);

-  // Add a column family that supports snapshot. Snapshot stays not supported.
-  ColumnFamilyOptions second;
-  CreateColumnFamilies({"second"}, {second});
-  auto* s3 = dbfull()->GetSnapshot();
-  ASSERT_TRUE(s3 == nullptr);
-  Close();
-}
-#endif  // !ROCKSDB_LITE
+//   // Add a column family that supports snapshot. Snapshot stays not
+//   supported. ColumnFamilyOptions second; CreateColumnFamilies({"second"},
+//   {second}); auto* s3 = dbfull()->GetSnapshot(); ASSERT_TRUE(s3 == nullptr);
+//   Close();
+// }
+// #endif  // !ROCKSDB_LITE

 class TestComparator : public Comparator {
  int Compare(const rocksdb::Slice& /*a*/,
@@ -2042,7 +2046,7 @@ TEST_P(ColumnFamilyTest, SameCFAutomaticManualCompactions) {
 }
 #endif  // !ROCKSDB_LITE

-#ifndef ROCKSDB_LITE  // Tailing interator not supported
+#ifndef ROCKSDB_LITE  // Tailing iterator not supported
 namespace {
 std::string IterStatus(Iterator* iter) {
  std::string result;
@@ -35,9 +35,9 @@ class CompactFilesTest : public testing::Test {
 class FlushedFileCollector : public EventListener {
 public:
  FlushedFileCollector() {}
-  ~FlushedFileCollector() {}
+  ~FlushedFileCollector() override {}

-  virtual void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
+  void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
    std::lock_guard<std::mutex> lock(mutex_);
    flushed_files_.push_back(info.file_path);
  }
@@ -256,9 +256,9 @@ TEST_F(CompactFilesTest, CapturingPendingFiles) {
 TEST_F(CompactFilesTest, CompactionFilterWithGetSv) {
  class FilterWithGet : public CompactionFilter {
   public:
-    virtual bool Filter(int /*level*/, const Slice& /*key*/,
-                        const Slice& /*value*/, std::string* /*new_value*/,
-                        bool* /*value_changed*/) const override {
+    bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+                std::string* /*new_value*/,
+                bool* /*value_changed*/) const override {
      if (db_ == nullptr) {
        return true;
      }
@@ -271,7 +271,7 @@ TEST_F(CompactFilesTest, CompactionFilterWithGetSv) {
      db_ = db;
    }

-    virtual const char* Name() const override { return "FilterWithGet"; }
+    const char* Name() const override { return "FilterWithGet"; }

   private:
    DB* db_;
@@ -357,6 +357,51 @@ TEST_F(CompactFilesTest, SentinelCompressionType) {
  }
 }

+TEST_F(CompactFilesTest, GetCompactionJobInfo) {
+  Options options;
+  options.create_if_missing = true;
+  // Disable RocksDB background compaction.
+  options.compaction_style = kCompactionStyleNone;
+  options.level0_slowdown_writes_trigger = 1000;
+  options.level0_stop_writes_trigger = 1000;
+  options.write_buffer_size = 65536;
+  options.max_write_buffer_number = 2;
+  options.compression = kNoCompression;
+  options.max_compaction_bytes = 5000;
+
+  // Add listener
+  FlushedFileCollector* collector = new FlushedFileCollector();
+  options.listeners.emplace_back(collector);
+
+  DB* db = nullptr;
+  DestroyDB(db_name_, options);
+  Status s = DB::Open(options, db_name_, &db);
+  assert(s.ok());
+  assert(db);
+
+  // create couple files
+  for (int i = 0; i < 500; ++i) {
+    db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
+  }
+  reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
+  auto l0_files_1 = collector->GetFlushedFiles();
+  CompactionOptions co;
+  co.compression = CompressionType::kLZ4Compression;
+  CompactionJobInfo compaction_job_info;
+  ASSERT_OK(
+      db->CompactFiles(co, l0_files_1, 0, -1, nullptr, &compaction_job_info));
+  ASSERT_EQ(compaction_job_info.base_input_level, 0);
+  ASSERT_EQ(compaction_job_info.cf_id, db->DefaultColumnFamily()->GetID());
+  ASSERT_EQ(compaction_job_info.cf_name, db->DefaultColumnFamily()->GetName());
+  ASSERT_EQ(compaction_job_info.compaction_reason,
+            CompactionReason::kManualCompaction);
+  ASSERT_EQ(compaction_job_info.compression, CompressionType::kLZ4Compression);
+  ASSERT_EQ(compaction_job_info.output_level, 0);
+  ASSERT_OK(compaction_job_info.status);
+  // no assertion failure
+  delete db;
+}
+
 }  // namespace rocksdb

 int main(int argc, char** argv) {
@@ -67,10 +67,11 @@ class CompactedDBImpl : public DBImpl {
  virtual Status EnableFileDeletions(bool /*force*/) override {
    return Status::NotSupported("Not supported in compacted db mode.");
  }
-  virtual Status GetLiveFiles(std::vector<std::string>&,
-                              uint64_t* /*manifest_file_size*/,
-                              bool /*flush_memtable*/ = true) override {
-    return Status::NotSupported("Not supported in compacted db mode.");
+  virtual Status GetLiveFiles(std::vector<std::string>& ret,
+                              uint64_t* manifest_file_size,
+                              bool /*flush_memtable*/) override {
+    return DBImpl::GetLiveFiles(ret, manifest_file_size,
+                                false /* flush_memtable */);
  }
  using DBImpl::Flush;
  virtual Status Flush(const FlushOptions& /*options*/,
@@ -250,6 +250,12 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
  if (max_subcompactions_ == 0) {
    max_subcompactions_ = immutable_cf_options_.max_subcompactions;
  }
+  if (!bottommost_level_) {
+    // Currently we only enable dictionary compression during compaction to the
+    // bottommost level.
+    output_compression_opts_.max_dict_bytes = 0;
+    output_compression_opts_.zstd_max_train_bytes = 0;
+  }

 #ifndef NDEBUG
  for (size_t i = 1; i < inputs_.size(); ++i) {
@@ -373,8 +379,8 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
        if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
          // We've advanced far enough
          if (user_cmp->Compare(user_key, f->smallest.user_key()) >= 0) {
-            // Key falls in this file's range, so definitely
-            // exists beyond output level
+            // Key falls in this file's range, so it may
+            // exist beyond output level
            return false;
          }
          break;
@@ -9,6 +9,23 @@
 #include "port/likely.h"
 #include "rocksdb/listener.h"
 #include "table/internal_iterator.h"
+#include "util/sync_point.h"
+
+#define DEFINITELY_IN_SNAPSHOT(seq, snapshot)                       \
+  ((seq) <= (snapshot) &&                                           \
+   (snapshot_checker_ == nullptr ||                                 \
+    LIKELY(snapshot_checker_->CheckInSnapshot((seq), (snapshot)) == \
+           SnapshotCheckerResult::kInSnapshot)))
+
+#define DEFINITELY_NOT_IN_SNAPSHOT(seq, snapshot)                     \
+  ((seq) > (snapshot) ||                                              \
+   (snapshot_checker_ != nullptr &&                                   \
+    UNLIKELY(snapshot_checker_->CheckInSnapshot((seq), (snapshot)) == \
+             SnapshotCheckerResult::kNotInSnapshot)))
+
+#define IN_EARLIEST_SNAPSHOT(seq) \
+  ((seq) <= earliest_snapshot_ && \
+   (snapshot_checker_ == nullptr || LIKELY(IsInEarliestSnapshot(seq))))

 namespace rocksdb {

@@ -18,17 +35,19 @@ CompactionIterator::CompactionIterator(
    SequenceNumber earliest_write_conflict_snapshot,
    const SnapshotChecker* snapshot_checker, Env* env,
    bool report_detailed_time, bool expect_valid_internal_key,
-    RangeDelAggregator* range_del_agg, const Compaction* compaction,
+    CompactionRangeDelAggregator* range_del_agg, const Compaction* compaction,
    const CompactionFilter* compaction_filter,
    const std::atomic<bool>* shutting_down,
-    const SequenceNumber preserve_deletes_seqnum)
+    const SequenceNumber preserve_deletes_seqnum,
+    SnapshotListFetchCallback* snap_list_callback)
    : CompactionIterator(
          input, cmp, merge_helper, last_sequence, snapshots,
          earliest_write_conflict_snapshot, snapshot_checker, env,
          report_detailed_time, expect_valid_internal_key, range_del_agg,
          std::unique_ptr<CompactionProxy>(
              compaction ? new CompactionProxy(compaction) : nullptr),
-          compaction_filter, shutting_down, preserve_deletes_seqnum) {}
+          compaction_filter, shutting_down, preserve_deletes_seqnum,
+          snap_list_callback) {}

 CompactionIterator::CompactionIterator(
    InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
@@ -36,11 +55,12 @@ CompactionIterator::CompactionIterator(
    SequenceNumber earliest_write_conflict_snapshot,
    const SnapshotChecker* snapshot_checker, Env* env,
    bool report_detailed_time, bool expect_valid_internal_key,
-    RangeDelAggregator* range_del_agg,
+    CompactionRangeDelAggregator* range_del_agg,
    std::unique_ptr<CompactionProxy> compaction,
    const CompactionFilter* compaction_filter,
    const std::atomic<bool>* shutting_down,
-    const SequenceNumber preserve_deletes_seqnum)
+    const SequenceNumber preserve_deletes_seqnum,
+    SnapshotListFetchCallback* snap_list_callback)
    : input_(input),
      cmp_(cmp),
      merge_helper_(merge_helper),
@@ -55,42 +75,21 @@ CompactionIterator::CompactionIterator(
      compaction_filter_(compaction_filter),
      shutting_down_(shutting_down),
      preserve_deletes_seqnum_(preserve_deletes_seqnum),
-      ignore_snapshots_(false),
      current_user_key_sequence_(0),
      current_user_key_snapshot_(0),
      merge_out_iter_(merge_helper_),
-      current_key_committed_(false) {
+      current_key_committed_(false),
+      snap_list_callback_(snap_list_callback) {
  assert(compaction_filter_ == nullptr || compaction_ != nullptr);
+  assert(snapshots_ != nullptr);
  bottommost_level_ =
      compaction_ == nullptr ? false : compaction_->bottommost_level();
  if (compaction_ != nullptr) {
    level_ptrs_ = std::vector<size_t>(compaction_->number_levels(), 0);
  }
-
-  if (snapshots_->size() == 0) {
-    // optimize for fast path if there are no snapshots
-    visible_at_tip_ = true;
-    earliest_snapshot_ = kMaxSequenceNumber;
-    latest_snapshot_ = 0;
-  } else {
-    visible_at_tip_ = false;
-    earliest_snapshot_ = snapshots_->at(0);
-    latest_snapshot_ = snapshots_->back();
-  }
-#ifndef NDEBUG
-  // findEarliestVisibleSnapshot assumes this ordering.
-  for (size_t i = 1; i < snapshots_->size(); ++i) {
-    assert(snapshots_->at(i - 1) <= snapshots_->at(i));
-  }
-#endif
-  if (compaction_filter_ != nullptr) {
-    if (compaction_filter_->IgnoreSnapshots()) {
-      ignore_snapshots_ = true;
-    }
-  } else {
-    ignore_snapshots_ = false;
-  }
+  ProcessSnapshotList();
  input_->SetPinnedItersMgr(&pinned_iters_mgr_);
+  TEST_SYNC_POINT_CALLBACK("CompactionIterator:AfterInit", compaction_.get());
 }

 CompactionIterator::~CompactionIterator() {
@@ -160,12 +159,7 @@ void CompactionIterator::Next() {
 void CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
                                              Slice* skip_until) {
  if (compaction_filter_ != nullptr &&
-      (ikey_.type == kTypeValue || ikey_.type == kTypeBlobIndex) &&
-      (visible_at_tip_ || ignore_snapshots_ ||
-       ikey_.sequence > latest_snapshot_ ||
-       (snapshot_checker_ != nullptr &&
-        UNLIKELY(!snapshot_checker_->IsInSnapshot(ikey_.sequence,
-                                                  latest_snapshot_))))) {
+      (ikey_.type == kTypeValue || ikey_.type == kTypeBlobIndex)) {
    // If the user has specified a compaction filter and the sequence
    // number is greater than any external snapshot, then invoke the
    // filter. If the return value of the compaction filter is true,
@@ -215,6 +209,28 @@ void CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
  }
 }

+void CompactionIterator::ProcessSnapshotList() {
+#ifndef NDEBUG
+  // findEarliestVisibleSnapshot assumes this ordering.
+  for (size_t i = 1; i < snapshots_->size(); ++i) {
+    assert(snapshots_->at(i - 1) < snapshots_->at(i));
+  }
+#endif
+  if (snapshots_->size() == 0) {
+    // optimize for fast path if there are no snapshots
+    visible_at_tip_ = true;
+    earliest_snapshot_iter_ = snapshots_->end();
+    earliest_snapshot_ = kMaxSequenceNumber;
+    latest_snapshot_ = 0;
+  } else {
+    visible_at_tip_ = false;
+    earliest_snapshot_iter_ = snapshots_->begin();
+    earliest_snapshot_ = snapshots_->at(0);
+    latest_snapshot_ = snapshots_->back();
+  }
+  released_snapshots_.clear();
+}
+
 void CompactionIterator::NextFromInput() {
  at_next_ = false;
  valid_ = false;
@@ -241,6 +257,7 @@ void CompactionIterator::NextFromInput() {
      valid_ = true;
      break;
    }
+    TEST_SYNC_POINT_CALLBACK("CompactionIterator:ProcessKV", &ikey_);

    // Update input statistics
    if (ikey_.type == kTypeDeletion || ikey_.type == kTypeSingleDeletion) {
@@ -261,6 +278,13 @@ void CompactionIterator::NextFromInput() {
    // compaction filter). ikey_.user_key is pointing to the copy.
    if (!has_current_user_key_ ||
        !cmp_->Equal(ikey_.user_key, current_user_key_)) {
+      num_keys_++;
+      // Use num_keys_ to reduce the overhead of reading current time
+      if (snap_list_callback_ && snapshots_->size() &&
+          snap_list_callback_->TimeToRefresh(num_keys_)) {
+        snap_list_callback_->Refresh(snapshots_, latest_snapshot_);
+        ProcessSnapshotList();
+      }
      // First occurrence of this user key
      // Copy key for output
      key_ = current_key_.SetInternalKey(key_, &ikey_);
@@ -269,9 +293,7 @@ void CompactionIterator::NextFromInput() {
      has_outputted_key_ = false;
      current_user_key_sequence_ = kMaxSequenceNumber;
      current_user_key_snapshot_ = 0;
-      current_key_committed_ =
-          (snapshot_checker_ == nullptr ||
-           snapshot_checker_->IsInSnapshot(ikey_.sequence, kMaxSequenceNumber));
+      current_key_committed_ = KeyCommitted(ikey_.sequence);

      // Apply the compaction filter to the first committed version of the user
      // key.
@@ -293,8 +315,7 @@ void CompactionIterator::NextFromInput() {
      // to query snapshot_checker_ in that case.
      if (UNLIKELY(!current_key_committed_)) {
        assert(snapshot_checker_ != nullptr);
-        current_key_committed_ =
-            snapshot_checker_->IsInSnapshot(ikey_.sequence, kMaxSequenceNumber);
+        current_key_committed_ = KeyCommitted(ikey_.sequence);
        // Apply the compaction filter to the first committed version of the
        // user key.
        if (current_key_committed_) {
@@ -378,10 +399,8 @@ void CompactionIterator::NextFromInput() {
          cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
        // Check whether the next key belongs to the same snapshot as the
        // SingleDelete.
-        if (prev_snapshot == 0 || next_ikey.sequence > prev_snapshot ||
-            (snapshot_checker_ != nullptr &&
-             UNLIKELY(!snapshot_checker_->IsInSnapshot(next_ikey.sequence,
-                                                       prev_snapshot)))) {
+        if (prev_snapshot == 0 ||
+            DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot)) {
          if (next_ikey.type == kTypeSingleDeletion) {
            // We encountered two SingleDeletes in a row.  This could be due to
            // unexpected user input.
@@ -393,11 +412,8 @@ void CompactionIterator::NextFromInput() {
            ++iter_stats_.num_record_drop_obsolete;
            ++iter_stats_.num_single_del_mismatch;
          } else if (has_outputted_key_ ||
-                     (ikey_.sequence <= earliest_write_conflict_snapshot_ &&
-                      (snapshot_checker_ == nullptr ||
-                       LIKELY(snapshot_checker_->IsInSnapshot(
-                           ikey_.sequence,
-                           earliest_write_conflict_snapshot_))))) {
+                     DEFINITELY_IN_SNAPSHOT(
+                         ikey_.sequence, earliest_write_conflict_snapshot_)) {
            // Found a matching value, we can drop the single delete and the
            // value.  It is safe to drop both records since we've already
            // outputted a key in this snapshot, or there is no earlier
@@ -407,7 +423,8 @@ void CompactionIterator::NextFromInput() {
            // is an unexpected Merge or Delete.  We will compact it out
            // either way. We will maintain counts of how many mismatches
            // happened
-            if (next_ikey.type != kTypeValue) {
+            if (next_ikey.type != kTypeValue &&
+                next_ikey.type != kTypeBlobIndex) {
              ++iter_stats_.num_single_del_mismatch;
            }

@@ -444,10 +461,7 @@ void CompactionIterator::NextFromInput() {
        // iteration. If the next key is corrupt, we return before the
        // comparison, so the value of has_current_user_key does not matter.
        has_current_user_key_ = false;
-        if (compaction_ != nullptr && ikey_.sequence <= earliest_snapshot_ &&
-            (snapshot_checker_ == nullptr ||
-             LIKELY(snapshot_checker_->IsInSnapshot(ikey_.sequence,
-                                                    earliest_snapshot_))) &&
+        if (compaction_ != nullptr && IN_EARLIEST_SNAPSHOT(ikey_.sequence) &&
            compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
                                                       &level_ptrs_)) {
          // Key doesn't exist outside of this range.
@@ -466,24 +480,37 @@ void CompactionIterator::NextFromInput() {
      if (valid_) {
        at_next_ = true;
      }
-    } else if (last_snapshot == current_user_key_snapshot_) {
+    } else if (last_snapshot == current_user_key_snapshot_ ||
+               (last_snapshot > 0 &&
+                last_snapshot < current_user_key_snapshot_)) {
      // If the earliest snapshot is which this key is visible in
      // is the same as the visibility of a previous instance of the
      // same key, then this kv is not visible in any snapshot.
      // Hidden by an newer entry for same user key
-      // TODO(noetzli): why not > ?
      //
      // Note: Dropping this key will not affect TransactionDB write-conflict
      // checking since there has already been a record returned for this key
      // in this snapshot.
      assert(last_sequence >= current_user_key_sequence_);
+
+      // Note2: if last_snapshot < current_user_key_snapshot, it can only
+      // mean last_snapshot is released between we process last value and
+      // this value, and findEarliestVisibleSnapshot returns the next snapshot
+      // as current_user_key_snapshot. In this case last value and current
+      // value are both in current_user_key_snapshot currently.
+      // Although last_snapshot is released we might still get a definitive
+      // response when key sequence number changes, e.g., when seq is determined
+      // too old and visible in all snapshots.
+      assert(last_snapshot == current_user_key_snapshot_ ||
+             (snapshot_checker_ != nullptr &&
+              snapshot_checker_->CheckInSnapshot(current_user_key_sequence_,
+                                                 last_snapshot) !=
+                  SnapshotCheckerResult::kNotInSnapshot));
+
      ++iter_stats_.num_record_drop_hidden;  // (A)
      input_->Next();
    } else if (compaction_ != nullptr && ikey_.type == kTypeDeletion &&
-               ikey_.sequence <= earliest_snapshot_ &&
-               (snapshot_checker_ == nullptr ||
-                LIKELY(snapshot_checker_->IsInSnapshot(ikey_.sequence,
-                                                       earliest_snapshot_))) &&
+               IN_EARLIEST_SNAPSHOT(ikey_.sequence) &&
               ikeyNotNeededForIncrementalSnapshot() &&
               compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
                                                          &level_ptrs_)) {
@@ -520,13 +547,10 @@ void CompactionIterator::NextFromInput() {
      input_->Next();
      // Skip over all versions of this key that happen to occur in the same snapshot
      // range as the delete
-      while (input_->Valid() &&
-             ParseInternalKey(input_->key(), &next_ikey) &&
+      while (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
             cmp_->Equal(ikey_.user_key, next_ikey.user_key) &&
-             (prev_snapshot == 0 || next_ikey.sequence > prev_snapshot ||
-              (snapshot_checker_ != nullptr &&
-               UNLIKELY(!snapshot_checker_->IsInSnapshot(next_ikey.sequence,
-                                                         prev_snapshot))))) {
+             (prev_snapshot == 0 ||
+              DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot))) {
        input_->Next();
      }
      // If you find you still need to output a row with this key, we need to output the
@@ -617,14 +641,9 @@ void CompactionIterator::PrepareOutput() {
  //
  // Can we do the same for levels above bottom level as long as
  // KeyNotExistsBeyondOutputLevel() return true?
-  if ((compaction_ != nullptr &&
-      !compaction_->allow_ingest_behind()) &&
-      ikeyNotNeededForIncrementalSnapshot() &&
-      bottommost_level_ && valid_ && ikey_.sequence <= earliest_snapshot_ &&
-      (snapshot_checker_ == nullptr || LIKELY(snapshot_checker_->IsInSnapshot(
-        ikey_.sequence, earliest_snapshot_))) &&
-      ikey_.type != kTypeMerge &&
-      !cmp_->Equal(compaction_->GetLargestUserKey(), ikey_.user_key)) {
+  if ((compaction_ != nullptr && !compaction_->allow_ingest_behind()) &&
+      ikeyNotNeededForIncrementalSnapshot() && bottommost_level_ && valid_ &&
+      IN_EARLIEST_SNAPSHOT(ikey_.sequence) && ikey_.type != kTypeMerge) {
    assert(ikey_.type != kTypeDeletion && ikey_.type != kTypeSingleDeletion);
    ikey_.sequence = 0;
    current_key_.UpdateInternalKey(0, ikey_.type);
@@ -642,12 +661,23 @@ inline SequenceNumber CompactionIterator::findEarliestVisibleSnapshot(
    *prev_snapshot = *std::prev(snapshots_iter);
    assert(*prev_snapshot < in);
  }
+  if (snapshot_checker_ == nullptr) {
+    return snapshots_iter != snapshots_->end()
+      ? *snapshots_iter : kMaxSequenceNumber;
+  }
+  bool has_released_snapshot = !released_snapshots_.empty();
  for (; snapshots_iter != snapshots_->end(); ++snapshots_iter) {
    auto cur = *snapshots_iter;
    assert(in <= cur);
-    if (snapshot_checker_ == nullptr ||
-        snapshot_checker_->IsInSnapshot(in, cur)) {
+    // Skip if cur is in released_snapshots.
+    if (has_released_snapshot && released_snapshots_.count(cur) > 0) {
+      continue;
+    }
+    auto res = snapshot_checker_->CheckInSnapshot(in, cur);
+    if (res == SnapshotCheckerResult::kInSnapshot) {
      return cur;
+    } else if (res == SnapshotCheckerResult::kSnapshotReleased) {
+      released_snapshots_.insert(cur);
    }
    *prev_snapshot = cur;
  }
@@ -661,4 +691,30 @@ inline bool CompactionIterator::ikeyNotNeededForIncrementalSnapshot() {
         (ikey_.sequence < preserve_deletes_seqnum_);
 }

+bool CompactionIterator::IsInEarliestSnapshot(SequenceNumber sequence) {
+  assert(snapshot_checker_ != nullptr);
+  assert(earliest_snapshot_ == kMaxSequenceNumber ||
+         (earliest_snapshot_iter_ != snapshots_->end() &&
+          *earliest_snapshot_iter_ == earliest_snapshot_));
+  auto in_snapshot =
+      snapshot_checker_->CheckInSnapshot(sequence, earliest_snapshot_);
+  while (UNLIKELY(in_snapshot == SnapshotCheckerResult::kSnapshotReleased)) {
+    // Avoid the the current earliest_snapshot_ being return as
+    // earliest visible snapshot for the next value. So if a value's sequence
+    // is zero-ed out by PrepareOutput(), the next value will be compact out.
+    released_snapshots_.insert(earliest_snapshot_);
+    earliest_snapshot_iter_++;
+
+    if (earliest_snapshot_iter_ == snapshots_->end()) {
+      earliest_snapshot_ = kMaxSequenceNumber;
+    } else {
+      earliest_snapshot_ = *earliest_snapshot_iter_;
+    }
+    in_snapshot =
+        snapshot_checker_->CheckInSnapshot(sequence, earliest_snapshot_);
+  }
+  assert(in_snapshot != SnapshotCheckerResult::kSnapshotReleased);
+  return in_snapshot == SnapshotCheckerResult::kInSnapshot;
+}
+
 }  // namespace rocksdb
@@ -7,6 +7,7 @@
 #include <algorithm>
 #include <deque>
 #include <string>
+#include <unordered_set>
 #include <vector>

 #include "db/compaction.h"
@@ -20,6 +21,54 @@

 namespace rocksdb {

+// This callback can be used to refresh the snapshot list from the db. It
+// includes logics to exponentially decrease the refresh rate to limit the
+// overhead of refresh.
+class SnapshotListFetchCallback {
+ public:
+  SnapshotListFetchCallback(Env* env, uint64_t snap_refresh_nanos,
+                            size_t every_nth_key = 1024)
+      : timer_(env, /*auto restart*/ true),
+        snap_refresh_nanos_(snap_refresh_nanos),
+        every_nth_key_minus_one_(every_nth_key - 1) {
+    assert(every_nth_key > 0);
+    assert((ceil(log2(every_nth_key)) == floor(log2(every_nth_key))));
+  }
+  // Refresh the snapshot list. snapshots will bre replacted with the new list.
+  // max is the upper bound. Note: this function will acquire the db_mutex_.
+  virtual void Refresh(std::vector<SequenceNumber>* snapshots,
+                       SequenceNumber max) = 0;
+  inline bool TimeToRefresh(const size_t key_index) {
+    assert(snap_refresh_nanos_ != 0);
+    // skip the key if key_index % every_nth_key (which is of power 2) is not 0.
+    if ((key_index & every_nth_key_minus_one_) != 0) {
+      return false;
+    }
+    const uint64_t elapsed = timer_.ElapsedNanos();
+    auto ret = elapsed > snap_refresh_nanos_;
+    // pre-compute the next time threshold
+    if (ret) {
+      // inc next refresh period exponentially (by x4)
+      auto next_refresh_threshold = snap_refresh_nanos_ << 2;
+      // make sure the shift has not overflown the highest 1 bit
+      snap_refresh_nanos_ =
+          std::max(snap_refresh_nanos_, next_refresh_threshold);
+    }
+    return ret;
+  }
+  static constexpr SnapshotListFetchCallback* kDisabled = nullptr;
+
+  virtual ~SnapshotListFetchCallback() {}
+
+ private:
+  // Time since the callback was created
+  StopWatchNano timer_;
+  // The delay before calling ::Refresh. To be increased exponentially.
+  uint64_t snap_refresh_nanos_;
+  // Skip evey nth key. Number n if of power 2. The math will require n-1.
+  const uint64_t every_nth_key_minus_one_;
+};
+
 class CompactionIterator {
 public:
  // A wrapper around Compaction. Has a much smaller interface, only what
@@ -64,11 +113,12 @@ class CompactionIterator {
                     SequenceNumber earliest_write_conflict_snapshot,
                     const SnapshotChecker* snapshot_checker, Env* env,
                     bool report_detailed_time, bool expect_valid_internal_key,
-                     RangeDelAggregator* range_del_agg,
+                     CompactionRangeDelAggregator* range_del_agg,
                     const Compaction* compaction = nullptr,
                     const CompactionFilter* compaction_filter = nullptr,
                     const std::atomic<bool>* shutting_down = nullptr,
-                     const SequenceNumber preserve_deletes_seqnum = 0);
+                     const SequenceNumber preserve_deletes_seqnum = 0,
+                     SnapshotListFetchCallback* snap_list_callback = nullptr);

  // Constructor with custom CompactionProxy, used for tests.
  CompactionIterator(InternalIterator* input, const Comparator* cmp,
@@ -77,11 +127,12 @@ class CompactionIterator {
                     SequenceNumber earliest_write_conflict_snapshot,
                     const SnapshotChecker* snapshot_checker, Env* env,
                     bool report_detailed_time, bool expect_valid_internal_key,
-                     RangeDelAggregator* range_del_agg,
+                     CompactionRangeDelAggregator* range_del_agg,
                     std::unique_ptr<CompactionProxy> compaction,
                     const CompactionFilter* compaction_filter = nullptr,
                     const std::atomic<bool>* shutting_down = nullptr,
-                     const SequenceNumber preserve_deletes_seqnum = 0);
+                     const SequenceNumber preserve_deletes_seqnum = 0,
+                     SnapshotListFetchCallback* snap_list_callback = nullptr);

  ~CompactionIterator();

@@ -109,6 +160,8 @@ class CompactionIterator {
 private:
  // Processes the input stream to find the next output
  void NextFromInput();
+  // Process snapshots_ and assign related variables
+  void ProcessSnapshotList();

  // Do last preparations before presenting the output to the callee. At this
  // point this only zeroes out the sequence number if possible for better
@@ -132,16 +185,31 @@ class CompactionIterator {
  // or seqnum be zero-ed out even if all other conditions for it are met.
  inline bool ikeyNotNeededForIncrementalSnapshot();

+  inline bool KeyCommitted(SequenceNumber sequence) {
+    return snapshot_checker_ == nullptr ||
+           snapshot_checker_->CheckInSnapshot(sequence, kMaxSequenceNumber) ==
+               SnapshotCheckerResult::kInSnapshot;
+  }
+
+  bool IsInEarliestSnapshot(SequenceNumber sequence);
+
  InternalIterator* input_;
  const Comparator* cmp_;
  MergeHelper* merge_helper_;
-  const std::vector<SequenceNumber>* snapshots_;
+  std::vector<SequenceNumber>* snapshots_;
+  // List of snapshots released during compaction.
+  // findEarliestVisibleSnapshot() find them out from return of
+  // snapshot_checker, and make sure they will not be returned as
+  // earliest visible snapshot of an older value.
+  // See WritePreparedTransactionTest::ReleaseSnapshotDuringCompaction3.
+  std::unordered_set<SequenceNumber> released_snapshots_;
+  std::vector<SequenceNumber>::const_iterator earliest_snapshot_iter_;
  const SequenceNumber earliest_write_conflict_snapshot_;
  const SnapshotChecker* const snapshot_checker_;
  Env* env_;
  bool report_detailed_time_;
  bool expect_valid_internal_key_;
-  RangeDelAggregator* range_del_agg_;
+  CompactionRangeDelAggregator* range_del_agg_;
  std::unique_ptr<CompactionProxy> compaction_;
  const CompactionFilter* compaction_filter_;
  const std::atomic<bool>* shutting_down_;
@@ -151,7 +219,6 @@ class CompactionIterator {
  bool visible_at_tip_;
  SequenceNumber earliest_snapshot_;
  SequenceNumber latest_snapshot_;
-  bool ignore_snapshots_;

  // State
  //
@@ -204,6 +271,9 @@ class CompactionIterator {
  // Used to avoid purging uncommitted values. The application can specify
  // uncommitted values by providing a SnapshotChecker object.
  bool current_key_committed_;
+  SnapshotListFetchCallback* snap_list_callback_;
+  // number of distinct keys processed
+  size_t num_keys_ = 0;

  bool IsShuttingDown() {
    // This is a best-effort facility, so memory_order_relaxed is sufficient.
@@ -112,39 +112,39 @@ class LoggingForwardVectorIterator : public InternalIterator {
    assert(keys_.size() == values_.size());
  }

-  virtual bool Valid() const override { return current_ < keys_.size(); }
+  bool Valid() const override { return current_ < keys_.size(); }

-  virtual void SeekToFirst() override {
+  void SeekToFirst() override {
    log.emplace_back(Action::Type::SEEK_TO_FIRST);
    current_ = 0;
  }
-  virtual void SeekToLast() override { assert(false); }
+  void SeekToLast() override { assert(false); }

-  virtual void Seek(const Slice& target) override {
+  void Seek(const Slice& target) override {
    log.emplace_back(Action::Type::SEEK, target.ToString());
    current_ = std::lower_bound(keys_.begin(), keys_.end(), target.ToString()) -
               keys_.begin();
  }

-  virtual void SeekForPrev(const Slice& /*target*/) override { assert(false); }
+  void SeekForPrev(const Slice& /*target*/) override { assert(false); }

-  virtual void Next() override {
+  void Next() override {
    assert(Valid());
    log.emplace_back(Action::Type::NEXT);
    current_++;
  }
-  virtual void Prev() override { assert(false); }
+  void Prev() override { assert(false); }

-  virtual Slice key() const override {
+  Slice key() const override {
    assert(Valid());
    return Slice(keys_[current_]);
  }
-  virtual Slice value() const override {
+  Slice value() const override {
    assert(Valid());
    return Slice(values_[current_]);
  }

-  virtual Status status() const override { return Status::OK(); }
+  Status status() const override { return Status::OK(); }

  std::vector<Action> log;

@@ -158,22 +158,20 @@ class FakeCompaction : public CompactionIterator::CompactionProxy {
 public:
  FakeCompaction() = default;

-  virtual int level(size_t /*compaction_input_level*/) const override {
-    return 0;
-  }
-  virtual bool KeyNotExistsBeyondOutputLevel(
+  int level(size_t /*compaction_input_level*/) const override { return 0; }
+  bool KeyNotExistsBeyondOutputLevel(
      const Slice& /*user_key*/,
      std::vector<size_t>* /*level_ptrs*/) const override {
    return is_bottommost_level || key_not_exists_beyond_output_level;
  }
-  virtual bool bottommost_level() const override { return is_bottommost_level; }
-  virtual int number_levels() const override { return 1; }
-  virtual Slice GetLargestUserKey() const override {
+  bool bottommost_level() const override { return is_bottommost_level; }
+  int number_levels() const override { return 1; }
+  Slice GetLargestUserKey() const override {
    return "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
  }
-  virtual bool allow_ingest_behind() const override { return false; }
+  bool allow_ingest_behind() const override { return false; }

-  virtual bool preserve_deletes() const override { return false; }
+  bool preserve_deletes() const override { return false; }

  bool key_not_exists_beyond_output_level = false;

@@ -190,13 +188,17 @@ class TestSnapshotChecker : public SnapshotChecker {
      : last_committed_sequence_(last_committed_sequence),
        snapshots_(snapshots) {}

-  bool IsInSnapshot(SequenceNumber seq,
-                    SequenceNumber snapshot_seq) const override {
+  SnapshotCheckerResult CheckInSnapshot(
+      SequenceNumber seq, SequenceNumber snapshot_seq) const override {
    if (snapshot_seq == kMaxSequenceNumber) {
-      return seq <= last_committed_sequence_;
+      return seq <= last_committed_sequence_
+                 ? SnapshotCheckerResult::kInSnapshot
+                 : SnapshotCheckerResult::kNotInSnapshot;
    }
    assert(snapshots_.count(snapshot_seq) > 0);
-    return seq <= snapshots_.at(snapshot_seq);
+    return seq <= snapshots_.at(snapshot_seq)
+               ? SnapshotCheckerResult::kInSnapshot
+               : SnapshotCheckerResult::kNotInSnapshot;
  }

 private:
@@ -221,10 +223,15 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
      MergeOperator* merge_op = nullptr, CompactionFilter* filter = nullptr,
      bool bottommost_level = false,
      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
-    std::unique_ptr<InternalIterator> range_del_iter(
+    std::unique_ptr<InternalIterator> unfragmented_range_del_iter(
        new test::VectorIterator(range_del_ks, range_del_vs));
-    range_del_agg_.reset(new RangeDelAggregator(icmp_, snapshots_));
-    ASSERT_OK(range_del_agg_->AddTombstones(std::move(range_del_iter)));
+    auto tombstone_list = std::make_shared<FragmentedRangeTombstoneList>(
+        std::move(unfragmented_range_del_iter), icmp_);
+    std::unique_ptr<FragmentedRangeTombstoneIterator> range_del_iter(
+        new FragmentedRangeTombstoneIterator(tombstone_list, icmp_,
+                                             kMaxSequenceNumber));
+    range_del_agg_.reset(new CompactionRangeDelAggregator(&icmp_, snapshots_));
+    range_del_agg_->AddTombstones(std::move(range_del_iter));

    std::unique_ptr<CompactionIterator::CompactionProxy> compaction;
    if (filter || bottommost_level) {
@@ -292,7 +299,7 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
  std::unique_ptr<MergeHelper> merge_helper_;
  std::unique_ptr<LoggingForwardVectorIterator> iter_;
  std::unique_ptr<CompactionIterator> c_iter_;
-  std::unique_ptr<RangeDelAggregator> range_del_agg_;
+  std::unique_ptr<CompactionRangeDelAggregator> range_del_agg_;
  std::unique_ptr<SnapshotChecker> snapshot_checker_;
  std::atomic<bool> shutting_down_{false};
  FakeCompaction* compaction_proxy_;
@@ -368,10 +375,9 @@ TEST_P(CompactionIteratorTest, RangeDeletionWithSnapshots) {

 TEST_P(CompactionIteratorTest, CompactionFilterSkipUntil) {
  class Filter : public CompactionFilter {
-    virtual Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
-                              const Slice& existing_value,
-                              std::string* /*new_value*/,
-                              std::string* skip_until) const override {
+    Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
+                      const Slice& existing_value, std::string* /*new_value*/,
+                      std::string* skip_until) const override {
      std::string k = key.ToString();
      std::string v = existing_value.ToString();
      // See InitIterators() call below for the sequence of keys and their
@@ -551,10 +557,9 @@ TEST_P(CompactionIteratorTest, ShuttingDownInMerge) {

 TEST_P(CompactionIteratorTest, SingleMergeOperand) {
  class Filter : public CompactionFilter {
-    virtual Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
-                              const Slice& existing_value,
-                              std::string* /*new_value*/,
-                              std::string* /*skip_until*/) const override {
+    Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
+                      const Slice& existing_value, std::string* /*new_value*/,
+                      std::string* /*skip_until*/) const override {
      std::string k = key.ToString();
      std::string v = existing_value.ToString();

@@ -36,7 +36,7 @@
 #include "db/memtable_list.h"
 #include "db/merge_context.h"
 #include "db/merge_helper.h"
-#include "db/range_del_aggregator_v2.h"
+#include "db/range_del_aggregator.h"
 #include "db/version_set.h"
 #include "monitoring/iostats_context_imp.h"
 #include "monitoring/perf_context_imp.h"
@@ -97,6 +97,8 @@ const char* GetCompactionReasonString(CompactionReason compaction_reason) {
      return "Flush";
    case CompactionReason::kExternalSstIngestion:
      return "ExternalSstIngestion";
+    case CompactionReason::kPeriodicCompaction:
+      return "PeriodicCompaction";
    case CompactionReason::kNumOfReasons:
      // fall through
    default:
@@ -157,7 +159,6 @@ struct CompactionJob::SubcompactionState {
  uint64_t overlapped_bytes = 0;
  // A flag determine whether the key has been seen in ShouldStopBefore()
  bool seen_key = false;
-  std::string compression_dict;

  SubcompactionState(Compaction* c, Slice* _start, Slice* _end,
                     uint64_t size = 0)
@@ -173,8 +174,7 @@ struct CompactionJob::SubcompactionState {
        approx_size(size),
        grandparent_index(0),
        overlapped_bytes(0),
-        seen_key(false),
-        compression_dict() {
+        seen_key(false) {
    assert(compaction != nullptr);
  }

@@ -197,7 +197,6 @@ struct CompactionJob::SubcompactionState {
    grandparent_index = std::move(o.grandparent_index);
    overlapped_bytes = std::move(o.overlapped_bytes);
    seen_key = std::move(o.seen_key);
-    compression_dict = std::move(o.compression_dict);
    return *this;
  }

@@ -315,7 +314,8 @@ CompactionJob::CompactionJob(
    SequenceNumber earliest_write_conflict_snapshot,
    const SnapshotChecker* snapshot_checker, std::shared_ptr<Cache> table_cache,
    EventLogger* event_logger, bool paranoid_file_checks, bool measure_io_stats,
-    const std::string& dbname, CompactionJobStats* compaction_job_stats)
+    const std::string& dbname, CompactionJobStats* compaction_job_stats,
+    Env::Priority thread_pri, SnapshotListFetchCallback* snap_list_callback)
    : job_id_(job_id),
      compact_(new CompactionState(compaction)),
      compaction_job_stats_(compaction_job_stats),
@@ -336,6 +336,7 @@ CompactionJob::CompactionJob(
      db_mutex_(db_mutex),
      db_error_handler_(db_error_handler),
      existing_snapshots_(std::move(existing_snapshots)),
+      snap_list_callback_(snap_list_callback),
      earliest_write_conflict_snapshot_(earliest_write_conflict_snapshot),
      snapshot_checker_(snapshot_checker),
      table_cache_(std::move(table_cache)),
@@ -343,7 +344,8 @@ CompactionJob::CompactionJob(
      bottommost_level_(false),
      paranoid_file_checks_(paranoid_file_checks),
      measure_io_stats_(measure_io_stats),
-      write_hint_(Env::WLTH_NOT_SET) {
+      write_hint_(Env::WLTH_NOT_SET),
+      thread_pri_(thread_pri) {
  assert(log_buffer_ != nullptr);
  const auto* cfd = compact_->compaction->column_family_data();
  ThreadStatusUtil::SetColumnFamily(cfd, cfd->ioptions()->env,
@@ -417,11 +419,10 @@ void CompactionJob::Prepare() {
  bottommost_level_ = c->bottommost_level();

  if (c->ShouldFormSubcompactions()) {
-    const uint64_t start_micros = env_->NowMicros();
-    GenSubcompactionBoundaries();
-    MeasureTime(stats_, SUBCOMPACTION_SETUP_TIME,
-                env_->NowMicros() - start_micros);
-
+    {
+      StopWatch sw(env_, stats_, SUBCOMPACTION_SETUP_TIME);
+      GenSubcompactionBoundaries();
+    }
    assert(sizes_.size() == boundaries_.size() + 1);

    for (size_t i = 0; i <= boundaries_.size(); i++) {
@@ -429,8 +430,8 @@ void CompactionJob::Prepare() {
      Slice* end = i == boundaries_.size() ? nullptr : &boundaries_[i];
      compact_->sub_compact_states.emplace_back(c, start, end, sizes_[i]);
    }
-    MeasureTime(stats_, NUM_SUBCOMPACTIONS_SCHEDULED,
-                compact_->sub_compact_states.size());
+    RecordInHistogram(stats_, NUM_SUBCOMPACTIONS_SCHEDULED,
+                      compact_->sub_compact_states.size());
  } else {
    compact_->sub_compact_states.emplace_back(c, nullptr, nullptr);
  }
@@ -604,7 +605,15 @@ Status CompactionJob::Run() {
  }

  compaction_stats_.micros = env_->NowMicros() - start_micros;
-  MeasureTime(stats_, COMPACTION_TIME, compaction_stats_.micros);
+  compaction_stats_.cpu_micros = 0;
+  for (size_t i = 0; i < compact_->sub_compact_states.size(); i++) {
+    compaction_stats_.cpu_micros +=
+        compact_->sub_compact_states[i].compaction_job_stats.cpu_micros;
+  }
+
+  RecordTimeToHistogram(stats_, COMPACTION_TIME, compaction_stats_.micros);
+  RecordTimeToHistogram(stats_, COMPACTION_CPU_TIME,
+                        compaction_stats_.cpu_micros);

  TEST_SYNC_POINT("CompactionJob::Run:BeforeVerify");

@@ -713,7 +722,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
  Status status = compact_->status;
  ColumnFamilyData* cfd = compact_->compaction->column_family_data();
  cfd->internal_stats()->AddCompactionStats(
-      compact_->compaction->output_level(), compaction_stats_);
+      compact_->compaction->output_level(), thread_pri_, compaction_stats_);

  if (status.ok()) {
    status = InstallCompactionResults(mutable_cf_options);
@@ -767,6 +776,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
  stream << "job" << job_id_ << "event"
         << "compaction_finished"
         << "compaction_time_micros" << compaction_stats_.micros
+         << "compaction_time_cpu_micros" << compaction_stats_.cpu_micros
         << "output_level" << compact_->compaction->output_level()
         << "num_output_files" << compact_->NumOutputFiles()
         << "total_output_size" << compact_->total_bytes << "num_input_records"
@@ -804,16 +814,35 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {

 void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  assert(sub_compact != nullptr);
+
+  uint64_t prev_cpu_micros = env_->NowCPUNanos() / 1000;
+
  ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
-  RangeDelAggregatorV2 range_del_agg_v2(&cfd->internal_comparator(),
-                                        kMaxSequenceNumber /* upper_bound */);
-  auto* range_del_agg =
-      range_del_agg_v2.DelegateToRangeDelAggregator(existing_snapshots_);
+
+  // Create compaction filter and fail the compaction if
+  // IgnoreSnapshots() = false because it is not supported anymore
+  const CompactionFilter* compaction_filter =
+      cfd->ioptions()->compaction_filter;
+  std::unique_ptr<CompactionFilter> compaction_filter_from_factory = nullptr;
+  if (compaction_filter == nullptr) {
+    compaction_filter_from_factory =
+        sub_compact->compaction->CreateCompactionFilter();
+    compaction_filter = compaction_filter_from_factory.get();
+  }
+  if (compaction_filter != nullptr && !compaction_filter->IgnoreSnapshots()) {
+    sub_compact->status = Status::NotSupported(
+        "CompactionFilter::IgnoreSnapshots() = false is not supported "
+        "anymore.");
+    return;
+  }
+
+  CompactionRangeDelAggregator range_del_agg(&cfd->internal_comparator(),
+                                             existing_snapshots_);

  // Although the v2 aggregator is what the level iterator(s) know about,
  // the AddTombstones calls will be propagated down to the v1 aggregator.
  std::unique_ptr<InternalIterator> input(versions_->MakeInputIterator(
-      sub_compact->compaction, &range_del_agg_v2, env_optiosn_for_read_));
+      sub_compact->compaction, &range_del_agg, env_optiosn_for_read_));

  AutoThreadOperationStageUpdater stage_updater(
      ThreadStatus::STAGE_COMPACTION_PROCESS_KV);
@@ -825,58 +854,19 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  uint64_t prev_fsync_nanos = 0;
  uint64_t prev_range_sync_nanos = 0;
  uint64_t prev_prepare_write_nanos = 0;
+  uint64_t prev_cpu_write_nanos = 0;
+  uint64_t prev_cpu_read_nanos = 0;
  if (measure_io_stats_) {
    prev_perf_level = GetPerfLevel();
-    SetPerfLevel(PerfLevel::kEnableTime);
+    SetPerfLevel(PerfLevel::kEnableTimeAndCPUTimeExceptForMutex);
    prev_write_nanos = IOSTATS(write_nanos);
    prev_fsync_nanos = IOSTATS(fsync_nanos);
    prev_range_sync_nanos = IOSTATS(range_sync_nanos);
    prev_prepare_write_nanos = IOSTATS(prepare_write_nanos);
+    prev_cpu_write_nanos = IOSTATS(cpu_write_nanos);
+    prev_cpu_read_nanos = IOSTATS(cpu_read_nanos);
  }

-  const MutableCFOptions* mutable_cf_options =
-      sub_compact->compaction->mutable_cf_options();
-
-  // To build compression dictionary, we sample the first output file, assuming
-  // it'll reach the maximum length. We optionally pass these samples through
-  // zstd's dictionary trainer, or just use them directly. Then, the dictionary
-  // is used for compressing subsequent output files in the same subcompaction.
-  const bool kUseZstdTrainer =
-      sub_compact->compaction->output_compression_opts().zstd_max_train_bytes >
-      0;
-  const size_t kSampleBytes =
-      kUseZstdTrainer
-          ? sub_compact->compaction->output_compression_opts()
-                .zstd_max_train_bytes
-          : sub_compact->compaction->output_compression_opts().max_dict_bytes;
-  const int kSampleLenShift = 6;  // 2^6 = 64-byte samples
-  std::set<size_t> sample_begin_offsets;
-  if (bottommost_level_ && kSampleBytes > 0) {
-    const size_t kMaxSamples = kSampleBytes >> kSampleLenShift;
-    const size_t kOutFileLen =
-        static_cast<size_t>(MaxFileSizeForLevel(*mutable_cf_options,
-            compact_->compaction->output_level(),
-            cfd->ioptions()->compaction_style,
-            compact_->compaction->GetInputBaseLevel(),
-            cfd->ioptions()->level_compaction_dynamic_level_bytes));
-    if (kOutFileLen != port::kMaxSizet) {
-      const size_t kOutFileNumSamples = kOutFileLen >> kSampleLenShift;
-      Random64 generator{versions_->NewFileNumber()};
-      for (size_t i = 0; i < kMaxSamples; ++i) {
-        sample_begin_offsets.insert(
-            static_cast<size_t>(generator.Uniform(kOutFileNumSamples))
-            << kSampleLenShift);
-      }
-    }
-  }
-
-  auto compaction_filter = cfd->ioptions()->compaction_filter;
-  std::unique_ptr<CompactionFilter> compaction_filter_from_factory = nullptr;
-  if (compaction_filter == nullptr) {
-    compaction_filter_from_factory =
-        sub_compact->compaction->CreateCompactionFilter();
-    compaction_filter = compaction_filter_from_factory.get();
-  }
  MergeHelper merge(
      env_, cfd->user_comparator(), cfd->ioptions()->merge_operator,
      compaction_filter, db_options_.info_log.get(),
@@ -902,8 +892,10 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
      input.get(), cfd->user_comparator(), &merge, versions_->LastSequence(),
      &existing_snapshots_, earliest_write_conflict_snapshot_,
      snapshot_checker_, env_, ShouldReportDetailedTime(env_, stats_), false,
-      range_del_agg, sub_compact->compaction, compaction_filter, shutting_down_,
-      preserve_deletes_seqnum_));
+      &range_del_agg, sub_compact->compaction, compaction_filter,
+      shutting_down_, preserve_deletes_seqnum_,
+      // Currently range_del_agg is incompatible with snapshot refresh feature.
+      range_del_agg.IsEmpty() ? snap_list_callback_ : nullptr));
  auto c_iter = sub_compact->c_iter.get();
  c_iter->SeekToFirst();
  if (c_iter->Valid() && sub_compact->compaction->output_level() != 0) {
@@ -914,12 +906,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
                                  sub_compact->current_output_file_size);
  }
  const auto& c_iter_stats = c_iter->iter_stats();
-  auto sample_begin_offset_iter = sample_begin_offsets.cbegin();
-  // data_begin_offset and dict_sample_data are only valid while generating
-  // dictionary from the first output file.
-  size_t data_begin_offset = 0;
-  std::string dict_sample_data;
-  dict_sample_data.reserve(kSampleBytes);

  while (status.ok() && !cfd->IsDropped() && c_iter->Valid()) {
    // Invariant: c_iter.status() is guaranteed to be OK if c_iter->Valid()
@@ -955,55 +941,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
        key, c_iter->ikey().sequence);
    sub_compact->num_output_records++;

-    if (sub_compact->outputs.size() == 1) {  // first output file
-      // Check if this key/value overlaps any sample intervals; if so, appends
-      // overlapping portions to the dictionary.
-      for (const auto& data_elmt : {key, value}) {
-        size_t data_end_offset = data_begin_offset + data_elmt.size();
-        while (sample_begin_offset_iter != sample_begin_offsets.cend() &&
-               *sample_begin_offset_iter < data_end_offset) {
-          size_t sample_end_offset =
-              *sample_begin_offset_iter + (1 << kSampleLenShift);
-          // Invariant: Because we advance sample iterator while processing the
-          // data_elmt containing the sample's last byte, the current sample
-          // cannot end before the current data_elmt.
-          assert(data_begin_offset < sample_end_offset);
-
-          size_t data_elmt_copy_offset, data_elmt_copy_len;
-          if (*sample_begin_offset_iter <= data_begin_offset) {
-            // The sample starts before data_elmt starts, so take bytes starting
-            // at the beginning of data_elmt.
-            data_elmt_copy_offset = 0;
-          } else {
-            // data_elmt starts before the sample starts, so take bytes starting
-            // at the below offset into data_elmt.
-            data_elmt_copy_offset =
-                *sample_begin_offset_iter - data_begin_offset;
-          }
-          if (sample_end_offset <= data_end_offset) {
-            // The sample ends before data_elmt ends, so take as many bytes as
-            // needed.
-            data_elmt_copy_len =
-                sample_end_offset - (data_begin_offset + data_elmt_copy_offset);
-          } else {
-            // data_elmt ends before the sample ends, so take all remaining
-            // bytes in data_elmt.
-            data_elmt_copy_len =
-                data_end_offset - (data_begin_offset + data_elmt_copy_offset);
-          }
-          dict_sample_data.append(&data_elmt.data()[data_elmt_copy_offset],
-                                  data_elmt_copy_len);
-          if (sample_end_offset > data_end_offset) {
-            // Didn't finish sample. Try to finish it with the next data_elmt.
-            break;
-          }
-          // Next sample may require bytes from same data_elmt.
-          sample_begin_offset_iter++;
-        }
-        data_begin_offset = data_end_offset;
-      }
-    }
-
    // Close output file if it is big enough. Two possibilities determine it's
    // time to close it: (1) the current key should be this file's last key, (2)
    // the next key should not be in this file.
@@ -1041,22 +978,10 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
      }
      CompactionIterationStats range_del_out_stats;
      status =
-          FinishCompactionOutputFile(input_status, sub_compact, range_del_agg,
+          FinishCompactionOutputFile(input_status, sub_compact, &range_del_agg,
                                     &range_del_out_stats, next_key);
      RecordDroppedKeys(range_del_out_stats,
                        &sub_compact->compaction_job_stats);
-      if (sub_compact->outputs.size() == 1) {
-        // Use samples from first output file to create dictionary for
-        // compression of subsequent files.
-        if (kUseZstdTrainer) {
-          sub_compact->compression_dict = ZSTD_TrainDictionary(
-              dict_sample_data, kSampleLenShift,
-              sub_compact->compaction->output_compression_opts()
-                  .max_dict_bytes);
-        } else {
-          sub_compact->compression_dict = std::move(dict_sample_data);
-        }
-      }
    }
  }

@@ -1092,8 +1017,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  }

  if (status.ok() && sub_compact->builder == nullptr &&
-      sub_compact->outputs.size() == 0 &&
-      !range_del_agg->IsEmpty()) {
+      sub_compact->outputs.size() == 0 && !range_del_agg.IsEmpty()) {
    // handle subcompaction containing only range deletions
    status = OpenCompactionOutputFile(sub_compact);
  }
@@ -1102,7 +1026,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  // close the output file.
  if (sub_compact->builder != nullptr) {
    CompactionIterationStats range_del_out_stats;
-    Status s = FinishCompactionOutputFile(status, sub_compact, range_del_agg,
+    Status s = FinishCompactionOutputFile(status, sub_compact, &range_del_agg,
                                          &range_del_out_stats);
    if (status.ok()) {
      status = s;
@@ -1110,6 +1034,9 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
    RecordDroppedKeys(range_del_out_stats, &sub_compact->compaction_job_stats);
  }

+  sub_compact->compaction_job_stats.cpu_micros =
+      env_->NowCPUNanos() / 1000 - prev_cpu_micros;
+
  if (measure_io_stats_) {
    sub_compact->compaction_job_stats.file_write_nanos +=
        IOSTATS(write_nanos) - prev_write_nanos;
@@ -1119,7 +1046,11 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
        IOSTATS(range_sync_nanos) - prev_range_sync_nanos;
    sub_compact->compaction_job_stats.file_prepare_write_nanos +=
        IOSTATS(prepare_write_nanos) - prev_prepare_write_nanos;
-    if (prev_perf_level != PerfLevel::kEnableTime) {
+    sub_compact->compaction_job_stats.cpu_micros -=
+        (IOSTATS(cpu_write_nanos) - prev_cpu_write_nanos +
+         IOSTATS(cpu_read_nanos) - prev_cpu_read_nanos) /
+        1000;
+    if (prev_perf_level != PerfLevel::kEnableTimeAndCPUTimeExceptForMutex) {
      SetPerfLevel(prev_perf_level);
    }
  }
@@ -1168,7 +1099,7 @@ void CompactionJob::RecordDroppedKeys(

 Status CompactionJob::FinishCompactionOutputFile(
    const Status& input_status, SubcompactionState* sub_compact,
-    RangeDelAggregator* range_del_agg,
+    CompactionRangeDelAggregator* range_del_agg,
    CompactionIterationStats* range_del_out_stats,
    const Slice* next_table_min_key /* = nullptr */) {
  AutoThreadOperationStageUpdater stage_updater(
@@ -1192,10 +1123,12 @@ Status CompactionJob::FinishCompactionOutputFile(
    Slice lower_bound_guard, upper_bound_guard;
    std::string smallest_user_key;
    const Slice *lower_bound, *upper_bound;
+    bool lower_bound_from_sub_compact = false;
    if (sub_compact->outputs.size() == 1) {
      // For the first output table, include range tombstones before the min key
      // but after the subcompaction boundary.
      lower_bound = sub_compact->start;
+      lower_bound_from_sub_compact = true;
    } else if (meta->smallest.size() > 0) {
      // For subsequent output tables, only include range tombstones from min
      // key onwards since the previous file was extended to contain range
@@ -1207,10 +1140,19 @@ Status CompactionJob::FinishCompactionOutputFile(
      lower_bound = nullptr;
    }
    if (next_table_min_key != nullptr) {
-      // This isn't the last file in the subcompaction, so extend until the next
-      // file starts.
+      // This may be the last file in the subcompaction in some cases, so we
+      // need to compare the end key of subcompaction with the next file start
+      // key. When the end key is chosen by the subcompaction, we know that
+      // it must be the biggest key in output file. Therefore, it is safe to
+      // use the smaller key as the upper bound of the output file, to ensure
+      // that there is no overlapping between different output files.
      upper_bound_guard = ExtractUserKey(*next_table_min_key);
-      upper_bound = &upper_bound_guard;
+      if (sub_compact->end != nullptr &&
+          ucmp->Compare(upper_bound_guard, *sub_compact->end) >= 0) {
+        upper_bound = sub_compact->end;
+      } else {
+        upper_bound = &upper_bound_guard;
+      }
    } else {
      // This is the last file in the subcompaction, so extend until the
      // subcompaction ends.
@@ -1220,11 +1162,6 @@ Status CompactionJob::FinishCompactionOutputFile(
    if (existing_snapshots_.size() > 0) {
      earliest_snapshot = existing_snapshots_[0];
    }
-    auto it = range_del_agg->NewIterator();
-    if (lower_bound != nullptr) {
-      it->Seek(*lower_bound);
-    }
-
    bool has_overlapping_endpoints;
    if (upper_bound != nullptr && meta->largest.size() > 0) {
      has_overlapping_endpoints =
@@ -1232,6 +1169,24 @@ Status CompactionJob::FinishCompactionOutputFile(
    } else {
      has_overlapping_endpoints = false;
    }
+
+    // The end key of the subcompaction must be bigger or equal to the upper
+    // bound. If the end of subcompaction is null or the upper bound is null,
+    // it means that this file is the last file in the compaction. So there
+    // will be no overlapping between this file and others.
+    assert(sub_compact->end == nullptr ||
+           upper_bound == nullptr ||
+           ucmp->Compare(*upper_bound , *sub_compact->end) <= 0);
+    auto it = range_del_agg->NewIterator(lower_bound, upper_bound,
+                                         has_overlapping_endpoints);
+    // Position the range tombstone output iterator. There may be tombstone
+    // fragments that are entirely out of range, so make sure that we do not
+    // include those.
+    if (lower_bound != nullptr) {
+      it->Seek(*lower_bound);
+    } else {
+      it->SeekToFirst();
+    }
    for (; it->Valid(); it->Next()) {
      auto tombstone = it->Tombstone();
      if (upper_bound != nullptr) {
@@ -1257,6 +1212,8 @@ Status CompactionJob::FinishCompactionOutputFile(
      }

      auto kv = tombstone.Serialize();
+      assert(lower_bound == nullptr ||
+             ucmp->Compare(*lower_bound, kv.second) < 0);
      sub_compact->builder->Add(kv.first.Encode(), kv.second);
      InternalKey smallest_candidate = std::move(kv.first);
      if (lower_bound != nullptr &&
@@ -1265,11 +1222,24 @@ Status CompactionJob::FinishCompactionOutputFile(
        // (the max key in the previous table or subcompaction) in order for
        // files to appear key-space partitioned.
        //
-        // Choose lowest seqnum so this file's smallest internal key comes
-        // after the previous file's/subcompaction's largest. The fake seqnum
-        // is OK because the read path's file-picking code only considers user
-        // key.
-        smallest_candidate = InternalKey(*lower_bound, 0, kTypeRangeDeletion);
+        // When lower_bound is chosen by a subcompaction, we know that
+        // subcompactions over smaller keys cannot contain any keys at
+        // lower_bound. We also know that smaller subcompactions exist, because
+        // otherwise the subcompaction woud be unbounded on the left. As a
+        // result, we know that no other files on the output level will contain
+        // actual keys at lower_bound (an output file may have a largest key of
+        // lower_bound@kMaxSequenceNumber, but this only indicates a large range
+        // tombstone was truncated). Therefore, it is safe to use the
+        // tombstone's sequence number, to ensure that keys at lower_bound at
+        // lower levels are covered by truncated tombstones.
+        //
+        // If lower_bound was chosen by the smallest data key in the file,
+        // choose lowest seqnum so this file's smallest internal key comes after
+        // the previous file's largest. The fake seqnum is OK because the read
+        // path's file-picking code only considers user key.
+        smallest_candidate = InternalKey(
+            *lower_bound, lower_bound_from_sub_compact ? tombstone.seq_ : 0,
+            kTypeRangeDeletion);
      }
      InternalKey largest_candidate = tombstone.SerializeEndKey();
      if (upper_bound != nullptr &&
@@ -1291,9 +1261,23 @@ Status CompactionJob::FinishCompactionOutputFile(
        largest_candidate =
            InternalKey(*upper_bound, kMaxSequenceNumber, kTypeRangeDeletion);
      }
+#ifndef NDEBUG
+      SequenceNumber smallest_ikey_seqnum = kMaxSequenceNumber;
+      if (meta->smallest.size() > 0) {
+        smallest_ikey_seqnum = GetInternalKeySeqno(meta->smallest.Encode());
+      }
+#endif
      meta->UpdateBoundariesForRange(smallest_candidate, largest_candidate,
                                     tombstone.seq_,
                                     cfd->internal_comparator());
+
+      // The smallest key in a file is used for range tombstone truncation, so
+      // it cannot have a seqnum of 0 (unless the smallest data key in a file
+      // has a seqnum of 0). Otherwise, the truncated tombstone may expose
+      // deleted keys at lower levels.
+      assert(smallest_ikey_seqnum == 0 ||
+             ExtractInternalKeyFooter(meta->smallest.Encode()) !=
+                 PackSequenceAndType(0, kTypeRangeDeletion));
    }
    meta->marked_for_compaction = sub_compact->builder->NeedCompact();
  }
@@ -1371,10 +1355,7 @@ Status CompactionJob::FinishCompactionOutputFile(
  auto sfm =
      static_cast<SstFileManagerImpl*>(db_options_.sst_file_manager.get());
  if (sfm && meta != nullptr && meta->fd.GetPathId() == 0) {
-    auto fn =
-        TableFileName(sub_compact->compaction->immutable_cf_options()->cf_paths,
-                      meta->fd.GetNumber(), meta->fd.GetPathId());
-    sfm->OnAddFile(fn);
+    sfm->OnAddFile(fname);
    if (sfm->IsMaxAllowedSpaceReached()) {
      // TODO(ajkr): should we return OK() if max space was reached by the final
      // compaction output file (similarly to how flush works when full)?
@@ -1496,7 +1477,7 @@ Status CompactionJob::OpenCompactionOutputFile(
      sub_compact->compaction->immutable_cf_options()->listeners;
  sub_compact->outfile.reset(
      new WritableFileWriter(std::move(writable_file), fname, env_options_,
-                             db_options_.statistics.get(), listeners));
+                             env_, db_options_.statistics.get(), listeners));

  // If the Column family flag is to only optimize filters for hits,
  // we can skip creating filters if this is the bottommost_level where
@@ -1504,20 +1485,20 @@ Status CompactionJob::OpenCompactionOutputFile(
  bool skip_filters =
      cfd->ioptions()->optimize_filters_for_hits && bottommost_level_;

-  uint64_t output_file_creation_time =
+  int64_t temp_current_time = 0;
+  auto get_time_status = env_->GetCurrentTime(&temp_current_time);
+  // Safe to proceed even if GetCurrentTime fails. So, log and proceed.
+  if (!get_time_status.ok()) {
+    ROCKS_LOG_WARN(db_options_.info_log,
+                   "Failed to get current time. Status: %s",
+                   get_time_status.ToString().c_str());
+  }
+  uint64_t current_time = static_cast<uint64_t>(temp_current_time);
+
+  uint64_t latest_key_time =
      sub_compact->compaction->MaxInputFileCreationTime();
-  if (output_file_creation_time == 0) {
-    int64_t _current_time = 0;
-    auto status = db_options_.env->GetCurrentTime(&_current_time);
-    // Safe to proceed even if GetCurrentTime fails. So, log and proceed.
-    if (!status.ok()) {
-      ROCKS_LOG_WARN(
-          db_options_.info_log,
-          "Failed to get current time to populate creation_time property. "
-          "Status: %s",
-          status.ToString().c_str());
-    }
-    output_file_creation_time = static_cast<uint64_t>(_current_time);
+  if (latest_key_time == 0) {
+    latest_key_time = current_time;
  }

  sub_compact->builder.reset(NewTableBuilder(
@@ -1525,9 +1506,11 @@ Status CompactionJob::OpenCompactionOutputFile(
      cfd->internal_comparator(), cfd->int_tbl_prop_collector_factories(),
      cfd->GetID(), cfd->GetName(), sub_compact->outfile.get(),
      sub_compact->compaction->output_compression(),
+      0 /*sample_for_compression */,
      sub_compact->compaction->output_compression_opts(),
-      sub_compact->compaction->output_level(), &sub_compact->compression_dict,
-      skip_filters, output_file_creation_time));
+      sub_compact->compaction->output_level(), skip_filters, latest_key_time,
+      0 /* oldest_key_time */, sub_compact->compaction->max_output_file_size(),
+      current_time));
  LogFlush(db_options_.info_log);
  return s;
 }
@@ -29,8 +29,8 @@
 #include "db/version_edit.h"
 #include "db/write_controller.h"
 #include "db/write_thread.h"
-#include "options/db_options.h"
 #include "options/cf_options.h"
+#include "options/db_options.h"
 #include "port/port.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/compaction_job_stats.h"
@@ -57,22 +57,20 @@ class VersionSet;

 class CompactionJob {
 public:
-  CompactionJob(int job_id, Compaction* compaction,
-                const ImmutableDBOptions& db_options,
-                const EnvOptions env_options, VersionSet* versions,
-                const std::atomic<bool>* shutting_down,
-                const SequenceNumber preserve_deletes_seqnum,
-                LogBuffer* log_buffer,
-                Directory* db_directory, Directory* output_directory,
-                Statistics* stats, InstrumentedMutex* db_mutex,
-                ErrorHandler* db_error_handler,
-                std::vector<SequenceNumber> existing_snapshots,
-                SequenceNumber earliest_write_conflict_snapshot,
-                const SnapshotChecker* snapshot_checker,
-                std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
-                bool paranoid_file_checks, bool measure_io_stats,
-                const std::string& dbname,
-                CompactionJobStats* compaction_job_stats);
+  CompactionJob(
+      int job_id, Compaction* compaction, const ImmutableDBOptions& db_options,
+      const EnvOptions env_options, VersionSet* versions,
+      const std::atomic<bool>* shutting_down,
+      const SequenceNumber preserve_deletes_seqnum, LogBuffer* log_buffer,
+      Directory* db_directory, Directory* output_directory, Statistics* stats,
+      InstrumentedMutex* db_mutex, ErrorHandler* db_error_handler,
+      std::vector<SequenceNumber> existing_snapshots,
+      SequenceNumber earliest_write_conflict_snapshot,
+      const SnapshotChecker* snapshot_checker,
+      std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
+      bool paranoid_file_checks, bool measure_io_stats,
+      const std::string& dbname, CompactionJobStats* compaction_job_stats,
+      Env::Priority thread_pri, SnapshotListFetchCallback* snap_list_callback);

  ~CompactionJob();

@@ -104,7 +102,7 @@ class CompactionJob {

  Status FinishCompactionOutputFile(
      const Status& input_status, SubcompactionState* sub_compact,
-      RangeDelAggregator* range_del_agg,
+      CompactionRangeDelAggregator* range_del_agg,
      CompactionIterationStats* range_del_out_stats,
      const Slice* next_table_min_key = nullptr);
  Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options);
@@ -152,6 +150,7 @@ class CompactionJob {
  // entirely within s1 and s2, then the earlier version of k1 can be safely
  // deleted because that version is not visible in any snapshot.
  std::vector<SequenceNumber> existing_snapshots_;
+  SnapshotListFetchCallback* snap_list_callback_;

  // This is the earliest snapshot that could be used for write-conflict
  // checking by a transaction.  For any user-key newer than this snapshot, we
@@ -172,6 +171,7 @@ class CompactionJob {
  // Stores the approx size of keys covered in the range of each subcompaction
  std::vector<uint64_t> sizes_;
  Env::WriteLifeTimeHint write_hint_;
+  Env::Priority thread_pri_;
 };

 }  // namespace rocksdb
@@ -113,7 +113,7 @@ class CompactionJobStatsTest : public testing::Test,
    Reopen(options);
  }

-  ~CompactionJobStatsTest() {
+  ~CompactionJobStatsTest() override {
    rocksdb::SyncPoint::GetInstance()->DisableProcessing();
    rocksdb::SyncPoint::GetInstance()->LoadDependency({});
    rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
@@ -426,7 +426,7 @@ class CompactionJobStatsChecker : public EventListener {
  // Once a compaction completed, this function will verify the returned
  // CompactionJobInfo with the oldest CompactionJobInfo added earlier
  // in "expected_stats_" which has not yet being used for verification.
-  virtual void OnCompactionCompleted(DB* /*db*/, const CompactionJobInfo& ci) {
+  void OnCompactionCompleted(DB* /*db*/, const CompactionJobInfo& ci) override {
    if (verify_next_comp_io_stats_) {
      ASSERT_GT(ci.stats.file_write_nanos, 0);
      ASSERT_GT(ci.stats.file_range_sync_nanos, 0);
@@ -523,7 +523,7 @@ class CompactionJobDeletionStatsChecker : public CompactionJobStatsChecker {
 public:
  // Verifies whether two CompactionJobStats match.
  void Verify(const CompactionJobStats& current_stats,
-              const CompactionJobStats& stats) {
+              const CompactionJobStats& stats) override {
    ASSERT_EQ(
      current_stats.num_input_deletion_records,
      stats.num_input_deletion_records);
@@ -5,7 +5,13 @@

 #ifndef ROCKSDB_LITE

+#ifndef __STDC_FORMAT_MACROS
+#define __STDC_FORMAT_MACROS
+#endif
+
+#include <inttypes.h>
 #include <algorithm>
+#include <array>
 #include <map>
 #include <string>
 #include <tuple>
@@ -172,7 +178,7 @@ class CompactionJobTest : public testing::Test {
        // This is how the key will look like once it's written in bottommost
        // file
        InternalKey bottommost_internal_key(
-            key, (key == "9999") ? sequence_number : 0, kTypeValue);
+            key, 0, kTypeValue);

        if (corrupt_id(k)) {
          test::CorruptKeyType(&internal_key);
@@ -194,6 +200,13 @@ class CompactionJobTest : public testing::Test {
  }

  void NewDB() {
+    DestroyDB(dbname_, Options());
+    EXPECT_OK(env_->CreateDirIfMissing(dbname_));
+    versions_.reset(new VersionSet(dbname_, &db_options_, env_options_,
+                                   table_cache_.get(), &write_buffer_manager_,
+                                   &write_controller_));
+    compaction_job_stats_.Reset();
+
    VersionEdit new_db;
    new_db.SetLogNumber(0);
    new_db.SetNextFile(2);
@@ -230,7 +243,10 @@ class CompactionJobTest : public testing::Test {
      const std::vector<std::vector<FileMetaData*>>& input_files,
      const stl_wrappers::KVMap& expected_results,
      const std::vector<SequenceNumber>& snapshots = {},
-      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
+      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber,
+      int output_level = 1, bool verify = true,
+      SnapshotListFetchCallback* snapshot_fetcher =
+          SnapshotListFetchCallback::kDisabled) {
    auto cfd = versions_->GetColumnFamilySet()->GetDefault();

    size_t num_input_files = 0;
@@ -247,7 +263,7 @@ class CompactionJobTest : public testing::Test {

    Compaction compaction(cfd->current()->storage_info(), *cfd->ioptions(),
                          *cfd->GetLatestMutableCFOptions(),
-                          compaction_input_files, 1, 1024 * 1024,
+                          compaction_input_files, output_level, 1024 * 1024,
                          10 * 1024 * 1024, 0, kNoCompression,
                          cfd->ioptions()->compression_opts, 0, {}, true);
    compaction.SetInputVersion(cfd->current());
@@ -262,7 +278,8 @@ class CompactionJobTest : public testing::Test {
        &shutting_down_, preserve_deletes_seqnum_, &log_buffer, nullptr,
        nullptr, nullptr, &mutex_, &error_handler_, snapshots,
        earliest_write_conflict_snapshot, snapshot_checker, table_cache_,
-        &event_logger, false, false, dbname_, &compaction_job_stats_);
+        &event_logger, false, false, dbname_, &compaction_job_stats_,
+        Env::Priority::USER, snapshot_fetcher);
    VerifyInitializationOfCompactionJobStats(compaction_job_stats_);

    compaction_job.Prepare();
@@ -274,15 +291,17 @@ class CompactionJobTest : public testing::Test {
    ASSERT_OK(compaction_job.Install(*cfd->GetLatestMutableCFOptions()));
    mutex_.Unlock();

-    if (expected_results.size() == 0) {
-      ASSERT_GE(compaction_job_stats_.elapsed_micros, 0U);
-      ASSERT_EQ(compaction_job_stats_.num_input_files, num_input_files);
-      ASSERT_EQ(compaction_job_stats_.num_output_files, 0U);
-    } else {
-      ASSERT_GE(compaction_job_stats_.elapsed_micros, 0U);
-      ASSERT_EQ(compaction_job_stats_.num_input_files, num_input_files);
-      ASSERT_EQ(compaction_job_stats_.num_output_files, 1U);
-      mock_table_factory_->AssertLatestFile(expected_results);
+    if (verify) {
+      if (expected_results.size() == 0) {
+        ASSERT_GE(compaction_job_stats_.elapsed_micros, 0U);
+        ASSERT_EQ(compaction_job_stats_.num_input_files, num_input_files);
+        ASSERT_EQ(compaction_job_stats_.num_output_files, 0U);
+      } else {
+        ASSERT_GE(compaction_job_stats_.elapsed_micros, 0U);
+        ASSERT_EQ(compaction_job_stats_.num_input_files, num_input_files);
+        ASSERT_EQ(compaction_job_stats_.num_output_files, 1U);
+        mock_table_factory_->AssertLatestFile(expected_results);
+      }
    }
  }

@@ -379,7 +398,7 @@ TEST_F(CompactionJobTest, SimpleOverwrite) {

  auto expected_results =
      mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "val2"},
-                          {KeyStr("b", 4U, kTypeValue), "val3"}});
+                          {KeyStr("b", 0U, kTypeValue), "val3"}});

  SetLastSequence(4U);
  auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -432,7 +451,7 @@ TEST_F(CompactionJobTest, SimpleMerge) {

  auto expected_results =
      mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "3,4,5"},
-                          {KeyStr("b", 2U, kTypeValue), "1,2"}});
+                          {KeyStr("b", 0U, kTypeValue), "1,2"}});

  SetLastSequence(5U);
  auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -456,8 +475,7 @@ TEST_F(CompactionJobTest, NonAssocMerge) {

  auto expected_results =
      mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "3,4,5"},
-                          {KeyStr("b", 2U, kTypeMerge), "2"},
-                          {KeyStr("b", 1U, kTypeMerge), "1"}});
+                          {KeyStr("b", 0U, kTypeValue), "1,2"}});

  SetLastSequence(5U);
  auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -484,7 +502,7 @@ TEST_F(CompactionJobTest, MergeOperandFilter) {

  auto expected_results =
      mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), test::EncodeInt(8U)},
-                          {KeyStr("b", 2U, kTypeMerge), test::EncodeInt(2U)}});
+                          {KeyStr("b", 0U, kTypeValue), test::EncodeInt(2U)}});

  SetLastSequence(5U);
  auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -747,7 +765,7 @@ TEST_F(CompactionJobTest, SingleDeleteZeroSeq) {
  AddMockFile(file2);

  auto expected_results = mock::MakeMockFile({
-      {KeyStr("dummy", 5U, kTypeValue), "val2"},
+      {KeyStr("dummy", 0U, kTypeValue), "val2"},
  });

  SetLastSequence(22U);
@@ -931,13 +949,112 @@ TEST_F(CompactionJobTest, CorruptionAfterDeletion) {
      mock::MakeMockFile({{test::KeyStr("A", 0U, kTypeValue), "val3"},
                          {test::KeyStr("a", 0U, kTypeValue, true), "val"},
                          {test::KeyStr("b", 0U, kTypeValue, true), "val"},
-                          {test::KeyStr("c", 1U, kTypeValue), "val2"}});
+                          {test::KeyStr("c", 0U, kTypeValue), "val2"}});

  SetLastSequence(6U);
  auto files = cfd_->current()->storage_info()->LevelFiles(0);
  RunCompaction({files}, expected_results);
 }

+// Test the snapshot fetcher in compaction
+TEST_F(CompactionJobTest, SnapshotRefresh) {
+  uint64_t time_seed = env_->NowMicros();
+  printf("time_seed is %" PRIu64 "\n", time_seed);  // would help to reproduce
+  Random64 rand(time_seed);
+  std::vector<SequenceNumber> db_snapshots;
+  class SnapshotListFetchCallbackTest : public SnapshotListFetchCallback {
+   public:
+    SnapshotListFetchCallbackTest(Env* env, Random64& rand,
+                                  std::vector<SequenceNumber>* snapshots)
+        : SnapshotListFetchCallback(env, 1 /*short time delay*/,
+                                    1 /*fetch after each key*/),
+          rand_(rand),
+          snapshots_(snapshots) {}
+    virtual void Refresh(std::vector<SequenceNumber>* snapshots,
+                         SequenceNumber) override {
+      assert(snapshots->size());
+      assert(snapshots_->size());
+      assert(snapshots_->size() == snapshots->size());
+      if (rand_.OneIn(2)) {
+        uint64_t release_index = rand_.Uniform(snapshots_->size());
+        snapshots_->erase(snapshots_->begin() + release_index);
+        *snapshots = *snapshots_;
+      }
+    }
+
+   private:
+    Random64 rand_;
+    std::vector<SequenceNumber>* snapshots_;
+  } snapshot_fetcher(env_, rand, &db_snapshots);
+
+  std::vector<std::pair<const std::string, std::string>> file1_kvs, file2_kvs;
+  std::array<ValueType, 4> types = {kTypeValue, kTypeDeletion,
+                                    kTypeSingleDeletion};
+  SequenceNumber last_seq = 0;
+  for (int i = 1; i < 100; i++) {
+    SequenceNumber seq = last_seq + 1;
+    last_seq = seq;
+    if (rand.OneIn(2)) {
+      auto type = types[rand.Uniform(types.size())];
+      file1_kvs.push_back(
+          {test::KeyStr("k" + ToString(i), seq, type), "v" + ToString(i)});
+    }
+  }
+  auto file1 = mock::MakeMockFile(file1_kvs);
+  for (int i = 1; i < 100; i++) {
+    SequenceNumber seq = last_seq + 1;
+    last_seq++;
+    if (rand.OneIn(2)) {
+      auto type = types[rand.Uniform(types.size())];
+      file2_kvs.push_back(
+          {test::KeyStr("k" + ToString(i), seq, type), "v" + ToString(i)});
+    }
+  }
+  auto file2 = mock::MakeMockFile(file2_kvs);
+  for (SequenceNumber i = 1; i < last_seq + 1; i++) {
+    if (rand.OneIn(5)) {
+      db_snapshots.push_back(i);
+    }
+  }
+
+  const bool kVerify = true;
+  const int output_level_0 = 0;
+  NewDB();
+  AddMockFile(file1);
+  AddMockFile(file2);
+  SetLastSequence(last_seq);
+  auto files = cfd_->current()->storage_info()->LevelFiles(0);
+  // put the output on L0 since it is easier to feed them again to the 2nd
+  // compaction
+  RunCompaction({files}, file1, db_snapshots, kMaxSequenceNumber,
+                output_level_0, !kVerify, &snapshot_fetcher);
+
+  // Now db_snapshots are changed. Run the compaction again without snapshot
+  // fetcher but with the updated snapshot list.
+  compaction_job_stats_.Reset();
+  files = cfd_->current()->storage_info()->LevelFiles(0);
+  RunCompaction({files}, file1, db_snapshots, kMaxSequenceNumber,
+                output_level_0 + 1, !kVerify);
+  // The result should be what we get if we run compaction without snapshot
+  // fetcher on the updated list of snapshots
+  auto expected = mock_table_factory_->output();
+
+  NewDB();
+  AddMockFile(file1);
+  AddMockFile(file2);
+  SetLastSequence(last_seq);
+  files = cfd_->current()->storage_info()->LevelFiles(0);
+  RunCompaction({files}, expected, db_snapshots, kMaxSequenceNumber,
+                output_level_0, !kVerify);
+  // The 2nd compaction above would get rid of useless delete markers. To get
+  // the output here exactly as what we got above after two compactions, we also
+  // run the compaction for 2nd time.
+  compaction_job_stats_.Reset();
+  files = cfd_->current()->storage_info()->LevelFiles(0);
+  RunCompaction({files}, expected, db_snapshots, kMaxSequenceNumber,
+                output_level_0 + 1, !kVerify);
+}
+
 }  // namespace rocksdb

 int main(int argc, char** argv) {
@@ -543,9 +543,9 @@ void CompactionPicker::GetGrandparents(
 Compaction* CompactionPicker::CompactRange(
    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
    VersionStorageInfo* vstorage, int input_level, int output_level,
-    uint32_t output_path_id, uint32_t max_subcompactions,
-    const InternalKey* begin, const InternalKey* end,
-    InternalKey** compaction_end, bool* manual_conflict) {
+    const CompactRangeOptions& compact_range_options, const InternalKey* begin,
+    const InternalKey* end, InternalKey** compaction_end, bool* manual_conflict,
+    uint64_t max_file_num_to_ignore) {
  // CompactionPickerFIFO has its own implementation of compact range
  assert(ioptions_.compaction_style != kCompactionStyleFIFO);

@@ -609,11 +609,13 @@ Compaction* CompactionPicker::CompactRange(
        output_level,
        MaxFileSizeForLevel(mutable_cf_options, output_level,
                            ioptions_.compaction_style),
-        /* max_compaction_bytes */ LLONG_MAX, output_path_id,
+        /* max_compaction_bytes */ LLONG_MAX,
+        compact_range_options.target_path_id,
        GetCompressionType(ioptions_, vstorage, mutable_cf_options,
                           output_level, 1),
        GetCompressionOptions(ioptions_, vstorage, output_level),
-        max_subcompactions, /* grandparents */ {}, /* is manual */ true);
+        compact_range_options.max_subcompactions, /* grandparents */ {},
+        /* is manual */ true);
    RegisterCompaction(c);
    return c;
  }
@@ -658,7 +660,43 @@ Compaction* CompactionPicker::CompactRange(
      }
    }
  }
-  assert(output_path_id < static_cast<uint32_t>(ioptions_.cf_paths.size()));
+  assert(compact_range_options.target_path_id <
+         static_cast<uint32_t>(ioptions_.cf_paths.size()));
+
+  // for BOTTOM LEVEL compaction only, use max_file_num_to_ignore to filter out
+  // files that are created during the current compaction.
+  if (compact_range_options.bottommost_level_compaction ==
+          BottommostLevelCompaction::kForceOptimized &&
+      max_file_num_to_ignore != port::kMaxUint64) {
+    assert(input_level == output_level);
+    // inputs_shrunk holds a continuous subset of input files which were all
+    // created before the current manual compaction
+    std::vector<FileMetaData*> inputs_shrunk;
+    size_t skip_input_index = inputs.size();
+    for (size_t i = 0; i < inputs.size(); ++i) {
+      if (inputs[i]->fd.GetNumber() < max_file_num_to_ignore) {
+        inputs_shrunk.push_back(inputs[i]);
+      } else if (!inputs_shrunk.empty()) {
+        // inputs[i] was created during the current manual compaction and
+        // need to be skipped
+        skip_input_index = i;
+        break;
+      }
+    }
+    if (inputs_shrunk.empty()) {
+      return nullptr;
+    }
+    if (inputs.size() != inputs_shrunk.size()) {
+      inputs.files.swap(inputs_shrunk);
+    }
+    // set covering_the_whole_range to false if there is any file that need to
+    // be compacted in the range of inputs[skip_input_index+1, inputs.size())
+    for (size_t i = skip_input_index + 1; i < inputs.size(); ++i) {
+      if (inputs[i]->fd.GetNumber() < max_file_num_to_ignore) {
+        covering_the_whole_range = false;
+      }
+    }
+  }

  InternalKey key_storage;
  InternalKey* next_smallest = &key_storage;
@@ -724,11 +762,12 @@ Compaction* CompactionPicker::CompactRange(
      MaxFileSizeForLevel(mutable_cf_options, output_level,
                          ioptions_.compaction_style, vstorage->base_level(),
                          ioptions_.level_compaction_dynamic_level_bytes),
-      mutable_cf_options.max_compaction_bytes, output_path_id,
+      mutable_cf_options.max_compaction_bytes,
+      compact_range_options.target_path_id,
      GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
                         vstorage->base_level()),
      GetCompressionOptions(ioptions_, vstorage, output_level),
-      max_subcompactions, std::move(grandparents),
+      compact_range_options.max_subcompactions, std::move(grandparents),
      /* is manual compaction */ true);

  TEST_SYNC_POINT_CALLBACK("CompactionPicker::CompactRange:Return", compaction);
@@ -1073,6 +1112,9 @@ bool LevelCompactionPicker::NeedsCompaction(
  if (!vstorage->ExpiredTtlFiles().empty()) {
    return true;
  }
+  if (!vstorage->FilesMarkedForPeriodicCompaction().empty()) {
+    return true;
+  }
  if (!vstorage->BottommostFilesMarkedForCompaction().empty()) {
    return true;
  }
@@ -1141,6 +1183,8 @@ class LevelCompactionBuilder {

  void PickExpiredTtlFiles();

+  void PickFilesMarkedForPeriodicCompaction();
+
  const std::string& cf_name_;
  VersionStorageInfo* vstorage_;
  CompactionPicker* compaction_picker_;
@@ -1203,6 +1247,39 @@ void LevelCompactionBuilder::PickExpiredTtlFiles() {
  start_level_inputs_.files.clear();
 }

+void LevelCompactionBuilder::PickFilesMarkedForPeriodicCompaction() {
+  if (vstorage_->FilesMarkedForPeriodicCompaction().empty()) {
+    return;
+  }
+
+  auto continuation = [&](std::pair<int, FileMetaData*> level_file) {
+    // If it's being compacted it has nothing to do here.
+    // If this assert() fails that means that some function marked some
+    // files as being_compacted, but didn't call ComputeCompactionScore()
+    assert(!level_file.second->being_compacted);
+    output_level_ = start_level_ = level_file.first;
+
+    if (start_level_ == 0 &&
+        !compaction_picker_->level0_compactions_in_progress()->empty()) {
+      return false;
+    }
+
+    start_level_inputs_.files = {level_file.second};
+    start_level_inputs_.level = start_level_;
+    return compaction_picker_->ExpandInputsToCleanCut(cf_name_, vstorage_,
+                                                      &start_level_inputs_);
+  };
+
+  for (auto& level_file : vstorage_->FilesMarkedForPeriodicCompaction()) {
+    if (continuation(level_file)) {
+      // found the compaction!
+      return;
+    }
+  }
+
+  start_level_inputs_.files.clear();
+}
+
 void LevelCompactionBuilder::SetupInitialFiles() {
  // Find the compactions by size on all levels.
  bool skipped_l0_to_base = false;
@@ -1256,7 +1333,6 @@ void LevelCompactionBuilder::SetupInitialFiles() {
  if (start_level_inputs_.empty()) {
    parent_index_ = base_index_ = -1;

-    // PickFilesMarkedForCompaction();
    compaction_picker_->PickFilesMarkedForCompaction(
        cf_name_, vstorage_, &start_level_, &output_level_, &start_level_inputs_);
    if (!start_level_inputs_.empty()) {
@@ -1264,7 +1340,10 @@ void LevelCompactionBuilder::SetupInitialFiles() {
      compaction_reason_ = CompactionReason::kFilesMarkedForCompaction;
      return;
    }
+  }

+  // Bottommost Files Compaction on deleting tombstones
+  if (start_level_inputs_.empty()) {
    size_t i;
    for (i = 0; i < vstorage_->BottommostFilesMarkedForCompaction().size();
         ++i) {
@@ -1285,11 +1364,23 @@ void LevelCompactionBuilder::SetupInitialFiles() {
      compaction_reason_ = CompactionReason::kBottommostFiles;
      return;
    }
+  }

-    assert(start_level_inputs_.empty());
+  // TTL Compaction
+  if (start_level_inputs_.empty()) {
    PickExpiredTtlFiles();
    if (!start_level_inputs_.empty()) {
      compaction_reason_ = CompactionReason::kTtl;
+      return;
+    }
+  }
+
+  // Periodic Compaction
+  if (start_level_inputs_.empty()) {
+    PickFilesMarkedForPeriodicCompaction();
+    if (!start_level_inputs_.empty()) {
+      compaction_reason_ = CompactionReason::kPeriodicCompaction;
+      return;
    }
  }
 }
@@ -58,9 +58,10 @@ class CompactionPicker {
  virtual Compaction* CompactRange(
      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
      VersionStorageInfo* vstorage, int input_level, int output_level,
-      uint32_t output_path_id, uint32_t max_subcompactions,
+      const CompactRangeOptions& compact_range_options,
      const InternalKey* begin, const InternalKey* end,
-      InternalKey** compaction_end, bool* manual_conflict);
+      InternalKey** compaction_end, bool* manual_conflict,
+      uint64_t max_file_num_to_ignore);

  // The maximum allowed output level.  Default value is NumberLevels() - 1.
  virtual int MaxOutputLevel() const { return NumberLevels() - 1; }
@@ -255,12 +256,12 @@ class NullCompactionPicker : public CompactionPicker {
                           const MutableCFOptions& /*mutable_cf_options*/,
                           VersionStorageInfo* /*vstorage*/,
                           int /*input_level*/, int /*output_level*/,
-                           uint32_t /*output_path_id*/,
-                           uint32_t /*max_subcompactions*/,
+                           const CompactRangeOptions& /*compact_range_options*/,
                           const InternalKey* /*begin*/,
                           const InternalKey* /*end*/,
                           InternalKey** /*compaction_end*/,
-                           bool* /*manual_conflict*/) override {
+                           bool* /*manual_conflict*/,
+                           uint64_t /*max_file_num_to_ignore*/) override {
    return nullptr;
  }

@@ -41,7 +41,7 @@ bool FIFOCompactionPicker::NeedsCompaction(
 Compaction* FIFOCompactionPicker::PickTTLCompaction(
    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
-  assert(mutable_cf_options.compaction_options_fifo.ttl > 0);
+  assert(mutable_cf_options.ttl > 0);

  const int kLevel0 = 0;
  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
@@ -63,7 +63,7 @@ Compaction* FIFOCompactionPicker::PickTTLCompaction(
  inputs[0].level = 0;

  // avoid underflow
-  if (current_time > mutable_cf_options.compaction_options_fifo.ttl) {
+  if (current_time > mutable_cf_options.ttl) {
    for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
      auto f = *ritr;
      if (f->fd.table_reader != nullptr &&
@@ -71,8 +71,7 @@ Compaction* FIFOCompactionPicker::PickTTLCompaction(
        auto creation_time =
            f->fd.table_reader->GetTableProperties()->creation_time;
        if (creation_time == 0 ||
-            creation_time >= (current_time -
-                              mutable_cf_options.compaction_options_fifo.ttl)) {
+            creation_time >= (current_time - mutable_cf_options.ttl)) {
          break;
        }
        total_size -= f->compensated_file_size;
@@ -201,7 +200,7 @@ Compaction* FIFOCompactionPicker::PickCompaction(
  assert(vstorage->num_levels() == 1);

  Compaction* c = nullptr;
-  if (mutable_cf_options.compaction_options_fifo.ttl > 0) {
+  if (mutable_cf_options.ttl > 0) {
    c = PickTTLCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
  }
  if (c == nullptr) {
@@ -214,9 +213,10 @@ Compaction* FIFOCompactionPicker::PickCompaction(
 Compaction* FIFOCompactionPicker::CompactRange(
    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
    VersionStorageInfo* vstorage, int input_level, int output_level,
-    uint32_t /*output_path_id*/, uint32_t /*max_subcompactions*/,
+    const CompactRangeOptions& /*compact_range_options*/,
    const InternalKey* /*begin*/, const InternalKey* /*end*/,
-    InternalKey** compaction_end, bool* /*manual_conflict*/) {
+    InternalKey** compaction_end, bool* /*manual_conflict*/,
+    uint64_t /*max_file_num_to_ignore*/) {
 #ifdef NDEBUG
  (void)input_level;
  (void)output_level;
@@ -27,9 +27,10 @@ class FIFOCompactionPicker : public CompactionPicker {
  virtual Compaction* CompactRange(
      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
      VersionStorageInfo* vstorage, int input_level, int output_level,
-      uint32_t output_path_id, uint32_t max_subcompactions,
+      const CompactRangeOptions& compact_range_options,
      const InternalKey* begin, const InternalKey* end,
-      InternalKey** compaction_end, bool* manual_conflict) override;
+      InternalKey** compaction_end, bool* manual_conflict,
+      uint64_t max_file_num_to_ignore) override;

  // The maximum allowed output level.  Always returns 0.
  virtual int MaxOutputLevel() const override { return 0; }
@@ -22,9 +22,7 @@ namespace rocksdb {
 class CountingLogger : public Logger {
 public:
  using Logger::Logv;
-  virtual void Logv(const char* /*format*/, va_list /*ap*/) override {
-    log_count++;
-  }
+  void Logv(const char* /*format*/, va_list /*ap*/) override { log_count++; }
  size_t log_count;
 };

@@ -59,14 +57,16 @@ class CompactionPickerTest : public testing::Test {
        log_buffer_(InfoLogLevel::INFO_LEVEL, &logger_),
        file_num_(1),
        vstorage_(nullptr) {
+    // ioptions_.compaction_pri = kMinOverlappingRatio has its own set of
+    // tests to cover.
+    ioptions_.compaction_pri = kByCompensatedSize;
    fifo_options_.max_table_files_size = 1;
    mutable_cf_options_.RefreshDerivedOptions(ioptions_);
    ioptions_.cf_paths.emplace_back("dummy",
                                    std::numeric_limits<uint64_t>::max());
  }

-  ~CompactionPickerTest() {
-  }
+  ~CompactionPickerTest() override {}

  void NewVersionStorage(int num_levels, CompactionStyle style) {
    DeleteVersionStorage();
@@ -85,8 +85,8 @@ class CompactionPickerTest : public testing::Test {

  void Add(int level, uint32_t file_number, const char* smallest,
           const char* largest, uint64_t file_size = 1, uint32_t path_id = 0,
-           SequenceNumber smallest_seq = 100,
-           SequenceNumber largest_seq = 100) {
+           SequenceNumber smallest_seq = 100, SequenceNumber largest_seq = 100,
+           size_t compensated_file_size = 0) {
    assert(level < vstorage_->num_levels());
    FileMetaData* f = new FileMetaData;
    f->fd = FileDescriptor(file_number, path_id, file_size);
@@ -94,7 +94,8 @@ class CompactionPickerTest : public testing::Test {
    f->largest = InternalKey(largest, largest_seq, kTypeValue);
    f->fd.smallest_seqno = smallest_seq;
    f->fd.largest_seqno = largest_seq;
-    f->compensated_file_size = file_size;
+    f->compensated_file_size =
+        (compensated_file_size != 0) ? compensated_file_size : file_size;
    f->refs = 0;
    vstorage_->AddFile(level, f);
    files_.emplace_back(f);
@@ -617,6 +618,35 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping3) {
  ASSERT_EQ(8U, compaction->input(0, 0)->fd.GetNumber());
 }

+TEST_F(CompactionPickerTest, CompactionPriMinOverlapping4) {
+  NewVersionStorage(6, kCompactionStyleLevel);
+  ioptions_.compaction_pri = kMinOverlappingRatio;
+  mutable_cf_options_.max_bytes_for_level_base = 10000000;
+  mutable_cf_options_.max_bytes_for_level_multiplier = 10;
+
+  // file 7 and 8 over lap with the same file, but file 8 is smaller so
+  // it will be picked.
+  // Overlaps with file 26, 27. And the file is compensated so will be
+  // picked up.
+  Add(2, 6U, "150", "167", 60000000U, 0, 100, 100, 180000000U);
+  Add(2, 7U, "168", "169", 60000000U);  // Overlaps with file 27
+  Add(2, 8U, "201", "300", 61000000U);  // Overlaps with file 28
+
+  Add(3, 26U, "160", "165", 60000000U);
+  // Boosted file size in output level is not considered.
+  Add(3, 27U, "166", "170", 60000000U, 0, 100, 100, 260000000U);
+  Add(3, 28U, "180", "400", 60000000U);
+  Add(3, 29U, "401", "500", 60000000U);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  // Picking file 8 because overlapping ratio is the biggest.
+  ASSERT_EQ(6U, compaction->input(0, 0)->fd.GetNumber());
+}
+
 // This test exhibits the bug where we don't properly reset parent_index in
 // PickCompaction()
 TEST_F(CompactionPickerTest, ParentIndexResetBug) {
@@ -395,8 +395,8 @@ Compaction* UniversalCompactionPicker::PickCompaction(
  }
 #endif
  // update statistics
-  MeasureTime(ioptions_.statistics, NUM_FILES_IN_SINGLE_COMPACTION,
-              c->inputs(0)->size());
+  RecordInHistogram(ioptions_.statistics, NUM_FILES_IN_SINGLE_COMPACTION,
+                    c->inputs(0)->size());

  RegisterCompaction(c);
  vstorage->ComputeCompactionScore(ioptions_, mutable_cf_options);
@@ -27,24 +27,24 @@ class KVIter : public Iterator {
 public:
  explicit KVIter(const stl_wrappers::KVMap* map)
      : map_(map), iter_(map_->end()) {}
-  virtual bool Valid() const override { return iter_ != map_->end(); }
-  virtual void SeekToFirst() override { iter_ = map_->begin(); }
-  virtual void SeekToLast() override {
+  bool Valid() const override { return iter_ != map_->end(); }
+  void SeekToFirst() override { iter_ = map_->begin(); }
+  void SeekToLast() override {
    if (map_->empty()) {
      iter_ = map_->end();
    } else {
      iter_ = map_->find(map_->rbegin()->first);
    }
  }
-  virtual void Seek(const Slice& k) override {
+  void Seek(const Slice& k) override {
    iter_ = map_->lower_bound(k.ToString());
  }
-  virtual void SeekForPrev(const Slice& k) override {
+  void SeekForPrev(const Slice& k) override {
    iter_ = map_->upper_bound(k.ToString());
    Prev();
  }
-  virtual void Next() override { ++iter_; }
-  virtual void Prev() override {
+  void Next() override { ++iter_; }
+  void Prev() override {
    if (iter_ == map_->begin()) {
      iter_ = map_->end();
      return;
@@ -52,9 +52,9 @@ class KVIter : public Iterator {
    --iter_;
  }

-  virtual Slice key() const override { return iter_->first; }
-  virtual Slice value() const override { return iter_->second; }
-  virtual Status status() const override { return Status::OK(); }
+  Slice key() const override { return iter_->first; }
+  Slice value() const override { return iter_->second; }
+  Status status() const override { return Status::OK(); }

 private:
  const stl_wrappers::KVMap* const map_;
@@ -171,9 +171,9 @@ class DoubleComparator : public Comparator {
 public:
  DoubleComparator() {}

-  virtual const char* Name() const override { return "DoubleComparator"; }
+  const char* Name() const override { return "DoubleComparator"; }

-  virtual int Compare(const Slice& a, const Slice& b) const override {
+  int Compare(const Slice& a, const Slice& b) const override {
 #ifndef CYGWIN
    double da = std::stod(a.ToString());
    double db = std::stod(b.ToString());
@@ -189,19 +189,19 @@ class DoubleComparator : public Comparator {
      return -1;
    }
  }
-  virtual void FindShortestSeparator(std::string* /*start*/,
-                                     const Slice& /*limit*/) const override {}
+  void FindShortestSeparator(std::string* /*start*/,
+                             const Slice& /*limit*/) const override {}

-  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
+  void FindShortSuccessor(std::string* /*key*/) const override {}
 };

 class HashComparator : public Comparator {
 public:
  HashComparator() {}

-  virtual const char* Name() const override { return "HashComparator"; }
+  const char* Name() const override { return "HashComparator"; }

-  virtual int Compare(const Slice& a, const Slice& b) const override {
+  int Compare(const Slice& a, const Slice& b) const override {
    uint32_t ha = Hash(a.data(), a.size(), 66);
    uint32_t hb = Hash(b.data(), b.size(), 66);
    if (ha == hb) {
@@ -212,19 +212,19 @@ class HashComparator : public Comparator {
      return -1;
    }
  }
-  virtual void FindShortestSeparator(std::string* /*start*/,
-                                     const Slice& /*limit*/) const override {}
+  void FindShortestSeparator(std::string* /*start*/,
+                             const Slice& /*limit*/) const override {}

-  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
+  void FindShortSuccessor(std::string* /*key*/) const override {}
 };

 class TwoStrComparator : public Comparator {
 public:
  TwoStrComparator() {}

-  virtual const char* Name() const override { return "TwoStrComparator"; }
+  const char* Name() const override { return "TwoStrComparator"; }

-  virtual int Compare(const Slice& a, const Slice& b) const override {
+  int Compare(const Slice& a, const Slice& b) const override {
    assert(a.size() >= 2);
    assert(b.size() >= 2);
    size_t size_a1 = static_cast<size_t>(a[0]);
@@ -244,10 +244,10 @@ class TwoStrComparator : public Comparator {
    }
    return a2.compare(b2);
  }
-  virtual void FindShortestSeparator(std::string* /*start*/,
-                                     const Slice& /*limit*/) const override {}
+  void FindShortestSeparator(std::string* /*start*/,
+                             const Slice& /*limit*/) const override {}

-  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
+  void FindShortSuccessor(std::string* /*key*/) const override {}
 };
 }  // namespace

@@ -272,7 +272,7 @@ class ComparatorDBTest
    EXPECT_OK(DestroyDB(dbname_, last_options_));
  }

-  ~ComparatorDBTest() {
+  ~ComparatorDBTest() override {
    delete db_;
    EXPECT_OK(DestroyDB(dbname_, last_options_));
    comparator = BytewiseComparator();
@@ -24,6 +24,8 @@
 #include "rocksdb/env.h"
 #include "rocksdb/table.h"
 #include "rocksdb/write_batch.h"
+#include "table/block_based_table_builder.h"
+#include "table/meta_blocks.h"
 #include "util/filename.h"
 #include "util/string_util.h"
 #include "util/testharness.h"
@@ -60,9 +62,9 @@ class CorruptionTest : public testing::Test {
    options_.create_if_missing = false;
  }

-  ~CorruptionTest() {
-     delete db_;
-     DestroyDB(dbname_, Options());
+  ~CorruptionTest() override {
+    delete db_;
+    DestroyDB(dbname_, Options());
  }

  void CloseDb() {
@@ -467,6 +469,39 @@ TEST_F(CorruptionTest, UnrelatedKeys) {
  ASSERT_EQ(Value(1000, &tmp2).ToString(), v);
 }

+TEST_F(CorruptionTest, RangeDeletionCorrupted) {
+  ASSERT_OK(
+      db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(), "a", "b"));
+  ASSERT_OK(db_->Flush(FlushOptions()));
+  std::vector<LiveFileMetaData> metadata;
+  db_->GetLiveFilesMetaData(&metadata);
+  ASSERT_EQ(static_cast<size_t>(1), metadata.size());
+  std::string filename = dbname_ + metadata[0].name;
+
+  std::unique_ptr<RandomAccessFile> file;
+  ASSERT_OK(options_.env->NewRandomAccessFile(filename, &file, EnvOptions()));
+  std::unique_ptr<RandomAccessFileReader> file_reader(
+      new RandomAccessFileReader(std::move(file), filename));
+
+  uint64_t file_size;
+  ASSERT_OK(options_.env->GetFileSize(filename, &file_size));
+
+  BlockHandle range_del_handle;
+  ASSERT_OK(FindMetaBlock(
+      file_reader.get(), file_size, kBlockBasedTableMagicNumber,
+      ImmutableCFOptions(options_), kRangeDelBlock, &range_del_handle));
+
+  ASSERT_OK(TryReopen());
+  CorruptFile(filename, static_cast<int>(range_del_handle.offset()), 1);
+  // The test case does not fail on TryReopen because failure to preload table
+  // handlers is not considered critical.
+  ASSERT_OK(TryReopen());
+  std::string val;
+  // However, it does fail on any read involving that file since that file
+  // cannot be opened with a corrupt range deletion meta-block.
+  ASSERT_TRUE(db_->Get(ReadOptions(), "a", &val).IsCorruption());
+}
+
 TEST_F(CorruptionTest, FileSystemStateCorrupted) {
  for (int iter = 0; iter < 2; ++iter) {
    Options options;
@@ -31,7 +31,7 @@ class CuckooTableDBTest : public testing::Test {
    Reopen();
  }

-  ~CuckooTableDBTest() {
+  ~CuckooTableDBTest() override {
    delete db_;
    EXPECT_OK(DestroyDB(dbname_, Options()));
  }
@@ -6,6 +6,7 @@
 // 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 <iostream>
 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
 #include "rocksdb/perf_context.h"
@@ -214,11 +215,11 @@ TEST_F(DBBasicTest, PutSingleDeleteGet) {
    ASSERT_EQ("v2", Get(1, "foo2"));
    ASSERT_OK(SingleDelete(1, "foo"));
    ASSERT_EQ("NOT_FOUND", Get(1, "foo"));
-    // Skip HashCuckooRep as it does not support single delete. FIFO and
-    // universal compaction do not apply to the test case. Skip MergePut
-    // because single delete does not get removed when it encounters a merge.
-  } while (ChangeOptions(kSkipHashCuckoo | kSkipFIFOCompaction |
-                         kSkipUniversalCompaction | kSkipMergePut));
+    // Ski FIFO and universal compaction because they do not apply to the test
+    // case. Skip MergePut because single delete does not get removed when it
+    // encounters a merge.
+  } while (ChangeOptions(kSkipFIFOCompaction | kSkipUniversalCompaction |
+                         kSkipMergePut));
 }

 TEST_F(DBBasicTest, EmptyFlush) {
@@ -236,11 +237,11 @@ TEST_F(DBBasicTest, EmptyFlush) {
    ASSERT_OK(Flush(1));

    ASSERT_EQ("[ ]", AllEntriesFor("a", 1));
-    // Skip HashCuckooRep as it does not support single delete. FIFO and
-    // universal compaction do not apply to the test case. Skip MergePut
-    // because merges cannot be combined with single deletions.
-  } while (ChangeOptions(kSkipHashCuckoo | kSkipFIFOCompaction |
-                         kSkipUniversalCompaction | kSkipMergePut));
+    // Skip FIFO and  universal compaction as they do not apply to the test
+    // case. Skip MergePut because merges cannot be combined with single
+    // deletions.
+  } while (ChangeOptions(kSkipFIFOCompaction | kSkipUniversalCompaction |
+                         kSkipMergePut));
 }

 TEST_F(DBBasicTest, GetFromVersions) {
@@ -264,11 +265,6 @@ TEST_F(DBBasicTest, GetSnapshot) {
      std::string key = (i == 0) ? std::string("foo") : std::string(200, 'x');
      ASSERT_OK(Put(1, key, "v1"));
      const Snapshot* s1 = db_->GetSnapshot();
-      if (option_config_ == kHashCuckoo) {
-        // Unsupported case.
-        ASSERT_TRUE(s1 == nullptr);
-        break;
-      }
      ASSERT_OK(Put(1, key, "v2"));
      ASSERT_EQ("v2", Get(1, key));
      ASSERT_EQ("v1", Get(1, key, s1));
@@ -509,7 +505,7 @@ TEST_F(DBBasicTest, Snapshot) {
    ASSERT_EQ(0U, GetNumSnapshots());
    ASSERT_EQ("0v4", Get(0, "foo"));
    ASSERT_EQ("1v4", Get(1, "foo"));
-  } while (ChangeOptions(kSkipHashCuckoo));
+  } while (ChangeOptions());
 }

 #endif  // ROCKSDB_LITE
@@ -565,8 +561,7 @@ TEST_F(DBBasicTest, CompactBetweenSnapshots) {
                           nullptr);
    ASSERT_EQ("sixth", Get(1, "foo"));
    ASSERT_EQ(AllEntriesFor("foo", 1), "[ sixth ]");
-    // skip HashCuckooRep as it does not support snapshot
-  } while (ChangeOptions(kSkipHashCuckoo | kSkipFIFOCompaction));
+  } while (ChangeOptions(kSkipFIFOCompaction));
 }

 TEST_F(DBBasicTest, DBOpen_Options) {
@@ -857,18 +852,17 @@ class TestEnv : public EnvWrapper {
      public:
        using Logger::Logv;
        TestLogger(TestEnv *env_ptr) : Logger() { env = env_ptr; }
-        ~TestLogger() {
+        ~TestLogger() override {
          if (!closed_) {
            CloseHelper();
          }
        }
-        virtual void Logv(const char* /*format*/, va_list /*ap*/) override{};
+        void Logv(const char* /*format*/, va_list /*ap*/) override{};

       protected:
-        virtual Status CloseImpl() override {
-          return CloseHelper();
-        }
-      private:
+        Status CloseImpl() override { return CloseHelper(); }
+
+       private:
        Status CloseHelper() {
          env->CloseCountInc();;
          return Status::IOError();
@@ -880,8 +874,8 @@ class TestEnv : public EnvWrapper {

    int GetCloseCount() { return close_count; }

-    virtual Status NewLogger(const std::string& /*fname*/,
-                             std::shared_ptr<Logger>* result) {
+    Status NewLogger(const std::string& /*fname*/,
+                     std::shared_ptr<Logger>* result) override {
      result->reset(new TestLogger(this));
      return Status::OK();
    }
@@ -956,6 +950,340 @@ TEST_F(DBBasicTest, DBCloseFlushError) {
  Destroy(options);
 }

+TEST_F(DBBasicTest, MultiGetMultiCF) {
+  Options options = CurrentOptions();
+  CreateAndReopenWithCF({"pikachu", "ilya", "muromec", "dobrynia", "nikitich",
+                         "alyosha", "popovich"},
+                        options);
+
+  for (int i = 0; i < 8; ++i) {
+    ASSERT_OK(Put(i, "cf" + std::to_string(i) + "_key",
+                  "cf" + std::to_string(i) + "_val"));
+  }
+
+  int get_sv_count = 0;
+  rocksdb::DBImpl* db = reinterpret_cast<DBImpl*>(db_);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::MultiGet::AfterRefSV", [&](void* /*arg*/) {
+        if (++get_sv_count == 2) {
+          // After MultiGet refs a couple of CFs, flush all CFs so MultiGet
+          // is forced to repeat the process
+          for (int i = 0; i < 8; ++i) {
+            ASSERT_OK(Flush(i));
+            ASSERT_OK(Put(i, "cf" + std::to_string(i) + "_key",
+                          "cf" + std::to_string(i) + "_val2"));
+          }
+        }
+        if (get_sv_count == 11) {
+          for (int i = 0; i < 8; ++i) {
+            auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(
+                            db->GetColumnFamilyHandle(i))
+                            ->cfd();
+            ASSERT_EQ(cfd->TEST_GetLocalSV()->Get(), SuperVersion::kSVInUse);
+          }
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  std::vector<int> cfs;
+  std::vector<std::string> keys;
+  std::vector<std::string> values;
+
+  for (int i = 0; i < 8; ++i) {
+    cfs.push_back(i);
+    keys.push_back("cf" + std::to_string(i) + "_key");
+  }
+
+  values = MultiGet(cfs, keys, nullptr);
+  ASSERT_EQ(values.size(), 8);
+  for (unsigned int j = 0; j < values.size(); ++j) {
+    ASSERT_EQ(values[j], "cf" + std::to_string(j) + "_val2");
+  }
+  for (int i = 0; i < 8; ++i) {
+    auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(
+                    reinterpret_cast<DBImpl*>(db_)->GetColumnFamilyHandle(i))
+                    ->cfd();
+    ASSERT_NE(cfd->TEST_GetLocalSV()->Get(), SuperVersion::kSVInUse);
+    ASSERT_NE(cfd->TEST_GetLocalSV()->Get(), SuperVersion::kSVObsolete);
+  }
+}
+
+TEST_F(DBBasicTest, MultiGetMultiCFMutex) {
+  Options options = CurrentOptions();
+  CreateAndReopenWithCF({"pikachu", "ilya", "muromec", "dobrynia", "nikitich",
+                         "alyosha", "popovich"},
+                        options);
+
+  for (int i = 0; i < 8; ++i) {
+    ASSERT_OK(Put(i, "cf" + std::to_string(i) + "_key",
+                  "cf" + std::to_string(i) + "_val"));
+  }
+
+  int get_sv_count = 0;
+  int retries = 0;
+  bool last_try = false;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::MultiGet::LastTry", [&](void* /*arg*/) {
+        last_try = true;
+        rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+      });
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::MultiGet::AfterRefSV", [&](void* /*arg*/) {
+        if (last_try) {
+          return;
+        }
+        if (++get_sv_count == 2) {
+          ++retries;
+          get_sv_count = 0;
+          for (int i = 0; i < 8; ++i) {
+            ASSERT_OK(Flush(i));
+            ASSERT_OK(Put(
+                i, "cf" + std::to_string(i) + "_key",
+                "cf" + std::to_string(i) + "_val" + std::to_string(retries)));
+          }
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  std::vector<int> cfs;
+  std::vector<std::string> keys;
+  std::vector<std::string> values;
+
+  for (int i = 0; i < 8; ++i) {
+    cfs.push_back(i);
+    keys.push_back("cf" + std::to_string(i) + "_key");
+  }
+
+  values = MultiGet(cfs, keys, nullptr);
+  ASSERT_TRUE(last_try);
+  ASSERT_EQ(values.size(), 8);
+  for (unsigned int j = 0; j < values.size(); ++j) {
+    ASSERT_EQ(values[j],
+              "cf" + std::to_string(j) + "_val" + std::to_string(retries));
+  }
+  for (int i = 0; i < 8; ++i) {
+    auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(
+                    reinterpret_cast<DBImpl*>(db_)->GetColumnFamilyHandle(i))
+                    ->cfd();
+    ASSERT_NE(cfd->TEST_GetLocalSV()->Get(), SuperVersion::kSVInUse);
+  }
+}
+
+TEST_F(DBBasicTest, MultiGetMultiCFSnapshot) {
+  Options options = CurrentOptions();
+  CreateAndReopenWithCF({"pikachu", "ilya", "muromec", "dobrynia", "nikitich",
+                         "alyosha", "popovich"},
+                        options);
+
+  for (int i = 0; i < 8; ++i) {
+    ASSERT_OK(Put(i, "cf" + std::to_string(i) + "_key",
+                  "cf" + std::to_string(i) + "_val"));
+  }
+
+  int get_sv_count = 0;
+  rocksdb::DBImpl* db = reinterpret_cast<DBImpl*>(db_);
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::MultiGet::AfterRefSV", [&](void* /*arg*/) {
+        if (++get_sv_count == 2) {
+          for (int i = 0; i < 8; ++i) {
+            ASSERT_OK(Flush(i));
+            ASSERT_OK(Put(i, "cf" + std::to_string(i) + "_key",
+                          "cf" + std::to_string(i) + "_val2"));
+          }
+        }
+        if (get_sv_count == 8) {
+          for (int i = 0; i < 8; ++i) {
+            auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(
+                            db->GetColumnFamilyHandle(i))
+                            ->cfd();
+            ASSERT_TRUE(
+                (cfd->TEST_GetLocalSV()->Get() == SuperVersion::kSVInUse) ||
+                (cfd->TEST_GetLocalSV()->Get() == SuperVersion::kSVObsolete));
+          }
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  std::vector<int> cfs;
+  std::vector<std::string> keys;
+  std::vector<std::string> values;
+
+  for (int i = 0; i < 8; ++i) {
+    cfs.push_back(i);
+    keys.push_back("cf" + std::to_string(i) + "_key");
+  }
+
+  const Snapshot* snapshot = db_->GetSnapshot();
+  values = MultiGet(cfs, keys, snapshot);
+  db_->ReleaseSnapshot(snapshot);
+  ASSERT_EQ(values.size(), 8);
+  for (unsigned int j = 0; j < values.size(); ++j) {
+    ASSERT_EQ(values[j], "cf" + std::to_string(j) + "_val");
+  }
+  for (int i = 0; i < 8; ++i) {
+    auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(
+                    reinterpret_cast<DBImpl*>(db_)->GetColumnFamilyHandle(i))
+                    ->cfd();
+    ASSERT_NE(cfd->TEST_GetLocalSV()->Get(), SuperVersion::kSVInUse);
+  }
+}
+
+TEST_F(DBBasicTest, MultiGetBatchedSimpleUnsorted) {
+  do {
+    CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
+    SetPerfLevel(kEnableCount);
+    ASSERT_OK(Put(1, "k1", "v1"));
+    ASSERT_OK(Put(1, "k2", "v2"));
+    ASSERT_OK(Put(1, "k3", "v3"));
+    ASSERT_OK(Put(1, "k4", "v4"));
+    ASSERT_OK(Delete(1, "k4"));
+    ASSERT_OK(Put(1, "k5", "v5"));
+    ASSERT_OK(Delete(1, "no_key"));
+
+    get_perf_context()->Reset();
+
+    std::vector<Slice> keys({"no_key", "k5", "k4", "k3", "k2", "k1"});
+    std::vector<PinnableSlice> values(keys.size());
+    std::vector<ColumnFamilyHandle*> cfs(keys.size(), handles_[1]);
+    std::vector<Status> s(keys.size());
+
+    db_->MultiGet(ReadOptions(), handles_[1], keys.size(), keys.data(),
+                  values.data(), s.data(), false);
+
+    ASSERT_EQ(values.size(), keys.size());
+    ASSERT_EQ(std::string(values[5].data(), values[5].size()), "v1");
+    ASSERT_EQ(std::string(values[4].data(), values[4].size()), "v2");
+    ASSERT_EQ(std::string(values[3].data(), values[3].size()), "v3");
+    ASSERT_EQ(std::string(values[1].data(), values[1].size()), "v5");
+    // four kv pairs * two bytes per value
+    ASSERT_EQ(8, (int)get_perf_context()->multiget_read_bytes);
+
+    ASSERT_TRUE(s[0].IsNotFound());
+    ASSERT_OK(s[1]);
+    ASSERT_TRUE(s[2].IsNotFound());
+    ASSERT_OK(s[3]);
+    ASSERT_OK(s[4]);
+    ASSERT_OK(s[5]);
+
+    SetPerfLevel(kDisable);
+  } while (ChangeCompactOptions());
+}
+
+TEST_F(DBBasicTest, MultiGetBatchedSimpleSorted) {
+  do {
+    CreateAndReopenWithCF({"pikachu"}, CurrentOptions());
+    SetPerfLevel(kEnableCount);
+    ASSERT_OK(Put(1, "k1", "v1"));
+    ASSERT_OK(Put(1, "k2", "v2"));
+    ASSERT_OK(Put(1, "k3", "v3"));
+    ASSERT_OK(Put(1, "k4", "v4"));
+    ASSERT_OK(Delete(1, "k4"));
+    ASSERT_OK(Put(1, "k5", "v5"));
+    ASSERT_OK(Delete(1, "no_key"));
+
+    get_perf_context()->Reset();
+
+    std::vector<Slice> keys({"k1", "k2", "k3", "k4", "k5", "no_key"});
+    std::vector<PinnableSlice> values(keys.size());
+    std::vector<ColumnFamilyHandle*> cfs(keys.size(), handles_[1]);
+    std::vector<Status> s(keys.size());
+
+    db_->MultiGet(ReadOptions(), handles_[1], keys.size(), keys.data(),
+                  values.data(), s.data(), true);
+
+    ASSERT_EQ(values.size(), keys.size());
+    ASSERT_EQ(std::string(values[0].data(), values[0].size()), "v1");
+    ASSERT_EQ(std::string(values[1].data(), values[1].size()), "v2");
+    ASSERT_EQ(std::string(values[2].data(), values[2].size()), "v3");
+    ASSERT_EQ(std::string(values[4].data(), values[4].size()), "v5");
+    // four kv pairs * two bytes per value
+    ASSERT_EQ(8, (int)get_perf_context()->multiget_read_bytes);
+
+    ASSERT_OK(s[0]);
+    ASSERT_OK(s[1]);
+    ASSERT_OK(s[2]);
+    ASSERT_TRUE(s[3].IsNotFound());
+    ASSERT_OK(s[4]);
+    ASSERT_TRUE(s[5].IsNotFound());
+
+    SetPerfLevel(kDisable);
+  } while (ChangeCompactOptions());
+}
+
+TEST_F(DBBasicTest, MultiGetBatchedMultiLevel) {
+  Options options = CurrentOptions();
+  options.disable_auto_compactions = true;
+  Reopen(options);
+  int num_keys = 0;
+
+  for (int i = 0; i < 128; ++i) {
+    ASSERT_OK(Put("key_" + std::to_string(i), "val_l2_" + std::to_string(i)));
+    num_keys++;
+    if (num_keys == 8) {
+      Flush();
+      num_keys = 0;
+    }
+  }
+  if (num_keys > 0) {
+    Flush();
+    num_keys = 0;
+  }
+  MoveFilesToLevel(2);
+
+  for (int i = 0; i < 128; i += 3) {
+    ASSERT_OK(Put("key_" + std::to_string(i), "val_l1_" + std::to_string(i)));
+    num_keys++;
+    if (num_keys == 8) {
+      Flush();
+      num_keys = 0;
+    }
+  }
+  if (num_keys > 0) {
+    Flush();
+    num_keys = 0;
+  }
+  MoveFilesToLevel(1);
+
+  for (int i = 0; i < 128; i += 5) {
+    ASSERT_OK(Put("key_" + std::to_string(i), "val_l0_" + std::to_string(i)));
+    num_keys++;
+    if (num_keys == 8) {
+      Flush();
+      num_keys = 0;
+    }
+  }
+  if (num_keys > 0) {
+    Flush();
+    num_keys = 0;
+  }
+  ASSERT_EQ(0, num_keys);
+
+  for (int i = 0; i < 128; i += 9) {
+    ASSERT_OK(Put("key_" + std::to_string(i), "val_mem_" + std::to_string(i)));
+  }
+
+  std::vector<std::string> keys;
+  std::vector<std::string> values;
+
+  for (int i = 64; i < 80; ++i) {
+    keys.push_back("key_" + std::to_string(i));
+  }
+
+  values = MultiGet(keys, nullptr);
+  ASSERT_EQ(values.size(), 16);
+  for (unsigned int j = 0; j < values.size(); ++j) {
+    int key = j + 64;
+    if (key % 9 == 0) {
+      ASSERT_EQ(values[j], "val_mem_" + std::to_string(key));
+    } else if (key % 5 == 0) {
+      ASSERT_EQ(values[j], "val_l0_" + std::to_string(key));
+    } else if (key % 3 == 0) {
+      ASSERT_EQ(values[j], "val_l1_" + std::to_string(key));
+    } else {
+      ASSERT_EQ(values[j], "val_l2_" + std::to_string(key));
+    }
+  }
+}
 }  // namespace rocksdb

 int main(int argc, char** argv) {
@@ -173,7 +173,7 @@ TEST_F(DBBlockCacheTest, TestWithoutCompressedBlockCache) {
  delete iter;
  iter = nullptr;

-  // Release interators and access cache again.
+  // Release iterators and access cache again.
  for (size_t i = 0; i < kNumBlocks - 1; i++) {
    iterators[i].reset();
    CheckCacheCounters(options, 0, 0, 0, 0);
@@ -346,14 +346,14 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
  options.statistics = rocksdb::CreateDBStatistics();
  BlockBasedTableOptions table_options;
  table_options.cache_index_and_filter_blocks = true;
-  // 200 bytes are enough to hold the first two blocks
-  std::shared_ptr<Cache> cache = NewLRUCache(200, 0, false);
+  // 500 bytes are enough to hold the first two blocks
+  std::shared_ptr<Cache> cache = NewLRUCache(500, 0, false);
  table_options.block_cache = cache;
  table_options.filter_policy.reset(NewBloomFilterPolicy(20, true));
  options.table_factory.reset(new BlockBasedTableFactory(table_options));
  CreateAndReopenWithCF({"pikachu"}, options);

-  ASSERT_OK(Put(1, "key", "val"));
+  ASSERT_OK(Put(1, "longer_key", "val"));
  // Create a new table
  ASSERT_OK(Flush(1));
  size_t index_bytes_insert =
@@ -367,7 +367,9 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
  cache->SetCapacity(index_bytes_insert + filter_bytes_insert);
  ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_EVICT), 0);
  ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_EVICT), 0);
-  ASSERT_OK(Put(1, "key2", "val"));
+  // Note that the second key needs to be no longer than the first one.
+  // Otherwise the second index block may not fit in cache.
+  ASSERT_OK(Put(1, "key", "val"));
  // Create a new table
  ASSERT_OK(Flush(1));
  // cache evicted old index and block entries
@@ -395,9 +397,9 @@ class MockCache : public LRUCache {
                 false /*strict_capacity_limit*/, 0.0 /*high_pri_pool_ratio*/) {
  }

-  virtual Status Insert(const Slice& key, void* value, size_t charge,
-                        void (*deleter)(const Slice& key, void* value),
-                        Handle** handle, Priority priority) override {
+  Status Insert(const Slice& key, void* value, size_t charge,
+                void (*deleter)(const Slice& key, void* value), Handle** handle,
+                Priority priority) override {
    if (priority == Priority::LOW) {
      low_pri_insert_count++;
    } else {
@@ -631,6 +633,74 @@ TEST_F(DBBlockCacheTest, CompressedCache) {
  }
 }

+TEST_F(DBBlockCacheTest, CacheCompressionDict) {
+  const int kNumFiles = 4;
+  const int kNumEntriesPerFile = 128;
+  const int kNumBytesPerEntry = 1024;
+
+  // Try all the available libraries that support dictionary compression
+  std::vector<CompressionType> compression_types;
+#ifdef ZLIB
+  compression_types.push_back(kZlibCompression);
+#endif  // ZLIB
+#if LZ4_VERSION_NUMBER >= 10400
+  compression_types.push_back(kLZ4Compression);
+  compression_types.push_back(kLZ4HCCompression);
+#endif  // LZ4_VERSION_NUMBER >= 10400
+#if ZSTD_VERSION_NUMBER >= 500
+  compression_types.push_back(kZSTD);
+#endif  // ZSTD_VERSION_NUMBER >= 500
+  Random rnd(301);
+  for (auto compression_type : compression_types) {
+    Options options = CurrentOptions();
+    options.compression = compression_type;
+    options.compression_opts.max_dict_bytes = 4096;
+    options.create_if_missing = true;
+    options.num_levels = 2;
+    options.statistics = rocksdb::CreateDBStatistics();
+    options.target_file_size_base = kNumEntriesPerFile * kNumBytesPerEntry;
+    BlockBasedTableOptions table_options;
+    table_options.cache_index_and_filter_blocks = true;
+    table_options.block_cache.reset(new MockCache());
+    options.table_factory.reset(new BlockBasedTableFactory(table_options));
+    DestroyAndReopen(options);
+
+    for (int i = 0; i < kNumFiles; ++i) {
+      ASSERT_EQ(i, NumTableFilesAtLevel(0, 0));
+      for (int j = 0; j < kNumEntriesPerFile; ++j) {
+        std::string value = RandomString(&rnd, kNumBytesPerEntry);
+        ASSERT_OK(Put(Key(j * kNumFiles + i), value.c_str()));
+      }
+      ASSERT_OK(Flush());
+    }
+    dbfull()->TEST_WaitForCompact();
+    ASSERT_EQ(0, NumTableFilesAtLevel(0));
+    ASSERT_EQ(kNumFiles, NumTableFilesAtLevel(1));
+
+    // Seek to a key in a file. It should cause the SST's dictionary meta-block
+    // to be read.
+    RecordCacheCounters(options);
+    ASSERT_EQ(0,
+              TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_MISS));
+    ASSERT_EQ(0, TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_ADD));
+    ASSERT_EQ(
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_BYTES_INSERT),
+        0);
+    ReadOptions read_options;
+    ASSERT_NE("NOT_FOUND", Get(Key(kNumFiles * kNumEntriesPerFile - 1)));
+    // Two blocks missed/added: dictionary and data block
+    // One block hit: index since it's prefetched
+    CheckCacheCounters(options, 2 /* expected_misses */, 1 /* expected_hits */,
+                       2 /* expected_inserts */, 0 /* expected_failures */);
+    ASSERT_EQ(1,
+              TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_MISS));
+    ASSERT_EQ(1, TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_ADD));
+    ASSERT_GT(
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_BYTES_INSERT),
+        0);
+  }
+}
+
 #endif  // ROCKSDB_LITE

 }  // namespace rocksdb
@@ -32,7 +32,7 @@ class DBBloomFilterTestWithParam
 public:
  DBBloomFilterTestWithParam() : DBTestBase("/db_bloom_filter_tests") {}

-  ~DBBloomFilterTestWithParam() {}
+  ~DBBloomFilterTestWithParam() override {}

  void SetUp() override {
    use_block_based_filter_ = std::get<0>(GetParam());
@@ -642,7 +642,7 @@ class WrappedBloom : public FilterPolicy {
  explicit WrappedBloom(int bits_per_key)
      : filter_(NewBloomFilterPolicy(bits_per_key)), counter_(0) {}

-  ~WrappedBloom() { delete filter_; }
+  ~WrappedBloom() override { delete filter_; }

  const char* Name() const override { return "WrappedRocksDbFilterPolicy"; }

@@ -786,6 +786,75 @@ TEST_F(DBBloomFilterTest, PrefixExtractorBlockFilter) {
  delete iter;
 }

+TEST_F(DBBloomFilterTest, MemtableWholeKeyBloomFilter) {
+  // regression test for #2743. the range delete tombstones in memtable should
+  // be added even when Get() skips searching due to its prefix bloom filter
+  const int kMemtableSize = 1 << 20;              // 1MB
+  const int kMemtablePrefixFilterSize = 1 << 13;  // 8KB
+  const int kPrefixLen = 4;
+  Options options = CurrentOptions();
+  options.memtable_prefix_bloom_size_ratio =
+      static_cast<double>(kMemtablePrefixFilterSize) / kMemtableSize;
+  options.prefix_extractor.reset(rocksdb::NewFixedPrefixTransform(kPrefixLen));
+  options.write_buffer_size = kMemtableSize;
+  options.memtable_whole_key_filtering = false;
+  Reopen(options);
+  std::string key1("AAAABBBB");
+  std::string key2("AAAACCCC");  // not in DB
+  std::string key3("AAAADDDD");
+  std::string key4("AAAAEEEE");
+  std::string value1("Value1");
+  std::string value3("Value3");
+  std::string value4("Value4");
+
+  ASSERT_OK(Put(key1, value1, WriteOptions()));
+
+  // check memtable bloom stats
+  ASSERT_EQ("NOT_FOUND", Get(key2));
+  ASSERT_EQ(0, get_perf_context()->bloom_memtable_miss_count);
+  // same prefix, bloom filter false positive
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_hit_count);
+
+  // enable whole key bloom filter
+  options.memtable_whole_key_filtering = true;
+  Reopen(options);
+  // check memtable bloom stats
+  ASSERT_OK(Put(key3, value3, WriteOptions()));
+  ASSERT_EQ("NOT_FOUND", Get(key2));
+  // whole key bloom filter kicks in and determines it's a miss
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_miss_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_hit_count);
+
+  // verify whole key filtering does not depend on prefix_extractor
+  options.prefix_extractor.reset();
+  Reopen(options);
+  // check memtable bloom stats
+  ASSERT_OK(Put(key4, value4, WriteOptions()));
+  ASSERT_EQ("NOT_FOUND", Get(key2));
+  // whole key bloom filter kicks in and determines it's a miss
+  ASSERT_EQ(2, get_perf_context()->bloom_memtable_miss_count);
+  ASSERT_EQ(1, get_perf_context()->bloom_memtable_hit_count);
+}
+
+TEST_F(DBBloomFilterTest, MemtablePrefixBloomOutOfDomain) {
+  constexpr size_t kPrefixSize = 8;
+  const std::string kKey = "key";
+  assert(kKey.size() < kPrefixSize);
+  Options options = CurrentOptions();
+  options.prefix_extractor.reset(NewFixedPrefixTransform(kPrefixSize));
+  options.memtable_prefix_bloom_size_ratio = 0.25;
+  Reopen(options);
+  ASSERT_OK(Put(kKey, "v"));
+  ASSERT_EQ("v", Get(kKey));
+  std::unique_ptr<Iterator> iter(dbfull()->NewIterator(ReadOptions()));
+  iter->Seek(kKey);
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(kKey, iter->key());
+  iter->SeekForPrev(kKey);
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ(kKey, iter->key());
+}
+
 #ifndef ROCKSDB_LITE
 class BloomStatsTestWithParam
    : public DBBloomFilterTest,
@@ -823,7 +892,7 @@ class BloomStatsTestWithParam
    DestroyAndReopen(options_);
  }

-  ~BloomStatsTestWithParam() {
+  ~BloomStatsTestWithParam() override {
    get_perf_context()->Reset();
    Destroy(options_);
  }
@@ -935,13 +1004,16 @@ TEST_P(BloomStatsTestWithParam, BloomStatsTestWithIter) {
  ASSERT_OK(iter->status());
  ASSERT_TRUE(iter->Valid());
  ASSERT_EQ(value3, iter->value().ToString());
-  ASSERT_EQ(2, get_perf_context()->bloom_sst_hit_count);
+  // The seek doesn't check block-based bloom filter because last index key
+  // starts with the same prefix we're seeking to.
+  uint64_t expected_hits = use_block_based_builder_ ? 1 : 2;
+  ASSERT_EQ(expected_hits, get_perf_context()->bloom_sst_hit_count);

  iter->Seek(key2);
  ASSERT_OK(iter->status());
  ASSERT_TRUE(!iter->Valid());
  ASSERT_EQ(1, get_perf_context()->bloom_sst_miss_count);
-  ASSERT_EQ(2, get_perf_context()->bloom_sst_hit_count);
+  ASSERT_EQ(expected_hits, get_perf_context()->bloom_sst_hit_count);
 }

 INSTANTIATE_TEST_CASE_P(BloomStatsTestWithParam, BloomStatsTestWithParam,
@@ -1266,6 +1338,8 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
    table_options.cache_index_and_filter_blocks = true;
    table_options.filter_policy.reset(
        NewBloomFilterPolicy(10, use_block_based_builder));
+    table_options.index_shortening = BlockBasedTableOptions::
+        IndexShorteningMode::kShortenSeparatorsAndSuccessor;
    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
    DestroyAndReopen(options);

@@ -63,33 +63,33 @@ INSTANTIATE_TEST_CASE_P(DBTestCompactionFilterWithCompactOption,

 class KeepFilter : public CompactionFilter {
 public:
-  virtual bool Filter(int /*level*/, const Slice& /*key*/,
-                      const Slice& /*value*/, std::string* /*new_value*/,
-                      bool* /*value_changed*/) const override {
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
    cfilter_count++;
    return false;
  }

-  virtual const char* Name() const override { return "KeepFilter"; }
+  const char* Name() const override { return "KeepFilter"; }
 };

 class DeleteFilter : public CompactionFilter {
 public:
-  virtual bool Filter(int /*level*/, const Slice& /*key*/,
-                      const Slice& /*value*/, std::string* /*new_value*/,
-                      bool* /*value_changed*/) const override {
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
    cfilter_count++;
    return true;
  }

-  virtual const char* Name() const override { return "DeleteFilter"; }
+  const char* Name() const override { return "DeleteFilter"; }
 };

 class DeleteISFilter : public CompactionFilter {
 public:
-  virtual bool Filter(int /*level*/, const Slice& key, const Slice& /*value*/,
-                      std::string* /*new_value*/,
-                      bool* /*value_changed*/) const override {
+  bool Filter(int /*level*/, const Slice& key, const Slice& /*value*/,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
    cfilter_count++;
    int i = std::stoi(key.ToString());
    if (i > 5 && i <= 105) {
@@ -98,20 +98,18 @@ class DeleteISFilter : public CompactionFilter {
    return false;
  }

-  virtual bool IgnoreSnapshots() const override { return true; }
+  bool IgnoreSnapshots() const override { return true; }

-  virtual const char* Name() const override { return "DeleteFilter"; }
+  const char* Name() const override { return "DeleteFilter"; }
 };

 // Skip x if floor(x/10) is even, use range skips. Requires that keys are
 // zero-padded to length 10.
 class SkipEvenFilter : public CompactionFilter {
 public:
-  virtual Decision FilterV2(int /*level*/, const Slice& key,
-                            ValueType /*value_type*/,
-                            const Slice& /*existing_value*/,
-                            std::string* /*new_value*/,
-                            std::string* skip_until) const override {
+  Decision FilterV2(int /*level*/, const Slice& key, ValueType /*value_type*/,
+                    const Slice& /*existing_value*/, std::string* /*new_value*/,
+                    std::string* skip_until) const override {
    cfilter_count++;
    int i = std::stoi(key.ToString());
    if (i / 10 % 2 == 0) {
@@ -124,22 +122,22 @@ class SkipEvenFilter : public CompactionFilter {
    return Decision::kKeep;
  }

-  virtual bool IgnoreSnapshots() const override { return true; }
+  bool IgnoreSnapshots() const override { return true; }

-  virtual const char* Name() const override { return "DeleteFilter"; }
+  const char* Name() const override { return "DeleteFilter"; }
 };

 class DelayFilter : public CompactionFilter {
 public:
  explicit DelayFilter(DBTestBase* d) : db_test(d) {}
-  virtual bool Filter(int /*level*/, const Slice& /*key*/,
-                      const Slice& /*value*/, std::string* /*new_value*/,
-                      bool* /*value_changed*/) const override {
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
    db_test->env_->addon_time_.fetch_add(1000);
    return true;
  }

-  virtual const char* Name() const override { return "DelayFilter"; }
+  const char* Name() const override { return "DelayFilter"; }

 private:
  DBTestBase* db_test;
@@ -149,13 +147,13 @@ class ConditionalFilter : public CompactionFilter {
 public:
  explicit ConditionalFilter(const std::string* filtered_value)
      : filtered_value_(filtered_value) {}
-  virtual bool Filter(int /*level*/, const Slice& /*key*/, const Slice& value,
-                      std::string* /*new_value*/,
-                      bool* /*value_changed*/) const override {
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& value,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
    return value.ToString() == *filtered_value_;
  }

-  virtual const char* Name() const override { return "ConditionalFilter"; }
+  const char* Name() const override { return "ConditionalFilter"; }

 private:
  const std::string* filtered_value_;
@@ -165,16 +163,15 @@ class ChangeFilter : public CompactionFilter {
 public:
  explicit ChangeFilter() {}

-  virtual bool Filter(int /*level*/, const Slice& /*key*/,
-                      const Slice& /*value*/, std::string* new_value,
-                      bool* value_changed) const override {
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+              std::string* new_value, bool* value_changed) const override {
    assert(new_value != nullptr);
    *new_value = NEW_VALUE;
    *value_changed = true;
    return false;
  }

-  virtual const char* Name() const override { return "ChangeFilter"; }
+  const char* Name() const override { return "ChangeFilter"; }
 };

 class KeepFilterFactory : public CompactionFilterFactory {
@@ -185,7 +182,7 @@ class KeepFilterFactory : public CompactionFilterFactory {
        check_context_cf_id_(check_context_cf_id),
        compaction_filter_created_(false) {}

-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& context) override {
    if (check_context_) {
      EXPECT_EQ(expect_full_compaction_.load(), context.is_full_compaction);
@@ -200,7 +197,7 @@ class KeepFilterFactory : public CompactionFilterFactory {

  bool compaction_filter_created() const { return compaction_filter_created_; }

-  virtual const char* Name() const override { return "KeepFilterFactory"; }
+  const char* Name() const override { return "KeepFilterFactory"; }
  bool check_context_;
  bool check_context_cf_id_;
  std::atomic_bool expect_full_compaction_;
@@ -211,7 +208,7 @@ class KeepFilterFactory : public CompactionFilterFactory {

 class DeleteFilterFactory : public CompactionFilterFactory {
 public:
-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& context) override {
    if (context.is_manual_compaction) {
      return std::unique_ptr<CompactionFilter>(new DeleteFilter());
@@ -220,13 +217,13 @@ class DeleteFilterFactory : public CompactionFilterFactory {
    }
  }

-  virtual const char* Name() const override { return "DeleteFilterFactory"; }
+  const char* Name() const override { return "DeleteFilterFactory"; }
 };

 // Delete Filter Factory which ignores snapshots
 class DeleteISFilterFactory : public CompactionFilterFactory {
 public:
-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& context) override {
    if (context.is_manual_compaction) {
      return std::unique_ptr<CompactionFilter>(new DeleteISFilter());
@@ -235,12 +232,12 @@ class DeleteISFilterFactory : public CompactionFilterFactory {
    }
  }

-  virtual const char* Name() const override { return "DeleteFilterFactory"; }
+  const char* Name() const override { return "DeleteFilterFactory"; }
 };

 class SkipEvenFilterFactory : public CompactionFilterFactory {
 public:
-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& context) override {
    if (context.is_manual_compaction) {
      return std::unique_ptr<CompactionFilter>(new SkipEvenFilter());
@@ -249,18 +246,18 @@ class SkipEvenFilterFactory : public CompactionFilterFactory {
    }
  }

-  virtual const char* Name() const override { return "SkipEvenFilterFactory"; }
+  const char* Name() const override { return "SkipEvenFilterFactory"; }
 };

 class DelayFilterFactory : public CompactionFilterFactory {
 public:
  explicit DelayFilterFactory(DBTestBase* d) : db_test(d) {}
-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& /*context*/) override {
    return std::unique_ptr<CompactionFilter>(new DelayFilter(db_test));
  }

-  virtual const char* Name() const override { return "DelayFilterFactory"; }
+  const char* Name() const override { return "DelayFilterFactory"; }

 private:
  DBTestBase* db_test;
@@ -271,15 +268,13 @@ class ConditionalFilterFactory : public CompactionFilterFactory {
  explicit ConditionalFilterFactory(const Slice& filtered_value)
      : filtered_value_(filtered_value.ToString()) {}

-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& /*context*/) override {
    return std::unique_ptr<CompactionFilter>(
        new ConditionalFilter(&filtered_value_));
  }

-  virtual const char* Name() const override {
-    return "ConditionalFilterFactory";
-  }
+  const char* Name() const override { return "ConditionalFilterFactory"; }

 private:
  std::string filtered_value_;
@@ -289,12 +284,12 @@ class ChangeFilterFactory : public CompactionFilterFactory {
 public:
  explicit ChangeFilterFactory() {}

-  virtual std::unique_ptr<CompactionFilter> CreateCompactionFilter(
+  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
      const CompactionFilter::Context& /*context*/) override {
    return std::unique_ptr<CompactionFilter>(new ChangeFilter());
  }

-  virtual const char* Name() const override { return "ChangeFilterFactory"; }
+  const char* Name() const override { return "ChangeFilterFactory"; }
 };

 #ifndef ROCKSDB_LITE
@@ -340,8 +335,8 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
  Arena arena;
  {
    InternalKeyComparator icmp(options.comparator);
-    RangeDelAggregatorV2 range_del_agg(&icmp,
-                                       kMaxSequenceNumber /* upper_bound */);
+    ReadRangeDelAggregator range_del_agg(&icmp,
+                                         kMaxSequenceNumber /* upper_bound */);
    ScopedArenaIterator iter(dbfull()->NewInternalIterator(
        &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
    iter->SeekToFirst();
@@ -357,7 +352,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
    }
  }
  ASSERT_EQ(total, 100000);
-  ASSERT_EQ(count, 1);
+  ASSERT_EQ(count, 0);

  // overwrite all the 100K keys once again.
  for (int i = 0; i < 100000; i++) {
@@ -430,8 +425,8 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
  count = 0;
  {
    InternalKeyComparator icmp(options.comparator);
-    RangeDelAggregatorV2 range_del_agg(&icmp,
-                                       kMaxSequenceNumber /* upper_bound */);
+    ReadRangeDelAggregator range_del_agg(&icmp,
+                                         kMaxSequenceNumber /* upper_bound */);
    ScopedArenaIterator iter(dbfull()->NewInternalIterator(
        &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
    iter->SeekToFirst();
@@ -648,8 +643,8 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
    int total = 0;
    Arena arena;
    InternalKeyComparator icmp(options.comparator);
-    RangeDelAggregatorV2 range_del_agg(&icmp,
-                                       kMaxSequenceNumber /* snapshots */);
+    ReadRangeDelAggregator range_del_agg(&icmp,
+                                         kMaxSequenceNumber /* snapshots */);
    ScopedArenaIterator iter(dbfull()->NewInternalIterator(
        &arena, &range_del_agg, kMaxSequenceNumber));
    iter->SeekToFirst();
@@ -664,7 +659,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
      iter->Next();
    }
    ASSERT_EQ(total, 700);
-    ASSERT_EQ(count, 1);
+    ASSERT_EQ(count, 0);
  }
 }
 #endif  // ROCKSDB_LITE
@@ -699,44 +694,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextCfId) {
 }

 #ifndef ROCKSDB_LITE
-// Compaction filters should only be applied to records that are newer than the
-// latest snapshot. This test inserts records and applies a delete filter.
-TEST_F(DBTestCompactionFilter, CompactionFilterSnapshot) {
-  Options options = CurrentOptions();
-  options.compaction_filter_factory = std::make_shared<DeleteFilterFactory>();
-  options.disable_auto_compactions = true;
-  options.create_if_missing = true;
-  DestroyAndReopen(options);
-
-  // Put some data.
-  const Snapshot* snapshot = nullptr;
-  for (int table = 0; table < 4; ++table) {
-    for (int i = 0; i < 10; ++i) {
-      Put(ToString(table * 100 + i), "val");
-    }
-    Flush();
-
-    if (table == 0) {
-      snapshot = db_->GetSnapshot();
-    }
-  }
-  assert(snapshot != nullptr);
-
-  cfilter_count = 0;
-  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-  // The filter should delete 10 records.
-  ASSERT_EQ(30U, cfilter_count);
-
-  // Release the snapshot and compact again -> now all records should be
-  // removed.
-  db_->ReleaseSnapshot(snapshot);
-  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-  ASSERT_EQ(0U, CountLiveFiles());
-}
-
-// Compaction filters should only be applied to records that are newer than the
-// latest snapshot. However, if the compaction filter asks to ignore snapshots
-// records newer than the snapshot will also be processed
+// Compaction filters aplies to all records, regardless snapshots.
 TEST_F(DBTestCompactionFilter, CompactionFilterIgnoreSnapshot) {
  std::string five = ToString(5);
  Options options = CurrentOptions();
@@ -874,6 +832,38 @@ TEST_F(DBTestCompactionFilter, SkipUntilWithBloomFilter) {
  EXPECT_EQ("v50", val);
 }

+class TestNotSupportedFilter : public CompactionFilter {
+ public:
+  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
+              std::string* /*new_value*/,
+              bool* /*value_changed*/) const override {
+    return true;
+  }
+
+  const char* Name() const override { return "NotSupported"; }
+  bool IgnoreSnapshots() const override { return false; }
+};
+
+TEST_F(DBTestCompactionFilter, IgnoreSnapshotsFalse) {
+  Options options = CurrentOptions();
+  options.compaction_filter = new TestNotSupportedFilter();
+  DestroyAndReopen(options);
+
+  Put("a", "v10");
+  Put("z", "v20");
+  Flush();
+
+  Put("a", "v10");
+  Put("z", "v20");
+  Flush();
+
+  // Comapction should fail because IgnoreSnapshots() = false
+  EXPECT_TRUE(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr)
+                  .IsNotSupported());
+
+  delete options.compaction_filter;
+}
+
 }  // namespace rocksdb

 int main(int argc, char** argv) {
@@ -10,8 +10,10 @@
 #include "db/db_test_util.h"
 #include "port/port.h"
 #include "port/stack_trace.h"
+#include "rocksdb/concurrent_task_limiter.h"
 #include "rocksdb/experimental.h"
 #include "rocksdb/utilities/convenience.h"
+#include "util/concurrent_task_limiter_impl.h"
 #include "util/fault_injection_test_env.h"
 #include "util/sync_point.h"

@@ -53,9 +55,9 @@ namespace {
 class FlushedFileCollector : public EventListener {
 public:
  FlushedFileCollector() {}
-  ~FlushedFileCollector() {}
+  ~FlushedFileCollector() override {}

-  virtual void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
+  void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
    std::lock_guard<std::mutex> lock(mutex_);
    flushed_files_.push_back(info.file_path);
  }
@@ -85,24 +87,23 @@ public:
    }
  }

-  ~CompactionStatsCollector() {}
+  ~CompactionStatsCollector() override {}

-  virtual void OnCompactionCompleted(DB* /* db */,
-      const CompactionJobInfo& info) override {
+  void OnCompactionCompleted(DB* /* db */,
+                             const CompactionJobInfo& info) override {
    int k = static_cast<int>(info.compaction_reason);
    int num_of_reasons = static_cast<int>(CompactionReason::kNumOfReasons);
    assert(k >= 0 && k < num_of_reasons);
    compaction_completed_[k]++;
  }

-  virtual void OnExternalFileIngested(DB* /* db */,
-      const ExternalFileIngestionInfo& /* info */) override {
+  void OnExternalFileIngested(
+      DB* /* db */, const ExternalFileIngestionInfo& /* info */) override {
    int k = static_cast<int>(CompactionReason::kExternalSstIngestion);
    compaction_completed_[k]++;
  }

-  virtual void OnFlushCompleted(DB* /* db */,
-      const FlushJobInfo& /* info */) override {
+  void OnFlushCompleted(DB* /* db */, const FlushJobInfo& /* info */) override {
    int k = static_cast<int>(CompactionReason::kFlush);
    compaction_completed_[k]++;
  }
@@ -352,7 +353,8 @@ TEST_P(DBCompactionTestWithParam, CompactionsPreserveDeletes) {
    CompactRangeOptions cro;
    cro.change_level = true;
    cro.target_level = 2;
-    cro.bottommost_level_compaction = BottommostLevelCompaction::kForce;
+    cro.bottommost_level_compaction =
+        BottommostLevelCompaction::kForceOptimized;

    dbfull()->TEST_WaitForFlushMemTable();
    dbfull()->CompactRange(cro, nullptr, nullptr);
@@ -412,6 +414,7 @@ TEST_F(DBCompactionTest, SkipStatsUpdateTest) {

  // Reopen the DB with stats-update disabled
  options.skip_stats_update_on_db_open = true;
+  options.max_open_files = 20;
  env_->random_file_open_counter_.store(0);
  Reopen(options);

@@ -437,7 +440,7 @@ TEST_F(DBCompactionTest, TestTableReaderForCompaction) {
  Options options = CurrentOptions();
  options.env = env_;
  options.new_table_reader_for_compaction_inputs = true;
-  options.max_open_files = 100;
+  options.max_open_files = 20;
  options.level0_file_num_compaction_trigger = 3;
  DestroyAndReopen(options);
  Random rnd(301);
@@ -466,15 +469,16 @@ TEST_F(DBCompactionTest, TestTableReaderForCompaction) {
      Flush();
      dbfull()->TEST_WaitForCompact();
      // preloading iterator issues one table cache lookup and create
-      // a new table reader.
-      ASSERT_EQ(num_table_cache_lookup, 1);
+      // a new table reader, if not preloaded.
+      int old_num_table_cache_lookup = num_table_cache_lookup;
+      ASSERT_GE(num_table_cache_lookup, 1);
      ASSERT_EQ(num_new_table_reader, 1);

      num_table_cache_lookup = 0;
      num_new_table_reader = 0;
      ASSERT_EQ(Key(k), Get(Key(k)));
      // lookup iterator from table cache and no need to create a new one.
-      ASSERT_EQ(num_table_cache_lookup, 1);
+      ASSERT_EQ(old_num_table_cache_lookup + num_table_cache_lookup, 2);
      ASSERT_EQ(num_new_table_reader, 0);
    }
  }
@@ -487,7 +491,10 @@ TEST_F(DBCompactionTest, TestTableReaderForCompaction) {
  // a new table reader. One file is created for flush and one for compaction.
  // Compaction inputs make no table cache look-up for data/range deletion
  // iterators
-  ASSERT_EQ(num_table_cache_lookup, 2);
+  // May preload table cache too.
+  ASSERT_GE(num_table_cache_lookup, 2);
+  int old_num_table_cache_lookup2 = num_table_cache_lookup;
+
  // Create new iterator for:
  // (1) 1 for verifying flush results
  // (2) 3 for compaction input files
@@ -497,7 +504,7 @@ TEST_F(DBCompactionTest, TestTableReaderForCompaction) {
  num_table_cache_lookup = 0;
  num_new_table_reader = 0;
  ASSERT_EQ(Key(1), Get(Key(1)));
-  ASSERT_EQ(num_table_cache_lookup, 1);
+  ASSERT_EQ(num_table_cache_lookup + old_num_table_cache_lookup2, 3);
  ASSERT_EQ(num_new_table_reader, 0);

  num_table_cache_lookup = 0;
@@ -505,18 +512,20 @@ TEST_F(DBCompactionTest, TestTableReaderForCompaction) {
  CompactRangeOptions cro;
  cro.change_level = true;
  cro.target_level = 2;
-  cro.bottommost_level_compaction = BottommostLevelCompaction::kForce;
+  cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
  db_->CompactRange(cro, nullptr, nullptr);
  // Only verifying compaction outputs issues one table cache lookup
  // for both data block and range deletion block).
-  ASSERT_EQ(num_table_cache_lookup, 1);
+  // May preload table cache too.
+  ASSERT_GE(num_table_cache_lookup, 1);
+  old_num_table_cache_lookup2 = num_table_cache_lookup;
  // One for compaction input, one for verifying compaction results.
  ASSERT_EQ(num_new_table_reader, 2);

  num_table_cache_lookup = 0;
  num_new_table_reader = 0;
  ASSERT_EQ(Key(1), Get(Key(1)));
-  ASSERT_EQ(num_table_cache_lookup, 1);
+  ASSERT_EQ(num_table_cache_lookup + old_num_table_cache_lookup2, 2);
  ASSERT_EQ(num_new_table_reader, 0);

  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
@@ -1594,7 +1603,8 @@ TEST_F(DBCompactionTest, DeleteFileRange) {
  // Note that we don't delete level 0 files
  compact_options.change_level = true;
  compact_options.target_level = 1;
-  ASSERT_OK(dbfull()->TEST_CompactRange(0, nullptr, nullptr));
+  ASSERT_OK(db_->CompactRange(compact_options, nullptr, nullptr));
+  dbfull()->TEST_WaitForCompact();

  ASSERT_OK(
      DeleteFilesInRange(db_, db_->DefaultColumnFamily(), nullptr, nullptr));
@@ -2251,6 +2261,8 @@ TEST_P(DBCompactionTestWithParam, ConvertCompactionStyle) {
  CompactRangeOptions compact_options;
  compact_options.change_level = true;
  compact_options.target_level = 0;
+  // cannot use kForceOptimized here because the compaction here is expected
+  // to generate one output file
  compact_options.bottommost_level_compaction =
      BottommostLevelCompaction::kForce;
  compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
@@ -3030,7 +3042,7 @@ TEST_P(DBCompactionTestWithParam, ForceBottommostLevelCompaction) {
  // then compacte the bottommost level L3=>L3 (non trivial move)
  compact_options = CompactRangeOptions();
  compact_options.bottommost_level_compaction =
-      BottommostLevelCompaction::kForce;
+      BottommostLevelCompaction::kForceOptimized;
  ASSERT_OK(db_->CompactRange(compact_options, nullptr, nullptr));
  ASSERT_EQ("0,0,0,1", FilesPerLevel(0));
  ASSERT_EQ(trivial_move, 4);
@@ -3339,7 +3351,7 @@ TEST_F(DBCompactionTest, CompactBottomLevelFilesWithDeletions) {
  options.level0_file_num_compaction_trigger = kNumLevelFiles;
  // inflate it a bit to account for key/metadata overhead
  options.target_file_size_base = 120 * kNumKeysPerFile * kValueSize / 100;
-  Reopen(options);
+  CreateAndReopenWithCF({"one"}, options);

  Random rnd(301);
  const Snapshot* snapshot = nullptr;
@@ -3370,10 +3382,12 @@ TEST_F(DBCompactionTest, CompactBottomLevelFilesWithDeletions) {
  // just need to bump seqnum so ReleaseSnapshot knows the newest key in the SST
  // files does not need to be preserved in case of a future snapshot.
  ASSERT_OK(Put(Key(0), "val"));
+  ASSERT_NE(kMaxSequenceNumber, dbfull()->bottommost_files_mark_threshold_);
  // release snapshot and wait for compactions to finish. Single-file
  // compactions should be triggered, which reduce the size of each bottom-level
  // file without changing file count.
  db_->ReleaseSnapshot(snapshot);
+  ASSERT_EQ(kMaxSequenceNumber, dbfull()->bottommost_files_mark_threshold_);
  rocksdb::SyncPoint::GetInstance()->SetCallBack(
      "LevelCompactionPicker::PickCompaction:Return", [&](void* arg) {
        Compaction* compaction = reinterpret_cast<Compaction*>(arg);
@@ -3506,6 +3520,234 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
 }

+TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
+  const int kNumKeysPerFile = 32;
+  const int kNumLevelFiles = 2;
+  const int kValueSize = 100;
+
+  Options options = CurrentOptions();
+  options.periodic_compaction_seconds = 48 * 60 * 60;  // 2 days
+  options.max_open_files = -1;       // needed for ttl compaction
+  env_->time_elapse_only_sleep_ = false;
+  options.env = env_;
+
+  env_->addon_time_.store(0);
+  DestroyAndReopen(options);
+
+  int periodic_compactions = 0;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "LevelCompactionPicker::PickCompaction:Return", [&](void* arg) {
+        Compaction* compaction = reinterpret_cast<Compaction*>(arg);
+        auto compaction_reason = compaction->compaction_reason();
+        if (compaction_reason == CompactionReason::kPeriodicCompaction) {
+          periodic_compactions++;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  Random rnd(301);
+  for (int i = 0; i < kNumLevelFiles; ++i) {
+    for (int j = 0; j < kNumKeysPerFile; ++j) {
+      ASSERT_OK(
+          Put(Key(i * kNumKeysPerFile + j), RandomString(&rnd, kValueSize)));
+    }
+    Flush();
+  }
+  dbfull()->TEST_WaitForCompact();
+
+  ASSERT_EQ("2", FilesPerLevel());
+  ASSERT_EQ(0, periodic_compactions);
+
+  // Add 50 hours and do a write
+  env_->addon_time_.fetch_add(50 * 60 * 60);
+  ASSERT_OK(Put("a", "1"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  // Assert that the files stay in the same level
+  ASSERT_EQ("3", FilesPerLevel());
+  // The two old files go through the periodic compaction process
+  ASSERT_EQ(2, periodic_compactions);
+
+  MoveFilesToLevel(1);
+  ASSERT_EQ("0,3", FilesPerLevel());
+
+  // Add another 50 hours and do another write
+  env_->addon_time_.fetch_add(50 * 60 * 60);
+  ASSERT_OK(Put("b", "2"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  ASSERT_EQ("1,3", FilesPerLevel());
+  // The three old files now go through the periodic compaction process. 2 + 3.
+  ASSERT_EQ(5, periodic_compactions);
+
+  // Add another 50 hours and do another write
+  env_->addon_time_.fetch_add(50 * 60 * 60);
+  ASSERT_OK(Put("c", "3"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  ASSERT_EQ("2,3", FilesPerLevel());
+  // The four old files now go through the periodic compaction process. 5 + 4.
+  ASSERT_EQ(9, periodic_compactions);
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+
+TEST_F(DBCompactionTest, LevelPeriodicCompactionWithOldDB) {
+  // This test makes sure that periodic compactions are working with a DB
+  // where file_creation_time of some files is 0.
+  // After compactions the new files are created with a valid file_creation_time
+
+  const int kNumKeysPerFile = 32;
+  const int kNumFiles = 4;
+  const int kValueSize = 100;
+
+  Options options = CurrentOptions();
+  options.max_open_files = -1;  // needed for ttl compaction
+  env_->time_elapse_only_sleep_ = false;
+  options.env = env_;
+
+  env_->addon_time_.store(0);
+  DestroyAndReopen(options);
+
+  int periodic_compactions = 0;
+  bool set_file_creation_time_to_zero = true;
+  bool set_creation_time_to_zero = true;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "LevelCompactionPicker::PickCompaction:Return", [&](void* arg) {
+        Compaction* compaction = reinterpret_cast<Compaction*>(arg);
+        auto compaction_reason = compaction->compaction_reason();
+        if (compaction_reason == CompactionReason::kPeriodicCompaction) {
+          periodic_compactions++;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "PropertyBlockBuilder::AddTableProperty:Start", [&](void* arg) {
+        TableProperties* props = reinterpret_cast<TableProperties*>(arg);
+        if (set_file_creation_time_to_zero) {
+          props->file_creation_time = 0;
+        }
+        if (set_creation_time_to_zero) {
+          props->creation_time = 0;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  Random rnd(301);
+  for (int i = 0; i < kNumFiles; ++i) {
+    for (int j = 0; j < kNumKeysPerFile; ++j) {
+      ASSERT_OK(
+          Put(Key(i * kNumKeysPerFile + j), RandomString(&rnd, kValueSize)));
+    }
+    Flush();
+    // Move the first two files to L2.
+    if (i == 1) {
+      MoveFilesToLevel(2);
+      set_creation_time_to_zero = false;
+    }
+  }
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  ASSERT_EQ("2,0,2", FilesPerLevel());
+  ASSERT_EQ(0, periodic_compactions);
+
+  Close();
+
+  set_file_creation_time_to_zero = false;
+  // Forward the clock by 2 days.
+  env_->addon_time_.fetch_add(2 * 24 * 60 * 60);
+  options.periodic_compaction_seconds = 1 * 24 * 60 * 60;  // 1 day
+
+  Reopen(options);
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+  ASSERT_EQ("2,0,2", FilesPerLevel());
+  // Make sure that all files go through periodic compaction.
+  ASSERT_EQ(kNumFiles, periodic_compactions);
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+
+TEST_F(DBCompactionTest, LevelPeriodicAndTtlCompaction) {
+  const int kNumKeysPerFile = 32;
+  const int kNumLevelFiles = 2;
+  const int kValueSize = 100;
+
+  Options options = CurrentOptions();
+  options.ttl = 10 * 60 * 60;  // 10 hours
+  options.periodic_compaction_seconds = 48 * 60 * 60;  // 2 days
+  options.max_open_files = -1;   // needed for both periodic and ttl compactions
+  env_->time_elapse_only_sleep_ = false;
+  options.env = env_;
+
+  env_->addon_time_.store(0);
+  DestroyAndReopen(options);
+
+  int periodic_compactions = 0;
+  int ttl_compactions = 0;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "LevelCompactionPicker::PickCompaction:Return", [&](void* arg) {
+        Compaction* compaction = reinterpret_cast<Compaction*>(arg);
+        auto compaction_reason = compaction->compaction_reason();
+        if (compaction_reason == CompactionReason::kPeriodicCompaction) {
+          periodic_compactions++;
+        } else if (compaction_reason == CompactionReason::kTtl) {
+          ttl_compactions++;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  Random rnd(301);
+  for (int i = 0; i < kNumLevelFiles; ++i) {
+    for (int j = 0; j < kNumKeysPerFile; ++j) {
+      ASSERT_OK(
+          Put(Key(i * kNumKeysPerFile + j), RandomString(&rnd, kValueSize)));
+    }
+    Flush();
+  }
+  dbfull()->TEST_WaitForCompact();
+
+  MoveFilesToLevel(3);
+
+  ASSERT_EQ("0,0,0,2", FilesPerLevel());
+  ASSERT_EQ(0, periodic_compactions);
+  ASSERT_EQ(0, ttl_compactions);
+
+  // Add some time greater than periodic_compaction_time.
+  env_->addon_time_.fetch_add(50 * 60 * 60);
+  ASSERT_OK(Put("a", "1"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  // Files in the bottom level go through periodic compactions.
+  ASSERT_EQ("1,0,0,2", FilesPerLevel());
+  ASSERT_EQ(2, periodic_compactions);
+  ASSERT_EQ(0, ttl_compactions);
+
+  // Add a little more time than ttl
+  env_->addon_time_.fetch_add(11 * 60 * 60);
+  ASSERT_OK(Put("b", "1"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  // Notice that the previous file in level 1 falls down to the bottom level
+  // due to ttl compactions, one level at a time.
+  // And bottom level files don't get picked up for ttl compactions.
+  ASSERT_EQ("1,0,0,3", FilesPerLevel());
+  ASSERT_EQ(2, periodic_compactions);
+  ASSERT_EQ(3, ttl_compactions);
+
+  // Add some time greater than periodic_compaction_time.
+  env_->addon_time_.fetch_add(50 * 60 * 60);
+  ASSERT_OK(Put("c", "1"));
+  Flush();
+  dbfull()->TEST_WaitForCompact();
+  // Previous L0 file falls one level at a time to bottom level due to ttl.
+  // And all 4 bottom files go through periodic compactions.
+  ASSERT_EQ("1,0,0,4", FilesPerLevel());
+  ASSERT_EQ(6, periodic_compactions);
+  ASSERT_EQ(6, ttl_compactions);
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+
+
 TEST_F(DBCompactionTest, CompactRangeDelayedByL0FileCount) {
  // Verify that, when `CompactRangeOptions::allow_write_stall == false`, manual
  // compaction only triggers flush after it's sure stall won't be triggered for
@@ -3890,6 +4132,192 @@ TEST_F(DBCompactionTest, CompactionHasEmptyOutput) {
  ASSERT_EQ(2, collector->num_ssts_creation_started());
 }

+TEST_F(DBCompactionTest, CompactionLimiter) {
+  const int kNumKeysPerFile = 10;
+  const int kMaxBackgroundThreads = 64;
+
+  struct CompactionLimiter {
+    std::string name;
+    int limit_tasks;
+    int max_tasks;
+    int tasks;
+    std::shared_ptr<ConcurrentTaskLimiter> limiter;
+  };
+
+  std::vector<CompactionLimiter> limiter_settings;
+  limiter_settings.push_back({"limiter_1", 1, 0, 0, nullptr});
+  limiter_settings.push_back({"limiter_2", 2, 0, 0, nullptr});
+  limiter_settings.push_back({"limiter_3", 3, 0, 0, nullptr});
+
+  for (auto& ls : limiter_settings) {
+    ls.limiter.reset(NewConcurrentTaskLimiter(ls.name, ls.limit_tasks));
+  }
+
+  std::shared_ptr<ConcurrentTaskLimiter> unique_limiter(
+    NewConcurrentTaskLimiter("unique_limiter", -1));
+
+  const char* cf_names[] = {"default", "0", "1", "2", "3", "4", "5",
+    "6", "7", "8", "9", "a", "b", "c", "d", "e", "f" };
+  const int cf_count = sizeof cf_names / sizeof cf_names[0];
+
+  std::unordered_map<std::string, CompactionLimiter*> cf_to_limiter;
+
+  Options options = CurrentOptions();
+  options.write_buffer_size = 110 * 1024;  // 110KB
+  options.arena_block_size = 4096;
+  options.num_levels = 3;
+  options.level0_file_num_compaction_trigger = 4;
+  options.level0_slowdown_writes_trigger = 64;
+  options.level0_stop_writes_trigger = 64;
+  options.max_background_jobs = kMaxBackgroundThreads; // Enough threads
+  options.memtable_factory.reset(new SpecialSkipListFactory(kNumKeysPerFile));
+  options.max_write_buffer_number = 10; // Enough memtables
+  DestroyAndReopen(options);
+
+  std::vector<Options> option_vector;
+  option_vector.reserve(cf_count);
+
+  for (int cf = 0; cf < cf_count; cf++) {
+    ColumnFamilyOptions cf_opt(options);
+    if (cf == 0) {
+      // "Default" CF does't use compaction limiter
+      cf_opt.compaction_thread_limiter = nullptr;
+    } else if (cf == 1) {
+      // "1" CF uses bypass compaction limiter
+      unique_limiter->SetMaxOutstandingTask(-1);
+      cf_opt.compaction_thread_limiter = unique_limiter;
+    } else {
+      // Assign limiter by mod
+      auto& ls = limiter_settings[cf % 3];
+      cf_opt.compaction_thread_limiter = ls.limiter;
+      cf_to_limiter[cf_names[cf]] = &ls;
+    }
+    option_vector.emplace_back(DBOptions(options), cf_opt);
+  }
+
+  for (int cf = 1; cf < cf_count; cf++) {
+    CreateColumnFamilies({cf_names[cf]}, option_vector[cf]);
+  }
+
+  ReopenWithColumnFamilies(std::vector<std::string>(cf_names,
+                                                    cf_names + cf_count),
+                           option_vector);
+
+  port::Mutex mutex;
+
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+    "DBImpl::BackgroundCompaction:BeforeCompaction", [&](void* arg) {
+      const auto& cf_name = static_cast<ColumnFamilyData*>(arg)->GetName();
+      auto iter = cf_to_limiter.find(cf_name);
+      if (iter != cf_to_limiter.end()) {
+        MutexLock l(&mutex);
+        ASSERT_GE(iter->second->limit_tasks, ++iter->second->tasks);
+        iter->second->max_tasks = std::max(iter->second->max_tasks,
+              iter->second->limit_tasks);
+      }
+    });
+
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+    "DBImpl::BackgroundCompaction:AfterCompaction", [&](void* arg) {
+      const auto& cf_name = static_cast<ColumnFamilyData*>(arg)->GetName();
+      auto iter = cf_to_limiter.find(cf_name);
+      if (iter != cf_to_limiter.end()) {
+        MutexLock l(&mutex);
+        ASSERT_GE(--iter->second->tasks, 0);
+      }
+    });
+
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  // Block all compact threads in thread pool.
+  const size_t kTotalFlushTasks = kMaxBackgroundThreads / 4;
+  const size_t kTotalCompactTasks = kMaxBackgroundThreads - kTotalFlushTasks;
+  env_->SetBackgroundThreads((int)kTotalFlushTasks, Env::HIGH);
+  env_->SetBackgroundThreads((int)kTotalCompactTasks, Env::LOW);
+
+  test::SleepingBackgroundTask sleeping_compact_tasks[kTotalCompactTasks];
+
+  // Block all compaction threads in thread pool.
+  for (size_t i = 0; i < kTotalCompactTasks; i++) {
+    env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask,
+                   &sleeping_compact_tasks[i], Env::LOW);
+    sleeping_compact_tasks[i].WaitUntilSleeping();
+  }
+
+  int keyIndex = 0;
+
+  for (int n = 0; n < options.level0_file_num_compaction_trigger; n++) {
+    for (int cf = 0; cf < cf_count; cf++) {
+      for (int i = 0; i < kNumKeysPerFile; i++) {
+        ASSERT_OK(Put(cf, Key(keyIndex++), ""));
+      }
+      // put extra key to trigger flush
+      ASSERT_OK(Put(cf, "", ""));
+    }
+
+    for (int cf = 0; cf < cf_count; cf++) {
+      dbfull()->TEST_WaitForFlushMemTable(handles_[cf]);
+    }
+  }
+
+  // Enough L0 files to trigger compaction
+  for (int cf = 0; cf < cf_count; cf++) {
+    ASSERT_EQ(NumTableFilesAtLevel(0, cf),
+      options.level0_file_num_compaction_trigger);
+  }
+
+  // Create more files for one column family, which triggers speed up
+  // condition, all compactions will be scheduled.
+  for (int num = 0; num < options.level0_file_num_compaction_trigger; num++) {
+    for (int i = 0; i < kNumKeysPerFile; i++) {
+      ASSERT_OK(Put(0, Key(i), ""));
+    }
+    // put extra key to trigger flush
+    ASSERT_OK(Put(0, "", ""));
+    dbfull()->TEST_WaitForFlushMemTable(handles_[0]);
+    ASSERT_EQ(options.level0_file_num_compaction_trigger + num + 1,
+              NumTableFilesAtLevel(0, 0));
+  }
+
+  // All CFs are pending compaction
+  ASSERT_EQ(cf_count, env_->GetThreadPoolQueueLen(Env::LOW));
+
+  // Unblock all compaction threads
+  for (size_t i = 0; i < kTotalCompactTasks; i++) {
+    sleeping_compact_tasks[i].WakeUp();
+    sleeping_compact_tasks[i].WaitUntilDone();
+  }
+
+  for (int cf = 0; cf < cf_count; cf++) {
+    dbfull()->TEST_WaitForFlushMemTable(handles_[cf]);
+  }
+
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  // Max outstanding compact tasks reached limit
+  for (auto& ls : limiter_settings) {
+    ASSERT_EQ(ls.limit_tasks, ls.max_tasks);
+    ASSERT_EQ(0, ls.limiter->GetOutstandingTask());
+  }
+
+  // test manual compaction under a fully throttled limiter
+  int cf_test = 1;
+  unique_limiter->SetMaxOutstandingTask(0);
+
+  // flush one more file to cf 1
+  for (int i = 0; i < kNumKeysPerFile; i++) {
+      ASSERT_OK(Put(cf_test, Key(keyIndex++), ""));
+  }
+  // put extra key to trigger flush
+  ASSERT_OK(Put(cf_test, "", ""));
+
+  dbfull()->TEST_WaitForFlushMemTable(handles_[cf_test]);
+  ASSERT_EQ(1, NumTableFilesAtLevel(0, cf_test));
+
+  Compact(cf_test, Key(0), Key(keyIndex));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+}
+
 INSTANTIATE_TEST_CASE_P(DBCompactionTestWithParam, DBCompactionTestWithParam,
                        ::testing::Values(std::make_tuple(1, true),
                                          std::make_tuple(1, false),
@@ -3991,14 +4419,14 @@ class NoopMergeOperator : public MergeOperator {
 public:
  NoopMergeOperator() {}

-  virtual bool FullMergeV2(const MergeOperationInput& /*merge_in*/,
-                           MergeOperationOutput* merge_out) const override {
+  bool FullMergeV2(const MergeOperationInput& /*merge_in*/,
+                   MergeOperationOutput* merge_out) const override {
    std::string val("bar");
    merge_out->new_value = val;
    return true;
  }

-  virtual const char* Name() const override { return "Noop"; }
+  const char* Name() const override { return "Noop"; }
 };

 TEST_F(DBCompactionTest, PartialManualCompaction) {
@@ -4027,7 +4455,7 @@ TEST_F(DBCompactionTest, PartialManualCompaction) {
      {{"max_compaction_bytes", std::to_string(max_compaction_bytes)}}));

  CompactRangeOptions cro;
-  cro.bottommost_level_compaction = BottommostLevelCompaction::kForce;
+  cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
  dbfull()->CompactRange(cro, nullptr, nullptr);
 }

@@ -4071,6 +4499,58 @@ TEST_F(DBCompactionTest, ManualCompactionFailsInReadOnlyMode) {
  Close();
 }

+// ManualCompactionBottomLevelOptimization tests the bottom level manual
+// compaction optimization to skip recompacting files created by Ln-1 to Ln
+// compaction
+TEST_F(DBCompactionTest, ManualCompactionBottomLevelOptimized) {
+  Options opts = CurrentOptions();
+  opts.num_levels = 3;
+  opts.level0_file_num_compaction_trigger = 5;
+  opts.compression = kNoCompression;
+  opts.merge_operator.reset(new NoopMergeOperator());
+  opts.target_file_size_base = 1024;
+  opts.max_bytes_for_level_multiplier = 2;
+  opts.disable_auto_compactions = true;
+  DestroyAndReopen(opts);
+  ColumnFamilyHandleImpl* cfh =
+      static_cast<ColumnFamilyHandleImpl*>(dbfull()->DefaultColumnFamily());
+  ColumnFamilyData* cfd = cfh->cfd();
+  InternalStats* internal_stats_ptr = cfd->internal_stats();
+  ASSERT_NE(internal_stats_ptr, nullptr);
+
+  Random rnd(301);
+  for (auto i = 0; i < 8; ++i) {
+    for (auto j = 0; j < 10; ++j) {
+      ASSERT_OK(
+          Put("foo" + std::to_string(i * 10 + j), RandomString(&rnd, 1024)));
+    }
+    Flush();
+  }
+
+  MoveFilesToLevel(2);
+
+  for (auto i = 0; i < 8; ++i) {
+    for (auto j = 0; j < 10; ++j) {
+      ASSERT_OK(
+          Put("bar" + std::to_string(i * 10 + j), RandomString(&rnd, 1024)));
+    }
+    Flush();
+  }
+  const std::vector<InternalStats::CompactionStats>& comp_stats =
+      internal_stats_ptr->TEST_GetCompactionStats();
+  int num = comp_stats[2].num_input_files_in_output_level;
+  ASSERT_EQ(num, 0);
+
+  CompactRangeOptions cro;
+  cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
+  dbfull()->CompactRange(cro, nullptr, nullptr);
+
+  const std::vector<InternalStats::CompactionStats>& comp_stats2 =
+      internal_stats_ptr->TEST_GetCompactionStats();
+  num = comp_stats2[2].num_input_files_in_output_level;
+  ASSERT_EQ(num, 0);
+}
+
 // FixFileIngestionCompactionDeadlock tests and verifies that compaction and
 // file ingestion do not cause deadlock in the event of write stall triggered
 // by number of L0 files reaching level0_stop_writes_trigger.
@@ -407,6 +407,144 @@ TEST_P(DBAtomicFlushTest, AtomicFlushRollbackSomeJobs) {
  Destroy(options);
 }

+TEST_P(DBAtomicFlushTest, FlushMultipleCFs_DropSomeBeforeRequestFlush) {
+  bool atomic_flush = GetParam();
+  if (!atomic_flush) {
+    return;
+  }
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.atomic_flush = atomic_flush;
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  CreateAndReopenWithCF({"pikachu", "eevee"}, options);
+  size_t num_cfs = handles_.size();
+  ASSERT_EQ(3, num_cfs);
+  WriteOptions wopts;
+  wopts.disableWAL = true;
+  std::vector<int> cf_ids;
+  for (size_t i = 0; i != num_cfs; ++i) {
+    int cf_id = static_cast<int>(i);
+    ASSERT_OK(Put(cf_id, "key", "value", wopts));
+    cf_ids.push_back(cf_id);
+  }
+  ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
+  ASSERT_TRUE(Flush(cf_ids).IsShutdownInProgress());
+  Destroy(options);
+}
+
+TEST_P(DBAtomicFlushTest,
+       FlushMultipleCFs_DropSomeAfterScheduleFlushBeforeFlushJobRun) {
+  bool atomic_flush = GetParam();
+  if (!atomic_flush) {
+    return;
+  }
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.atomic_flush = atomic_flush;
+
+  CreateAndReopenWithCF({"pikachu", "eevee"}, options);
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  SyncPoint::GetInstance()->LoadDependency(
+      {{"DBImpl::AtomicFlushMemTables:AfterScheduleFlush",
+        "DBAtomicFlushTest::BeforeDropCF"},
+       {"DBAtomicFlushTest::AfterDropCF",
+        "DBImpl::BackgroundCallFlush:start"}});
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  size_t num_cfs = handles_.size();
+  ASSERT_EQ(3, num_cfs);
+  WriteOptions wopts;
+  wopts.disableWAL = true;
+  for (size_t i = 0; i != num_cfs; ++i) {
+    int cf_id = static_cast<int>(i);
+    ASSERT_OK(Put(cf_id, "key", "value", wopts));
+  }
+  port::Thread user_thread([&]() {
+    TEST_SYNC_POINT("DBAtomicFlushTest::BeforeDropCF");
+    ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
+    TEST_SYNC_POINT("DBAtomicFlushTest::AfterDropCF");
+  });
+  FlushOptions flush_opts;
+  flush_opts.wait = true;
+  ASSERT_OK(dbfull()->Flush(flush_opts, handles_));
+  user_thread.join();
+  for (size_t i = 0; i != num_cfs; ++i) {
+    int cf_id = static_cast<int>(i);
+    ASSERT_EQ("value", Get(cf_id, "key"));
+  }
+
+  ReopenWithColumnFamilies({kDefaultColumnFamilyName, "eevee"}, options);
+  num_cfs = handles_.size();
+  ASSERT_EQ(2, num_cfs);
+  for (size_t i = 0; i != num_cfs; ++i) {
+    int cf_id = static_cast<int>(i);
+    ASSERT_EQ("value", Get(cf_id, "key"));
+  }
+  Destroy(options);
+}
+
+TEST_P(DBAtomicFlushTest, TriggerFlushAndClose) {
+  bool atomic_flush = GetParam();
+  if (!atomic_flush) {
+    return;
+  }
+  const int kNumKeysTriggerFlush = 4;
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.atomic_flush = atomic_flush;
+  options.memtable_factory.reset(
+      new SpecialSkipListFactory(kNumKeysTriggerFlush));
+  CreateAndReopenWithCF({"pikachu"}, options);
+
+  for (int i = 0; i != kNumKeysTriggerFlush; ++i) {
+    ASSERT_OK(Put(0, "key" + std::to_string(i), "value" + std::to_string(i)));
+  }
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  SyncPoint::GetInstance()->EnableProcessing();
+  ASSERT_OK(Put(0, "key", "value"));
+  Close();
+
+  ReopenWithColumnFamilies({kDefaultColumnFamilyName, "pikachu"}, options);
+  ASSERT_EQ("value", Get(0, "key"));
+}
+
+TEST_P(DBAtomicFlushTest, PickMemtablesRaceWithBackgroundFlush) {
+  bool atomic_flush = GetParam();
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.atomic_flush = atomic_flush;
+  options.max_write_buffer_number = 4;
+  // Set min_write_buffer_number_to_merge to be greater than 1, so that
+  // a column family with one memtable in the imm will not cause IsFlushPending
+  // to return true when flush_requested_ is false.
+  options.min_write_buffer_number_to_merge = 2;
+  CreateAndReopenWithCF({"pikachu"}, options);
+  ASSERT_EQ(2, handles_.size());
+  ASSERT_OK(dbfull()->PauseBackgroundWork());
+  ASSERT_OK(Put(0, "key00", "value00"));
+  ASSERT_OK(Put(1, "key10", "value10"));
+  FlushOptions flush_opts;
+  flush_opts.wait = false;
+  ASSERT_OK(dbfull()->Flush(flush_opts, handles_));
+  ASSERT_OK(Put(0, "key01", "value01"));
+  // Since max_write_buffer_number is 4, the following flush won't cause write
+  // stall.
+  ASSERT_OK(dbfull()->Flush(flush_opts));
+  ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
+  ASSERT_OK(dbfull()->DestroyColumnFamilyHandle(handles_[1]));
+  handles_[1] = nullptr;
+  ASSERT_OK(dbfull()->ContinueBackgroundWork());
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[0]));
+  delete handles_[0];
+  handles_.clear();
+}
+
 INSTANTIATE_TEST_CASE_P(DBFlushDirectIOTest, DBFlushDirectIOTest,
                        testing::Bool());

@@ -31,7 +31,7 @@
 #include "db/log_writer.h"
 #include "db/logs_with_prep_tracker.h"
 #include "db/pre_release_callback.h"
-#include "db/range_del_aggregator_v2.h"
+#include "db/range_del_aggregator.h"
 #include "db/read_callback.h"
 #include "db/snapshot_checker.h"
 #include "db/snapshot_impl.h"
@@ -63,8 +63,10 @@ namespace rocksdb {

 class Arena;
 class ArenaWrappedDBIter;
+class InMemoryStatsHistoryIterator;
 class MemTable;
 class TableCache;
+class TaskLimiterToken;
 class Version;
 class VersionEdit;
 class VersionSet;
@@ -122,6 +124,25 @@ class DBImpl : public DB {
      const std::vector<Slice>& keys,
      std::vector<std::string>* values) override;

+  // This MultiGet is a batched version, which may be faster than calling Get
+  // multiple times, especially if the keys have some spatial locality that
+  // enables them to be queried in the same SST files/set of files. The larger
+  // the batch size, the more scope for batching and performance improvement
+  // The values and statuses parameters are arrays with number of elements
+  // equal to keys.size(). This allows the storage for those to be alloacted
+  // by the caller on the stack for small batches
+  virtual void MultiGet(const ReadOptions& options,
+                        ColumnFamilyHandle* column_family,
+                        const size_t num_keys, const Slice* keys,
+                        PinnableSlice* values, Status* statuses,
+                        const bool sorted_input = false) override;
+
+  void MultiGetImpl(
+      const ReadOptions& options, ColumnFamilyHandle* column_family,
+      autovector<KeyContext, MultiGetContext::MAX_BATCH_SIZE>& key_context,
+      bool sorted_input, ReadCallback* callback = nullptr,
+      bool* is_blob_index = nullptr);
+
  virtual Status CreateColumnFamily(const ColumnFamilyOptions& cf_options,
                                    const std::string& column_family,
                                    ColumnFamilyHandle** handle) override;
@@ -176,10 +197,9 @@ class DBImpl : public DB {
  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,
-                                   uint8_t include_flags
-                                   = INCLUDE_FILES) override;
+  virtual void GetApproximateSizes(
+      ColumnFamilyHandle* column_family, const Range* range, int n,
+      uint64_t* sizes, uint8_t include_flags = INCLUDE_FILES) override;
  using DB::GetApproximateMemTableStats;
  virtual void GetApproximateMemTableStats(ColumnFamilyHandle* column_family,
                                           const Range& range,
@@ -191,13 +211,13 @@ class DBImpl : public DB {
                              const Slice* begin, const Slice* end) override;

  using DB::CompactFiles;
-  virtual Status CompactFiles(const CompactionOptions& compact_options,
-                              ColumnFamilyHandle* column_family,
-                              const std::vector<std::string>& input_file_names,
-                              const int output_level,
-                              const int output_path_id = -1,
-                              std::vector<std::string>* const output_file_names
-                              = nullptr) override;
+  virtual Status CompactFiles(
+      const CompactionOptions& compact_options,
+      ColumnFamilyHandle* column_family,
+      const std::vector<std::string>& input_file_names, const int output_level,
+      const int output_path_id = -1,
+      std::vector<std::string>* const output_file_names = nullptr,
+      CompactionJobInfo* compaction_job_info = nullptr) override;

  virtual Status PauseBackgroundWork() override;
  virtual Status ContinueBackgroundWork() override;
@@ -233,10 +253,19 @@ class DBImpl : public DB {
      const FlushOptions& options,
      const std::vector<ColumnFamilyHandle*>& column_families) override;
  virtual Status FlushWAL(bool sync) override;
-  bool TEST_WALBufferIsEmpty();
+  bool TEST_WALBufferIsEmpty(bool lock = true);
  virtual Status SyncWAL() override;
+  virtual Status LockWAL() override;
+  virtual Status UnlockWAL() override;

  virtual SequenceNumber GetLatestSequenceNumber() const override;
+  virtual SequenceNumber GetLastPublishedSequence() const {
+    if (last_seq_same_as_publish_seq_) {
+      return versions_->LastSequence();
+    } else {
+      return versions_->LastPublishedSequence();
+    }
+  }
  // REQUIRES: joined the main write queue if two_write_queues is disabled, and
  // the second write queue otherwise.
  virtual void SetLastPublishedSequence(SequenceNumber seq);
@@ -275,9 +304,8 @@ class DBImpl : public DB {
  // Status::NotFound() will be returned if the current DB does not have
  // any column family match the specified name.
  // TODO(yhchiang): output parameter is placed in the end in this codebase.
-  virtual void GetColumnFamilyMetaData(
-      ColumnFamilyHandle* column_family,
-      ColumnFamilyMetaData* metadata) override;
+  virtual void GetColumnFamilyMetaData(ColumnFamilyHandle* column_family,
+                                       ColumnFamilyMetaData* metadata) override;

  Status SuggestCompactRange(ColumnFamilyHandle* column_family,
                             const Slice* begin, const Slice* end) override;
@@ -338,6 +366,10 @@ class DBImpl : public DB {
      const std::vector<std::string>& external_files,
      const IngestExternalFileOptions& ingestion_options) override;

+  using DB::IngestExternalFiles;
+  virtual Status IngestExternalFiles(
+      const std::vector<IngestExternalFileArg>& args) override;
+
  virtual Status VerifyChecksum() override;

  using DB::StartTrace;
@@ -363,19 +395,22 @@ class DBImpl : public DB {

  virtual Status GetDbIdentity(std::string& identity) const override;

+  // max_file_num_to_ignore allows bottom level compaction to filter out newly
+  // compacted SST files. Setting max_file_num_to_ignore to kMaxUint64 will
+  // disable the filtering
  Status RunManualCompaction(ColumnFamilyData* cfd, int input_level,
-                             int output_level, uint32_t output_path_id,
-                             uint32_t max_subcompactions,
+                             int output_level,
+                             const CompactRangeOptions& compact_range_options,
                             const Slice* begin, const Slice* end,
-                             bool exclusive,
-                             bool disallow_trivial_move = false);
+                             bool exclusive, bool disallow_trivial_move,
+                             uint64_t max_file_num_to_ignore);

  // 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, RangeDelAggregatorV2* range_del_agg,
-      SequenceNumber sequence, ColumnFamilyHandle* column_family = nullptr);
+      Arena* arena, RangeDelAggregator* range_del_agg, SequenceNumber sequence,
+      ColumnFamilyHandle* column_family = nullptr);

  LogsWithPrepTracker* logs_with_prep_tracker() {
    return &logs_with_prep_tracker_;
@@ -414,8 +449,8 @@ class DBImpl : public DB {

  // Return the maximum overlapping data (in bytes) at next level for any
  // file at a level >= 1.
-  int64_t TEST_MaxNextLevelOverlappingBytes(ColumnFamilyHandle* column_family =
-                                                nullptr);
+  int64_t TEST_MaxNextLevelOverlappingBytes(
+      ColumnFamilyHandle* column_family = nullptr);

  // Return the current manifest file no.
  uint64_t TEST_Current_Manifest_FileNo();
@@ -470,7 +505,10 @@ class DBImpl : public DB {
  int TEST_BGCompactionsAllowed() const;
  int TEST_BGFlushesAllowed() const;
  size_t TEST_GetWalPreallocateBlockSize(uint64_t write_buffer_size) const;
-  void TEST_WaitForTimedTaskRun(std::function<void()> callback) const;
+  void TEST_WaitForDumpStatsRun(std::function<void()> callback) const;
+  void TEST_WaitForPersistStatsRun(std::function<void()> callback) const;
+  bool TEST_IsPersistentStatsEnabled() const;
+  size_t TEST_EstiamteStatsHistorySize() const;

 #endif  // NDEBUG

@@ -523,6 +561,13 @@ class DBImpl : public DB {

  const SnapshotList& snapshots() const { return snapshots_; }

+  void LoadSnapshots(std::vector<SequenceNumber>* snap_vector,
+                     SequenceNumber* oldest_write_conflict_snapshot,
+                     const SequenceNumber& max_seq) const {
+    InstrumentedMutexLock l(mutex());
+    snapshots().GetAll(snap_vector, oldest_write_conflict_snapshot, max_seq);
+  }
+
  const ImmutableDBOptions& immutable_db_options() const {
    return immutable_db_options_;
  }
@@ -578,12 +623,9 @@ class DBImpl : public DB {

  const WriteController& write_controller() { return write_controller_; }

-  InternalIterator* NewInternalIterator(const ReadOptions&,
-                                        ColumnFamilyData* cfd,
-                                        SuperVersion* super_version,
-                                        Arena* arena,
-                                        RangeDelAggregatorV2* range_del_agg,
-                                        SequenceNumber sequence);
+  InternalIterator* NewInternalIterator(
+      const ReadOptions&, ColumnFamilyData* cfd, SuperVersion* super_version,
+      Arena* arena, RangeDelAggregator* range_del_agg, SequenceNumber sequence);

  // hollow transactions shell used for recovery.
  // these will then be passed to TransactionDB so that
@@ -695,10 +737,16 @@ class DBImpl : public DB {

  void SetSnapshotChecker(SnapshotChecker* snapshot_checker);

+  // Fill JobContext with snapshot information needed by flush and compaction.
+  void GetSnapshotContext(JobContext* job_context,
+                          std::vector<SequenceNumber>* snapshot_seqs,
+                          SequenceNumber* earliest_write_conflict_snapshot,
+                          SnapshotChecker** snapshot_checker);
+
  // Not thread-safe.
  void SetRecoverableStatePreReleaseCallback(PreReleaseCallback* callback);

-  InstrumentedMutex* mutex() { return &mutex_; }
+  InstrumentedMutex* mutex() const { return &mutex_; }

  Status NewDB();

@@ -713,6 +761,17 @@ class DBImpl : public DB {
  static Status CreateAndNewDirectory(Env* env, const std::string& dirname,
                                      std::unique_ptr<Directory>* directory);

+  // Given a time window, return an iterator for accessing stats history
+  Status GetStatsHistory(
+      uint64_t start_time, uint64_t end_time,
+      std::unique_ptr<StatsHistoryIterator>* stats_iterator) override;
+
+  // find stats map from stats_history_ with smallest timestamp in
+  // the range of [start_time, end_time)
+  bool FindStatsByTime(uint64_t start_time, uint64_t end_time,
+                       uint64_t* new_time,
+                       std::map<std::string, uint64_t>* stats_map);
+
 protected:
  Env* const env_;
  const std::string dbname_;
@@ -728,6 +787,46 @@ class DBImpl : public DB {
  std::unique_ptr<Tracer> tracer_;
  InstrumentedMutex trace_mutex_;

+  // State below is protected by mutex_
+  // With two_write_queues enabled, some of the variables that accessed during
+  // WriteToWAL need different synchronization: log_empty_, alive_log_files_,
+  // logs_, logfile_number_. Refer to the definition of each variable below for
+  // more description.
+  mutable InstrumentedMutex mutex_;
+
+  ColumnFamilyHandleImpl* default_cf_handle_;
+  InternalStats* default_cf_internal_stats_;
+
+  // only used for dynamically adjusting max_total_wal_size. it is a sum of
+  // [write_buffer_size * max_write_buffer_number] over all column families
+  uint64_t max_total_in_memory_state_;
+  // If true, we have only one (default) column family. We use this to optimize
+  // some code-paths
+  bool single_column_family_mode_;
+
+  // The options to access storage files
+  const EnvOptions env_options_;
+
+  // Additonal options for compaction and flush
+  EnvOptions env_options_for_compaction_;
+
+  std::unique_ptr<ColumnFamilyMemTablesImpl> column_family_memtables_;
+
+  // Increase the sequence number after writing each batch, whether memtable is
+  // disabled for that or not. Otherwise the sequence number is increased after
+  // writing each key into memtable. This implies that when disable_memtable is
+  // set, the seq is not increased at all.
+  //
+  // Default: false
+  const bool seq_per_batch_;
+  // This determines during recovery whether we expect one writebatch per
+  // recovered transaction, or potentially multiple writebatches per
+  // transaction. For WriteUnprepared, this is set to false, since multiple
+  // batches can exist per transaction.
+  //
+  // Default: true
+  const bool batch_per_txn_;
+
  // Except in DB::Open(), WriteOptionsFile can only be called when:
  // Persist options to options file.
  // If need_mutex_lock = false, the method will lock DB mutex.
@@ -748,13 +847,12 @@ class DBImpl : public DB {
                              const MutableCFOptions& mutable_cf_options,
                              int job_id, TableProperties prop);

-  void NotifyOnCompactionBegin(ColumnFamilyData* cfd,
-                               Compaction *c, const Status &st,
-                               const CompactionJobStats& job_stats,
-                               int job_id);
+  void NotifyOnCompactionBegin(ColumnFamilyData* cfd, Compaction* c,
+                               const Status& st,
+                               const CompactionJobStats& job_stats, int job_id);

-  void NotifyOnCompactionCompleted(ColumnFamilyData* cfd,
-                                   Compaction *c, const Status &st,
+  void NotifyOnCompactionCompleted(ColumnFamilyData* cfd, Compaction* c,
+                                   const Status& st,
                                   const CompactionJobStats& job_stats,
                                   int job_id);
  void NotifyOnMemTableSealed(ColumnFamilyData* cfd,
@@ -815,6 +913,14 @@ class DBImpl : public DB {
  // Actual implementation of Close()
  Status CloseImpl();

+  // Recover the descriptor from persistent storage.  May do a significant
+  // amount of work to recover recently logged updates.  Any changes to
+  // be made to the descriptor are added to *edit.
+  virtual Status Recover(
+      const std::vector<ColumnFamilyDescriptor>& column_families,
+      bool read_only = false, bool error_if_log_file_exist = false,
+      bool error_if_data_exists_in_logs = false);
+
 private:
  friend class DB;
  friend class ErrorHandler;
@@ -835,6 +941,7 @@ class DBImpl : public DB {
  friend class CompactedDBImpl;
  friend class DBTest_ConcurrentFlushWAL_Test;
  friend class DBTest_MixedSlowdownOptionsStop_Test;
+  friend class DBCompactionTest_CompactBottomLevelFilesWithDeletions_Test;
 #ifndef NDEBUG
  friend class DBTest2_ReadCallbackTest_Test;
  friend class WriteCallbackTest_WriteWithCallbackTest_Test;
@@ -862,13 +969,6 @@ class DBImpl : public DB {
  struct PrepickedCompaction;
  struct PurgeFileInfo;

-  // Recover the descriptor from persistent storage.  May do a significant
-  // amount of work to recover recently logged updates.  Any changes to
-  // be made to the descriptor are added to *edit.
-  Status Recover(const std::vector<ColumnFamilyDescriptor>& column_families,
-                 bool read_only = false, bool error_if_log_file_exist = false,
-                 bool error_if_data_exists_in_logs = false);
-
  Status ResumeImpl();

  void MaybeIgnoreError(Status* s) const;
@@ -912,11 +1012,14 @@ class DBImpl : public DB {
  // Flush the in-memory write buffer to storage.  Switches to a new
  // log-file/memtable and writes a new descriptor iff successful. Then
  // installs a new super version for the column family.
-  Status FlushMemTableToOutputFile(ColumnFamilyData* cfd,
-                                   const MutableCFOptions& mutable_cf_options,
-                                   bool* madeProgress, JobContext* job_context,
-                                   SuperVersionContext* superversion_context,
-                                   LogBuffer* log_buffer);
+  Status FlushMemTableToOutputFile(
+      ColumnFamilyData* cfd, const MutableCFOptions& mutable_cf_options,
+      bool* madeProgress, JobContext* job_context,
+      SuperVersionContext* superversion_context,
+      std::vector<SequenceNumber>& snapshot_seqs,
+      SequenceNumber earliest_write_conflict_snapshot,
+      SnapshotChecker* snapshot_checker, LogBuffer* log_buffer,
+      Env::Priority thread_pri);

  // Argument required by background flush thread.
  struct BGFlushArg {
@@ -939,19 +1042,26 @@ class DBImpl : public DB {
    SuperVersionContext* superversion_context_;
  };

+  // Argument passed to flush thread.
+  struct FlushThreadArg {
+    DBImpl* db_;
+
+    Env::Priority thread_pri_;
+  };
+
  // Flush the memtables of (multiple) column families to multiple files on
  // persistent storage.
  Status FlushMemTablesToOutputFiles(
      const autovector<BGFlushArg>& bg_flush_args, bool* made_progress,
-      JobContext* job_context, LogBuffer* log_buffer);
+      JobContext* job_context, LogBuffer* log_buffer, Env::Priority thread_pri);

  Status AtomicFlushMemTablesToOutputFiles(
      const autovector<BGFlushArg>& bg_flush_args, bool* made_progress,
-      JobContext* job_context, LogBuffer* log_buffer);
+      JobContext* job_context, LogBuffer* log_buffer, Env::Priority thread_pri);

  // REQUIRES: log_numbers are sorted in ascending order
-  Status RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
-                         SequenceNumber* next_sequence, bool read_only);
+  virtual Status RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
+                                 SequenceNumber* next_sequence, bool read_only);

  // The following two methods are used to flush a memtable to
  // storage. The first one is used at database RecoveryTime (when the
@@ -1053,7 +1163,8 @@ class DBImpl : public DB {
                          const std::vector<std::string>& input_file_names,
                          std::vector<std::string>* const output_file_names,
                          const int output_level, int output_path_id,
-                          JobContext* job_context, LogBuffer* log_buffer);
+                          JobContext* job_context, LogBuffer* log_buffer,
+                          CompactionJobInfo* compaction_job_info);

  // Wait for current IngestExternalFile() calls to finish.
  // REQUIRES: mutex_ held
@@ -1089,35 +1200,50 @@ class DBImpl : public DB {
  // Runs a pre-chosen universal compaction involving bottom level in a
  // separate, bottom-pri thread pool.
  static void BGWorkBottomCompaction(void* arg);
-  static void BGWorkFlush(void* db);
+  static void BGWorkFlush(void* arg);
  static void BGWorkPurge(void* arg);
-  static void UnscheduleCallback(void* arg);
+  static void UnscheduleCompactionCallback(void* arg);
+  static void UnscheduleFlushCallback(void* arg);
  void BackgroundCallCompaction(PrepickedCompaction* prepicked_compaction,
-                                Env::Priority bg_thread_pri);
-  void BackgroundCallFlush();
+                                Env::Priority thread_pri);
+  void BackgroundCallFlush(Env::Priority thread_pri);
  void BackgroundCallPurge();
  Status BackgroundCompaction(bool* madeProgress, JobContext* job_context,
                              LogBuffer* log_buffer,
-                              PrepickedCompaction* prepicked_compaction);
+                              PrepickedCompaction* prepicked_compaction,
+                              Env::Priority thread_pri);
  Status BackgroundFlush(bool* madeProgress, JobContext* job_context,
-                         LogBuffer* log_buffer, FlushReason* reason);
+                         LogBuffer* log_buffer, FlushReason* reason,
+                         Env::Priority thread_pri);

  bool EnoughRoomForCompaction(ColumnFamilyData* cfd,
                               const std::vector<CompactionInputFiles>& inputs,
                               bool* sfm_bookkeeping, LogBuffer* log_buffer);

+  // Request compaction tasks token from compaction thread limiter.
+  // It always succeeds if force = true or limiter is disable.
+  bool RequestCompactionToken(ColumnFamilyData* cfd, bool force,
+                              std::unique_ptr<TaskLimiterToken>* token,
+                              LogBuffer* log_buffer);
+
  // Schedule background tasks
  void StartTimedTasks();

  void PrintStatistics();

+  size_t EstiamteStatsHistorySize() const;
+
+  // persist stats to column family "_persistent_stats"
+  void PersistStats();
+
  // dump rocksdb.stats to LOG
  void DumpStats();

  // Return the minimum empty level that could hold the total data in the
  // input level. Return the input level, if such level could not be found.
  int FindMinimumEmptyLevelFitting(ColumnFamilyData* cfd,
-      const MutableCFOptions& mutable_cf_options, int level);
+                                   const MutableCFOptions& mutable_cf_options,
+                                   int level);

  // Move the files in the input level to the target level.
  // If target_level < 0, automatically calculate the minimum level that could
@@ -1129,10 +1255,15 @@ class DBImpl : public DB {
  ColumnFamilyData* PopFirstFromCompactionQueue();
  FlushRequest PopFirstFromFlushQueue();

+  // Pick the first unthrottled compaction with task token from queue.
+  ColumnFamilyData* PickCompactionFromQueue(
+      std::unique_ptr<TaskLimiterToken>* token, LogBuffer* log_buffer);
+
  // helper function to call after some of the logs_ were synced
  void MarkLogsSynced(uint64_t up_to, bool synced_dir, const Status& status);

-  SnapshotImpl* GetSnapshotImpl(bool is_write_conflict_boundary);
+  SnapshotImpl* GetSnapshotImpl(bool is_write_conflict_boundary,
+                                bool lock = true);

  uint64_t GetMaxTotalWalSize() const;

@@ -1148,17 +1279,13 @@ class DBImpl : public DB {
  // Lock over the persistent DB state.  Non-nullptr iff successfully acquired.
  FileLock* db_lock_;

+  // In addition to mutex_, log_write_mutex_ protected writes to stats_history_
+  InstrumentedMutex stats_history_mutex_;
  // In addition to mutex_, log_write_mutex_ protected writes to logs_ and
  // logfile_number_. With two_write_queues it also protects alive_log_files_,
  // and log_empty_. Refer to the definition of each variable below for more
  // details.
  InstrumentedMutex log_write_mutex_;
-  // State below is protected by mutex_
-  // With two_write_queues enabled, some of the variables that accessed during
-  // WriteToWAL need different synchronization: log_empty_, alive_log_files_,
-  // logs_, logfile_number_. Refer to the definition of each variable below for
-  // more description.
-  mutable InstrumentedMutex mutex_;

  std::atomic<bool> shutting_down_;
  // This condition variable is signaled on these conditions:
@@ -1190,12 +1317,9 @@ class DBImpl : public DB {
  // read and writes are protected by log_write_mutex_ instead. This is to avoid
  // expesnive mutex_ lock during WAL write, which update log_empty_.
  bool log_empty_;
-  ColumnFamilyHandleImpl* default_cf_handle_;
-  InternalStats* default_cf_internal_stats_;
-  std::unique_ptr<ColumnFamilyMemTablesImpl> column_family_memtables_;
+
  struct LogFileNumberSize {
-    explicit LogFileNumberSize(uint64_t _number)
-        : number(_number) {}
+    explicit LogFileNumberSize(uint64_t _number) : number(_number) {}
    void AddSize(uint64_t new_size) { size += new_size; }
    uint64_t number;
    uint64_t size = 0;
@@ -1258,18 +1382,19 @@ class DBImpl : public DB {
  WriteBatch cached_recoverable_state_;
  std::atomic<bool> cached_recoverable_state_empty_ = {true};
  std::atomic<uint64_t> total_log_size_;
-  // only used for dynamically adjusting max_total_wal_size. it is a sum of
-  // [write_buffer_size * max_write_buffer_number] over all column families
-  uint64_t max_total_in_memory_state_;
-  // If true, we have only one (default) column family. We use this to optimize
-  // some code-paths
-  bool single_column_family_mode_;
+
  // If this is non-empty, we need to delete these log files in background
  // threads. Protected by db mutex.
  autovector<log::Writer*> logs_to_free_;

  bool is_snapshot_supported_;

+  std::map<uint64_t, std::map<std::string, uint64_t>> stats_history_;
+
+  std::map<std::string, uint64_t> stats_slice_;
+
+  bool stats_slice_initialized_ = false;
+
  // Class to maintain directories for all database paths other than main one.
  class Directories {
   public:
@@ -1355,7 +1480,7 @@ class DBImpl : public DB {
  // compacted. Consumers of these queues are flush and compaction threads. When
  // column family is put on this queue, we increase unscheduled_flushes_ and
  // unscheduled_compactions_. When these variables are bigger than zero, that
-  // means we need to schedule background threads for compaction and thread.
+  // means we need to schedule background threads for flush and compaction.
  // Once the background threads are scheduled, we decrease unscheduled_flushes_
  // and unscheduled_compactions_. That way we keep track of number of
  // compaction and flush threads we need to schedule. This scheduling is done
@@ -1406,15 +1531,15 @@ class DBImpl : public DB {
    uint32_t output_path_id;
    Status status;
    bool done;
-    bool in_progress;             // compaction request being processed?
-    bool incomplete;              // only part of requested range compacted
-    bool exclusive;               // current behavior of only one manual
-    bool disallow_trivial_move;   // Force actual compaction to run
-    const InternalKey* begin;     // nullptr means beginning of key range
-    const InternalKey* end;       // nullptr means end of key range
-    InternalKey* manual_end;      // how far we are compacting
-    InternalKey tmp_storage;      // Used to keep track of compaction progress
-    InternalKey tmp_storage1;     // Used to keep track of compaction progress
+    bool in_progress;            // compaction request being processed?
+    bool incomplete;             // only part of requested range compacted
+    bool exclusive;              // current behavior of only one manual
+    bool disallow_trivial_move;  // Force actual compaction to run
+    const InternalKey* begin;    // nullptr means beginning of key range
+    const InternalKey* end;      // nullptr means end of key range
+    InternalKey* manual_end;     // how far we are compacting
+    InternalKey tmp_storage;     // Used to keep track of compaction progress
+    InternalKey tmp_storage1;    // Used to keep track of compaction progress
  };
  struct PrepickedCompaction {
    // background compaction takes ownership of `compaction`.
@@ -1422,6 +1547,8 @@ class DBImpl : public DB {
    // caller retains ownership of `manual_compaction_state` as it is reused
    // across background compactions.
    ManualCompactionState* manual_compaction_state;  // nullptr if non-manual
+    // task limiter token is requested during compaction picking.
+    std::unique_ptr<TaskLimiterToken> task_token;
  };
  std::deque<ManualCompactionState*> manual_compaction_dequeue_;

@@ -1461,7 +1588,7 @@ class DBImpl : public DB {
  std::atomic<bool> has_unpersisted_data_;

  // if an attempt was made to flush all column families that
-  // the oldest log depends on but uncommited data in the oldest
+  // the oldest log depends on but uncommitted data in the oldest
  // log prevents the log from being released.
  // We must attempt to free the dependent memtables again
  // at a later time after the transaction in the oldest
@@ -1474,12 +1601,6 @@ class DBImpl : public DB {

  std::string db_absolute_path_;

-  // The options to access storage files
-  const EnvOptions env_options_;
-
-  // Additonal options for compaction and flush
-  EnvOptions env_options_for_compaction_;
-
  // Number of running IngestExternalFile() calls.
  // REQUIRES: mutex held
  int num_running_ingest_file_;
@@ -1503,6 +1624,10 @@ class DBImpl : public DB {
  // Indicate DB was opened successfully
  bool opened_successfully_;

+  // The min threshold to triggere bottommost compaction for removing
+  // garbages, among all column families.
+  SequenceNumber bottommost_files_mark_threshold_ = kMaxSequenceNumber;
+
  LogsWithPrepTracker logs_with_prep_tracker_;

  // Callback for compaction to check if a key is visible to a snapshot.
@@ -1513,10 +1638,14 @@ class DBImpl : public DB {
  // Only to be set during initialization
  std::unique_ptr<PreReleaseCallback> recoverable_state_pre_release_callback_;

-  // handle for scheduling jobs at fixed intervals
+  // handle for scheduling stats dumping at fixed intervals
  // REQUIRES: mutex locked
  std::unique_ptr<rocksdb::RepeatableThread> thread_dump_stats_;

+  // handle for scheduling stats snapshoting at fixed intervals
+  // REQUIRES: mutex locked
+  std::unique_ptr<rocksdb::RepeatableThread> thread_persist_stats_;
+
  // No copying allowed
  DBImpl(const DBImpl&);
  void operator=(const DBImpl&);
@@ -1530,14 +1659,13 @@ class DBImpl : public DB {
  // state needs flush or compaction.
  void InstallSuperVersionAndScheduleWork(
      ColumnFamilyData* cfd, SuperVersionContext* sv_context,
-      const MutableCFOptions& mutable_cf_options,
-      FlushReason flush_reason = FlushReason::kOthers);
+      const MutableCFOptions& mutable_cf_options);

 #ifndef ROCKSDB_LITE
  using DB::GetPropertiesOfAllTables;
-  virtual Status GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
-                                          TablePropertiesCollection* props)
-      override;
+  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;
@@ -1556,33 +1684,35 @@ class DBImpl : public DB {
  bool ShouldntRunManualCompaction(ManualCompactionState* m);
  bool HaveManualCompaction(ColumnFamilyData* cfd);
  bool MCOverlap(ManualCompactionState* m, ManualCompactionState* m1);
+#ifndef ROCKSDB_LITE
+  void BuildCompactionJobInfo(const ColumnFamilyData* cfd, Compaction* c,
+                              const Status& st,
+                              const CompactionJobStats& compaction_job_stats,
+                              const int job_id, const Version* current,
+                              CompactionJobInfo* compaction_job_info) const;
+  // Reserve the next 'num' file numbers for to-be-ingested external SST files,
+  // and return the current file_number in 'next_file_number'.
+  // Write a version edit to the MANIFEST.
+  Status ReserveFileNumbersBeforeIngestion(
+      ColumnFamilyData* cfd, uint64_t num,
+      std::list<uint64_t>::iterator* pending_output_elem,
+      uint64_t* next_file_number);
+#endif  //! ROCKSDB_LITE

  bool ShouldPurge(uint64_t file_number) const;
  void MarkAsGrabbedForPurge(uint64_t file_number);

  size_t GetWalPreallocateBlockSize(uint64_t write_buffer_size) const;
-  Env::WriteLifeTimeHint CalculateWALWriteHint() {
-    return Env::WLTH_SHORT;
-  }
+  Env::WriteLifeTimeHint CalculateWALWriteHint() { return Env::WLTH_SHORT; }
+
+  Status CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
+                   size_t preallocate_block_size, log::Writer** new_log);

  // When set, we use a separate queue for writes that dont write to memtable.
  // In 2PC these are the writes at Prepare phase.
  const bool two_write_queues_;
  const bool manual_wal_flush_;
-  // Increase the sequence number after writing each batch, whether memtable is
-  // disabled for that or not. Otherwise the sequence number is increased after
-  // writing each key into memtable. This implies that when disable_memtable is
-  // set, the seq is not increased at all.
-  //
-  // Default: false
-  const bool seq_per_batch_;
-  // This determines during recovery whether we expect one writebatch per
-  // recovered transaction, or potentially multiple writebatches per
-  // transaction. For WriteUnprepared, this is set to false, since multiple
-  // batches can exist per transaction.
-  //
-  // Default: true
-  const bool batch_per_txn_;
+
  // LastSequence also indicates last published sequence visibile to the
  // readers. Otherwise LastPublishedSequence should be used.
  const bool last_seq_same_as_publish_seq_;
@@ -1613,19 +1743,19 @@ class DBImpl : public DB {

  ErrorHandler error_handler_;

-  // True if the DB is committing atomic flush.
-  // TODO (yanqin) the current impl assumes that the entire DB belongs to
-  // a single atomic flush group. In the future we need to add a new class
-  // (struct) similar to the following to make it more general.
-  // struct AtomicFlushGroup {
-  //   bool commit_in_progress_;
-  //   std::vector<MemTableList*> imm_lists;
-  // };
-  bool atomic_flush_commit_in_progress_;
+  // Conditional variable to coordinate installation of atomic flush results.
+  // With atomic flush, each bg thread installs the result of flushing multiple
+  // column families, and different threads can flush different column
+  // families. It's difficult to rely on one thread to perform batch
+  // installation for all threads. This is different from the non-atomic flush
+  // case.
+  // atomic_flush_install_cv_ makes sure that threads install atomic flush
+  // results sequentially. Flush results of memtables with lower IDs get
+  // installed to MANIFEST first.
+  InstrumentedCondVar atomic_flush_install_cv_;
 };

-extern Options SanitizeOptions(const std::string& db,
-                               const Options& src);
+extern Options SanitizeOptions(const std::string& db, const Options& src);

 extern DBOptions SanitizeOptions(const std::string& db, const DBOptions& src);

@@ -26,10 +26,16 @@ void DBImpl::TEST_SwitchWAL() {
  SwitchWAL(&write_context);
 }

-bool DBImpl::TEST_WALBufferIsEmpty() {
-  InstrumentedMutexLock wl(&log_write_mutex_);
+bool DBImpl::TEST_WALBufferIsEmpty(bool lock) {
+  if (lock) {
+    log_write_mutex_.Lock();
+  }
  log::Writer* cur_log_writer = logs_.back().writer;
-  return cur_log_writer->TEST_BufferIsEmpty();
+  auto res = cur_log_writer->TEST_BufferIsEmpty();
+  if (lock) {
+    log_write_mutex_.Unlock();
+  }
+  return res;
 }

 int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
@@ -87,8 +93,9 @@ Status DBImpl::TEST_CompactRange(int level, const Slice* begin,
       cfd->ioptions()->compaction_style == kCompactionStyleFIFO)
          ? level
          : level + 1;
-  return RunManualCompaction(cfd, level, output_level, 0, 0, begin, end, true,
-                             disallow_trivial_move);
+  return RunManualCompaction(cfd, level, output_level, CompactRangeOptions(),
+                             begin, end, true, disallow_trivial_move,
+                             port::kMaxUint64 /*max_file_num_to_ignore*/);
 }

 Status DBImpl::TEST_SwitchMemtable(ColumnFamilyData* cfd) {
@@ -101,7 +108,7 @@ Status DBImpl::TEST_SwitchMemtable(ColumnFamilyData* cfd) {
 }

 Status DBImpl::TEST_FlushMemTable(bool wait, bool allow_write_stall,
-    ColumnFamilyHandle* cfh) {
+                                  ColumnFamilyHandle* cfh) {
  FlushOptions fo;
  fo.wait = wait;
  fo.allow_write_stall = allow_write_stall;
@@ -143,13 +150,9 @@ Status DBImpl::TEST_WaitForCompact(bool wait_unscheduled) {
  return error_handler_.GetBGError();
 }

-void DBImpl::TEST_LockMutex() {
-  mutex_.Lock();
-}
+void DBImpl::TEST_LockMutex() { mutex_.Lock(); }

-void DBImpl::TEST_UnlockMutex() {
-  mutex_.Unlock();
-}
+void DBImpl::TEST_UnlockMutex() { mutex_.Unlock(); }

 void* DBImpl::TEST_BeginWrite() {
  auto w = new WriteThread::Writer();
@@ -243,10 +246,24 @@ size_t DBImpl::TEST_GetWalPreallocateBlockSize(
  return GetWalPreallocateBlockSize(write_buffer_size);
 }

-void DBImpl::TEST_WaitForTimedTaskRun(std::function<void()> callback) const {
+void DBImpl::TEST_WaitForDumpStatsRun(std::function<void()> callback) const {
  if (thread_dump_stats_ != nullptr) {
    thread_dump_stats_->TEST_WaitForRun(callback);
  }
 }
+
+void DBImpl::TEST_WaitForPersistStatsRun(std::function<void()> callback) const {
+  if (thread_persist_stats_ != nullptr) {
+    thread_persist_stats_->TEST_WaitForRun(callback);
+  }
+}
+
+bool DBImpl::TEST_IsPersistentStatsEnabled() const {
+  return thread_persist_stats_ && thread_persist_stats_->IsRunning();
+}
+
+size_t DBImpl::TEST_EstiamteStatsHistorySize() const {
+  return EstiamteStatsHistorySize();
+}
 }  // namespace rocksdb
 #endif  // NDEBUG
@@ -108,7 +108,7 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
  versions_->AddLiveFiles(&job_context->sst_live);
  if (doing_the_full_scan) {
    InfoLogPrefix info_log_prefix(!immutable_db_options_.db_log_dir.empty(),
-        dbname_);
+                                  dbname_);
    std::set<std::string> paths;
    for (size_t path_id = 0; path_id < immutable_db_options_.db_paths.size();
         path_id++) {
@@ -152,8 +152,7 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
        }

        // TODO(icanadi) clean up this mess to avoid having one-off "/" prefixes
-        job_context->full_scan_candidate_files.emplace_back(
-            "/" + file, path);
+        job_context->full_scan_candidate_files.emplace_back("/" + file, path);
      }
    }

@@ -163,8 +162,8 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
      env_->GetChildren(immutable_db_options_.wal_dir,
                        &log_files);  // Ignore errors
      for (const std::string& log_file : log_files) {
-        job_context->full_scan_candidate_files.emplace_back(log_file,
-            immutable_db_options_.wal_dir);
+        job_context->full_scan_candidate_files.emplace_back(
+            log_file, immutable_db_options_.wal_dir);
      }
    }
    // Add info log files in db_log_dir
@@ -174,8 +173,8 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
      // Ignore errors
      env_->GetChildren(immutable_db_options_.db_log_dir, &info_log_files);
      for (std::string& log_file : info_log_files) {
-        job_context->full_scan_candidate_files.emplace_back(log_file,
-            immutable_db_options_.db_log_dir);
+        job_context->full_scan_candidate_files.emplace_back(
+            log_file, immutable_db_options_.db_log_dir);
      }
    }
  }
@@ -260,14 +259,14 @@ void DBImpl::DeleteObsoleteFileImpl(int job_id, const std::string& fname,
                                    const std::string& path_to_sync,
                                    FileType type, uint64_t number) {
  Status file_deletion_status;
-  if (type == kTableFile) {
+  if (type == kTableFile || type == kLogFile) {
    file_deletion_status =
-        DeleteSSTFile(&immutable_db_options_, fname, path_to_sync);
+        DeleteDBFile(&immutable_db_options_, fname, path_to_sync);
  } else {
    file_deletion_status = env_->DeleteFile(fname);
  }
  TEST_SYNC_POINT_CALLBACK("DBImpl::DeleteObsoleteFileImpl:AfterDeletion",
-      &file_deletion_status);
+                           &file_deletion_status);
  if (file_deletion_status.ok()) {
    ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
                    "[JOB %d] Delete %s type=%d #%" PRIu64 " -- %s\n", job_id,
@@ -322,7 +321,8 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
  const char* kDumbDbName = "";
  for (auto& file : state.sst_delete_files) {
    candidate_files.emplace_back(
-        MakeTableFileName(kDumbDbName, file.metadata->fd.GetNumber()), file.path);
+        MakeTableFileName(kDumbDbName, file.metadata->fd.GetNumber()),
+        file.path);
    if (file.metadata->table_reader_handle) {
      table_cache_->Release(file.metadata->table_reader_handle);
    }
@@ -332,7 +332,7 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
  for (auto file_num : state.log_delete_files) {
    if (file_num > 0) {
      candidate_files.emplace_back(LogFileName(kDumbDbName, file_num),
-          immutable_db_options_.wal_dir);
+                                   immutable_db_options_.wal_dir);
    }
  }
  for (const auto& filename : state.manifest_delete_files) {
@@ -465,13 +465,12 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
    } else {
      dir_to_sync =
          (type == kLogFile) ? immutable_db_options_.wal_dir : dbname_;
-      fname = dir_to_sync
-            + (
-                (!dir_to_sync.empty() && dir_to_sync.back() == '/') ||
-                (!to_delete.empty() && to_delete.front() == '/')
-                ? "" : "/"
-              )
-            + to_delete;
+      fname = dir_to_sync +
+              ((!dir_to_sync.empty() && dir_to_sync.back() == '/') ||
+                       (!to_delete.empty() && to_delete.front() == '/')
+                   ? ""
+                   : "/") +
+              to_delete;
    }

 #ifndef ROCKSDB_LITE
@@ -482,6 +481,11 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
    }
 #endif  // !ROCKSDB_LITE

+    for (const auto w : state.logs_to_free) {
+      // TODO: maybe check the return value of Close.
+      w->Close();
+    }
+
    Status file_deletion_status;
    if (schedule_only) {
      InstrumentedMutexLock guard_lock(&mutex_);
@@ -23,8 +23,7 @@
 #include "util/sync_point.h"

 namespace rocksdb {
-Options SanitizeOptions(const std::string& dbname,
-                        const Options& src) {
+Options SanitizeOptions(const std::string& dbname, const Options& src) {
  auto db_options = SanitizeOptions(dbname, DBOptions(src));
  ImmutableDBOptions immutable_db_options(db_options);
  auto cf_options =
@@ -42,6 +41,8 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
      max_max_open_files = 0x400000;
    }
    ClipToRange(&result.max_open_files, 20, max_max_open_files);
+    TEST_SYNC_POINT_CALLBACK("SanitizeOptions::AfterChangeMaxOpenFiles",
+                             &result.max_open_files);
  }

  if (result.info_log == nullptr) {
@@ -56,10 +57,9 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
    result.write_buffer_manager.reset(
        new WriteBufferManager(result.db_write_buffer_size));
  }
-  auto bg_job_limits = DBImpl::GetBGJobLimits(result.max_background_flushes,
-                                              result.max_background_compactions,
-                                              result.max_background_jobs,
-                                              true /* parallelize_compactions */);
+  auto bg_job_limits = DBImpl::GetBGJobLimits(
+      result.max_background_flushes, result.max_background_compactions,
+      result.max_background_jobs, true /* parallelize_compactions */);
  result.env->IncBackgroundThreadsIfNeeded(bg_job_limits.max_compactions,
                                           Env::Priority::LOW);
  result.env->IncBackgroundThreadsIfNeeded(bg_job_limits.max_flushes,
@@ -107,14 +107,12 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
    result.db_paths.emplace_back(dbname, std::numeric_limits<uint64_t>::max());
  }

-  if (result.use_direct_reads &&
-      result.compaction_readahead_size == 0) {
+  if (result.use_direct_reads && result.compaction_readahead_size == 0) {
    TEST_SYNC_POINT_CALLBACK("SanitizeOptions:direct_io", nullptr);
    result.compaction_readahead_size = 1024 * 1024 * 2;
  }

-  if (result.compaction_readahead_size > 0 ||
-      result.use_direct_reads) {
+  if (result.compaction_readahead_size > 0 || result.use_direct_reads) {
    result.new_table_reader_for_compaction_inputs = true;
  }

@@ -178,7 +176,7 @@ static Status ValidateOptions(
      return s;
    }

-    if (cfd.options.ttl > 0 || cfd.options.compaction_options_fifo.ttl > 0) {
+    if (cfd.options.ttl > 0) {
      if (db_options.max_open_files != -1) {
        return Status::NotSupported(
            "TTL is only supported when files are always "
@@ -190,6 +188,20 @@ static Status ValidateOptions(
            "TTL is only supported in Block-Based Table format. ");
      }
    }
+
+    if (cfd.options.periodic_compaction_seconds > 0) {
+      if (db_options.max_open_files != -1) {
+        return Status::NotSupported(
+            "Periodic Compaction is only supported when files are always "
+            "kept open (set max_open_files = -1). ");
+      }
+      if (cfd.options.table_factory->Name() !=
+          BlockBasedTableFactory().Name()) {
+        return Status::NotSupported(
+            "Periodic Compaction is only supported in "
+            "Block-Based Table format. ");
+      }
+    }
  }

  if (db_options.db_paths.size() > 4) {
@@ -218,7 +230,7 @@ static Status ValidateOptions(

  return Status::OK();
 }
-} // namespace
+}  // namespace
 Status DBImpl::NewDB() {
  VersionEdit new_db;
  new_db.SetLogNumber(0);
@@ -239,7 +251,7 @@ Status DBImpl::NewDB() {
    file->SetPreallocationBlockSize(
        immutable_db_options_.manifest_preallocation_size);
    std::unique_ptr<WritableFileWriter> file_writer(new WritableFileWriter(
-        std::move(file), manifest, env_options, nullptr /* stats */,
+        std::move(file), manifest, env_options, env_, nullptr /* stats */,
        immutable_db_options_.listeners));
    log::Writer log(std::move(file_writer), 0, false);
    std::string record;
@@ -258,9 +270,8 @@ Status DBImpl::NewDB() {
  return s;
 }

-Status DBImpl::CreateAndNewDirectory(
-    Env* env, const std::string& dirname,
-    std::unique_ptr<Directory>* directory) {
+Status DBImpl::CreateAndNewDirectory(Env* env, const std::string& dirname,
+                                     std::unique_ptr<Directory>* directory) {
  // We call CreateDirIfMissing() as the directory may already exist (if we
  // are reopening a DB), when this happens we don't want creating the
  // directory to cause an error. However, we need to check if creating the
@@ -341,8 +352,8 @@ Status DBImpl::Recover(
      }
    } else if (s.ok()) {
      if (immutable_db_options_.error_if_exists) {
-        return Status::InvalidArgument(
-            dbname_, "exists (error_if_exists is true)");
+        return Status::InvalidArgument(dbname_,
+                                       "exists (error_if_exists is true)");
      }
    } else {
      // Unexpected error reading file
@@ -424,7 +435,10 @@ Status DBImpl::Recover(
    // produced by an older version of rocksdb.
    std::vector<std::string> filenames;
    s = env_->GetChildren(immutable_db_options_.wal_dir, &filenames);
-    if (!s.ok()) {
+    if (s.IsNotFound()) {
+      return Status::InvalidArgument("wal_dir not found",
+                                     immutable_db_options_.wal_dir);
+    } else if (!s.ok()) {
      return s;
    }

@@ -479,6 +493,28 @@ Status DBImpl::Recover(
    }
  }

+  if (read_only) {
+    // If we are opening as read-only, we need to update options_file_number_
+    // to reflect the most recent OPTIONS file. It does not matter for regular
+    // read-write db instance because options_file_number_ will later be
+    // updated to versions_->NewFileNumber() in RenameTempFileToOptionsFile.
+    std::vector<std::string> file_names;
+    if (s.ok()) {
+      s = env_->GetChildren(GetName(), &file_names);
+    }
+    if (s.ok()) {
+      uint64_t number = 0;
+      uint64_t options_file_number = 0;
+      FileType type;
+      for (const auto& fname : file_names) {
+        if (ParseFileName(fname, &number, &type) && type == kOptionsFile) {
+          options_file_number = std::max(number, options_file_number);
+        }
+      }
+      versions_->options_file_number_ = options_file_number;
+    }
+  }
+
  return s;
 }

@@ -490,7 +526,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
    Logger* info_log;
    const char* fname;
    Status* status;  // nullptr if immutable_db_options_.paranoid_checks==false
-    virtual void Corruption(size_t bytes, const Status& s) override {
+    void Corruption(size_t bytes, const Status& s) override {
      ROCKS_LOG_WARN(info_log, "%s%s: dropping %d bytes; %s",
                     (this->status == nullptr ? "(ignoring error) " : ""),
                     fname, static_cast<int>(bytes), s.ToString().c_str());
@@ -527,10 +563,9 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
    std::map<std::string, uint32_t> cf_name_id_map;
    std::map<uint32_t, uint64_t> cf_lognumber_map;
    for (auto cfd : *versions_->GetColumnFamilySet()) {
-      cf_name_id_map.insert(
-        std::make_pair(cfd->GetName(), cfd->GetID()));
+      cf_name_id_map.insert(std::make_pair(cfd->GetName(), cfd->GetID()));
      cf_lognumber_map.insert(
-        std::make_pair(cfd->GetID(), cfd->GetLogNumber()));
+          std::make_pair(cfd->GetID(), cfd->GetLogNumber()));
    }

    immutable_db_options_.wal_filter->ColumnFamilyLogNumberMap(cf_lognumber_map,
@@ -559,7 +594,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,

    ROCKS_LOG_INFO(immutable_db_options_.info_log,
                   "Recovering log #%" PRIu64 " mode %d", log_number,
-                   immutable_db_options_.wal_recovery_mode);
+                   static_cast<int>(immutable_db_options_.wal_recovery_mode));
    auto logFileDropped = [this, &fname]() {
      uint64_t bytes;
      if (env_->GetFileSize(fname, &bytes).ok()) {
@@ -608,8 +643,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
    // to be skipped instead of propagating bad information (like overly
    // large sequence numbers).
    log::Reader reader(immutable_db_options_.info_log, std::move(file_reader),
-                       &reporter, true /*checksum*/, log_number,
-                       false /* retry_after_eof */);
+                       &reporter, true /*checksum*/, log_number);

    // Determine if we should tolerate incomplete records at the tail end of the
    // Read all the records and add to a memtable
@@ -703,7 +737,8 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
                " mode %d log filter %s returned "
                "more records (%d) than original (%d) which is not allowed. "
                "Aborting recovery.",
-                log_number, immutable_db_options_.wal_recovery_mode,
+                log_number,
+                static_cast<int>(immutable_db_options_.wal_recovery_mode),
                immutable_db_options_.wal_filter->Name(), new_count,
                original_count);
            status = Status::NotSupported(
@@ -880,8 +915,8 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
      // VersionSet::next_file_number_ always to be strictly greater than any
      // log number
      versions_->MarkFileNumberUsed(max_log_number + 1);
-      status = versions_->LogAndApply(
-          cfd, *cfd->GetLatestMutableCFOptions(), edit, &mutex_);
+      status = versions_->LogAndApply(cfd, *cfd->GetLatestMutableCFOptions(),
+                                      edit, &mutex_);
      if (!status.ok()) {
        // Recovery failed
        break;
@@ -994,15 +1029,21 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
      if (use_custom_gc_ && snapshot_checker == nullptr) {
        snapshot_checker = DisableGCSnapshotChecker::Instance();
      }
+      std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
+          range_del_iters;
+      auto range_del_iter =
+          mem->NewRangeTombstoneIterator(ro, kMaxSequenceNumber);
+      if (range_del_iter != nullptr) {
+        range_del_iters.emplace_back(range_del_iter);
+      }
      s = BuildTable(
          dbname_, env_, *cfd->ioptions(), mutable_cf_options,
          env_options_for_compaction_, cfd->table_cache(), iter.get(),
-          std::unique_ptr<InternalIterator>(
-              mem->NewRangeTombstoneIterator(ro, versions_->LastSequence())),
-          &meta, cfd->internal_comparator(),
+          std::move(range_del_iters), &meta, cfd->internal_comparator(),
          cfd->int_tbl_prop_collector_factories(), cfd->GetID(), cfd->GetName(),
          snapshot_seqs, earliest_write_conflict_snapshot, snapshot_checker,
          GetCompressionFlush(*cfd->ioptions(), mutable_cf_options),
+          mutable_cf_options.sample_for_compression,
          cfd->ioptions()->compression_opts, paranoid_file_checks,
          cfd->internal_stats(), TableFileCreationReason::kRecovery,
          &event_logger_, job_id, Env::IO_HIGH, nullptr /* table_properties */,
@@ -1032,9 +1073,9 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
  stats.micros = env_->NowMicros() - start_micros;
  stats.bytes_written = meta.fd.GetFileSize();
  stats.num_output_files = 1;
-  cfd->internal_stats()->AddCompactionStats(level, stats);
-  cfd->internal_stats()->AddCFStats(
-      InternalStats::BYTES_FLUSHED, meta.fd.GetFileSize());
+  cfd->internal_stats()->AddCompactionStats(level, Env::Priority::USER, stats);
+  cfd->internal_stats()->AddCFStats(InternalStats::BYTES_FLUSHED,
+                                    meta.fd.GetFileSize());
  RecordTick(stats_, COMPACT_WRITE_BYTES, meta.fd.GetFileSize());
  return s;
 }
@@ -1065,6 +1106,45 @@ Status DB::Open(const DBOptions& db_options, const std::string& dbname,
                      !kSeqPerBatch, kBatchPerTxn);
 }

+Status DBImpl::CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
+                         size_t preallocate_block_size, log::Writer** new_log) {
+  Status s;
+  std::unique_ptr<WritableFile> lfile;
+
+  DBOptions db_options =
+      BuildDBOptions(immutable_db_options_, mutable_db_options_);
+  EnvOptions opt_env_options =
+      env_->OptimizeForLogWrite(env_options_, db_options);
+  std::string log_fname =
+      LogFileName(immutable_db_options_.wal_dir, log_file_num);
+
+  if (recycle_log_number) {
+    ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                   "reusing log %" PRIu64 " from recycle list\n",
+                   recycle_log_number);
+    std::string old_log_fname =
+        LogFileName(immutable_db_options_.wal_dir, recycle_log_number);
+    s = env_->ReuseWritableFile(log_fname, old_log_fname, &lfile,
+                                opt_env_options);
+  } else {
+    s = NewWritableFile(env_, log_fname, &lfile, opt_env_options);
+  }
+
+  if (s.ok()) {
+    lfile->SetWriteLifeTimeHint(CalculateWALWriteHint());
+    lfile->SetPreallocationBlockSize(preallocate_block_size);
+
+    const auto& listeners = immutable_db_options_.listeners;
+    std::unique_ptr<WritableFileWriter> file_writer(
+        new WritableFileWriter(std::move(lfile), log_fname, opt_env_options,
+                               env_, nullptr /* stats */, listeners));
+    *new_log = new log::Writer(std::move(file_writer), log_file_num,
+                               immutable_db_options_.recycle_log_file_num > 0,
+                               immutable_db_options_.manual_wal_flush);
+  }
+  return s;
+}
+
 Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
                    const std::vector<ColumnFamilyDescriptor>& column_families,
                    std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
@@ -1125,40 +1205,23 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
    return s;
  }
  impl->mutex_.Lock();
-  auto write_hint = impl->CalculateWALWriteHint();
  // Handles create_if_missing, error_if_exists
  s = impl->Recover(column_families);
  if (s.ok()) {
    uint64_t new_log_number = impl->versions_->NewFileNumber();
-    std::unique_ptr<WritableFile> lfile;
-    EnvOptions soptions(db_options);
-    EnvOptions opt_env_options =
-        impl->immutable_db_options_.env->OptimizeForLogWrite(
-            soptions, BuildDBOptions(impl->immutable_db_options_,
-                                     impl->mutable_db_options_));
-    std::string log_fname =
-        LogFileName(impl->immutable_db_options_.wal_dir, new_log_number);
-    s = NewWritableFile(impl->immutable_db_options_.env, log_fname, &lfile,
-                        opt_env_options);
+    log::Writer* new_log = nullptr;
+    const size_t preallocate_block_size =
+        impl->GetWalPreallocateBlockSize(max_write_buffer_size);
+    s = impl->CreateWAL(new_log_number, 0 /*recycle_log_number*/,
+                        preallocate_block_size, &new_log);
    if (s.ok()) {
-      lfile->SetWriteLifeTimeHint(write_hint);
-      lfile->SetPreallocationBlockSize(
-          impl->GetWalPreallocateBlockSize(max_write_buffer_size));
-      {
-        InstrumentedMutexLock wl(&impl->log_write_mutex_);
-        impl->logfile_number_ = new_log_number;
-        const auto& listeners = impl->immutable_db_options_.listeners;
-        std::unique_ptr<WritableFileWriter> file_writer(
-            new WritableFileWriter(std::move(lfile), log_fname, opt_env_options,
-                                   nullptr /* stats */, listeners));
-        impl->logs_.emplace_back(
-            new_log_number,
-            new log::Writer(
-                std::move(file_writer), new_log_number,
-                impl->immutable_db_options_.recycle_log_file_num > 0,
-                impl->immutable_db_options_.manual_wal_flush));
-      }
+      InstrumentedMutexLock wl(&impl->log_write_mutex_);
+      impl->logfile_number_ = new_log_number;
+      assert(new_log != nullptr);
+      impl->logs_.emplace_back(new_log_number, new_log);
+    }

+    if (s.ok()) {
      // set column family handles
      for (auto cf : column_families) {
        auto cfd =
@@ -1227,7 +1290,8 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
          !cfd->mem()->IsMergeOperatorSupported()) {
        s = Status::InvalidArgument(
            "The memtable of column family %s does not support merge operator "
-            "its options.merge_operator is non-null", cfd->GetName().c_str());
+            "its options.merge_operator is non-null",
+            cfd->GetName().c_str());
      }
      if (!s.ok()) {
        break;
@@ -1289,7 +1353,8 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
 #endif  // !ROCKSDB_LITE

  if (s.ok()) {
-    ROCKS_LOG_INFO(impl->immutable_db_options_.info_log, "DB pointer %p", impl);
+    ROCKS_LOG_HEADER(impl->immutable_db_options_.info_log, "DB pointer %p",
+                     impl);
    LogFlush(impl->immutable_db_options_.info_log);
    assert(impl->TEST_WALBufferIsEmpty());
    // If the assert above fails then we need to FlushWAL before returning
@@ -9,7 +9,6 @@
 #include "db/db_impl.h"
 #include "db/db_iter.h"
 #include "db/merge_context.h"
-#include "db/range_del_aggregator.h"
 #include "monitoring/perf_context_imp.h"

 namespace rocksdb {
@@ -61,7 +60,7 @@ Status DBImplReadOnly::Get(const ReadOptions& read_options,
  RecordTick(stats_, NUMBER_KEYS_READ);
  size_t size = pinnable_val->size();
  RecordTick(stats_, BYTES_READ, size);
-  MeasureTime(stats_, BYTES_PER_READ, size);
+  RecordInHistogram(stats_, BYTES_PER_READ, size);
  PERF_COUNTER_ADD(get_read_bytes, size);
  return s;
 }
@@ -7,9 +7,9 @@

 #ifndef ROCKSDB_LITE

-#include "db/db_impl.h"
-#include <vector>
 #include <string>
+#include <vector>
+#include "db/db_impl.h"

 namespace rocksdb {

@@ -77,8 +77,8 @@ class DBImplReadOnly : public DBImpl {
      ColumnFamilyHandle* /*column_family*/,
      const std::vector<std::string>& /*input_file_names*/,
      const int /*output_level*/, const int /*output_path_id*/ = -1,
-      std::vector<std::string>* const /*output_file_names*/ = nullptr
-      ) override {
+      std::vector<std::string>* const /*output_file_names*/ = nullptr,
+      CompactionJobInfo* /*compaction_job_info*/ = nullptr) override {
    return Status::NotSupported("Not supported operation in read only mode.");
  }

@@ -89,10 +89,11 @@ class DBImplReadOnly : public DBImpl {
  virtual Status EnableFileDeletions(bool /*force*/) override {
    return Status::NotSupported("Not supported operation in read only mode.");
  }
-  virtual Status GetLiveFiles(std::vector<std::string>&,
-                              uint64_t* /*manifest_file_size*/,
-                              bool /*flush_memtable*/ = true) override {
-    return Status::NotSupported("Not supported operation in read only mode.");
+  virtual Status GetLiveFiles(std::vector<std::string>& ret,
+                              uint64_t* manifest_file_size,
+                              bool /*flush_memtable*/) override {
+    return DBImpl::GetLiveFiles(ret, manifest_file_size,
+                                false /* flush_memtable */);
  }

  using DBImpl::Flush;
@@ -121,6 +122,6 @@ class DBImplReadOnly : public DBImpl {
  DBImplReadOnly(const DBImplReadOnly&);
  void operator=(const DBImplReadOnly&);
 };
-}
+}  // namespace rocksdb

 #endif  // !ROCKSDB_LITE
@@ -0,0 +1,524 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/db_impl_secondary.h"
+
+#ifndef __STDC_FORMAT_MACROS
+#define __STDC_FORMAT_MACROS
+#endif
+#include <inttypes.h>
+
+#include "db/db_iter.h"
+#include "db/merge_context.h"
+#include "monitoring/perf_context_imp.h"
+#include "util/auto_roll_logger.h"
+
+namespace rocksdb {
+
+#ifndef ROCKSDB_LITE
+
+DBImplSecondary::DBImplSecondary(const DBOptions& db_options,
+                                 const std::string& dbname)
+    : DBImpl(db_options, dbname) {
+  ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                 "Opening the db in secondary mode");
+  LogFlush(immutable_db_options_.info_log);
+}
+
+DBImplSecondary::~DBImplSecondary() {}
+
+Status DBImplSecondary::Recover(
+    const std::vector<ColumnFamilyDescriptor>& column_families,
+    bool /*readonly*/, bool /*error_if_log_file_exist*/,
+    bool /*error_if_data_exists_in_logs*/) {
+  mutex_.AssertHeld();
+
+  Status s;
+  s = static_cast<ReactiveVersionSet*>(versions_.get())
+          ->Recover(column_families, &manifest_reader_, &manifest_reporter_,
+                    &manifest_reader_status_);
+  if (!s.ok()) {
+    return s;
+  }
+  if (immutable_db_options_.paranoid_checks && s.ok()) {
+    s = CheckConsistency();
+  }
+  // Initial max_total_in_memory_state_ before recovery logs.
+  max_total_in_memory_state_ = 0;
+  for (auto cfd : *versions_->GetColumnFamilySet()) {
+    auto* mutable_cf_options = cfd->GetLatestMutableCFOptions();
+    max_total_in_memory_state_ += mutable_cf_options->write_buffer_size *
+                                  mutable_cf_options->max_write_buffer_number;
+  }
+  if (s.ok()) {
+    default_cf_handle_ = new ColumnFamilyHandleImpl(
+        versions_->GetColumnFamilySet()->GetDefault(), this, &mutex_);
+    default_cf_internal_stats_ = default_cf_handle_->cfd()->internal_stats();
+    single_column_family_mode_ =
+        versions_->GetColumnFamilySet()->NumberOfColumnFamilies() == 1;
+
+    // Recover from all newer log files than the ones named in the
+    // descriptor.
+    std::vector<std::string> filenames;
+    s = env_->GetChildren(immutable_db_options_.wal_dir, &filenames);
+    if (s.IsNotFound()) {
+      return Status::InvalidArgument("Failed to open wal_dir",
+                                     immutable_db_options_.wal_dir);
+    } else if (!s.ok()) {
+      return s;
+    }
+
+    std::vector<uint64_t> logs;
+    // if log_readers_ is non-empty, it means we have applied all logs with log
+    // numbers smaller than the smallest log in log_readers_, so there is no
+    // need to pass these logs to RecoverLogFiles
+    uint64_t log_number_min = 0;
+    if (log_readers_.size() > 0) {
+      log_number_min = log_readers_.begin()->first;
+    }
+    for (size_t i = 0; i < filenames.size(); i++) {
+      uint64_t number;
+      FileType type;
+      if (ParseFileName(filenames[i], &number, &type) && type == kLogFile &&
+          number >= log_number_min) {
+        logs.push_back(number);
+      }
+    }
+
+    if (!logs.empty()) {
+      // Recover in the order in which the logs were generated
+      std::sort(logs.begin(), logs.end());
+      SequenceNumber next_sequence(kMaxSequenceNumber);
+      s = RecoverLogFiles(logs, &next_sequence, true /*read_only*/);
+    }
+  }
+
+  // TODO: update options_file_number_ needed?
+
+  return s;
+}
+
+// try to find log reader using log_number from log_readers_ map, initialize
+// if it doesn't exist
+Status DBImplSecondary::MaybeInitLogReader(
+    uint64_t log_number, log::FragmentBufferedReader** log_reader) {
+  auto iter = log_readers_.find(log_number);
+  // make sure the log file is still present
+  if (iter == log_readers_.end() ||
+      iter->second->reader_->GetLogNumber() != log_number) {
+    // delete the obsolete log reader if log number mismatch
+    if (iter != log_readers_.end()) {
+      log_readers_.erase(iter);
+    }
+    // initialize log reader from log_number
+    // TODO: min_log_number_to_keep_2pc check needed?
+    // Open the log file
+    std::string fname = LogFileName(immutable_db_options_.wal_dir, log_number);
+    ROCKS_LOG_INFO(immutable_db_options_.info_log,
+                   "Recovering log #%" PRIu64 " mode %d", log_number,
+                   static_cast<int>(immutable_db_options_.wal_recovery_mode));
+
+    std::unique_ptr<SequentialFileReader> file_reader;
+    {
+      std::unique_ptr<SequentialFile> file;
+      Status status = env_->NewSequentialFile(
+          fname, &file, env_->OptimizeForLogRead(env_options_));
+      if (!status.ok()) {
+        *log_reader = nullptr;
+        return status;
+      }
+      file_reader.reset(new SequentialFileReader(std::move(file), fname));
+    }
+
+    // Create the log reader.
+    LogReaderContainer* log_reader_container = new LogReaderContainer(
+        env_, immutable_db_options_.info_log, std::move(fname),
+        std::move(file_reader), log_number);
+    log_readers_.insert(std::make_pair(
+        log_number, std::unique_ptr<LogReaderContainer>(log_reader_container)));
+  }
+  iter = log_readers_.find(log_number);
+  assert(iter != log_readers_.end());
+  *log_reader = iter->second->reader_;
+  return Status::OK();
+}
+
+// After manifest recovery, replay WALs and refresh log_readers_ if necessary
+// REQUIRES: log_numbers are sorted in ascending order
+Status DBImplSecondary::RecoverLogFiles(
+    const std::vector<uint64_t>& log_numbers, SequenceNumber* next_sequence,
+    bool /*read_only*/) {
+  mutex_.AssertHeld();
+  Status status;
+  for (auto log_number : log_numbers) {
+    log::FragmentBufferedReader* reader = nullptr;
+    status = MaybeInitLogReader(log_number, &reader);
+    if (!status.ok()) {
+      return status;
+    }
+    assert(reader != nullptr);
+  }
+  for (auto log_number : log_numbers) {
+    auto it  = log_readers_.find(log_number);
+    assert(it != log_readers_.end());
+    log::FragmentBufferedReader* reader = it->second->reader_;
+    // Manually update the file number allocation counter in VersionSet.
+    versions_->MarkFileNumberUsed(log_number);
+
+    // Determine if we should tolerate incomplete records at the tail end of the
+    // Read all the records and add to a memtable
+    std::string scratch;
+    Slice record;
+    WriteBatch batch;
+
+    while (reader->ReadRecord(&record, &scratch,
+                              immutable_db_options_.wal_recovery_mode) &&
+           status.ok()) {
+      if (record.size() < WriteBatchInternal::kHeader) {
+        reader->GetReporter()->Corruption(
+            record.size(), Status::Corruption("log record too small"));
+        continue;
+      }
+      WriteBatchInternal::SetContents(&batch, record);
+      // do not check sequence number because user may toggle disableWAL
+      // between writes which breaks sequence number continuity guarantee
+
+      // 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 --
+      // we just ignore the update.
+      // That's why we set ignore missing column families to true
+      // passing null flush_scheduler will disable memtable flushing which is
+      // needed for secondary instances
+      bool has_valid_writes = false;
+      status = WriteBatchInternal::InsertInto(
+          &batch, column_family_memtables_.get(), nullptr /* flush_scheduler */,
+          true, log_number, this, false /* concurrent_memtable_writes */,
+          next_sequence, &has_valid_writes, seq_per_batch_, batch_per_txn_);
+      if (!status.ok()) {
+        // We are treating this as a failure while reading since we read valid
+        // blocks that do not form coherent data
+        reader->GetReporter()->Corruption(record.size(), status);
+        continue;
+      }
+    }
+
+    if (!status.ok()) {
+      return status;
+    }
+
+    auto last_sequence = *next_sequence - 1;
+    if ((*next_sequence != kMaxSequenceNumber) &&
+        (versions_->LastSequence() <= last_sequence)) {
+      versions_->SetLastAllocatedSequence(last_sequence);
+      versions_->SetLastPublishedSequence(last_sequence);
+      versions_->SetLastSequence(last_sequence);
+    }
+  }
+  // remove logreaders from map after successfully recovering the WAL
+  if (log_readers_.size() > 1) {
+    auto eraseIter = log_readers_.begin();
+    std::advance(eraseIter, log_readers_.size() - 1);
+    log_readers_.erase(log_readers_.begin(), eraseIter);
+  }
+  return status;
+}
+
+// Implementation of the DB interface
+Status DBImplSecondary::Get(const ReadOptions& read_options,
+                            ColumnFamilyHandle* column_family, const Slice& key,
+                            PinnableSlice* value) {
+  return GetImpl(read_options, column_family, key, value);
+}
+
+Status DBImplSecondary::GetImpl(const ReadOptions& read_options,
+                                ColumnFamilyHandle* column_family,
+                                const Slice& key, PinnableSlice* pinnable_val) {
+  assert(pinnable_val != nullptr);
+  PERF_CPU_TIMER_GUARD(get_cpu_nanos, env_);
+  StopWatch sw(env_, stats_, DB_GET);
+  PERF_TIMER_GUARD(get_snapshot_time);
+
+  auto cfh = static_cast<ColumnFamilyHandleImpl*>(column_family);
+  ColumnFamilyData* cfd = cfh->cfd();
+  if (tracer_) {
+    InstrumentedMutexLock lock(&trace_mutex_);
+    if (tracer_) {
+      tracer_->Get(column_family, key);
+    }
+  }
+  // Acquire SuperVersion
+  SuperVersion* super_version = GetAndRefSuperVersion(cfd);
+  SequenceNumber snapshot = versions_->LastSequence();
+  MergeContext merge_context;
+  SequenceNumber max_covering_tombstone_seq = 0;
+  Status s;
+  LookupKey lkey(key, snapshot);
+  PERF_TIMER_STOP(get_snapshot_time);
+
+  bool done = false;
+  if (super_version->mem->Get(lkey, pinnable_val->GetSelf(), &s, &merge_context,
+                              &max_covering_tombstone_seq, read_options)) {
+    done = true;
+    pinnable_val->PinSelf();
+    RecordTick(stats_, MEMTABLE_HIT);
+  } else if ((s.ok() || s.IsMergeInProgress()) &&
+             super_version->imm->Get(
+                 lkey, pinnable_val->GetSelf(), &s, &merge_context,
+                 &max_covering_tombstone_seq, read_options)) {
+    done = true;
+    pinnable_val->PinSelf();
+    RecordTick(stats_, MEMTABLE_HIT);
+  }
+  if (!done && !s.ok() && !s.IsMergeInProgress()) {
+    ReturnAndCleanupSuperVersion(cfd, super_version);
+    return s;
+  }
+  if (!done) {
+    PERF_TIMER_GUARD(get_from_output_files_time);
+    super_version->current->Get(read_options, lkey, pinnable_val, &s,
+                                &merge_context, &max_covering_tombstone_seq);
+    RecordTick(stats_, MEMTABLE_MISS);
+  }
+  {
+    PERF_TIMER_GUARD(get_post_process_time);
+    ReturnAndCleanupSuperVersion(cfd, super_version);
+    RecordTick(stats_, NUMBER_KEYS_READ);
+    size_t size = pinnable_val->size();
+    RecordTick(stats_, BYTES_READ, size);
+    RecordTimeToHistogram(stats_, BYTES_PER_READ, size);
+    PERF_COUNTER_ADD(get_read_bytes, size);
+  }
+  return s;
+}
+
+Iterator* DBImplSecondary::NewIterator(const ReadOptions& read_options,
+                                       ColumnFamilyHandle* column_family) {
+  if (read_options.managed) {
+    return NewErrorIterator(
+        Status::NotSupported("Managed iterator is not supported anymore."));
+  }
+  if (read_options.read_tier == kPersistedTier) {
+    return NewErrorIterator(Status::NotSupported(
+        "ReadTier::kPersistedData is not yet supported in iterators."));
+  }
+  Iterator* result = nullptr;
+  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfd = cfh->cfd();
+  ReadCallback* read_callback = nullptr;  // No read callback provided.
+  if (read_options.tailing) {
+    return NewErrorIterator(Status::NotSupported(
+        "tailing iterator not supported in secondary mode"));
+  } else if (read_options.snapshot != nullptr) {
+    // TODO (yanqin) support snapshot.
+    return NewErrorIterator(
+        Status::NotSupported("snapshot not supported in secondary mode"));
+  } else {
+    auto snapshot = versions_->LastSequence();
+    result = NewIteratorImpl(read_options, cfd, snapshot, read_callback);
+  }
+  return result;
+}
+
+ArenaWrappedDBIter* DBImplSecondary::NewIteratorImpl(
+    const ReadOptions& read_options, ColumnFamilyData* cfd,
+    SequenceNumber snapshot, ReadCallback* read_callback) {
+  assert(nullptr != cfd);
+  SuperVersion* super_version = cfd->GetReferencedSuperVersion(&mutex_);
+  auto db_iter = NewArenaWrappedDbIterator(
+      env_, read_options, *cfd->ioptions(), super_version->mutable_cf_options,
+      snapshot,
+      super_version->mutable_cf_options.max_sequential_skip_in_iterations,
+      super_version->version_number, read_callback);
+  auto internal_iter =
+      NewInternalIterator(read_options, cfd, super_version, db_iter->GetArena(),
+                          db_iter->GetRangeDelAggregator(), snapshot);
+  db_iter->SetIterUnderDBIter(internal_iter);
+  return db_iter;
+}
+
+Status DBImplSecondary::NewIterators(
+    const ReadOptions& read_options,
+    const std::vector<ColumnFamilyHandle*>& column_families,
+    std::vector<Iterator*>* iterators) {
+  if (read_options.managed) {
+    return Status::NotSupported("Managed iterator is not supported anymore.");
+  }
+  if (read_options.read_tier == kPersistedTier) {
+    return Status::NotSupported(
+        "ReadTier::kPersistedData is not yet supported in iterators.");
+  }
+  ReadCallback* read_callback = nullptr;  // No read callback provided.
+  if (iterators == nullptr) {
+    return Status::InvalidArgument("iterators not allowed to be nullptr");
+  }
+  iterators->clear();
+  iterators->reserve(column_families.size());
+  if (read_options.tailing) {
+    return Status::NotSupported(
+        "tailing iterator not supported in secondary mode");
+  } else if (read_options.snapshot != nullptr) {
+    // TODO (yanqin) support snapshot.
+    return Status::NotSupported("snapshot not supported in secondary mode");
+  } else {
+    SequenceNumber read_seq = versions_->LastSequence();
+    for (auto cfh : column_families) {
+      ColumnFamilyData* cfd = static_cast<ColumnFamilyHandleImpl*>(cfh)->cfd();
+      iterators->push_back(
+          NewIteratorImpl(read_options, cfd, read_seq, read_callback));
+    }
+  }
+  return Status::OK();
+}
+
+Status DBImplSecondary::TryCatchUpWithPrimary() {
+  assert(versions_.get() != nullptr);
+  assert(manifest_reader_.get() != nullptr);
+  Status s;
+  std::unordered_set<ColumnFamilyData*> cfds_changed;
+  InstrumentedMutexLock lock_guard(&mutex_);
+  s = static_cast<ReactiveVersionSet*>(versions_.get())
+          ->ReadAndApply(&mutex_, &manifest_reader_, &cfds_changed);
+  if (s.ok()) {
+    SuperVersionContext sv_context(true /* create_superversion */);
+    for (auto cfd : cfds_changed) {
+      sv_context.NewSuperVersion();
+      cfd->InstallSuperVersion(&sv_context, &mutex_);
+    }
+    sv_context.Clean();
+  }
+  return s;
+}
+
+Status DB::OpenAsSecondary(const Options& options, const std::string& dbname,
+                           const std::string& secondary_path, DB** dbptr) {
+  *dbptr = nullptr;
+
+  DBOptions db_options(options);
+  ColumnFamilyOptions cf_options(options);
+  std::vector<ColumnFamilyDescriptor> column_families;
+  column_families.emplace_back(kDefaultColumnFamilyName, cf_options);
+  std::vector<ColumnFamilyHandle*> handles;
+
+  Status s = DB::OpenAsSecondary(db_options, dbname, secondary_path,
+                                 column_families, &handles, dbptr);
+  if (s.ok()) {
+    assert(handles.size() == 1);
+    delete handles[0];
+  }
+  return s;
+}
+
+Status DB::OpenAsSecondary(
+    const DBOptions& db_options, const std::string& dbname,
+    const std::string& secondary_path,
+    const std::vector<ColumnFamilyDescriptor>& column_families,
+    std::vector<ColumnFamilyHandle*>* handles, DB** dbptr) {
+  *dbptr = nullptr;
+  if (db_options.max_open_files != -1) {
+    // TODO (yanqin) maybe support max_open_files != -1 by creating hard links
+    // on SST files so that db secondary can still have access to old SSTs
+    // while primary instance may delete original.
+    return Status::InvalidArgument("require max_open_files to be -1");
+  }
+
+  DBOptions tmp_opts(db_options);
+  if (nullptr == tmp_opts.info_log) {
+    Env* env = tmp_opts.env;
+    assert(env != nullptr);
+    std::string secondary_abs_path;
+    env->GetAbsolutePath(secondary_path, &secondary_abs_path);
+    std::string fname = InfoLogFileName(secondary_path, secondary_abs_path,
+                                        tmp_opts.db_log_dir);
+
+    env->CreateDirIfMissing(secondary_path);
+    if (tmp_opts.log_file_time_to_roll > 0 || tmp_opts.max_log_file_size > 0) {
+      AutoRollLogger* result = new AutoRollLogger(
+          env, secondary_path, tmp_opts.db_log_dir, tmp_opts.max_log_file_size,
+          tmp_opts.log_file_time_to_roll, tmp_opts.info_log_level);
+      Status s = result->GetStatus();
+      if (!s.ok()) {
+        delete result;
+      } else {
+        tmp_opts.info_log.reset(result);
+      }
+    }
+    if (nullptr == tmp_opts.info_log) {
+      env->RenameFile(
+          fname, OldInfoLogFileName(secondary_path, env->NowMicros(),
+                                    secondary_abs_path, tmp_opts.db_log_dir));
+      Status s = env->NewLogger(fname, &(tmp_opts.info_log));
+      if (tmp_opts.info_log != nullptr) {
+        tmp_opts.info_log->SetInfoLogLevel(tmp_opts.info_log_level);
+      }
+    }
+  }
+
+  assert(tmp_opts.info_log != nullptr);
+
+  handles->clear();
+  DBImplSecondary* impl = new DBImplSecondary(tmp_opts, dbname);
+  impl->versions_.reset(new ReactiveVersionSet(
+      dbname, &impl->immutable_db_options_, impl->env_options_,
+      impl->table_cache_.get(), impl->write_buffer_manager_,
+      &impl->write_controller_));
+  impl->column_family_memtables_.reset(
+      new ColumnFamilyMemTablesImpl(impl->versions_->GetColumnFamilySet()));
+  impl->mutex_.Lock();
+  Status s = impl->Recover(column_families, true, false, false);
+  if (s.ok()) {
+    for (auto cf : column_families) {
+      auto cfd =
+          impl->versions_->GetColumnFamilySet()->GetColumnFamily(cf.name);
+      if (nullptr == cfd) {
+        s = Status::InvalidArgument("Column family not found: ", cf.name);
+        break;
+      }
+      handles->push_back(new ColumnFamilyHandleImpl(cfd, impl, &impl->mutex_));
+    }
+  }
+  SuperVersionContext sv_context(true /* create_superversion */);
+  if (s.ok()) {
+    for (auto cfd : *impl->versions_->GetColumnFamilySet()) {
+      sv_context.NewSuperVersion();
+      cfd->InstallSuperVersion(&sv_context, &impl->mutex_);
+    }
+  }
+  impl->mutex_.Unlock();
+  sv_context.Clean();
+  if (s.ok()) {
+    *dbptr = impl;
+    for (auto h : *handles) {
+      impl->NewThreadStatusCfInfo(
+          reinterpret_cast<ColumnFamilyHandleImpl*>(h)->cfd());
+    }
+  } else {
+    for (auto h : *handles) {
+      delete h;
+    }
+    handles->clear();
+    delete impl;
+  }
+  return s;
+}
+#else   // !ROCKSDB_LITE
+
+Status DB::OpenAsSecondary(const Options& /*options*/,
+                           const std::string& /*name*/,
+                           const std::string& /*secondary_path*/,
+                           DB** /*dbptr*/) {
+  return Status::NotSupported("Not supported in ROCKSDB_LITE.");
+}
+
+Status DB::OpenAsSecondary(
+    const DBOptions& /*db_options*/, const std::string& /*dbname*/,
+    const std::string& /*secondary_path*/,
+    const std::vector<ColumnFamilyDescriptor>& /*column_families*/,
+    std::vector<ColumnFamilyHandle*>* /*handles*/, DB** /*dbptr*/) {
+  return Status::NotSupported("Not supported in ROCKSDB_LITE.");
+}
+#endif  // !ROCKSDB_LITE
+
+}  // namespace rocksdb
@@ -0,0 +1,212 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#pragma once
+
+#ifndef ROCKSDB_LITE
+
+#include <string>
+#include <vector>
+#include "db/db_impl.h"
+
+namespace rocksdb {
+
+class LogReaderContainer {
+ public:
+  LogReaderContainer()
+      : reader_(nullptr), reporter_(nullptr), status_(nullptr) {}
+  LogReaderContainer(Env* env, std::shared_ptr<Logger> info_log,
+                     std::string fname,
+                     std::unique_ptr<SequentialFileReader>&& file_reader,
+                     uint64_t log_number) {
+    LogReporter* reporter = new LogReporter();
+    status_ = new Status();
+    reporter->env = env;
+    reporter->info_log = info_log.get();
+    reporter->fname = std::move(fname);
+    reporter->status = status_;
+    reporter_ = reporter;
+    // We intentially make log::Reader do checksumming even if
+    // paranoid_checks==false so that corruptions cause entire commits
+    // to be skipped instead of propagating bad information (like overly
+    // large sequence numbers).
+    reader_ = new log::FragmentBufferedReader(info_log, std::move(file_reader),
+                                              reporter, true /*checksum*/,
+                                              log_number);
+  }
+  log::FragmentBufferedReader* reader_;
+  log::Reader::Reporter* reporter_;
+  Status* status_;
+  ~LogReaderContainer() {
+    delete reader_;
+    delete reporter_;
+    delete status_;
+  }
+ private:
+  struct LogReporter : public log::Reader::Reporter {
+    Env* env;
+    Logger* info_log;
+    std::string fname;
+    Status* status;  // nullptr if immutable_db_options_.paranoid_checks==false
+    void Corruption(size_t bytes, const Status& s) override {
+      ROCKS_LOG_WARN(info_log, "%s%s: dropping %d bytes; %s",
+                     (this->status == nullptr ? "(ignoring error) " : ""),
+                     fname.c_str(), static_cast<int>(bytes),
+                     s.ToString().c_str());
+      if (this->status != nullptr && this->status->ok()) {
+        *this->status = s;
+      }
+    }
+  };
+};
+
+class DBImplSecondary : public DBImpl {
+ public:
+  DBImplSecondary(const DBOptions& options, const std::string& dbname);
+  ~DBImplSecondary() override;
+
+  Status Recover(const std::vector<ColumnFamilyDescriptor>& column_families,
+                 bool read_only, bool error_if_log_file_exist,
+                 bool error_if_data_exists_in_logs) override;
+
+  // Implementations of the DB interface
+  using DB::Get;
+  Status Get(const ReadOptions& options, ColumnFamilyHandle* column_family,
+             const Slice& key, PinnableSlice* value) override;
+
+  Status GetImpl(const ReadOptions& options, ColumnFamilyHandle* column_family,
+                 const Slice& key, PinnableSlice* value);
+
+  using DBImpl::NewIterator;
+  Iterator* NewIterator(const ReadOptions&,
+                        ColumnFamilyHandle* column_family) override;
+
+  ArenaWrappedDBIter* NewIteratorImpl(const ReadOptions& read_options,
+                                      ColumnFamilyData* cfd,
+                                      SequenceNumber snapshot,
+                                      ReadCallback* read_callback);
+
+  Status NewIterators(const ReadOptions& options,
+                      const std::vector<ColumnFamilyHandle*>& column_families,
+                      std::vector<Iterator*>* iterators) override;
+
+  using DBImpl::Put;
+  Status Put(const WriteOptions& /*options*/,
+             ColumnFamilyHandle* /*column_family*/, const Slice& /*key*/,
+             const Slice& /*value*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::Merge;
+  Status Merge(const WriteOptions& /*options*/,
+               ColumnFamilyHandle* /*column_family*/, const Slice& /*key*/,
+               const Slice& /*value*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::Delete;
+  Status Delete(const WriteOptions& /*options*/,
+                ColumnFamilyHandle* /*column_family*/,
+                const Slice& /*key*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::SingleDelete;
+  Status SingleDelete(const WriteOptions& /*options*/,
+                      ColumnFamilyHandle* /*column_family*/,
+                      const Slice& /*key*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  Status Write(const WriteOptions& /*options*/,
+               WriteBatch* /*updates*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::CompactRange;
+  Status CompactRange(const CompactRangeOptions& /*options*/,
+                      ColumnFamilyHandle* /*column_family*/,
+                      const Slice* /*begin*/, const Slice* /*end*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::CompactFiles;
+  Status CompactFiles(
+      const CompactionOptions& /*compact_options*/,
+      ColumnFamilyHandle* /*column_family*/,
+      const std::vector<std::string>& /*input_file_names*/,
+      const int /*output_level*/, const int /*output_path_id*/ = -1,
+      std::vector<std::string>* const /*output_file_names*/ = nullptr,
+      CompactionJobInfo* /*compaction_job_info*/ = nullptr) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  Status DisableFileDeletions() override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  Status EnableFileDeletions(bool /*force*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  Status GetLiveFiles(std::vector<std::string>&,
+                      uint64_t* /*manifest_file_size*/,
+                      bool /*flush_memtable*/ = true) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::Flush;
+  Status Flush(const FlushOptions& /*options*/,
+               ColumnFamilyHandle* /*column_family*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DBImpl::SyncWAL;
+  Status SyncWAL() override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  using DB::IngestExternalFile;
+  Status IngestExternalFile(
+      ColumnFamilyHandle* /*column_family*/,
+      const std::vector<std::string>& /*external_files*/,
+      const IngestExternalFileOptions& /*ingestion_options*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
+  // Try to catch up with the primary by reading as much as possible from the
+  // log files until there is nothing more to read or encounters an error. If
+  // the amount of information in the log files to process is huge, this
+  // method can take long time due to all the I/O and CPU costs.
+  Status TryCatchUpWithPrimary() override;
+
+  Status MaybeInitLogReader(uint64_t log_number,
+                            log::FragmentBufferedReader** log_reader);
+
+ private:
+  friend class DB;
+
+  // No copying allowed
+  DBImplSecondary(const DBImplSecondary&);
+  void operator=(const DBImplSecondary&);
+
+  using DBImpl::Recover;
+
+  Status RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
+                         SequenceNumber* next_sequence,
+                         bool read_only) override;
+
+  std::unique_ptr<log::FragmentBufferedReader> manifest_reader_;
+  std::unique_ptr<log::Reader::Reporter> manifest_reporter_;
+  std::unique_ptr<Status> manifest_reader_status_;
+
+  // cache log readers for each log number, used for continue WAL replay
+  // after recovery
+  std::map<uint64_t, std::unique_ptr<LogReaderContainer>> log_readers_;
+};
+
+}  // namespace rocksdb
+
+#endif  // !ROCKSDB_LITE
@@ -146,17 +146,6 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,

    if (write_thread_.CompleteParallelMemTableWriter(&w)) {
      // we're responsible for exit batch group
-      for (auto* writer : *(w.write_group)) {
-        if (!writer->CallbackFailed() && writer->pre_release_callback) {
-          assert(writer->sequence != kMaxSequenceNumber);
-          Status ws = writer->pre_release_callback->Callback(writer->sequence,
-                                                             disable_memtable);
-          if (!ws.ok()) {
-            status = ws;
-            break;
-          }
-        }
-      }
      // TODO(myabandeh): propagate status to write_group
      auto last_sequence = w.write_group->last_sequence;
      versions_->SetLastSequence(last_sequence);
@@ -279,7 +268,7 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
                        concurrent_update);
      RecordTick(stats_, WRITE_DONE_BY_OTHER, write_done_by_other);
    }
-    MeasureTime(stats_, BYTES_PER_WRITE, total_byte_size);
+    RecordInHistogram(stats_, BYTES_PER_WRITE, total_byte_size);

    if (write_options.disableWAL) {
      has_unpersisted_data_.store(true, std::memory_order_relaxed);
@@ -309,6 +298,35 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
    const SequenceNumber current_sequence = last_sequence + 1;
    last_sequence += seq_inc;

+    // PreReleaseCallback is called after WAL write and before memtable write
+    if (status.ok()) {
+      SequenceNumber next_sequence = current_sequence;
+      // Note: the logic for advancing seq here must be consistent with the
+      // logic in WriteBatchInternal::InsertInto(write_group...) as well as
+      // with WriteBatchInternal::InsertInto(write_batch...) that is called on
+      // the merged batch during recovery from the WAL.
+      for (auto* writer : write_group) {
+        if (writer->CallbackFailed()) {
+          continue;
+        }
+        writer->sequence = next_sequence;
+        if (writer->pre_release_callback) {
+          Status ws = writer->pre_release_callback->Callback(
+              writer->sequence, disable_memtable, writer->log_used);
+          if (!ws.ok()) {
+            status = ws;
+            break;
+          }
+        }
+        if (seq_per_batch_) {
+          assert(writer->batch_cnt);
+          next_sequence += writer->batch_cnt;
+        } else if (writer->ShouldWriteToMemtable()) {
+          next_sequence += WriteBatchInternal::Count(writer->batch);
+        }
+      }
+    }
+
    if (status.ok()) {
      PERF_TIMER_GUARD(write_memtable_time);

@@ -320,23 +338,6 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
            0 /*recovery_log_number*/, this, parallel, seq_per_batch_,
            batch_per_txn_);
      } else {
-        SequenceNumber next_sequence = current_sequence;
-        // Note: the logic for advancing seq here must be consistent with the
-        // logic in WriteBatchInternal::InsertInto(write_group...) as well as
-        // with WriteBatchInternal::InsertInto(write_batch...) that is called on
-        // the merged batch during recovery from the WAL.
-        for (auto* writer : write_group) {
-          if (writer->CallbackFailed()) {
-            continue;
-          }
-          writer->sequence = next_sequence;
-          if (seq_per_batch_) {
-            assert(writer->batch_cnt);
-            next_sequence += writer->batch_cnt;
-          } else if (writer->ShouldWriteToMemtable()) {
-            next_sequence += WriteBatchInternal::Count(writer->batch);
-          }
-        }
        write_group.last_sequence = last_sequence;
        write_thread_.LaunchParallelMemTableWriters(&write_group);
        in_parallel_group = true;
@@ -388,17 +389,8 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
  }
  if (should_exit_batch_group) {
    if (status.ok()) {
-      for (auto* writer : write_group) {
-        if (!writer->CallbackFailed() && writer->pre_release_callback) {
-          assert(writer->sequence != kMaxSequenceNumber);
-          Status ws = writer->pre_release_callback->Callback(writer->sequence,
-                                                             disable_memtable);
-          if (!ws.ok()) {
-            status = ws;
-            break;
-          }
-        }
-      }
+      // Note: if we are to resume after non-OK statuses we need to revisit how
+      // we reacts to non-OK statuses here.
      versions_->SetLastSequence(last_sequence);
    }
    MemTableInsertStatusCheck(w.status);
@@ -471,11 +463,11 @@ Status DBImpl::PipelinedWriteImpl(const WriteOptions& write_options,
    RecordTick(stats_, NUMBER_KEYS_WRITTEN, total_count);
    stats->AddDBStats(InternalStats::BYTES_WRITTEN, total_byte_size);
    RecordTick(stats_, BYTES_WRITTEN, total_byte_size);
-    MeasureTime(stats_, BYTES_PER_WRITE, total_byte_size);
+    RecordInHistogram(stats_, BYTES_PER_WRITE, total_byte_size);

    PERF_TIMER_STOP(write_pre_and_post_process_time);

-    if (w.ShouldWriteToWAL()) {
+    if (w.status.ok() && !write_options.disableWAL) {
      PERF_TIMER_GUARD(write_wal_time);
      stats->AddDBStats(InternalStats::WRITE_DONE_BY_SELF, 1);
      RecordTick(stats_, WRITE_DONE_BY_SELF, 1);
@@ -504,7 +496,7 @@ Status DBImpl::PipelinedWriteImpl(const WriteOptions& write_options,
  WriteThread::WriteGroup memtable_write_group;
  if (w.state == WriteThread::STATE_MEMTABLE_WRITER_LEADER) {
    PERF_TIMER_GUARD(write_memtable_time);
-    assert(w.status.ok());
+    assert(w.ShouldWriteToMemtable());
    write_thread_.EnterAsMemTableWriter(&w, &memtable_write_group);
    if (memtable_write_group.size > 1 &&
        immutable_db_options_.allow_concurrent_memtable_write) {
@@ -602,7 +594,7 @@ Status DBImpl::WriteImplWALOnly(const WriteOptions& write_options,
                      concurrent_update);
    RecordTick(stats_, WRITE_DONE_BY_OTHER, write_done_by_other);
  }
-  MeasureTime(stats_, BYTES_PER_WRITE, total_byte_size);
+  RecordInHistogram(stats_, BYTES_PER_WRITE, total_byte_size);

  PERF_TIMER_STOP(write_pre_and_post_process_time);

@@ -657,8 +649,8 @@ Status DBImpl::WriteImplWALOnly(const WriteOptions& write_options,
      if (!writer->CallbackFailed() && writer->pre_release_callback) {
        assert(writer->sequence != kMaxSequenceNumber);
        const bool DISABLE_MEMTABLE = true;
-        Status ws = writer->pre_release_callback->Callback(writer->sequence,
-                                                           DISABLE_MEMTABLE);
+        Status ws = writer->pre_release_callback->Callback(
+            writer->sequence, DISABLE_MEMTABLE, writer->log_used);
        if (!ws.ok()) {
          status = ws;
          break;
@@ -876,7 +868,7 @@ Status DBImpl::WriteToWAL(const WriteThread::WriteGroup& write_group,
  status = WriteToWAL(*merged_batch, log_writer, log_used, &log_size);
  if (to_be_cached_state) {
    cached_recoverable_state_ = *to_be_cached_state;
-      cached_recoverable_state_empty_ = false;
+    cached_recoverable_state_empty_ = false;
  }

  if (status.ok() && need_log_sync) {
@@ -952,7 +944,7 @@ Status DBImpl::ConcurrentWriteToWAL(const WriteThread::WriteGroup& write_group,
  status = WriteToWAL(*merged_batch, log_writer, log_used, &log_size);
  if (to_be_cached_state) {
    cached_recoverable_state_ = *to_be_cached_state;
-      cached_recoverable_state_empty_ = false;
+    cached_recoverable_state_empty_ = false;
  }
  log_write_mutex_.Unlock();

@@ -1001,8 +993,9 @@ Status DBImpl::WriteRecoverableState() {
      const bool DISABLE_MEMTABLE = true;
      for (uint64_t sub_batch_seq = seq + 1;
           sub_batch_seq < next_seq && status.ok(); sub_batch_seq++) {
+        uint64_t const no_log_num = 0;
        status = recoverable_state_pre_release_callback_->Callback(
-            sub_batch_seq, !DISABLE_MEMTABLE);
+            sub_batch_seq, !DISABLE_MEMTABLE, no_log_num);
      }
    }
    if (status.ok()) {
@@ -1048,22 +1041,22 @@ Status DBImpl::SwitchWAL(WriteContext* write_context) {
  auto oldest_alive_log = alive_log_files_.begin()->number;
  bool flush_wont_release_oldest_log = false;
  if (allow_2pc()) {
-    auto oldest_log_with_uncommited_prep =
+    auto oldest_log_with_uncommitted_prep =
        logs_with_prep_tracker_.FindMinLogContainingOutstandingPrep();

-    assert(oldest_log_with_uncommited_prep == 0 ||
-           oldest_log_with_uncommited_prep >= oldest_alive_log);
-    if (oldest_log_with_uncommited_prep > 0 &&
-        oldest_log_with_uncommited_prep == oldest_alive_log) {
+    assert(oldest_log_with_uncommitted_prep == 0 ||
+           oldest_log_with_uncommitted_prep >= oldest_alive_log);
+    if (oldest_log_with_uncommitted_prep > 0 &&
+        oldest_log_with_uncommitted_prep == oldest_alive_log) {
      if (unable_to_release_oldest_log_) {
        // we already attempted to flush all column families dependent on
-        // the oldest alive log but the log still contained uncommited
+        // the oldest alive log but the log still contained uncommitted
        // transactions so there is still nothing that we can do.
        return status;
      } else {
        ROCKS_LOG_WARN(
            immutable_db_options_.info_log,
-            "Unable to release oldest log due to uncommited transaction");
+            "Unable to release oldest log due to uncommitted transaction");
        unable_to_release_oldest_log_ = true;
        flush_wont_release_oldest_log = true;
      }
@@ -1072,16 +1065,17 @@ Status DBImpl::SwitchWAL(WriteContext* write_context) {
  if (!flush_wont_release_oldest_log) {
    // we only mark this log as getting flushed if we have successfully
    // flushed all data in this log. If this log contains outstanding prepared
-    // transactions then we cannot flush this log until those transactions are commited.
+    // transactions then we cannot flush this log until those transactions are
+    // commited.
    unable_to_release_oldest_log_ = false;
    alive_log_files_.begin()->getting_flushed = true;
  }

-  ROCKS_LOG_INFO(immutable_db_options_.info_log,
-                 "Flushing all column families with data in WAL number %" PRIu64
-                 ". Total log size is %" PRIu64
-                 " while max_total_wal_size is %" PRIu64,
-                 oldest_alive_log, total_log_size_.load(), GetMaxTotalWalSize());
+  ROCKS_LOG_INFO(
+      immutable_db_options_.info_log,
+      "Flushing all column families with data in WAL number %" PRIu64
+      ". Total log size is %" PRIu64 " while max_total_wal_size is %" PRIu64,
+      oldest_alive_log, total_log_size_.load(), GetMaxTotalWalSize());
  // no need to refcount because drop is happening in write thread, so can't
  // happen while we're in the write thread
  autovector<ColumnFamilyData*> cfds;
@@ -1133,7 +1127,7 @@ Status DBImpl::HandleWriteBufferFull(WriteContext* write_context) {
  ROCKS_LOG_INFO(
      immutable_db_options_.info_log,
      "Flushing column family with largest mem table size. Write buffer is "
-      "using %" PRIu64 " bytes out of a total of %" PRIu64 ".",
+      "using %" ROCKSDB_PRIszt " bytes out of a total of %" ROCKSDB_PRIszt ".",
      write_buffer_manager_->memory_usage(),
      write_buffer_manager_->buffer_size());
  // no need to refcount because drop is happening in write thread, so can't
@@ -1424,58 +1418,19 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
  // Log this later after lock release. It may be outdated, e.g., if background
  // flush happens before logging, but that should be ok.
  int num_imm_unflushed = cfd->imm()->NumNotFlushed();
-  DBOptions db_options =
-      BuildDBOptions(immutable_db_options_, mutable_db_options_);
  const auto preallocate_block_size =
-    GetWalPreallocateBlockSize(mutable_cf_options.write_buffer_size);
-  auto write_hint = CalculateWALWriteHint();
+      GetWalPreallocateBlockSize(mutable_cf_options.write_buffer_size);
  mutex_.Unlock();
-  {
-    std::string log_fname =
-        LogFileName(immutable_db_options_.wal_dir, new_log_number);
-    if (creating_new_log) {
-      EnvOptions opt_env_opt =
-          env_->OptimizeForLogWrite(env_options_, db_options);
-      if (recycle_log_number) {
-        ROCKS_LOG_INFO(immutable_db_options_.info_log,
-                       "reusing log %" PRIu64 " from recycle list\n",
-                       recycle_log_number);
-        std::string old_log_fname =
-            LogFileName(immutable_db_options_.wal_dir, recycle_log_number);
-        s = env_->ReuseWritableFile(log_fname, old_log_fname, &lfile,
-                                    opt_env_opt);
-      } else {
-        s = NewWritableFile(env_, log_fname, &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.
-
-        // use preallocate_block_size instead
-        // of calling GetWalPreallocateBlockSize()
-        lfile->SetPreallocationBlockSize(preallocate_block_size);
-        lfile->SetWriteLifeTimeHint(write_hint);
-        std::unique_ptr<WritableFileWriter> file_writer(new WritableFileWriter(
-            std::move(lfile), log_fname, opt_env_opt, nullptr /* stats */,
-            immutable_db_options_.listeners));
-        new_log = new log::Writer(
-            std::move(file_writer), new_log_number,
-            immutable_db_options_.recycle_log_file_num > 0, manual_wal_flush_);
-      }
-    }
-
-    if (s.ok()) {
-      SequenceNumber seq = versions_->LastSequence();
-      new_mem = cfd->ConstructNewMemtable(mutable_cf_options, seq);
-      context->superversion_context.NewSuperVersion();
-    }
-
-#ifndef ROCKSDB_LITE
-    // PLEASE NOTE: We assume that there are no failable operations
-    // after lock is acquired below since we are already notifying
-    // client about mem table becoming immutable.
-    NotifyOnMemTableSealed(cfd, memtable_info);
-#endif //ROCKSDB_LITE
+  if (creating_new_log) {
+    // TODO: Write buffer size passed in should be max of all CF's instead
+    // of mutable_cf_options.write_buffer_size.
+    s = CreateWAL(new_log_number, recycle_log_number, preallocate_block_size,
+                  &new_log);
+  }
+  if (s.ok()) {
+    SequenceNumber seq = versions_->LastSequence();
+    new_mem = cfd->ConstructNewMemtable(mutable_cf_options, seq);
+    context->superversion_context.NewSuperVersion();
  }
  ROCKS_LOG_INFO(immutable_db_options_.info_log,
                 "[%s] New memtable created with log file: #%" PRIu64
@@ -1484,10 +1439,7 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
  mutex_.Lock();
  if (s.ok() && creating_new_log) {
    log_write_mutex_.Lock();
-    logfile_number_ = new_log_number;
    assert(new_log != nullptr);
-    log_empty_ = true;
-    log_dir_synced_ = false;
    if (!logs_.empty()) {
      // Alway flush the buffer of the last log before switching to a new one
      log::Writer* cur_log_writer = logs_.back().writer;
@@ -1495,21 +1447,41 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
      if (!s.ok()) {
        ROCKS_LOG_WARN(immutable_db_options_.info_log,
                       "[%s] Failed to switch from #%" PRIu64 " to #%" PRIu64
-                       "  WAL file -- %s\n",
+                       "  WAL file\n",
                       cfd->GetName().c_str(), cur_log_writer->get_log_number(),
                       new_log_number);
      }
    }
-    logs_.emplace_back(logfile_number_, new_log);
-    alive_log_files_.push_back(LogFileNumberSize(logfile_number_));
+    if (s.ok()) {
+      logfile_number_ = new_log_number;
+      log_empty_ = true;
+      log_dir_synced_ = false;
+      logs_.emplace_back(logfile_number_, new_log);
+      alive_log_files_.push_back(LogFileNumberSize(logfile_number_));
+    }
    log_write_mutex_.Unlock();
  }

  if (!s.ok()) {
    // how do we fail if we're not creating new log?
    assert(creating_new_log);
-    assert(!new_mem);
-    assert(!new_log);
+    if (new_mem) {
+      delete new_mem;
+    }
+    if (new_log) {
+      delete new_log;
+    }
+    SuperVersion* new_superversion =
+        context->superversion_context.new_superversion.release();
+    if (new_superversion != nullptr) {
+      delete new_superversion;
+    }
+    // We may have lost data from the WritableFileBuffer in-memory buffer for
+    // the current log, so treat it as a fatal error and set bg_error
+    error_handler_.SetBGError(s, BackgroundErrorReason::kMemTable);
+    // Read back bg_error in order to get the right severity
+    s = error_handler_.GetBGError();
+
    if (two_write_queues_) {
      nonmem_write_thread_.ExitUnbatched(&nonmem_w);
    }
@@ -1536,6 +1508,13 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
  cfd->SetMemtable(new_mem);
  InstallSuperVersionAndScheduleWork(cfd, &context->superversion_context,
                                     mutable_cf_options);
+#ifndef ROCKSDB_LITE
+  mutex_.Unlock();
+  // Notify client that memtable is sealed, now that we have successfully
+  // installed a new memtable
+  NotifyOnMemTableSealed(cfd, memtable_info);
+  mutex_.Lock();
+#endif  // ROCKSDB_LITE
  if (two_write_queues_) {
    nonmem_write_thread_.ExitUnbatched(&nonmem_w);
  }
@@ -1544,13 +1523,13 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {

 size_t DBImpl::GetWalPreallocateBlockSize(uint64_t write_buffer_size) const {
  mutex_.AssertHeld();
-  size_t bsize = static_cast<size_t>(
-    write_buffer_size / 10 + write_buffer_size);
+  size_t bsize =
+      static_cast<size_t>(write_buffer_size / 10 + write_buffer_size);
  // Some users might set very high write_buffer_size and rely on
  // max_total_wal_size or other parameters to control the WAL size.
  if (mutable_db_options_.max_total_wal_size > 0) {
-    bsize = std::min<size_t>(bsize, static_cast<size_t>(
-      mutable_db_options_.max_total_wal_size));
+    bsize = std::min<size_t>(
+        bsize, static_cast<size_t>(mutable_db_options_.max_total_wal_size));
  }
  if (immutable_db_options_.db_write_buffer_size > 0) {
    bsize = std::min<size_t>(bsize, immutable_db_options_.db_write_buffer_size);
@@ -22,12 +22,14 @@
 #include "rocksdb/merge_operator.h"
 #include "rocksdb/options.h"
 #include "table/internal_iterator.h"
+#include "table/iterator_wrapper.h"
 #include "util/arena.h"
 #include "util/filename.h"
 #include "util/logging.h"
 #include "util/mutexlock.h"
 #include "util/string_util.h"
 #include "util/trace_replay.h"
+#include "util/user_comparator_wrapper.h"

 namespace rocksdb {

@@ -117,29 +119,29 @@ class DBIter final: public Iterator {
         uint64_t max_sequential_skip_in_iterations,
         ReadCallback* read_callback, DBImpl* db_impl, ColumnFamilyData* cfd,
         bool allow_blob)
-      : arena_mode_(arena_mode),
-        env_(_env),
+      : env_(_env),
        logger_(cf_options.info_log),
        user_comparator_(cmp),
        merge_operator_(cf_options.merge_operator),
        iter_(iter),
+        read_callback_(read_callback),
        sequence_(s),
-        direction_(kForward),
-        valid_(false),
-        current_entry_is_merged_(false),
        statistics_(cf_options.statistics),
        num_internal_keys_skipped_(0),
        iterate_lower_bound_(read_options.iterate_lower_bound),
        iterate_upper_bound_(read_options.iterate_upper_bound),
+        direction_(kForward),
+        valid_(false),
+        current_entry_is_merged_(false),
        prefix_same_as_start_(read_options.prefix_same_as_start),
        pin_thru_lifetime_(read_options.pin_data),
        total_order_seek_(read_options.total_order_seek),
-        range_del_agg_(&cf_options.internal_comparator, s),
-        read_callback_(read_callback),
-        db_impl_(db_impl),
-        cfd_(cfd),
        allow_blob_(allow_blob),
        is_blob_(false),
+        arena_mode_(arena_mode),
+        range_del_agg_(&cf_options.internal_comparator, s),
+        db_impl_(db_impl),
+        cfd_(cfd),
        start_seqnum_(read_options.iter_start_seqnum) {
    RecordTick(statistics_, NO_ITERATOR_CREATED);
    prefix_extractor_ = mutable_cf_options.prefix_extractor.get();
@@ -148,11 +150,11 @@ class DBIter final: public Iterator {
    if (pin_thru_lifetime_) {
      pinned_iters_mgr_.StartPinning();
    }
-    if (iter_) {
-      iter_->SetPinnedItersMgr(&pinned_iters_mgr_);
+    if (iter_.iter()) {
+      iter_.iter()->SetPinnedItersMgr(&pinned_iters_mgr_);
    }
  }
-  virtual ~DBIter() {
+  ~DBIter() override {
    // Release pinned data if any
    if (pinned_iters_mgr_.PinningEnabled()) {
      pinned_iters_mgr_.ReleasePinnedData();
@@ -160,32 +162,27 @@ class DBIter final: public Iterator {
    RecordTick(statistics_, NO_ITERATOR_DELETED);
    ResetInternalKeysSkippedCounter();
    local_stats_.BumpGlobalStatistics(statistics_);
-    if (!arena_mode_) {
-      delete iter_;
-    } else {
-      iter_->~InternalIterator();
-    }
+    iter_.DeleteIter(arena_mode_);
  }
  virtual void SetIter(InternalIterator* iter) {
-    assert(iter_ == nullptr);
-    iter_ = iter;
-    iter_->SetPinnedItersMgr(&pinned_iters_mgr_);
+    assert(iter_.iter() == nullptr);
+    iter_.Set(iter);
+    iter_.iter()->SetPinnedItersMgr(&pinned_iters_mgr_);
  }
-  virtual RangeDelAggregatorV2* GetRangeDelAggregator() {
+  virtual ReadRangeDelAggregator* GetRangeDelAggregator() {
    return &range_del_agg_;
  }

-  virtual bool Valid() const override { return valid_; }
-  virtual Slice key() const override {
+  bool Valid() const override { return valid_; }
+  Slice key() const override {
    assert(valid_);
    if(start_seqnum_ > 0) {
      return saved_key_.GetInternalKey();
    } else {
      return saved_key_.GetUserKey();
    }
-
  }
-  virtual Slice value() const override {
+  Slice value() const override {
    assert(valid_);
    if (current_entry_is_merged_) {
      // If pinned_value_ is set then the result of merge operator is one of
@@ -194,12 +191,12 @@ class DBIter final: public Iterator {
    } else if (direction_ == kReverse) {
      return pinned_value_;
    } else {
-      return iter_->value();
+      return iter_.value();
    }
  }
-  virtual Status status() const override {
+  Status status() const override {
    if (status_.ok()) {
-      return iter_->status();
+      return iter_.status();
    } else {
      assert(!valid_);
      return status_;
@@ -210,14 +207,13 @@ class DBIter final: public Iterator {
    return is_blob_;
  }

-  virtual Status GetProperty(std::string prop_name,
-                             std::string* prop) override {
+  Status GetProperty(std::string prop_name, std::string* prop) override {
    if (prop == nullptr) {
      return Status::InvalidArgument("prop is nullptr");
    }
    if (prop_name == "rocksdb.iterator.super-version-number") {
      // First try to pass the value returned from inner iterator.
-      return iter_->GetProperty(prop_name, prop);
+      return iter_.iter()->GetProperty(prop_name, prop);
    } else if (prop_name == "rocksdb.iterator.is-key-pinned") {
      if (valid_) {
        *prop = (pin_thru_lifetime_ && saved_key_.IsKeyPinned()) ? "1" : "0";
@@ -232,14 +228,19 @@ class DBIter final: public Iterator {
    return Status::InvalidArgument("Unidentified property.");
  }

-  virtual void Next() override;
-  virtual void Prev() override;
-  virtual void Seek(const Slice& target) override;
-  virtual void SeekForPrev(const Slice& target) override;
-  virtual void SeekToFirst() override;
-  virtual void SeekToLast() override;
+  inline void Next() final override;
+  inline void Prev() final override;
+  inline void Seek(const Slice& target) final override;
+  inline void SeekForPrev(const Slice& target) final override;
+  inline void SeekToFirst() final override;
+  inline void SeekToLast() final override;
  Env* env() { return env_; }
-  void set_sequence(uint64_t s) { sequence_ = s; }
+  void set_sequence(uint64_t s) {
+    sequence_ = s;
+    if (read_callback_) {
+      read_callback_->Refresh(s);
+    }
+  }
  void set_valid(bool v) { valid_ = v; }

 private:
@@ -253,13 +254,13 @@ class DBIter final: public Iterator {
  bool FindValueForCurrentKeyUsingSeek();
  bool FindUserKeyBeforeSavedKey();
  inline bool FindNextUserEntry(bool skipping, bool prefix_check);
-  bool FindNextUserEntryInternal(bool skipping, bool prefix_check);
+  inline bool FindNextUserEntryInternal(bool skipping, bool prefix_check);
  bool ParseKey(ParsedInternalKey* key);
  bool MergeValuesNewToOld();

  void PrevInternal();
  bool TooManyInternalKeysSkipped(bool increment = true);
-  bool IsVisible(SequenceNumber sequence);
+  inline bool IsVisible(SequenceNumber sequence);

  // CanReseekToSkip() returns whether the iterator can use the optimization
  // where it reseek by sequence number to get the next key when there are too
@@ -267,12 +268,6 @@ class DBIter final: public Iterator {
  // sequence number does not guarantee that it is visible.
  inline bool CanReseekToSkip();

-  // MaxVisibleSequenceNumber() returns the maximum visible sequence number
-  // for this snapshot. This sequence number may be greater than snapshot
-  // seqno because uncommitted data written to DB for write unprepared will
-  // have a higher sequence number.
-  inline SequenceNumber MaxVisibleSequenceNumber();
-
  // Temporarily pin the blocks that we encounter until ReleaseTempPinnedData()
  // is called
  void TempPinData() {
@@ -306,15 +301,16 @@ class DBIter final: public Iterator {
  }

  const SliceTransform* prefix_extractor_;
-  bool arena_mode_;
  Env* const env_;
  Logger* logger_;
-  const Comparator* const user_comparator_;
+  UserComparatorWrapper user_comparator_;
  const MergeOperator* const merge_operator_;
-  InternalIterator* iter_;
+  IteratorWrapper iter_;
+  ReadCallback* read_callback_;
+  // Max visible sequence number. It is normally the snapshot seq unless we have
+  // uncommitted data in db as in WriteUnCommitted.
  SequenceNumber sequence_;

-  Status status_;
  IterKey saved_key_;
  // Reusable internal key data structure. This is only used inside one function
  // and should not be used across functions. Reusing this object can reduce
@@ -322,9 +318,6 @@ class DBIter final: public Iterator {
  ParsedInternalKey ikey_;
  std::string saved_value_;
  Slice pinned_value_;
-  Direction direction_;
-  bool valid_;
-  bool current_entry_is_merged_;
  // for prefix seek mode to support prev()
  Statistics* statistics_;
  uint64_t max_skip_;
@@ -332,23 +325,29 @@ class DBIter final: public Iterator {
  uint64_t num_internal_keys_skipped_;
  const Slice* iterate_lower_bound_;
  const Slice* iterate_upper_bound_;
+
  IterKey prefix_start_buf_;
+
+  Status status_;
  Slice prefix_start_key_;
+  Direction direction_;
+  bool valid_;
+  bool current_entry_is_merged_;
  const bool prefix_same_as_start_;
  // Means that we will pin all data blocks we read as long the Iterator
  // is not deleted, will be true if ReadOptions::pin_data is true
  const bool pin_thru_lifetime_;
  const bool total_order_seek_;
-  // List of operands for merge operator.
-  MergeContext merge_context_;
-  RangeDelAggregatorV2 range_del_agg_;
-  LocalStatistics local_stats_;
-  PinnedIteratorsManager pinned_iters_mgr_;
-  ReadCallback* read_callback_;
-  DBImpl* db_impl_;
-  ColumnFamilyData* cfd_;
  bool allow_blob_;
  bool is_blob_;
+  bool arena_mode_;
+  // List of operands for merge operator.
+  MergeContext merge_context_;
+  ReadRangeDelAggregator range_del_agg_;
+  LocalStatistics local_stats_;
+  PinnedIteratorsManager pinned_iters_mgr_;
+  DBImpl* db_impl_;
+  ColumnFamilyData* cfd_;
  // for diff snapshots we want the lower bound on the seqnum;
  // if this value > 0 iterator will return internal keys
  SequenceNumber start_seqnum_;
@@ -359,11 +358,11 @@ class DBIter final: public Iterator {
 };

 inline bool DBIter::ParseKey(ParsedInternalKey* ikey) {
-  if (!ParseInternalKey(iter_->key(), ikey)) {
+  if (!ParseInternalKey(iter_.key(), ikey)) {
    status_ = Status::Corruption("corrupted internal key in DBIter");
    valid_ = false;
    ROCKS_LOG_ERROR(logger_, "corrupted internal key in DBIter: %s",
-                    iter_->key().ToString(true).c_str());
+                    iter_.key().ToString(true).c_str());
    return false;
  } else {
    return true;
@@ -374,28 +373,30 @@ void DBIter::Next() {
  assert(valid_);
  assert(status_.ok());

+  PERF_CPU_TIMER_GUARD(iter_next_cpu_nanos, env_);
  // Release temporarily pinned blocks from last operation
  ReleaseTempPinnedData();
-  ResetInternalKeysSkippedCounter();
+  local_stats_.skip_count_ += num_internal_keys_skipped_;
+  local_stats_.skip_count_--;
+  num_internal_keys_skipped_ = 0;
  bool ok = true;
  if (direction_ == kReverse) {
    if (!ReverseToForward()) {
      ok = false;
    }
-  } else if (iter_->Valid() && !current_entry_is_merged_) {
+  } else if (!current_entry_is_merged_) {
    // If the current value is not a merge, the iter position is the
    // current key, which is already returned. We can safely issue a
    // Next() without checking the current key.
    // If the current key is a merge, very likely iter already points
    // to the next internal position.
-    iter_->Next();
+    assert(iter_.Valid());
+    iter_.Next();
    PERF_COUNTER_ADD(internal_key_skipped_count, 1);
  }

-  if (statistics_ != nullptr) {
-    local_stats_.next_count_++;
-  }
-  if (ok && iter_->Valid()) {
+  local_stats_.next_count_++;
+  if (ok && iter_.Valid()) {
    FindNextUserEntry(true /* skipping the current user key */,
                      prefix_same_as_start_);
  } else {
@@ -427,9 +428,9 @@ inline bool DBIter::FindNextUserEntry(bool skipping, bool prefix_check) {
 }

 // Actual implementation of DBIter::FindNextUserEntry()
-bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
+inline bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
  // Loop until we hit an acceptable entry to yield
-  assert(iter_->Valid());
+  assert(iter_.Valid());
  assert(status_.ok());
  assert(direction_ == kForward);
  current_entry_is_merged_ = false;
@@ -454,7 +455,7 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
    }

    if (iterate_upper_bound_ != nullptr &&
-        user_comparator_->Compare(ikey_.user_key, *iterate_upper_bound_) >= 0) {
+        user_comparator_.Compare(ikey_.user_key, *iterate_upper_bound_) >= 0) {
      break;
    }

@@ -469,8 +470,8 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
    }

    if (IsVisible(ikey_.sequence)) {
-      if (skipping && user_comparator_->Compare(ikey_.user_key,
-                                                saved_key_.GetUserKey()) <= 0) {
+      if (skipping && user_comparator_.Compare(ikey_.user_key,
+                                               saved_key_.GetUserKey()) <= 0) {
        num_skipped++;  // skip this entry
        PERF_COUNTER_ADD(internal_key_skipped_count, 1);
      } else {
@@ -491,8 +492,8 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
              return true;
            } else {
              saved_key_.SetUserKey(
-                ikey_.user_key,
-                !pin_thru_lifetime_ || !iter_->IsKeyPinned() /* copy */);
+                  ikey_.user_key, !pin_thru_lifetime_ ||
+                                      !iter_.iter()->IsKeyPinned() /* copy */);
              skipping = true;
              PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
            }
@@ -512,14 +513,16 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
              } else {
                // this key and all previous versions shouldn't be included,
                // skipping
-                saved_key_.SetUserKey(ikey_.user_key,
-                  !pin_thru_lifetime_ || !iter_->IsKeyPinned() /* copy */);
+                saved_key_.SetUserKey(
+                    ikey_.user_key,
+                    !pin_thru_lifetime_ ||
+                        !iter_.iter()->IsKeyPinned() /* copy */);
                skipping = true;
              }
            } else {
              saved_key_.SetUserKey(
-                  ikey_.user_key,
-                  !pin_thru_lifetime_ || !iter_->IsKeyPinned() /* copy */);
+                  ikey_.user_key, !pin_thru_lifetime_ ||
+                                      !iter_.iter()->IsKeyPinned() /* copy */);
              if (range_del_agg_.ShouldDelete(
                      ikey_, RangeDelPositioningMode::kForwardTraversal)) {
                // Arrange to skip all upcoming entries for this key since
@@ -549,7 +552,7 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
          case kTypeMerge:
            saved_key_.SetUserKey(
                ikey_.user_key,
-                !pin_thru_lifetime_ || !iter_->IsKeyPinned() /* copy */);
+                !pin_thru_lifetime_ || !iter_.iter()->IsKeyPinned() /* copy */);
            if (range_del_agg_.ShouldDelete(
                    ikey_, RangeDelPositioningMode::kForwardTraversal)) {
              // Arrange to skip all upcoming entries for this key since
@@ -577,13 +580,13 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
      // If this happens too many times in a row for the same user key, we want
      // to seek to the target sequence number.
      int cmp =
-          user_comparator_->Compare(ikey_.user_key, saved_key_.GetUserKey());
+          user_comparator_.Compare(ikey_.user_key, saved_key_.GetUserKey());
      if (cmp == 0 || (skipping && cmp <= 0)) {
        num_skipped++;
      } else {
        saved_key_.SetUserKey(
            ikey_.user_key,
-            !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+            !iter_.iter()->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
        skipping = false;
        num_skipped = 0;
      }
@@ -612,20 +615,20 @@ bool DBIter::FindNextUserEntryInternal(bool skipping, bool prefix_check) {
                          ParsedInternalKey(saved_key_.GetUserKey(), sequence_,
                                            kValueTypeForSeek));
      }
-      iter_->Seek(last_key);
+      iter_.Seek(last_key);
      RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
    } else {
-      iter_->Next();
+      iter_.Next();
    }
-  } while (iter_->Valid());
+  } while (iter_.Valid());

  valid_ = false;
-  return iter_->status().ok();
+  return iter_.status().ok();
 }

 // Merge values of the same user key starting from the current iter_ position
 // Scan from the newer entries to older entries.
-// PRE: iter_->key() points to the first merge type entry
+// PRE: iter_.key() points to the first merge type entry
 //      saved_key_ stores the user key
 // POST: saved_value_ has the merged value for the user key
 //       iter_ points to the next entry (or invalid)
@@ -641,19 +644,19 @@ bool DBIter::MergeValuesNewToOld() {
  TempPinData();
  merge_context_.Clear();
  // Start the merge process by pushing the first operand
-  merge_context_.PushOperand(iter_->value(),
-                             iter_->IsValuePinned() /* operand_pinned */);
+  merge_context_.PushOperand(
+      iter_.value(), iter_.iter()->IsValuePinned() /* operand_pinned */);
  TEST_SYNC_POINT("DBIter::MergeValuesNewToOld:PushedFirstOperand");

  ParsedInternalKey ikey;
  Status s;
-  for (iter_->Next(); iter_->Valid(); iter_->Next()) {
+  for (iter_.Next(); iter_.Valid(); iter_.Next()) {
    TEST_SYNC_POINT("DBIter::MergeValuesNewToOld:SteppedToNextOperand");
    if (!ParseKey(&ikey)) {
      return false;
    }

-    if (!user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
+    if (!user_comparator_.Equal(ikey.user_key, saved_key_.GetUserKey())) {
      // hit the next user key, stop right here
      break;
    } else if (kTypeDeletion == ikey.type || kTypeSingleDeletion == ikey.type ||
@@ -661,12 +664,12 @@ bool DBIter::MergeValuesNewToOld() {
                   ikey, RangeDelPositioningMode::kForwardTraversal)) {
      // hit a delete with the same user key, stop right here
      // iter_ is positioned after delete
-      iter_->Next();
+      iter_.Next();
      break;
    } else if (kTypeValue == ikey.type) {
      // hit a put, merge the put value with operands and store the
      // final result in saved_value_. We are done!
-      const Slice val = iter_->value();
+      const Slice val = iter_.value();
      s = MergeHelper::TimedFullMerge(
          merge_operator_, ikey.user_key, &val, merge_context_.GetOperands(),
          &saved_value_, logger_, statistics_, env_, &pinned_value_, true);
@@ -676,8 +679,8 @@ bool DBIter::MergeValuesNewToOld() {
        return false;
      }
      // iter_ is positioned after put
-      iter_->Next();
-      if (!iter_->status().ok()) {
+      iter_.Next();
+      if (!iter_.status().ok()) {
        valid_ = false;
        return false;
      }
@@ -685,8 +688,8 @@ bool DBIter::MergeValuesNewToOld() {
    } else if (kTypeMerge == ikey.type) {
      // hit a merge, add the value as an operand and run associative merge.
      // when complete, add result to operands and continue.
-      merge_context_.PushOperand(iter_->value(),
-                                 iter_->IsValuePinned() /* operand_pinned */);
+      merge_context_.PushOperand(
+          iter_.value(), iter_.iter()->IsValuePinned() /* operand_pinned */);
      PERF_COUNTER_ADD(internal_merge_count, 1);
    } else if (kTypeBlobIndex == ikey.type) {
      if (!allow_blob_) {
@@ -705,7 +708,7 @@ bool DBIter::MergeValuesNewToOld() {
    }
  }

-  if (!iter_->status().ok()) {
+  if (!iter_.status().ok()) {
    valid_ = false;
    return false;
  }
@@ -731,6 +734,8 @@ bool DBIter::MergeValuesNewToOld() {
 void DBIter::Prev() {
  assert(valid_);
  assert(status_.ok());
+
+  PERF_CPU_TIMER_GUARD(iter_prev_cpu_nanos, env_);
  ReleaseTempPinnedData();
  ResetInternalKeysSkippedCounter();
  bool ok = true;
@@ -752,33 +757,32 @@ void DBIter::Prev() {
 }

 bool DBIter::ReverseToForward() {
-  assert(iter_->status().ok());
+  assert(iter_.status().ok());

  // When moving backwards, iter_ is positioned on _previous_ key, which may
  // not exist or may have different prefix than the current key().
  // If that's the case, seek iter_ to current key.
-  if ((prefix_extractor_ != nullptr && !total_order_seek_) || !iter_->Valid()) {
+  if ((prefix_extractor_ != nullptr && !total_order_seek_) || !iter_.Valid()) {
    IterKey last_key;
    last_key.SetInternalKey(ParsedInternalKey(
        saved_key_.GetUserKey(), kMaxSequenceNumber, kValueTypeForSeek));
-    iter_->Seek(last_key.GetInternalKey());
+    iter_.Seek(last_key.GetInternalKey());
  }

  direction_ = kForward;
  // Skip keys less than the current key() (a.k.a. saved_key_).
-  while (iter_->Valid()) {
+  while (iter_.Valid()) {
    ParsedInternalKey ikey;
    if (!ParseKey(&ikey)) {
      return false;
    }
-    if (user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) >=
-        0) {
+    if (user_comparator_.Compare(ikey.user_key, saved_key_.GetUserKey()) >= 0) {
      return true;
    }
-    iter_->Next();
+    iter_.Next();
  }

-  if (!iter_->status().ok()) {
+  if (!iter_.status().ok()) {
    valid_ = false;
    return false;
  }
@@ -788,14 +792,14 @@ bool DBIter::ReverseToForward() {

 // Move iter_ to the key before saved_key_.
 bool DBIter::ReverseToBackward() {
-  assert(iter_->status().ok());
+  assert(iter_.status().ok());

  // When current_entry_is_merged_ is true, iter_ may be positioned on the next
  // key, which may not exist or may have prefix different from current.
  // If that's the case, seek to saved_key_.
  if (current_entry_is_merged_ &&
      ((prefix_extractor_ != nullptr && !total_order_seek_) ||
-       !iter_->Valid())) {
+       !iter_.Valid())) {
    IterKey last_key;
    // Using kMaxSequenceNumber and kValueTypeForSeek
    // (not kValueTypeForSeekForPrev) to seek to a key strictly smaller
@@ -803,15 +807,15 @@ bool DBIter::ReverseToBackward() {
    last_key.SetInternalKey(ParsedInternalKey(
        saved_key_.GetUserKey(), kMaxSequenceNumber, kValueTypeForSeek));
    if (prefix_extractor_ != nullptr && !total_order_seek_) {
-      iter_->SeekForPrev(last_key.GetInternalKey());
+      iter_.SeekForPrev(last_key.GetInternalKey());
    } else {
      // Some iterators may not support SeekForPrev(), so we avoid using it
      // when prefix seek mode is disabled. This is somewhat expensive
      // (an extra Prev(), as well as an extra change of direction of iter_),
      // so we may need to reconsider it later.
-      iter_->Seek(last_key.GetInternalKey());
-      if (!iter_->Valid() && iter_->status().ok()) {
-        iter_->SeekToLast();
+      iter_.Seek(last_key.GetInternalKey());
+      if (!iter_.Valid() && iter_.status().ok()) {
+        iter_.SeekToLast();
      }
    }
  }
@@ -821,10 +825,10 @@ bool DBIter::ReverseToBackward() {
 }

 void DBIter::PrevInternal() {
-  while (iter_->Valid()) {
+  while (iter_.Valid()) {
    saved_key_.SetUserKey(
-        ExtractUserKey(iter_->key()),
-        !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+        ExtractUserKey(iter_.key()),
+        !iter_.iter()->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);

    if (prefix_extractor_ && prefix_same_as_start_ &&
        prefix_extractor_->Transform(saved_key_.GetUserKey())
@@ -835,8 +839,8 @@ void DBIter::PrevInternal() {
    }

    if (iterate_lower_bound_ != nullptr &&
-        user_comparator_->Compare(saved_key_.GetUserKey(),
-                                  *iterate_lower_bound_) < 0) {
+        user_comparator_.Compare(saved_key_.GetUserKey(),
+                                 *iterate_lower_bound_) < 0) {
      // We've iterated earlier than the user-specified lower bound.
      valid_ = false;
      return;
@@ -878,7 +882,7 @@ void DBIter::PrevInternal() {
 // POST: iter_ is positioned on one of the entries equal to saved_key_, or on
 //       the entry just before them, or on the entry just after them.
 bool DBIter::FindValueForCurrentKey() {
-  assert(iter_->Valid());
+  assert(iter_.Valid());
  merge_context_.Clear();
  current_entry_is_merged_ = false;
  // last entry before merge (could be kTypeDeletion, kTypeSingleDeletion or
@@ -890,14 +894,14 @@ bool DBIter::FindValueForCurrentKey() {
  ReleaseTempPinnedData();
  TempPinData();
  size_t num_skipped = 0;
-  while (iter_->Valid()) {
+  while (iter_.Valid()) {
    ParsedInternalKey ikey;
    if (!ParseKey(&ikey)) {
      return false;
    }

    if (!IsVisible(ikey.sequence) ||
-        !user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
+        !user_comparator_.Equal(ikey.user_key, saved_key_.GetUserKey())) {
      break;
    }
    if (TooManyInternalKeysSkipped()) {
@@ -920,8 +924,8 @@ bool DBIter::FindValueForCurrentKey() {
          last_key_entry_type = kTypeRangeDeletion;
          PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
        } else {
-          assert(iter_->IsValuePinned());
-          pinned_value_ = iter_->value();
+          assert(iter_.iter()->IsValuePinned());
+          pinned_value_ = iter_.value();
        }
        merge_context_.Clear();
        last_not_merge_type = last_key_entry_type;
@@ -942,7 +946,8 @@ bool DBIter::FindValueForCurrentKey() {
        } else {
          assert(merge_operator_ != nullptr);
          merge_context_.PushOperandBack(
-              iter_->value(), iter_->IsValuePinned() /* operand_pinned */);
+              iter_.value(),
+              iter_.iter()->IsValuePinned() /* operand_pinned */);
          PERF_COUNTER_ADD(internal_merge_count, 1);
        }
        break;
@@ -951,11 +956,11 @@ bool DBIter::FindValueForCurrentKey() {
    }

    PERF_COUNTER_ADD(internal_key_skipped_count, 1);
-    iter_->Prev();
+    iter_.Prev();
    ++num_skipped;
  }

-  if (!iter_->status().ok()) {
+  if (!iter_.status().ok()) {
    valid_ = false;
    return false;
  }
@@ -1035,22 +1040,22 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
  std::string last_key;
  AppendInternalKey(&last_key, ParsedInternalKey(saved_key_.GetUserKey(),
                                                 sequence_, kValueTypeForSeek));
-  iter_->Seek(last_key);
+  iter_.Seek(last_key);
  RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);

  // In case read_callback presents, the value we seek to may not be visible.
  // Find the next value that's visible.
  ParsedInternalKey ikey;
  while (true) {
-    if (!iter_->Valid()) {
+    if (!iter_.Valid()) {
      valid_ = false;
-      return iter_->status().ok();
+      return iter_.status().ok();
    }

    if (!ParseKey(&ikey)) {
      return false;
    }
-    if (!user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
+    if (!user_comparator_.Equal(ikey.user_key, saved_key_.GetUserKey())) {
      // No visible values for this key, even though FindValueForCurrentKey()
      // has seen some. This is possible if we're using a tailing iterator, and
      // the entries were discarded in a compaction.
@@ -1062,7 +1067,7 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
      break;
    }

-    iter_->Next();
+    iter_.Next();
  }

  if (ikey.type == kTypeDeletion || ikey.type == kTypeSingleDeletion ||
@@ -1080,8 +1085,8 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
    return false;
  }
  if (ikey.type == kTypeValue || ikey.type == kTypeBlobIndex) {
-    assert(iter_->IsValuePinned());
-    pinned_value_ = iter_->value();
+    assert(iter_.iter()->IsValuePinned());
+    pinned_value_ = iter_.value();
    valid_ = true;
    return true;
  }
@@ -1091,13 +1096,13 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
  assert(ikey.type == kTypeMerge);
  current_entry_is_merged_ = true;
  merge_context_.Clear();
-  merge_context_.PushOperand(iter_->value(),
-                             iter_->IsValuePinned() /* operand_pinned */);
+  merge_context_.PushOperand(
+      iter_.value(), iter_.iter()->IsValuePinned() /* operand_pinned */);
  while (true) {
-    iter_->Next();
+    iter_.Next();

-    if (!iter_->Valid()) {
-      if (!iter_->status().ok()) {
+    if (!iter_.Valid()) {
+      if (!iter_.status().ok()) {
        valid_ = false;
        return false;
      }
@@ -1106,7 +1111,7 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
    if (!ParseKey(&ikey)) {
      return false;
    }
-    if (!user_comparator_->Equal(ikey.user_key, saved_key_.GetUserKey())) {
+    if (!user_comparator_.Equal(ikey.user_key, saved_key_.GetUserKey())) {
      break;
    }

@@ -1115,7 +1120,7 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
            ikey, RangeDelPositioningMode::kForwardTraversal)) {
      break;
    } else if (ikey.type == kTypeValue) {
-      const Slice val = iter_->value();
+      const Slice val = iter_.value();
      Status s = MergeHelper::TimedFullMerge(
          merge_operator_, saved_key_.GetUserKey(), &val,
          merge_context_.GetOperands(), &saved_value_, logger_, statistics_,
@@ -1128,8 +1133,8 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
      valid_ = true;
      return true;
    } else if (ikey.type == kTypeMerge) {
-      merge_context_.PushOperand(iter_->value(),
-                                 iter_->IsValuePinned() /* operand_pinned */);
+      merge_context_.PushOperand(
+          iter_.value(), iter_.iter()->IsValuePinned() /* operand_pinned */);
      PERF_COUNTER_ADD(internal_merge_count, 1);
    } else if (ikey.type == kTypeBlobIndex) {
      if (!allow_blob_) {
@@ -1161,13 +1166,13 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
  // Make sure we leave iter_ in a good state. If it's valid and we don't care
  // about prefixes, that's already good enough. Otherwise it needs to be
  // seeked to the current key.
-  if ((prefix_extractor_ != nullptr && !total_order_seek_) || !iter_->Valid()) {
+  if ((prefix_extractor_ != nullptr && !total_order_seek_) || !iter_.Valid()) {
    if (prefix_extractor_ != nullptr && !total_order_seek_) {
-      iter_->SeekForPrev(last_key);
+      iter_.SeekForPrev(last_key);
    } else {
-      iter_->Seek(last_key);
-      if (!iter_->Valid() && iter_->status().ok()) {
-        iter_->SeekToLast();
+      iter_.Seek(last_key);
+      if (!iter_.Valid() && iter_.status().ok()) {
+        iter_.SeekToLast();
      }
    }
    RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
@@ -1182,13 +1187,13 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
 bool DBIter::FindUserKeyBeforeSavedKey() {
  assert(status_.ok());
  size_t num_skipped = 0;
-  while (iter_->Valid()) {
+  while (iter_.Valid()) {
    ParsedInternalKey ikey;
    if (!ParseKey(&ikey)) {
      return false;
    }

-    if (user_comparator_->Compare(ikey.user_key, saved_key_.GetUserKey()) < 0) {
+    if (user_comparator_.Compare(ikey.user_key, saved_key_.GetUserKey()) < 0) {
      return true;
    }

@@ -1210,19 +1215,19 @@ bool DBIter::FindUserKeyBeforeSavedKey() {
          saved_key_.GetUserKey(), kMaxSequenceNumber, kValueTypeForSeek));
      // It would be more efficient to use SeekForPrev() here, but some
      // iterators may not support it.
-      iter_->Seek(last_key.GetInternalKey());
+      iter_.Seek(last_key.GetInternalKey());
      RecordTick(statistics_, NUMBER_OF_RESEEKS_IN_ITERATION);
-      if (!iter_->Valid()) {
+      if (!iter_.Valid()) {
        break;
      }
    } else {
      ++num_skipped;
    }

-    iter_->Prev();
+    iter_.Prev();
  }

-  if (!iter_->status().ok()) {
+  if (!iter_.status().ok()) {
    valid_ = false;
    return false;
  }
@@ -1243,30 +1248,25 @@ bool DBIter::TooManyInternalKeysSkipped(bool increment) {
 }

 bool DBIter::IsVisible(SequenceNumber sequence) {
-  return sequence <= MaxVisibleSequenceNumber() &&
-         (read_callback_ == nullptr || read_callback_->IsVisible(sequence));
+  if (read_callback_ == nullptr) {
+    return sequence <= sequence_;
+  } else {
+    return read_callback_->IsVisible(sequence);
+  }
 }

 bool DBIter::CanReseekToSkip() {
-  return read_callback_ == nullptr ||
-         read_callback_->MaxUnpreparedSequenceNumber() == 0;
-}
-
-SequenceNumber DBIter::MaxVisibleSequenceNumber() {
-  if (read_callback_ == nullptr) {
-    return sequence_;
-  }
-
-  return std::max(sequence_, read_callback_->MaxUnpreparedSequenceNumber());
+  return read_callback_ == nullptr || read_callback_->CanReseekToSkip();
 }

 void DBIter::Seek(const Slice& target) {
+  PERF_CPU_TIMER_GUARD(iter_seek_cpu_nanos, env_);
  StopWatch sw(env_, statistics_, DB_SEEK);
  status_ = Status::OK();
  ReleaseTempPinnedData();
  ResetInternalKeysSkippedCounter();

-  SequenceNumber seq = MaxVisibleSequenceNumber();
+  SequenceNumber seq = sequence_;
  saved_key_.Clear();
  saved_key_.SetInternalKey(target, seq);

@@ -1277,19 +1277,19 @@ void DBIter::Seek(const Slice& target) {
 #endif  // ROCKSDB_LITE

  if (iterate_lower_bound_ != nullptr &&
-      user_comparator_->Compare(saved_key_.GetUserKey(),
-                                *iterate_lower_bound_) < 0) {
+      user_comparator_.Compare(saved_key_.GetUserKey(), *iterate_lower_bound_) <
+          0) {
    saved_key_.Clear();
    saved_key_.SetInternalKey(*iterate_lower_bound_, seq);
  }

  {
    PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->Seek(saved_key_.GetInternalKey());
+    iter_.Seek(saved_key_.GetInternalKey());
    range_del_agg_.InvalidateRangeDelMapPositions();
  }
  RecordTick(statistics_, NUMBER_DB_SEEK);
-  if (iter_->Valid()) {
+  if (iter_.Valid()) {
    if (prefix_extractor_ && prefix_same_as_start_) {
      prefix_start_key_ = prefix_extractor_->Transform(target);
    }
@@ -1318,6 +1318,7 @@ void DBIter::Seek(const Slice& target) {
 }

 void DBIter::SeekForPrev(const Slice& target) {
+  PERF_CPU_TIMER_GUARD(iter_seek_cpu_nanos, env_);
  StopWatch sw(env_, statistics_, DB_SEEK);
  status_ = Status::OK();
  ReleaseTempPinnedData();
@@ -1328,15 +1329,15 @@ void DBIter::SeekForPrev(const Slice& target) {
                            kValueTypeForSeekForPrev);

  if (iterate_upper_bound_ != nullptr &&
-      user_comparator_->Compare(saved_key_.GetUserKey(),
-                                *iterate_upper_bound_) >= 0) {
+      user_comparator_.Compare(saved_key_.GetUserKey(),
+                               *iterate_upper_bound_) >= 0) {
    saved_key_.Clear();
    saved_key_.SetInternalKey(*iterate_upper_bound_, kMaxSequenceNumber);
  }

  {
    PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->SeekForPrev(saved_key_.GetInternalKey());
+    iter_.SeekForPrev(saved_key_.GetInternalKey());
    range_del_agg_.InvalidateRangeDelMapPositions();
  }

@@ -1347,7 +1348,7 @@ void DBIter::SeekForPrev(const Slice& target) {
 #endif  // ROCKSDB_LITE

  RecordTick(statistics_, NUMBER_DB_SEEK);
-  if (iter_->Valid()) {
+  if (iter_.Valid()) {
    if (prefix_extractor_ && prefix_same_as_start_) {
      prefix_start_key_ = prefix_extractor_->Transform(target);
    }
@@ -1378,6 +1379,7 @@ void DBIter::SeekToFirst() {
    Seek(*iterate_lower_bound_);
    return;
  }
+  PERF_CPU_TIMER_GUARD(iter_seek_cpu_nanos, env_);
  // Don't use iter_::Seek() if we set a prefix extractor
  // because prefix seek will be used.
  if (prefix_extractor_ != nullptr && !total_order_seek_) {
@@ -1391,15 +1393,15 @@ void DBIter::SeekToFirst() {

  {
    PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->SeekToFirst();
+    iter_.SeekToFirst();
    range_del_agg_.InvalidateRangeDelMapPositions();
  }

  RecordTick(statistics_, NUMBER_DB_SEEK);
-  if (iter_->Valid()) {
+  if (iter_.Valid()) {
    saved_key_.SetUserKey(
-        ExtractUserKey(iter_->key()),
-        !iter_->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
+        ExtractUserKey(iter_.key()),
+        !iter_.iter()->IsKeyPinned() || !pin_thru_lifetime_ /* copy */);
    FindNextUserEntry(false /* not skipping */, false /* no prefix check */);
    if (statistics_ != nullptr) {
      if (valid_) {
@@ -1422,13 +1424,14 @@ void DBIter::SeekToLast() {
  if (iterate_upper_bound_ != nullptr) {
    // Seek to last key strictly less than ReadOptions.iterate_upper_bound.
    SeekForPrev(*iterate_upper_bound_);
-    if (Valid() && user_comparator_->Equal(*iterate_upper_bound_, key())) {
+    if (Valid() && user_comparator_.Equal(*iterate_upper_bound_, key())) {
      ReleaseTempPinnedData();
      PrevInternal();
    }
    return;
  }

+  PERF_CPU_TIMER_GUARD(iter_seek_cpu_nanos, env_);
  // Don't use iter_::Seek() if we set a prefix extractor
  // because prefix seek will be used.
  if (prefix_extractor_ != nullptr && !total_order_seek_) {
@@ -1442,7 +1445,7 @@ void DBIter::SeekToLast() {

  {
    PERF_TIMER_GUARD(seek_internal_seek_time);
-    iter_->SeekToLast();
+    iter_.SeekToLast();
    range_del_agg_.InvalidateRangeDelMapPositions();
  }
  PrevInternal();
@@ -1479,7 +1482,7 @@ Iterator* NewDBIterator(Env* env, const ReadOptions& read_options,

 ArenaWrappedDBIter::~ArenaWrappedDBIter() { db_iter_->~DBIter(); }

-RangeDelAggregatorV2* ArenaWrappedDBIter::GetRangeDelAggregator() {
+ReadRangeDelAggregator* ArenaWrappedDBIter::GetRangeDelAggregator() {
  return db_iter_->GetRangeDelAggregator();
 }

@@ -1549,6 +1552,9 @@ Status ArenaWrappedDBIter::Refresh() {
    new (&arena_) Arena();

    SuperVersion* sv = cfd_->GetReferencedSuperVersion(db_impl_->mutex());
+    if (read_callback_) {
+      read_callback_->Refresh(latest_seq);
+    }
    Init(env, read_options_, *(cfd_->ioptions()), sv->mutable_cf_options,
         latest_seq, sv->mutable_cf_options.max_sequential_skip_in_iterations,
         cur_sv_number, read_callback_, db_impl_, cfd_, allow_blob_,
@@ -12,7 +12,7 @@
 #include <string>
 #include "db/db_impl.h"
 #include "db/dbformat.h"
-#include "db/range_del_aggregator_v2.h"
+#include "db/range_del_aggregator.h"
 #include "options/cf_options.h"
 #include "rocksdb/db.h"
 #include "rocksdb/iterator.h"
@@ -48,7 +48,7 @@ class ArenaWrappedDBIter : public Iterator {
  // Get the arena to be used to allocate memory for DBIter to be wrapped,
  // as well as child iterators in it.
  virtual Arena* GetArena() { return &arena_; }
-  virtual RangeDelAggregatorV2* GetRangeDelAggregator();
+  virtual ReadRangeDelAggregator* GetRangeDelAggregator();

  // Set the internal iterator wrapped inside the DB Iterator. Usually it is
  // a merging iterator.
@@ -116,12 +116,12 @@ class TestIterator : public InternalIterator {
  // Number of operations done on this iterator since construction.
  size_t steps() const { return steps_; }

-  virtual bool Valid() const override {
+  bool Valid() const override {
    assert(initialized_);
    return valid_;
  }

-  virtual void SeekToFirst() override {
+  void SeekToFirst() override {
    assert(initialized_);
    ++steps_;
    DeleteCurrentIfNeeded();
@@ -129,7 +129,7 @@ class TestIterator : public InternalIterator {
    iter_ = 0;
  }

-  virtual void SeekToLast() override {
+  void SeekToLast() override {
    assert(initialized_);
    ++steps_;
    DeleteCurrentIfNeeded();
@@ -137,7 +137,7 @@ class TestIterator : public InternalIterator {
    iter_ = data_.size() - 1;
  }

-  virtual void Seek(const Slice& target) override {
+  void Seek(const Slice& target) override {
    assert(initialized_);
    SeekToFirst();
    ++steps_;
@@ -154,13 +154,13 @@ class TestIterator : public InternalIterator {
    }
  }

-  virtual void SeekForPrev(const Slice& target) override {
+  void SeekForPrev(const Slice& target) override {
    assert(initialized_);
    DeleteCurrentIfNeeded();
    SeekForPrevImpl(target, &cmp);
  }

-  virtual void Next() override {
+  void Next() override {
    assert(initialized_);
    assert(valid_);
    assert(iter_ < data_.size());
@@ -174,7 +174,7 @@ class TestIterator : public InternalIterator {
    valid_ = iter_ < data_.size();
  }

-  virtual void Prev() override {
+  void Prev() override {
    assert(initialized_);
    assert(valid_);
    assert(iter_ < data_.size());
@@ -188,23 +188,23 @@ class TestIterator : public InternalIterator {
    }
  }

-  virtual Slice key() const override {
+  Slice key() const override {
    assert(initialized_);
    return data_[iter_].first;
  }

-  virtual Slice value() const override {
+  Slice value() const override {
    assert(initialized_);
    return data_[iter_].second;
  }

-  virtual Status status() const override {
+  Status status() const override {
    assert(initialized_);
    return Status::OK();
  }

-  virtual bool IsKeyPinned() const override { return true; }
-  virtual bool IsValuePinned() const override { return true; }
+  bool IsKeyPinned() const override { return true; }
+  bool IsValuePinned() const override { return true; }

 private:
  bool initialized_;
@@ -15,12 +15,16 @@
 #include "port/stack_trace.h"
 #include "rocksdb/iostats_context.h"
 #include "rocksdb/perf_context.h"
+#include "table/flush_block_policy.h"

 namespace rocksdb {

 // A dumb ReadCallback which saying every key is committed.
 class DummyReadCallback : public ReadCallback {
-  bool IsVisible(SequenceNumber /*seq*/) override { return true; }
+ public:
+  DummyReadCallback() : ReadCallback(kMaxSequenceNumber) {}
+  bool IsVisibleFullCheck(SequenceNumber /*seq*/) override { return true; }
+  void SetSnapshot(SequenceNumber seq) { max_visible_seq_ = seq; }
 };

 // Test param:
@@ -40,40 +44,20 @@ class DBIteratorTest : public DBTestBase,
                             ? read_options.snapshot->GetSequenceNumber()
                             : db_->GetLatestSequenceNumber();
    bool use_read_callback = GetParam();
-    ReadCallback* read_callback = use_read_callback ? &read_callback_ : nullptr;
+    DummyReadCallback* read_callback = nullptr;
+    if (use_read_callback) {
+      read_callback = new DummyReadCallback();
+      read_callback->SetSnapshot(seq);
+      InstrumentedMutexLock lock(&mutex_);
+      read_callbacks_.push_back(
+          std::unique_ptr<DummyReadCallback>(read_callback));
+    }
    return dbfull()->NewIteratorImpl(read_options, cfd, seq, read_callback);
  }

 private:
-  DummyReadCallback read_callback_;
-};
-
-class FlushBlockEveryKeyPolicy : public FlushBlockPolicy {
- public:
-  virtual bool Update(const Slice& /*key*/, const Slice& /*value*/) override {
-    if (!start_) {
-      start_ = true;
-      return false;
-    }
-    return true;
-  }
- private:
-  bool start_ = false;
-};
-
-class FlushBlockEveryKeyPolicyFactory : public FlushBlockPolicyFactory {
- public:
-  explicit FlushBlockEveryKeyPolicyFactory() {}
-
-  const char* Name() const override {
-    return "FlushBlockEveryKeyPolicyFactory";
-  }
-
-  FlushBlockPolicy* NewFlushBlockPolicy(
-      const BlockBasedTableOptions& /*table_options*/,
-      const BlockBuilder& /*data_block_builder*/) const override {
-    return new FlushBlockEveryKeyPolicy;
-  }
+  InstrumentedMutex mutex_;
+  std::vector<std::unique_ptr<DummyReadCallback>> read_callbacks_;
 };

 TEST_P(DBIteratorTest, IteratorProperty) {
@@ -179,9 +163,7 @@ TEST_P(DBIteratorTest, NonBlockingIteration) {

    // This test verifies block cache behaviors, which is not used by plain
    // table format.
-    // Exclude kHashCuckoo as it does not support iteration currently
-  } while (ChangeOptions(kSkipPlainTable | kSkipNoSeekToLast | kSkipHashCuckoo |
-                         kSkipMmapReads));
+  } while (ChangeOptions(kSkipPlainTable | kSkipNoSeekToLast | kSkipMmapReads));
 }

 TEST_P(DBIteratorTest, IterSeekBeforePrev) {
@@ -765,8 +747,7 @@ TEST_P(DBIteratorTest, IterWithSnapshot) {
    }
    db_->ReleaseSnapshot(snapshot);
    delete iter;
-    // skip as HashCuckooRep does not support snapshot
-  } while (ChangeOptions(kSkipHashCuckoo));
+  } while (ChangeOptions());
 }

 TEST_P(DBIteratorTest, IteratorPinsRef) {
@@ -1025,48 +1006,48 @@ TEST_P(DBIteratorTest, DBIteratorBoundMultiSeek) {
 #endif

 TEST_P(DBIteratorTest, DBIteratorBoundOptimizationTest) {
-  int upper_bound_hits = 0;
-  Options options = CurrentOptions();
-  rocksdb::SyncPoint::GetInstance()->SetCallBack(
-      "BlockBasedTable::BlockEntryIteratorState::KeyReachedUpperBound",
-      [&upper_bound_hits](void* arg) {
-        assert(arg != nullptr);
-        upper_bound_hits += (*static_cast<bool*>(arg) ? 1 : 0);
-      });
-  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
-  options.env = env_;
-  options.create_if_missing = true;
-  options.prefix_extractor = nullptr;
-  BlockBasedTableOptions table_options;
-  table_options.flush_block_policy_factory =
-    std::make_shared<FlushBlockEveryKeyPolicyFactory>();
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  for (auto format_version : {2, 3, 4}) {
+    int upper_bound_hits = 0;
+    Options options = CurrentOptions();
+    rocksdb::SyncPoint::GetInstance()->SetCallBack(
+        "BlockBasedTableIterator:out_of_bound",
+        [&upper_bound_hits](void*) { upper_bound_hits++; });
+    rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+    options.env = env_;
+    options.create_if_missing = true;
+    options.prefix_extractor = nullptr;
+    BlockBasedTableOptions table_options;
+    table_options.format_version = format_version;
+    table_options.flush_block_policy_factory =
+        std::make_shared<FlushBlockEveryKeyPolicyFactory>();
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));

-  DestroyAndReopen(options);
-  ASSERT_OK(Put("foo1", "bar1"));
-  ASSERT_OK(Put("foo2", "bar2"));
-  ASSERT_OK(Put("foo4", "bar4"));
-  ASSERT_OK(Flush());
+    DestroyAndReopen(options);
+    ASSERT_OK(Put("foo1", "bar1"));
+    ASSERT_OK(Put("foo2", "bar2"));
+    ASSERT_OK(Put("foo4", "bar4"));
+    ASSERT_OK(Flush());

-  Slice ub("foo3");
-  ReadOptions ro;
-  ro.iterate_upper_bound = &ub;
+    Slice ub("foo3");
+    ReadOptions ro;
+    ro.iterate_upper_bound = &ub;

-  std::unique_ptr<Iterator> iter(NewIterator(ro));
+    std::unique_ptr<Iterator> iter(NewIterator(ro));

-  iter->Seek("foo");
-  ASSERT_TRUE(iter->Valid());
-  ASSERT_EQ(iter->key().compare(Slice("foo1")), 0);
-  ASSERT_EQ(upper_bound_hits, 0);
+    iter->Seek("foo");
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ(iter->key().compare(Slice("foo1")), 0);
+    ASSERT_EQ(upper_bound_hits, 0);

-  iter->Next();
-  ASSERT_TRUE(iter->Valid());
-  ASSERT_EQ(iter->key().compare(Slice("foo2")), 0);
-  ASSERT_EQ(upper_bound_hits, 0);
+    iter->Next();
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ(iter->key().compare(Slice("foo2")), 0);
+    ASSERT_EQ(upper_bound_hits, 0);

-  iter->Next();
-  ASSERT_FALSE(iter->Valid());
-  ASSERT_EQ(upper_bound_hits, 1);
+    iter->Next();
+    ASSERT_FALSE(iter->Valid());
+    ASSERT_EQ(upper_bound_hits, 1);
+  }
 }
 // TODO(3.13): fix the issue of Seek() + Prev() which might not necessary
 //             return the biggest key which is smaller than the seek key.
@@ -1962,8 +1943,8 @@ TEST_P(DBIteratorTest, ReadAhead) {
  delete iter;
  int64_t num_file_closes_readahead =
      TestGetTickerCount(options, NO_FILE_CLOSES);
-  ASSERT_EQ(num_file_opens + 3, num_file_opens_readahead);
-  ASSERT_EQ(num_file_closes + 3, num_file_closes_readahead);
+  ASSERT_EQ(num_file_opens, num_file_opens_readahead);
+  ASSERT_EQ(num_file_closes, num_file_closes_readahead);
  ASSERT_GT(bytes_read_readahead, bytes_read);
  ASSERT_GT(bytes_read_readahead, read_options.readahead_size * 3);

@@ -2478,15 +2459,12 @@ class DBIteratorWithReadCallbackTest : public DBIteratorTest {};
 TEST_F(DBIteratorWithReadCallbackTest, ReadCallback) {
  class TestReadCallback : public ReadCallback {
   public:
-    explicit TestReadCallback(SequenceNumber last_visible_seq)
-        : last_visible_seq_(last_visible_seq) {}
+    explicit TestReadCallback(SequenceNumber _max_visible_seq)
+        : ReadCallback(_max_visible_seq) {}

-    bool IsVisible(SequenceNumber seq) override {
-      return seq <= last_visible_seq_;
+    bool IsVisibleFullCheck(SequenceNumber seq) override {
+      return seq <= max_visible_seq_;
    }
-
-   private:
-    SequenceNumber last_visible_seq_;
  };

  ASSERT_OK(Put("foo", "v1"));
@@ -249,22 +249,20 @@ TEST_F(DBTestXactLogIterator, TransactionLogIteratorBlobs) {
  auto res = OpenTransactionLogIter(0)->GetBatch();
  struct Handler : public WriteBatch::Handler {
    std::string seen;
-    virtual Status PutCF(uint32_t cf, const Slice& key,
-                         const Slice& value) override {
+    Status PutCF(uint32_t cf, const Slice& key, const Slice& value) override {
      seen += "Put(" + ToString(cf) + ", " + key.ToString() + ", " +
              ToString(value.size()) + ")";
      return Status::OK();
    }
-    virtual Status MergeCF(uint32_t cf, const Slice& key,
-                           const Slice& value) override {
+    Status MergeCF(uint32_t cf, const Slice& key, const Slice& value) override {
      seen += "Merge(" + ToString(cf) + ", " + key.ToString() + ", " +
              ToString(value.size()) + ")";
      return Status::OK();
    }
-    virtual void LogData(const Slice& blob) override {
+    void LogData(const Slice& blob) override {
      seen += "LogData(" + blob.ToString() + ")";
    }
-    virtual Status DeleteCF(uint32_t cf, const Slice& key) override {
+    Status DeleteCF(uint32_t cf, const Slice& key) override {
      seen += "Delete(" + ToString(cf) + ", " + key.ToString() + ")";
      return Status::OK();
    }
@@ -8,6 +8,7 @@

 #include "db/db_test_util.h"
 #include "db/memtable.h"
+#include "db/range_del_aggregator.h"
 #include "port/stack_trace.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/slice_transform.h"
@@ -24,13 +25,13 @@ class MockMemTableRep : public MemTableRep {
  explicit MockMemTableRep(Allocator* allocator, MemTableRep* rep)
      : MemTableRep(allocator), rep_(rep), num_insert_with_hint_(0) {}

-  virtual KeyHandle Allocate(const size_t len, char** buf) override {
+  KeyHandle Allocate(const size_t len, char** buf) override {
    return rep_->Allocate(len, buf);
  }

-  virtual void Insert(KeyHandle handle) override { rep_->Insert(handle); }
+  void Insert(KeyHandle handle) override { rep_->Insert(handle); }

-  virtual void InsertWithHint(KeyHandle handle, void** hint) override {
+  void InsertWithHint(KeyHandle handle, void** hint) override {
    num_insert_with_hint_++;
    EXPECT_NE(nullptr, hint);
    last_hint_in_ = *hint;
@@ -38,21 +39,18 @@ class MockMemTableRep : public MemTableRep {
    last_hint_out_ = *hint;
  }

-  virtual bool Contains(const char* key) const override {
-    return rep_->Contains(key);
-  }
+  bool Contains(const char* key) const override { return rep_->Contains(key); }

-  virtual void Get(const LookupKey& k, void* callback_args,
-                   bool (*callback_func)(void* arg,
-                                         const char* entry)) override {
+  void Get(const LookupKey& k, void* callback_args,
+           bool (*callback_func)(void* arg, const char* entry)) override {
    rep_->Get(k, callback_args, callback_func);
  }

-  virtual size_t ApproximateMemoryUsage() override {
+  size_t ApproximateMemoryUsage() override {
    return rep_->ApproximateMemoryUsage();
  }

-  virtual Iterator* GetIterator(Arena* arena) override {
+  Iterator* GetIterator(Arena* arena) override {
    return rep_->GetIterator(arena);
  }

@@ -69,10 +67,10 @@ class MockMemTableRep : public MemTableRep {

 class MockMemTableRepFactory : public MemTableRepFactory {
 public:
-  virtual MemTableRep* CreateMemTableRep(const MemTableRep::KeyComparator& cmp,
-                                         Allocator* allocator,
-                                         const SliceTransform* transform,
-                                         Logger* logger) override {
+  MemTableRep* CreateMemTableRep(const MemTableRep::KeyComparator& cmp,
+                                 Allocator* allocator,
+                                 const SliceTransform* transform,
+                                 Logger* logger) override {
    SkipListFactory factory;
    MemTableRep* skiplist_rep =
        factory.CreateMemTableRep(cmp, allocator, transform, logger);
@@ -80,16 +78,16 @@ class MockMemTableRepFactory : public MemTableRepFactory {
    return mock_rep_;
  }

-  virtual MemTableRep* CreateMemTableRep(const MemTableRep::KeyComparator& cmp,
-                                         Allocator* allocator,
-                                         const SliceTransform* transform,
-                                         Logger* logger,
-                                         uint32_t column_family_id) override {
+  MemTableRep* CreateMemTableRep(const MemTableRep::KeyComparator& cmp,
+                                 Allocator* allocator,
+                                 const SliceTransform* transform,
+                                 Logger* logger,
+                                 uint32_t column_family_id) override {
    last_column_family_id_ = column_family_id;
    return CreateMemTableRep(cmp, allocator, transform, logger);
  }

-  virtual const char* Name() const override { return "MockMemTableRepFactory"; }
+  const char* Name() const override { return "MockMemTableRepFactory"; }

  MockMemTableRep* rep() { return mock_rep_; }

@@ -105,9 +103,9 @@ class MockMemTableRepFactory : public MemTableRepFactory {

 class TestPrefixExtractor : public SliceTransform {
 public:
-  virtual const char* Name() const override { return "TestPrefixExtractor"; }
+  const char* Name() const override { return "TestPrefixExtractor"; }

-  virtual Slice Transform(const Slice& key) const override {
+  Slice Transform(const Slice& key) const override {
    const char* p = separator(key);
    if (p == nullptr) {
      return Slice();
@@ -115,11 +113,11 @@ class TestPrefixExtractor : public SliceTransform {
    return Slice(key.data(), p - key.data() + 1);
  }

-  virtual bool InDomain(const Slice& key) const override {
+  bool InDomain(const Slice& key) const override {
    return separator(key) != nullptr;
  }

-  virtual bool InRange(const Slice& /*key*/) const override { return false; }
+  bool InRange(const Slice& /*key*/) const override { return false; }

 private:
  const char* separator(const Slice& key) const {
@@ -135,7 +133,8 @@ TEST_F(DBMemTableTest, DuplicateSeq) {
  MergeContext merge_context;
  Options options;
  InternalKeyComparator ikey_cmp(options.comparator);
-  RangeDelAggregator range_del_agg(ikey_cmp, {} /* snapshots */);
+  ReadRangeDelAggregator range_del_agg(&ikey_cmp,
+                                       kMaxSequenceNumber /* upper_bound */);

  // Create a MemTable
  InternalKeyComparator cmp(BytewiseComparator());
@@ -18,10 +18,13 @@ class TestReadCallback : public ReadCallback {
 public:
  TestReadCallback(SnapshotChecker* snapshot_checker,
                   SequenceNumber snapshot_seq)
-      : snapshot_checker_(snapshot_checker), snapshot_seq_(snapshot_seq) {}
+      : ReadCallback(snapshot_seq),
+        snapshot_checker_(snapshot_checker),
+        snapshot_seq_(snapshot_seq) {}

-  bool IsVisible(SequenceNumber seq) override {
-    return snapshot_checker_->IsInSnapshot(seq, snapshot_seq_);
+  bool IsVisibleFullCheck(SequenceNumber seq) override {
+    return snapshot_checker_->CheckInSnapshot(seq, snapshot_seq_) ==
+           SnapshotCheckerResult::kInSnapshot;
  }

 private:
@@ -331,8 +334,7 @@ TEST_P(MergeOperatorPinningTest, Randomized) {

    VerifyDBFromMap(true_data);

-    // Skip HashCuckoo since it does not support merge operators
-  } while (ChangeOptions(kSkipMergePut | kSkipHashCuckoo));
+  } while (ChangeOptions(kSkipMergePut));
 }

 class MergeOperatorHook : public MergeOperator {
@@ -340,15 +342,15 @@ class MergeOperatorHook : public MergeOperator {
  explicit MergeOperatorHook(std::shared_ptr<MergeOperator> _merge_op)
      : merge_op_(_merge_op) {}

-  virtual bool FullMergeV2(const MergeOperationInput& merge_in,
-                           MergeOperationOutput* merge_out) const override {
+  bool FullMergeV2(const MergeOperationInput& merge_in,
+                   MergeOperationOutput* merge_out) const override {
    before_merge_();
    bool res = merge_op_->FullMergeV2(merge_in, merge_out);
    after_merge_();
    return res;
  }

-  virtual const char* Name() const override { return merge_op_->Name(); }
+  const char* Name() const override { return merge_op_->Name(); }

  std::shared_ptr<MergeOperator> merge_op_;
  std::function<void()> before_merge_ = []() {};
@@ -547,8 +549,15 @@ TEST_F(DBMergeOperatorTest, SnapshotCheckerAndReadCallback) {
  DestroyAndReopen(options);

  class TestSnapshotChecker : public SnapshotChecker {
-    bool IsInSnapshot(SequenceNumber seq,
-                      SequenceNumber snapshot_seq) const override {
+   public:
+    SnapshotCheckerResult CheckInSnapshot(
+        SequenceNumber seq, SequenceNumber snapshot_seq) const override {
+      return IsInSnapshot(seq, snapshot_seq)
+                 ? SnapshotCheckerResult::kInSnapshot
+                 : SnapshotCheckerResult::kNotInSnapshot;
+    }
+
+    bool IsInSnapshot(SequenceNumber seq, SequenceNumber snapshot_seq) const {
      switch (snapshot_seq) {
        case 0:
          return seq == 0;
@@ -18,12 +18,15 @@
 #include "rocksdb/cache.h"
 #include "rocksdb/convenience.h"
 #include "rocksdb/rate_limiter.h"
+#include "rocksdb/stats_history.h"
 #include "util/random.h"
 #include "util/sync_point.h"
 #include "util/testutil.h"

 namespace rocksdb {

+const int kMicrosInSec = 1000000;
+
 class DBOptionsTest : public DBTestBase {
 public:
  DBOptionsTest() : DBTestBase("/db_options_test") {}
@@ -508,10 +511,11 @@ TEST_F(DBOptionsTest, SetStatsDumpPeriodSec) {

  for (int i = 0; i < 20; i++) {
    int num = rand() % 5000 + 1;
-    ASSERT_OK(dbfull()->SetDBOptions(
-        {{"stats_dump_period_sec", std::to_string(num)}}));
+    ASSERT_OK(
+        dbfull()->SetDBOptions({{"stats_dump_period_sec", ToString(num)}}));
    ASSERT_EQ(num, dbfull()->GetDBOptions().stats_dump_period_sec);
  }
+  Close();
 }

 TEST_F(DBOptionsTest, RunStatsDumpPeriodSec) {
@@ -523,6 +527,17 @@ TEST_F(DBOptionsTest, RunStatsDumpPeriodSec) {
  mock_env->set_current_time(0); // in seconds
  options.env = mock_env.get();
  int counter = 0;
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
+#if defined(OS_MACOSX) && !defined(NDEBUG)
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "InstrumentedCondVar::TimedWaitInternal", [&](void* arg) {
+        uint64_t time_us = *reinterpret_cast<uint64_t*>(arg);
+        if (time_us < mock_env->RealNowMicros()) {
+          *reinterpret_cast<uint64_t*>(arg) = mock_env->RealNowMicros() + 1000;
+        }
+      });
+#endif  // OS_MACOSX && !NDEBUG
  rocksdb::SyncPoint::GetInstance()->SetCallBack(
      "DBImpl::DumpStats:1", [&](void* /*arg*/) {
        counter++;
@@ -530,17 +545,258 @@ TEST_F(DBOptionsTest, RunStatsDumpPeriodSec) {
  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
  Reopen(options);
  ASSERT_EQ(5, dbfull()->GetDBOptions().stats_dump_period_sec);
-  dbfull()->TEST_WaitForTimedTaskRun([&] { mock_env->set_current_time(5); });
+  dbfull()->TEST_WaitForDumpStatsRun([&] { mock_env->set_current_time(5); });
  ASSERT_GE(counter, 1);

  // Test cacel job through SetOptions
  ASSERT_OK(dbfull()->SetDBOptions({{"stats_dump_period_sec", "0"}}));
  int old_val = counter;
-  env_->SleepForMicroseconds(10000000);
+  for (int i = 6; i < 20; ++i) {
+    dbfull()->TEST_WaitForDumpStatsRun([&] { mock_env->set_current_time(i); });
+  }
  ASSERT_EQ(counter, old_val);
  Close();
 }

+// Test persistent stats background thread scheduling and cancelling
+TEST_F(DBOptionsTest, StatsPersistScheduling) {
+  Options options;
+  options.create_if_missing = true;
+  options.stats_persist_period_sec = 5;
+  std::unique_ptr<rocksdb::MockTimeEnv> mock_env;
+  mock_env.reset(new rocksdb::MockTimeEnv(env_));
+  mock_env->set_current_time(0);  // in seconds
+  options.env = mock_env.get();
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
+#if defined(OS_MACOSX) && !defined(NDEBUG)
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "InstrumentedCondVar::TimedWaitInternal", [&](void* arg) {
+        uint64_t time_us = *reinterpret_cast<uint64_t*>(arg);
+        if (time_us < mock_env->RealNowMicros()) {
+          *reinterpret_cast<uint64_t*>(arg) = mock_env->RealNowMicros() + 1000;
+        }
+      });
+#endif  // OS_MACOSX && !NDEBUG
+  int counter = 0;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::PersistStats:Entry", [&](void* /*arg*/) { counter++; });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  Reopen(options);
+  ASSERT_EQ(5, dbfull()->GetDBOptions().stats_persist_period_sec);
+  dbfull()->TEST_WaitForPersistStatsRun([&] { mock_env->set_current_time(5); });
+  ASSERT_GE(counter, 1);
+
+  // Test cacel job through SetOptions
+  ASSERT_TRUE(dbfull()->TEST_IsPersistentStatsEnabled());
+  ASSERT_OK(dbfull()->SetDBOptions({{"stats_persist_period_sec", "0"}}));
+  ASSERT_FALSE(dbfull()->TEST_IsPersistentStatsEnabled());
+  Close();
+}
+
+// Test enabling persistent stats for the first time
+TEST_F(DBOptionsTest, PersistentStatsFreshInstall) {
+  Options options;
+  options.create_if_missing = true;
+  options.stats_persist_period_sec = 0;
+  std::unique_ptr<rocksdb::MockTimeEnv> mock_env;
+  mock_env.reset(new rocksdb::MockTimeEnv(env_));
+  mock_env->set_current_time(0);  // in seconds
+  options.env = mock_env.get();
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
+#if defined(OS_MACOSX) && !defined(NDEBUG)
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "InstrumentedCondVar::TimedWaitInternal", [&](void* arg) {
+        uint64_t time_us = *reinterpret_cast<uint64_t*>(arg);
+        if (time_us < mock_env->RealNowMicros()) {
+          *reinterpret_cast<uint64_t*>(arg) = mock_env->RealNowMicros() + 1000;
+        }
+      });
+#endif  // OS_MACOSX && !NDEBUG
+  int counter = 0;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::PersistStats:Entry", [&](void* /*arg*/) { counter++; });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+  Reopen(options);
+  ASSERT_OK(dbfull()->SetDBOptions({{"stats_persist_period_sec", "5"}}));
+  ASSERT_EQ(5, dbfull()->GetDBOptions().stats_persist_period_sec);
+  dbfull()->TEST_WaitForPersistStatsRun([&] { mock_env->set_current_time(5); });
+  ASSERT_GE(counter, 1);
+  Close();
+}
+
+TEST_F(DBOptionsTest, SetOptionsStatsPersistPeriodSec) {
+  Options options;
+  options.create_if_missing = true;
+  options.stats_persist_period_sec = 5;
+  options.env = env_;
+  Reopen(options);
+  ASSERT_EQ(5, dbfull()->GetDBOptions().stats_persist_period_sec);
+
+  ASSERT_OK(dbfull()->SetDBOptions({{"stats_persist_period_sec", "12345"}}));
+  ASSERT_EQ(12345, dbfull()->GetDBOptions().stats_persist_period_sec);
+  ASSERT_NOK(dbfull()->SetDBOptions({{"stats_persist_period_sec", "abcde"}}));
+  ASSERT_EQ(12345, dbfull()->GetDBOptions().stats_persist_period_sec);
+}
+
+TEST_F(DBOptionsTest, GetStatsHistory) {
+  Options options;
+  options.create_if_missing = true;
+  options.stats_persist_period_sec = 5;
+  options.statistics = rocksdb::CreateDBStatistics();
+  std::unique_ptr<rocksdb::MockTimeEnv> mock_env;
+  mock_env.reset(new rocksdb::MockTimeEnv(env_));
+  mock_env->set_current_time(0);  // in seconds
+  options.env = mock_env.get();
+#if defined(OS_MACOSX) && !defined(NDEBUG)
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "InstrumentedCondVar::TimedWaitInternal", [&](void* arg) {
+        uint64_t time_us = *reinterpret_cast<uint64_t*>(arg);
+        if (time_us < mock_env->RealNowMicros()) {
+          *reinterpret_cast<uint64_t*>(arg) = mock_env->RealNowMicros() + 1000;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+#endif  // OS_MACOSX && !NDEBUG
+
+  CreateColumnFamilies({"pikachu"}, options);
+  ASSERT_OK(Put("foo", "bar"));
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+
+  int mock_time = 1;
+  // Wait for stats persist to finish
+  dbfull()->TEST_WaitForPersistStatsRun([&] { mock_env->set_current_time(5); });
+  std::unique_ptr<StatsHistoryIterator> stats_iter;
+  db_->GetStatsHistory(0, 6 * kMicrosInSec, &stats_iter);
+  ASSERT_TRUE(stats_iter != nullptr);
+  // disabled stats snapshots
+  ASSERT_OK(dbfull()->SetDBOptions({{"stats_persist_period_sec", "0"}}));
+  size_t stats_count = 0;
+  for (; stats_iter->Valid(); stats_iter->Next()) {
+    auto stats_map = stats_iter->GetStatsMap();
+    stats_count += stats_map.size();
+  }
+  ASSERT_GT(stats_count, 0);
+  // Wait a bit and verify no more stats are found
+  for (mock_time = 6; mock_time < 20; ++mock_time) {
+    dbfull()->TEST_WaitForPersistStatsRun(
+        [&] { mock_env->set_current_time(mock_time); });
+  }
+  db_->GetStatsHistory(0, 20 * kMicrosInSec, &stats_iter);
+  ASSERT_TRUE(stats_iter != nullptr);
+  size_t stats_count_new = 0;
+  for (; stats_iter->Valid(); stats_iter->Next()) {
+    stats_count_new += stats_iter->GetStatsMap().size();
+  }
+  ASSERT_EQ(stats_count_new, stats_count);
+  Close();
+}
+
+TEST_F(DBOptionsTest, InMemoryStatsHistoryPurging) {
+  Options options;
+  options.create_if_missing = true;
+  options.statistics = rocksdb::CreateDBStatistics();
+  options.stats_persist_period_sec = 1;
+  std::unique_ptr<rocksdb::MockTimeEnv> mock_env;
+  mock_env.reset(new rocksdb::MockTimeEnv(env_));
+  mock_env->set_current_time(0);  // in seconds
+  options.env = mock_env.get();
+#if defined(OS_MACOSX) && !defined(NDEBUG)
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+  rocksdb::SyncPoint::GetInstance()->ClearAllCallBacks();
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "InstrumentedCondVar::TimedWaitInternal", [&](void* arg) {
+        uint64_t time_us = *reinterpret_cast<uint64_t*>(arg);
+        if (time_us < mock_env->RealNowMicros()) {
+          *reinterpret_cast<uint64_t*>(arg) = mock_env->RealNowMicros() + 1000;
+        }
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+#endif  // OS_MACOSX && !NDEBUG
+
+  CreateColumnFamilies({"pikachu"}, options);
+  ASSERT_OK(Put("foo", "bar"));
+  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  // some random operation to populate statistics
+  ASSERT_OK(Delete("foo"));
+  ASSERT_OK(Put("sol", "sol"));
+  ASSERT_OK(Put("epic", "epic"));
+  ASSERT_OK(Put("ltd", "ltd"));
+  ASSERT_EQ("sol", Get("sol"));
+  ASSERT_EQ("epic", Get("epic"));
+  ASSERT_EQ("ltd", Get("ltd"));
+  Iterator* iterator = db_->NewIterator(ReadOptions());
+  for (iterator->SeekToFirst(); iterator->Valid(); iterator->Next()) {
+    ASSERT_TRUE(iterator->key() == iterator->value());
+  }
+  delete iterator;
+  ASSERT_OK(Flush());
+  ASSERT_OK(Delete("sol"));
+  db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  int mock_time = 1;
+  // Wait for stats persist to finish
+  for (; mock_time < 5; ++mock_time) {
+    dbfull()->TEST_WaitForPersistStatsRun(
+        [&] { mock_env->set_current_time(mock_time); });
+  }
+
+  // second round of ops
+  ASSERT_OK(Put("saigon", "saigon"));
+  ASSERT_OK(Put("noodle talk", "noodle talk"));
+  ASSERT_OK(Put("ping bistro", "ping bistro"));
+  iterator = db_->NewIterator(ReadOptions());
+  for (iterator->SeekToFirst(); iterator->Valid(); iterator->Next()) {
+    ASSERT_TRUE(iterator->key() == iterator->value());
+  }
+  delete iterator;
+  ASSERT_OK(Flush());
+  db_->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  for (; mock_time < 10; ++mock_time) {
+    dbfull()->TEST_WaitForPersistStatsRun(
+        [&] { mock_env->set_current_time(mock_time); });
+  }
+  std::unique_ptr<StatsHistoryIterator> stats_iter;
+  db_->GetStatsHistory(0, 10 * kMicrosInSec, &stats_iter);
+  ASSERT_TRUE(stats_iter != nullptr);
+  size_t stats_count = 0;
+  int slice_count = 0;
+  for (; stats_iter->Valid(); stats_iter->Next()) {
+    slice_count++;
+    auto stats_map = stats_iter->GetStatsMap();
+    stats_count += stats_map.size();
+  }
+  size_t stats_history_size = dbfull()->TEST_EstiamteStatsHistorySize();
+  ASSERT_GE(slice_count, 9);
+  ASSERT_GE(stats_history_size, 12000);
+  // capping memory cost at 12000 bytes since one slice is around 10000~12000
+  ASSERT_OK(dbfull()->SetDBOptions({{"stats_history_buffer_size", "12000"}}));
+  ASSERT_EQ(12000, dbfull()->GetDBOptions().stats_history_buffer_size);
+  // Wait for stats persist to finish
+  for (; mock_time < 20; ++mock_time) {
+    dbfull()->TEST_WaitForPersistStatsRun(
+        [&] { mock_env->set_current_time(mock_time); });
+  }
+  db_->GetStatsHistory(0, 20 * kMicrosInSec, &stats_iter);
+  ASSERT_TRUE(stats_iter != nullptr);
+  size_t stats_count_reopen = 0;
+  slice_count = 0;
+  for (; stats_iter->Valid(); stats_iter->Next()) {
+    slice_count++;
+    auto stats_map = stats_iter->GetStatsMap();
+    stats_count_reopen += stats_map.size();
+  }
+  size_t stats_history_size_reopen = dbfull()->TEST_EstiamteStatsHistorySize();
+  // only one slice can fit under the new stats_history_buffer_size
+  ASSERT_LT(slice_count, 2);
+  ASSERT_TRUE(stats_history_size_reopen < 12000 &&
+              stats_history_size_reopen > 0);
+  ASSERT_TRUE(stats_count_reopen < stats_count && stats_count_reopen > 0);
+  Close();
+}
+
 static void assert_candidate_files_empty(DBImpl* dbfull, const bool empty) {
  dbfull->TEST_LockMutex();
  JobContext job_context(0);
@@ -628,9 +884,9 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
  env_->time_elapse_only_sleep_ = false;
  options.env = env_;

-  // Test dynamically changing compaction_options_fifo.ttl
+  // Test dynamically changing ttl.
  env_->addon_time_.store(0);
-  options.compaction_options_fifo.ttl = 1 * 60 * 60;  // 1 hour
+  options.ttl = 1 * 60 * 60;  // 1 hour
  ASSERT_OK(TryReopen(options));

  Random rnd(301);
@@ -648,13 +904,13 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
  env_->addon_time_.fetch_add(61);

  // No files should be compacted as ttl is set to 1 hour.
-  ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.ttl, 3600);
+  ASSERT_EQ(dbfull()->GetOptions().ttl, 3600);
  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
  ASSERT_EQ(NumTableFilesAtLevel(0), 10);

  // Set ttl to 1 minute. So all files should get deleted.
-  ASSERT_OK(dbfull()->SetOptions({{"compaction_options_fifo", "{ttl=60;}"}}));
-  ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.ttl, 60);
+  ASSERT_OK(dbfull()->SetOptions({{"ttl", "60"}}));
+  ASSERT_EQ(dbfull()->GetOptions().ttl, 60);
  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ(NumTableFilesAtLevel(0), 0);
@@ -662,7 +918,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
  // Test dynamically changing compaction_options_fifo.max_table_files_size
  env_->addon_time_.store(0);
  options.compaction_options_fifo.max_table_files_size = 500 << 10;  // 00KB
-  options.compaction_options_fifo.ttl = 0;
+  options.ttl = 0;
  DestroyAndReopen(options);

  for (int i = 0; i < 10; i++) {
@@ -692,7 +948,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {

  // Test dynamically changing compaction_options_fifo.allow_compaction
  options.compaction_options_fifo.max_table_files_size = 500 << 10;  // 500KB
-  options.compaction_options_fifo.ttl = 0;
+  options.ttl = 0;
  options.compaction_options_fifo.allow_compaction = false;
  options.level0_file_num_compaction_trigger = 6;
  DestroyAndReopen(options);
@@ -245,15 +245,14 @@ void GetExpectedTableProperties(
  const int kDeletionCount = kTableCount * kDeletionsPerTable;
  const int kMergeCount = kTableCount * kMergeOperandsPerTable;
  const int kRangeDeletionCount = kTableCount * kRangeDeletionsPerTable;
-  const int kKeyCount = kPutCount + kDeletionCount + kMergeCount;
+  const int kKeyCount = kPutCount + kDeletionCount + kMergeCount + kRangeDeletionCount;
  const int kAvgSuccessorSize = kKeySize / 5;
  const int kEncodingSavePerKey = kKeySize / 4;
-  expected_tp->raw_key_size =
-      (kKeyCount + kRangeDeletionCount) * (kKeySize + 8);
+  expected_tp->raw_key_size = kKeyCount * (kKeySize + 8);
  expected_tp->raw_value_size =
      (kPutCount + kMergeCount + kRangeDeletionCount) * kValueSize;
  expected_tp->num_entries = kKeyCount;
-  expected_tp->num_deletions = kDeletionCount;
+  expected_tp->num_deletions = kDeletionCount + kRangeDeletionCount;
  expected_tp->num_merge_operands = kMergeCount;
  expected_tp->num_range_deletions = kRangeDeletionCount;
  expected_tp->num_data_blocks =
@@ -393,7 +392,15 @@ TEST_F(DBPropertiesTest, ReadLatencyHistogramByLevel) {
  options.target_file_size_base = 98 << 10;
  options.max_write_buffer_number = 2;
  options.statistics = rocksdb::CreateDBStatistics();
-  options.max_open_files = 100;
+  options.max_open_files = 11;  // Make sure no proloading of table readers
+
+  // RocksDB sanitize max open files to at least 20. Modify it back.
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "SanitizeOptions::AfterChangeMaxOpenFiles", [&](void* arg) {
+        int* max_open_files = static_cast<int*>(arg);
+        *max_open_files = 11;
+      });
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();

  BlockBasedTableOptions table_options;
  table_options.no_block_cache = true;
@@ -441,6 +448,7 @@ TEST_F(DBPropertiesTest, ReadLatencyHistogramByLevel) {

  // Reopen and issue iterating. See thee latency tracked
  ReopenWithColumnFamilies({"default", "pikachu"}, options);
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
  ASSERT_TRUE(dbfull()->GetProperty("rocksdb.cf-file-histogram", &prop));
  ASSERT_EQ(std::string::npos, prop.find("** Level 0 read latency histogram"));
  ASSERT_EQ(std::string::npos, prop.find("** Level 1 read latency histogram"));
@@ -1086,7 +1094,7 @@ class CountingUserTblPropCollector : public TablePropertiesCollector {
    return Status::OK();
  }

-  virtual UserCollectedProperties GetReadableProperties() const override {
+  UserCollectedProperties GetReadableProperties() const override {
    return UserCollectedProperties{};
  }

@@ -1102,7 +1110,7 @@ class CountingUserTblPropCollectorFactory
      uint32_t expected_column_family_id)
      : expected_column_family_id_(expected_column_family_id),
        num_created_(0) {}
-  virtual TablePropertiesCollector* CreateTablePropertiesCollector(
+  TablePropertiesCollector* CreateTablePropertiesCollector(
      TablePropertiesCollectorFactory::Context context) override {
    EXPECT_EQ(expected_column_family_id_, context.column_family_id);
    num_created_++;
@@ -1150,7 +1158,7 @@ class CountingDeleteTabPropCollector : public TablePropertiesCollector {
 class CountingDeleteTabPropCollectorFactory
    : public TablePropertiesCollectorFactory {
 public:
-  virtual TablePropertiesCollector* CreateTablePropertiesCollector(
+  TablePropertiesCollector* CreateTablePropertiesCollector(
      TablePropertiesCollectorFactory::Context /*context*/) override {
    return new CountingDeleteTabPropCollector();
  }
@@ -1424,7 +1432,7 @@ TEST_F(DBPropertiesTest, EstimateOldestKeyTime) {
  }

  options.compaction_style = kCompactionStyleFIFO;
-  options.compaction_options_fifo.ttl = 300;
+  options.ttl = 300;
  options.compaction_options_fifo.allow_compaction = false;
  DestroyAndReopen(options);

@@ -72,8 +72,8 @@ TEST_F(DBRangeDelTest, CompactionOutputHasOnlyRangeTombstone) {
    // Skip cuckoo memtables, which do not support snapshots. Skip non-leveled
    // compactions as the above assertions about the number of files in a level
    // do not hold true.
-  } while (ChangeOptions(kRangeDelSkipConfigs | kSkipHashCuckoo |
-                         kSkipUniversalCompaction | kSkipFIFOCompaction));
+  } while (ChangeOptions(kRangeDelSkipConfigs | kSkipUniversalCompaction |
+                         kSkipFIFOCompaction));
 }

 TEST_F(DBRangeDelTest, CompactionOutputFilesExactlyFilled) {
@@ -438,9 +438,10 @@ TEST_F(DBRangeDelTest, ValidUniversalSubcompactionBoundaries) {
  ASSERT_OK(dbfull()->RunManualCompaction(
      reinterpret_cast<ColumnFamilyHandleImpl*>(db_->DefaultColumnFamily())
          ->cfd(),
-      1 /* input_level */, 2 /* output_level */, 0 /* output_path_id */,
-      0 /* max_subcompactions */, nullptr /* begin */, nullptr /* end */,
-      true /* exclusive */, true /* disallow_trivial_move */));
+      1 /* input_level */, 2 /* output_level */, CompactRangeOptions(),
+      nullptr /* begin */, nullptr /* end */, true /* exclusive */,
+      true /* disallow_trivial_move */,
+      port::kMaxUint64 /* max_file_num_to_ignore */));
 }
 #endif  // ROCKSDB_LITE

@@ -490,6 +491,30 @@ TEST_F(DBRangeDelTest, CompactionRemovesCoveredMergeOperands) {
  ASSERT_EQ(expected, actual);
 }

+TEST_F(DBRangeDelTest, PutDeleteRangeMergeFlush) {
+  // Test the sequence of operations: (1) Put, (2) DeleteRange, (3) Merge, (4)
+  // Flush. The `CompactionIterator` previously had a bug where we forgot to
+  // check for covering range tombstones when processing the (1) Put, causing
+  // it to reappear after the flush.
+  Options opts = CurrentOptions();
+  opts.merge_operator = MergeOperators::CreateUInt64AddOperator();
+  Reopen(opts);
+
+  std::string val;
+  PutFixed64(&val, 1);
+  ASSERT_OK(db_->Put(WriteOptions(), "key", val));
+  ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                             "key", "key_"));
+  ASSERT_OK(db_->Merge(WriteOptions(), "key", val));
+  ASSERT_OK(db_->Flush(FlushOptions()));
+
+  ReadOptions read_opts;
+  std::string expected, actual;
+  ASSERT_OK(db_->Get(read_opts, "key", &actual));
+  PutFixed64(&expected, 1);
+  ASSERT_EQ(expected, actual);
+}
+
 // NumTableFilesAtLevel() is not supported in ROCKSDB_LITE
 #ifndef ROCKSDB_LITE
 TEST_F(DBRangeDelTest, ObsoleteTombstoneCleanup) {
@@ -645,8 +670,7 @@ TEST_F(DBRangeDelTest, GetCoveredKeyFromSst) {
    std::string value;
    ASSERT_TRUE(db_->Get(read_opts, "key", &value).IsNotFound());
    db_->ReleaseSnapshot(snapshot);
-    // Cuckoo memtables do not support snapshots.
-  } while (ChangeOptions(kRangeDelSkipConfigs | kSkipHashCuckoo));
+  } while (ChangeOptions(kRangeDelSkipConfigs));
 }

 TEST_F(DBRangeDelTest, GetCoveredMergeOperandFromMemtable) {
@@ -1041,11 +1065,16 @@ TEST_F(DBRangeDelTest, RangeTombstoneEndKeyAsSstableUpperBound) {
    //   L2:
    //     [key000000#1,1, key000000#1,1]
    //     [key000002#6,1, key000004#72057594037927935,15]
+    //
+    // At the same time, verify the compaction does not cause the key at the
+    // endpoint (key000002#6,1) to disappear.
+    ASSERT_EQ(value, Get(Key(2)));
    auto begin_str = Key(3);
    const rocksdb::Slice begin = begin_str;
    dbfull()->TEST_CompactRange(1, &begin, nullptr);
    ASSERT_EQ(1, NumTableFilesAtLevel(1));
    ASSERT_EQ(2, NumTableFilesAtLevel(2));
+    ASSERT_EQ(value, Get(Key(2)));
  }

  {
@@ -1107,14 +1136,14 @@ class MockMergeOperator : public MergeOperator {
  // Mock non-associative operator. Non-associativity is expressed by lack of
  // implementation for any `PartialMerge*` functions.
 public:
-  virtual bool FullMergeV2(const MergeOperationInput& merge_in,
-                           MergeOperationOutput* merge_out) const override {
+  bool FullMergeV2(const MergeOperationInput& merge_in,
+                   MergeOperationOutput* merge_out) const override {
    assert(merge_out != nullptr);
    merge_out->new_value = merge_in.operand_list.back().ToString();
    return true;
  }

-  virtual const char* Name() const override { return "MockMergeOperator"; }
+  const char* Name() const override { return "MockMergeOperator"; }
 };

 TEST_F(DBRangeDelTest, KeyAtOverlappingEndpointReappears) {
@@ -1133,6 +1162,14 @@ TEST_F(DBRangeDelTest, KeyAtOverlappingEndpointReappears) {
  options.target_file_size_base = kFileBytes;
  Reopen(options);

+  // Push dummy data to L3 so that our actual test files on L0-L2
+  // will not be considered "bottommost" level, otherwise compaction
+  // may prevent us from creating overlapping user keys
+  // as on the bottommost layer MergeHelper
+  ASSERT_OK(db_->Merge(WriteOptions(), "key", "dummy"));
+  ASSERT_OK(db_->Flush(FlushOptions()));
+  MoveFilesToLevel(3);
+
  Random rnd(301);
  const Snapshot* snapshot = nullptr;
  for (int i = 0; i < kNumFiles; ++i) {
@@ -1154,8 +1191,9 @@ TEST_F(DBRangeDelTest, KeyAtOverlappingEndpointReappears) {
  std::string value;
  ASSERT_TRUE(db_->Get(ReadOptions(), "key", &value).IsNotFound());

-  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr /* begin_key */,
-                              nullptr /* end_key */));
+  dbfull()->TEST_CompactRange(0 /* level */, nullptr /* begin */,
+                              nullptr /* end */, nullptr /* column_family */,
+                              true /* disallow_trivial_move */);
  ASSERT_EQ(0, NumTableFilesAtLevel(0));
  // Now we have multiple files at L1 all containing a single user key, thus
  // guaranteeing overlap in the file endpoints.
@@ -0,0 +1,524 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file. See the AUTHORS file for names of contributors.
+
+#include "db/db_impl_secondary.h"
+#include "db/db_test_util.h"
+#include "port/stack_trace.h"
+#include "util/fault_injection_test_env.h"
+#include "util/sync_point.h"
+
+namespace rocksdb {
+
+#ifndef ROCKSDB_LITE
+class DBSecondaryTest : public DBTestBase {
+ public:
+  DBSecondaryTest()
+      : DBTestBase("/db_secondary_test"),
+        secondary_path_(),
+        handles_secondary_(),
+        db_secondary_(nullptr) {
+    secondary_path_ =
+        test::PerThreadDBPath(env_, "/db_secondary_test_secondary");
+  }
+
+  ~DBSecondaryTest() override {
+    CloseSecondary();
+    if (getenv("KEEP_DB") != nullptr) {
+      fprintf(stdout, "Secondary DB is still at %s\n", secondary_path_.c_str());
+    } else {
+      Options options;
+      options.env = env_;
+      EXPECT_OK(DestroyDB(secondary_path_, options));
+    }
+  }
+
+ protected:
+  Status ReopenAsSecondary(const Options& options) {
+    return DB::OpenAsSecondary(options, dbname_, secondary_path_, &db_);
+  }
+
+  void OpenSecondary(const Options& options);
+
+  void OpenSecondaryWithColumnFamilies(
+      const std::vector<std::string>& column_families, const Options& options);
+
+  void CloseSecondary() {
+    for (auto h : handles_secondary_) {
+      db_secondary_->DestroyColumnFamilyHandle(h);
+    }
+    handles_secondary_.clear();
+    delete db_secondary_;
+    db_secondary_ = nullptr;
+  }
+
+  DBImplSecondary* db_secondary_full() {
+    return static_cast<DBImplSecondary*>(db_secondary_);
+  }
+
+  void CheckFileTypeCounts(const std::string& dir, int expected_log,
+                           int expected_sst, int expected_manifest) const;
+
+  std::string secondary_path_;
+  std::vector<ColumnFamilyHandle*> handles_secondary_;
+  DB* db_secondary_;
+};
+
+void DBSecondaryTest::OpenSecondary(const Options& options) {
+  Status s =
+      DB::OpenAsSecondary(options, dbname_, secondary_path_, &db_secondary_);
+  ASSERT_OK(s);
+}
+
+void DBSecondaryTest::OpenSecondaryWithColumnFamilies(
+    const std::vector<std::string>& column_families, const Options& options) {
+  std::vector<ColumnFamilyDescriptor> cf_descs;
+  cf_descs.emplace_back(kDefaultColumnFamilyName, options);
+  for (const auto& cf_name : column_families) {
+    cf_descs.emplace_back(cf_name, options);
+  }
+  Status s = DB::OpenAsSecondary(options, dbname_, secondary_path_, cf_descs,
+                                 &handles_secondary_, &db_secondary_);
+  ASSERT_OK(s);
+}
+
+void DBSecondaryTest::CheckFileTypeCounts(const std::string& dir,
+                                          int expected_log, int expected_sst,
+                                          int expected_manifest) const {
+  std::vector<std::string> filenames;
+  env_->GetChildren(dir, &filenames);
+
+  int log_cnt = 0, sst_cnt = 0, manifest_cnt = 0;
+  for (auto file : filenames) {
+    uint64_t number;
+    FileType type;
+    if (ParseFileName(file, &number, &type)) {
+      log_cnt += (type == kLogFile);
+      sst_cnt += (type == kTableFile);
+      manifest_cnt += (type == kDescriptorFile);
+    }
+  }
+  ASSERT_EQ(expected_log, log_cnt);
+  ASSERT_EQ(expected_sst, sst_cnt);
+  ASSERT_EQ(expected_manifest, manifest_cnt);
+}
+
+TEST_F(DBSecondaryTest, ReopenAsSecondary) {
+  Options options;
+  options.env = env_;
+  Reopen(options);
+  ASSERT_OK(Put("foo", "foo_value"));
+  ASSERT_OK(Put("bar", "bar_value"));
+  ASSERT_OK(dbfull()->Flush(FlushOptions()));
+  Close();
+
+  ASSERT_OK(ReopenAsSecondary(options));
+  ASSERT_EQ("foo_value", Get("foo"));
+  ASSERT_EQ("bar_value", Get("bar"));
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  auto db1 = static_cast<DBImplSecondary*>(db_);
+  ASSERT_NE(nullptr, db1);
+  Iterator* iter = db1->NewIterator(ropts);
+  ASSERT_NE(nullptr, iter);
+  size_t count = 0;
+  for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+    if (0 == count) {
+      ASSERT_EQ("bar", iter->key().ToString());
+      ASSERT_EQ("bar_value", iter->value().ToString());
+    } else if (1 == count) {
+      ASSERT_EQ("foo", iter->key().ToString());
+      ASSERT_EQ("foo_value", iter->value().ToString());
+    }
+    ++count;
+  }
+  delete iter;
+  ASSERT_EQ(2, count);
+}
+
+TEST_F(DBSecondaryTest, OpenAsSecondary) {
+  Options options;
+  options.env = env_;
+  options.level0_file_num_compaction_trigger = 4;
+  Reopen(options);
+  for (int i = 0; i < 3; ++i) {
+    ASSERT_OK(Put("foo", "foo_value" + std::to_string(i)));
+    ASSERT_OK(Put("bar", "bar_value" + std::to_string(i)));
+    ASSERT_OK(Flush());
+  }
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  const auto verify_db_func = [&](const std::string& foo_val,
+                                  const std::string& bar_val) {
+    std::string value;
+    ASSERT_OK(db_secondary_->Get(ropts, "foo", &value));
+    ASSERT_EQ(foo_val, value);
+    ASSERT_OK(db_secondary_->Get(ropts, "bar", &value));
+    ASSERT_EQ(bar_val, value);
+    Iterator* iter = db_secondary_->NewIterator(ropts);
+    ASSERT_NE(nullptr, iter);
+    iter->Seek("foo");
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ("foo", iter->key().ToString());
+    ASSERT_EQ(foo_val, iter->value().ToString());
+    iter->Seek("bar");
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ("bar", iter->key().ToString());
+    ASSERT_EQ(bar_val, iter->value().ToString());
+    size_t count = 0;
+    for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+      ++count;
+    }
+    ASSERT_EQ(2, count);
+    delete iter;
+  };
+
+  verify_db_func("foo_value2", "bar_value2");
+
+  ASSERT_OK(Put("foo", "new_foo_value"));
+  ASSERT_OK(Put("bar", "new_bar_value"));
+  ASSERT_OK(Flush());
+
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  verify_db_func("new_foo_value", "new_bar_value");
+}
+
+TEST_F(DBSecondaryTest, OpenAsSecondaryWALTailing) {
+  Options options;
+  options.env = env_;
+  options.level0_file_num_compaction_trigger = 4;
+  Reopen(options);
+  for (int i = 0; i < 3; ++i) {
+    ASSERT_OK(Put("foo", "foo_value" + std::to_string(i)));
+    ASSERT_OK(Put("bar", "bar_value" + std::to_string(i)));
+  }
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  const auto verify_db_func = [&](const std::string& foo_val,
+                                  const std::string& bar_val) {
+    std::string value;
+    ASSERT_OK(db_secondary_->Get(ropts, "foo", &value));
+    ASSERT_EQ(foo_val, value);
+    ASSERT_OK(db_secondary_->Get(ropts, "bar", &value));
+    ASSERT_EQ(bar_val, value);
+    Iterator* iter = db_secondary_->NewIterator(ropts);
+    ASSERT_NE(nullptr, iter);
+    iter->Seek("foo");
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ("foo", iter->key().ToString());
+    ASSERT_EQ(foo_val, iter->value().ToString());
+    iter->Seek("bar");
+    ASSERT_TRUE(iter->Valid());
+    ASSERT_EQ("bar", iter->key().ToString());
+    ASSERT_EQ(bar_val, iter->value().ToString());
+    size_t count = 0;
+    for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+      ++count;
+    }
+    ASSERT_EQ(2, count);
+    delete iter;
+  };
+
+  verify_db_func("foo_value2", "bar_value2");
+}
+
+TEST_F(DBSecondaryTest, OpenWithNonExistColumnFamily) {
+  Options options;
+  options.env = env_;
+  CreateAndReopenWithCF({"pikachu"}, options);
+
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  std::vector<ColumnFamilyDescriptor> cf_descs;
+  cf_descs.emplace_back(kDefaultColumnFamilyName, options1);
+  cf_descs.emplace_back("pikachu", options1);
+  cf_descs.emplace_back("eevee", options1);
+  Status s = DB::OpenAsSecondary(options1, dbname_, secondary_path_, cf_descs,
+                                 &handles_secondary_, &db_secondary_);
+  ASSERT_NOK(s);
+}
+
+TEST_F(DBSecondaryTest, OpenWithSubsetOfColumnFamilies) {
+  Options options;
+  options.env = env_;
+  CreateAndReopenWithCF({"pikachu"}, options);
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+  ASSERT_EQ(0, handles_secondary_.size());
+  ASSERT_NE(nullptr, db_secondary_);
+
+  ASSERT_OK(Put(0 /*cf*/, "foo", "foo_value"));
+  ASSERT_OK(Put(1 /*cf*/, "foo", "foo_value"));
+  ASSERT_OK(Flush(0 /*cf*/));
+  ASSERT_OK(Flush(1 /*cf*/));
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  std::string value;
+  ASSERT_OK(db_secondary_->Get(ropts, "foo", &value));
+  ASSERT_EQ("foo_value", value);
+}
+
+TEST_F(DBSecondaryTest, SwitchToNewManifestDuringOpen) {
+  Options options;
+  options.env = env_;
+  Reopen(options);
+  Close();
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  SyncPoint::GetInstance()->LoadDependency(
+      {{"ReactiveVersionSet::MaybeSwitchManifest:AfterGetCurrentManifestPath:0",
+        "VersionSet::ProcessManifestWrites:BeforeNewManifest"},
+       {"VersionSet::ProcessManifestWrites:AfterNewManifest",
+        "ReactiveVersionSet::MaybeSwitchManifest:AfterGetCurrentManifestPath:"
+        "1"}});
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  // Make sure db calls RecoverLogFiles so as to trigger a manifest write,
+  // which causes the db to switch to a new MANIFEST upon start.
+  port::Thread ro_db_thread([&]() {
+    Options options1;
+    options1.env = env_;
+    options1.max_open_files = -1;
+    OpenSecondary(options1);
+    CloseSecondary();
+  });
+  Reopen(options);
+  ro_db_thread.join();
+}
+
+TEST_F(DBSecondaryTest, MissingTableFileDuringOpen) {
+  Options options;
+  options.env = env_;
+  options.level0_file_num_compaction_trigger = 4;
+  Reopen(options);
+  for (int i = 0; i != options.level0_file_num_compaction_trigger; ++i) {
+    ASSERT_OK(Put("foo", "foo_value" + std::to_string(i)));
+    ASSERT_OK(Put("bar", "bar_value" + std::to_string(i)));
+    ASSERT_OK(dbfull()->Flush(FlushOptions()));
+  }
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  std::string value;
+  ASSERT_OK(db_secondary_->Get(ropts, "foo", &value));
+  ASSERT_EQ("foo_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            value);
+  ASSERT_OK(db_secondary_->Get(ropts, "bar", &value));
+  ASSERT_EQ("bar_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            value);
+  Iterator* iter = db_secondary_->NewIterator(ropts);
+  ASSERT_NE(nullptr, iter);
+  iter->Seek("bar");
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ("bar", iter->key().ToString());
+  ASSERT_EQ("bar_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            iter->value().ToString());
+  iter->Seek("foo");
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ("foo", iter->key().ToString());
+  ASSERT_EQ("foo_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            iter->value().ToString());
+  size_t count = 0;
+  for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+    ++count;
+  }
+  ASSERT_EQ(2, count);
+  delete iter;
+}
+
+TEST_F(DBSecondaryTest, MissingTableFile) {
+  int table_files_not_exist = 0;
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  SyncPoint::GetInstance()->SetCallBack(
+      "ReactiveVersionSet::ReadAndApply:AfterLoadTableHandlers",
+      [&](void* arg) {
+        Status s = *reinterpret_cast<Status*>(arg);
+        if (s.IsPathNotFound()) {
+          ++table_files_not_exist;
+        } else if (!s.ok()) {
+          assert(false);  // Should not reach here
+        }
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+  Options options;
+  options.env = env_;
+  options.level0_file_num_compaction_trigger = 4;
+  Reopen(options);
+
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+
+  for (int i = 0; i != options.level0_file_num_compaction_trigger; ++i) {
+    ASSERT_OK(Put("foo", "foo_value" + std::to_string(i)));
+    ASSERT_OK(Put("bar", "bar_value" + std::to_string(i)));
+    ASSERT_OK(dbfull()->Flush(FlushOptions()));
+  }
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  ASSERT_NE(nullptr, db_secondary_full());
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  std::string value;
+  ASSERT_NOK(db_secondary_->Get(ropts, "foo", &value));
+  ASSERT_NOK(db_secondary_->Get(ropts, "bar", &value));
+
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  ASSERT_EQ(options.level0_file_num_compaction_trigger, table_files_not_exist);
+  ASSERT_OK(db_secondary_->Get(ropts, "foo", &value));
+  ASSERT_EQ("foo_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            value);
+  ASSERT_OK(db_secondary_->Get(ropts, "bar", &value));
+  ASSERT_EQ("bar_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            value);
+  Iterator* iter = db_secondary_->NewIterator(ropts);
+  ASSERT_NE(nullptr, iter);
+  iter->Seek("bar");
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ("bar", iter->key().ToString());
+  ASSERT_EQ("bar_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            iter->value().ToString());
+  iter->Seek("foo");
+  ASSERT_TRUE(iter->Valid());
+  ASSERT_EQ("foo", iter->key().ToString());
+  ASSERT_EQ("foo_value" +
+                std::to_string(options.level0_file_num_compaction_trigger - 1),
+            iter->value().ToString());
+  size_t count = 0;
+  for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+    ++count;
+  }
+  ASSERT_EQ(2, count);
+  delete iter;
+}
+
+TEST_F(DBSecondaryTest, PrimaryDropColumnFamily) {
+  Options options;
+  options.env = env_;
+  const std::string kCfName1 = "pikachu";
+  CreateAndReopenWithCF({kCfName1}, options);
+
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondaryWithColumnFamilies({kCfName1}, options1);
+  ASSERT_EQ(2, handles_secondary_.size());
+
+  ASSERT_OK(Put(1 /*cf*/, "foo", "foo_val_1"));
+  ASSERT_OK(Flush(1 /*cf*/));
+
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  ReadOptions ropts;
+  ropts.verify_checksums = true;
+  std::string value;
+  ASSERT_OK(db_secondary_->Get(ropts, handles_secondary_[1], "foo", &value));
+  ASSERT_EQ("foo_val_1", value);
+
+  ASSERT_OK(dbfull()->DropColumnFamily(handles_[1]));
+  Close();
+  CheckFileTypeCounts(dbname_, 1, 0, 1);
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  value.clear();
+  ASSERT_OK(db_secondary_->Get(ropts, handles_secondary_[1], "foo", &value));
+  ASSERT_EQ("foo_val_1", value);
+}
+
+TEST_F(DBSecondaryTest, SwitchManifest) {
+  Options options;
+  options.env = env_;
+  options.level0_file_num_compaction_trigger = 4;
+  Reopen(options);
+
+  Options options1;
+  options1.env = env_;
+  options1.max_open_files = -1;
+  OpenSecondary(options1);
+
+  const int kNumFiles = options.level0_file_num_compaction_trigger - 1;
+  // Keep it smaller than 10 so that key0, key1, ..., key9 are sorted as 0, 1,
+  // ..., 9.
+  const int kNumKeys = 10;
+  // Create two sst
+  for (int i = 0; i != kNumFiles; ++i) {
+    for (int j = 0; j != kNumKeys; ++j) {
+      ASSERT_OK(Put("key" + std::to_string(j), "value_" + std::to_string(i)));
+    }
+    ASSERT_OK(Flush());
+  }
+
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  const auto& range_scan_db = [&]() {
+    ReadOptions tmp_ropts;
+    tmp_ropts.total_order_seek = true;
+    tmp_ropts.verify_checksums = true;
+    std::unique_ptr<Iterator> iter(db_secondary_->NewIterator(tmp_ropts));
+    int cnt = 0;
+    for (iter->SeekToFirst(); iter->Valid(); iter->Next(), ++cnt) {
+      ASSERT_EQ("key" + std::to_string(cnt), iter->key().ToString());
+      ASSERT_EQ("value_" + std::to_string(kNumFiles - 1),
+                iter->value().ToString());
+    }
+  };
+
+  range_scan_db();
+
+  // While secondary instance still keeps old MANIFEST open, we close primary,
+  // restart primary, performs full compaction, close again, restart again so
+  // that next time secondary tries to catch up with primary, the secondary
+  // will skip the MANIFEST in middle.
+  Reopen(options);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  Reopen(options);
+  ASSERT_OK(dbfull()->SetOptions({{"disable_auto_compactions", "false"}}));
+
+  ASSERT_OK(db_secondary_->TryCatchUpWithPrimary());
+  range_scan_db();
+}
+#endif  //! ROCKSDB_LITE
+
+}  // namespace rocksdb
+
+int main(int argc, char** argv) {
+  rocksdb::port::InstallStackTraceHandler();
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}
@@ -25,9 +25,9 @@ class DBSSTTest : public DBTestBase {
 class FlushedFileCollector : public EventListener {
 public:
  FlushedFileCollector() {}
-  ~FlushedFileCollector() {}
+  ~FlushedFileCollector() override {}

-  virtual void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
+  void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
    std::lock_guard<std::mutex> lock(mutex_);
    flushed_files_.push_back(info.file_path);
  }
@@ -356,6 +356,14 @@ TEST_F(DBSSTTest, RateLimitedDelete) {
  env_->time_elapse_only_sleep_ = true;
  Options options = CurrentOptions();
  options.disable_auto_compactions = true;
+  // Need to disable stats dumping and persisting which also use
+  // RepeatableThread, one of whose member variables is of type
+  // InstrumentedCondVar. The callback for
+  // InstrumentedCondVar::TimedWaitInternal can be triggered by stats dumping
+  // and persisting threads and cause time_spent_deleting measurement to become
+  // incorrect.
+  options.stats_dump_period_sec = 0;
+  options.stats_persist_period_sec = 0;
  options.env = env_;

  int64_t rate_bytes_per_sec = 1024 * 10;  // 10 Kbs / Sec
@@ -367,14 +375,16 @@ TEST_F(DBSSTTest, RateLimitedDelete) {
  auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());
  sfm->delete_scheduler()->SetMaxTrashDBRatio(1.1);

+  WriteOptions wo;
+  wo.disableWAL = true;
  ASSERT_OK(TryReopen(options));
  // Create 4 files in L0
  for (char v = 'a'; v <= 'd'; v++) {
-    ASSERT_OK(Put("Key2", DummyString(1024, v)));
-    ASSERT_OK(Put("Key3", DummyString(1024, v)));
-    ASSERT_OK(Put("Key4", DummyString(1024, v)));
-    ASSERT_OK(Put("Key1", DummyString(1024, v)));
-    ASSERT_OK(Put("Key4", DummyString(1024, v)));
+    ASSERT_OK(Put("Key2", DummyString(1024, v), wo));
+    ASSERT_OK(Put("Key3", DummyString(1024, v), wo));
+    ASSERT_OK(Put("Key4", DummyString(1024, v), wo));
+    ASSERT_OK(Put("Key1", DummyString(1024, v), wo));
+    ASSERT_OK(Put("Key4", DummyString(1024, v), wo));
    ASSERT_OK(Flush());
  }
  // We created 4 sst files in L0
@@ -385,6 +395,7 @@ TEST_F(DBSSTTest, RateLimitedDelete) {

  // Compaction will move the 4 files in L0 to trash and create 1 L1 file
  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
  ASSERT_EQ("0,1", FilesPerLevel(0));

  uint64_t delete_start_time = env_->NowMicros();
@@ -407,6 +418,57 @@ TEST_F(DBSSTTest, RateLimitedDelete) {
  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
 }

+TEST_F(DBSSTTest, RateLimitedWALDelete) {
+  Destroy(last_options_);
+
+  std::vector<uint64_t> penalties;
+  rocksdb::SyncPoint::GetInstance()->SetCallBack(
+      "DeleteScheduler::BackgroundEmptyTrash:Wait",
+      [&](void* arg) { penalties.push_back(*(static_cast<uint64_t*>(arg))); });
+
+  env_->no_slowdown_ = true;
+  env_->time_elapse_only_sleep_ = true;
+  Options options = CurrentOptions();
+  options.disable_auto_compactions = true;
+  options.env = env_;
+
+  int64_t rate_bytes_per_sec = 1024 * 10;  // 10 Kbs / Sec
+  Status s;
+  options.sst_file_manager.reset(
+      NewSstFileManager(env_, nullptr, "", 0, false, &s, 0));
+  ASSERT_OK(s);
+  options.sst_file_manager->SetDeleteRateBytesPerSecond(rate_bytes_per_sec);
+  auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());
+  sfm->delete_scheduler()->SetMaxTrashDBRatio(2.1);
+
+  ASSERT_OK(TryReopen(options));
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  // Create 4 files in L0
+  for (char v = 'a'; v <= 'd'; v++) {
+    ASSERT_OK(Put("Key2", DummyString(1024, v)));
+    ASSERT_OK(Put("Key3", DummyString(1024, v)));
+    ASSERT_OK(Put("Key4", DummyString(1024, v)));
+    ASSERT_OK(Put("Key1", DummyString(1024, v)));
+    ASSERT_OK(Put("Key4", DummyString(1024, v)));
+    ASSERT_OK(Flush());
+  }
+  // We created 4 sst files in L0
+  ASSERT_EQ("4", FilesPerLevel(0));
+
+  // Compaction will move the 4 files in L0 to trash and create 1 L1 file
+  CompactRangeOptions cro;
+  cro.bottommost_level_compaction = BottommostLevelCompaction::kForce;
+  ASSERT_OK(db_->CompactRange(cro, nullptr, nullptr));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_EQ("0,1", FilesPerLevel(0));
+
+  sfm->WaitForEmptyTrash();
+  ASSERT_EQ(penalties.size(), 8);
+
+  rocksdb::SyncPoint::GetInstance()->DisableProcessing();
+}
+
 TEST_F(DBSSTTest, OpenDBWithExistingTrash) {
  Options options = CurrentOptions();

@@ -445,7 +507,6 @@ TEST_F(DBSSTTest, DeleteSchedulerMultipleDBPaths) {
  rocksdb::SyncPoint::GetInstance()->SetCallBack(
      "DeleteScheduler::DeleteFile",
      [&](void* /*arg*/) { bg_delete_file++; });
-  rocksdb::SyncPoint::GetInstance()->EnableProcessing();

  Options options = CurrentOptions();
  options.disable_auto_compactions = true;
@@ -463,10 +524,14 @@ TEST_F(DBSSTTest, DeleteSchedulerMultipleDBPaths) {
  auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());

  DestroyAndReopen(options);
+  rocksdb::SyncPoint::GetInstance()->EnableProcessing();
+
+  WriteOptions wo;
+  wo.disableWAL = true;

  // Create 4 files in L0
  for (int i = 0; i < 4; i++) {
-    ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'A')));
+    ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'A'), wo));
    ASSERT_OK(Flush());
  }
  // We created 4 sst files in L0
@@ -482,7 +547,7 @@ TEST_F(DBSSTTest, DeleteSchedulerMultipleDBPaths) {

  // Create 4 files in L0
  for (int i = 4; i < 8; i++) {
-    ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'B')));
+    ASSERT_OK(Put("Key" + ToString(i), DummyString(1024, 'B'), wo));
    ASSERT_OK(Flush());
  }
  ASSERT_EQ("4,1", FilesPerLevel(0));
@@ -500,7 +565,7 @@ TEST_F(DBSSTTest, DeleteSchedulerMultipleDBPaths) {
  // Compaction will delete both files and regenerate a file in L1 in second
  // db path. The deleted files should still be cleaned up via delete scheduler.
  compact_options.bottommost_level_compaction =
-      BottommostLevelCompaction::kForce;
+      BottommostLevelCompaction::kForceOptimized;
  ASSERT_OK(db_->CompactRange(compact_options, nullptr, nullptr));
  ASSERT_EQ("0,1", FilesPerLevel(0));

@@ -537,14 +602,27 @@ TEST_F(DBSSTTest, DestroyDBWithRateLimitedDelete) {
  // Close DB and destroy it using DeleteScheduler
  Close();

+  int num_sst_files = 0;
+  int num_wal_files = 0;
+  std::vector<std::string> db_files;
+  env_->GetChildren(dbname_, &db_files);
+  for (std::string f : db_files) {
+    if (f.substr(f.find_last_of(".") + 1) == "sst") {
+      num_sst_files++;
+    } else if (f.substr(f.find_last_of(".") + 1) == "log") {
+      num_wal_files++;
+    }
+  }
+  ASSERT_GT(num_sst_files, 0);
+  ASSERT_GT(num_wal_files, 0);
+
  auto sfm = static_cast<SstFileManagerImpl*>(options.sst_file_manager.get());

  sfm->SetDeleteRateBytesPerSecond(1024 * 1024);
  sfm->delete_scheduler()->SetMaxTrashDBRatio(1.1);
  ASSERT_OK(DestroyDB(dbname_, options));
  sfm->WaitForEmptyTrash();
-  // We have deleted the 4 sst files in the delete_scheduler
-  ASSERT_EQ(bg_delete_file, 4);
+  ASSERT_EQ(bg_delete_file, num_sst_files + num_wal_files);
 }

 TEST_F(DBSSTTest, DBWithMaxSpaceAllowed) {
@@ -46,7 +46,7 @@ TEST_F(DBStatisticsTest, CompressionStatsTest) {
  Options options = CurrentOptions();
  options.compression = type;
  options.statistics = rocksdb::CreateDBStatistics();
-  options.statistics->stats_level_ = StatsLevel::kExceptTimeForMutex;
+  options.statistics->set_stats_level(StatsLevel::kExceptTimeForMutex);
  DestroyAndReopen(options);

  int kNumKeysWritten = 100000;
@@ -105,7 +105,7 @@ TEST_F(DBStatisticsTest, MutexWaitStats) {
  Options options = CurrentOptions();
  options.create_if_missing = true;
  options.statistics = rocksdb::CreateDBStatistics();
-  options.statistics->stats_level_ = StatsLevel::kAll;
+  options.statistics->set_stats_level(StatsLevel::kAll);
  CreateAndReopenWithCF({"pikachu"}, options);
  const uint64_t kMutexWaitDelay = 100;
  ThreadStatusUtil::TEST_SetStateDelay(ThreadStatus::STATE_MUTEX_WAIT,