Update version.h to 7.10.2

Update HISTORY.md for 7.10.2
Fix bug in WAL streaming uncompression (#11198 )
2026-11-07 03:22:34 -05:00 · 2023-02-23 10:00:36 -08:00 · 2023-02-10 14:59:29 -08:00 · 2023-02-10 14:58:13 -08:00 · 2023-02-02 17:28:08 -08:00 · 2023-02-02 08:51:12 -08:00
629 changed files with 10843 additions and 10928 deletions
@@ -152,15 +152,15 @@ commands:
    steps:
      - run:
          name: "Test low-variance benchmarks"
-          command: ./tools/benchmark_ci.py --db_dir /tmp/rocksdb-benchmark-datadir --output_dir /tmp/benchmark-results --num_keys 20000000
+          command: ./tools/benchmark_ci.py --db_dir /tmp/rocksdb-benchmark-datadir --output_dir /tmp/benchmark-results --num_keys 10000000
          environment:
            LD_LIBRARY_PATH: /usr/local/lib
            # How long to run parts of the test(s)
-            DURATION_RO: 300
-            DURATION_RW: 500
+            DURATION_RO: 400
+            DURATION_RW: 700
            # Keep threads within physical capacity of server (much lower than default)
            NUM_THREADS: 1
-            MAX_BACKGROUND_JOBS: 4
+            MAX_BACKGROUND_JOBS: 3
            # Don't run a couple of "optional" initial tests
            CI_TESTS_ONLY: "true"
            # Reduce configured size of levels to ensure more levels in the leveled compaction LSM tree
@@ -170,11 +170,7 @@ commands:
            # The benchmark host has 32GB memory
            # The following values are tailored to work with that
            # Note, tests may not exercise the targeted issues if the memory is increased on new test hosts.
-            COMPRESSION_TYPE: "none"
-            CACHE_INDEX_AND_FILTER_BLOCKS: 1
-            MIN_LEVEL_TO_COMPRESS: 3
-            CACHE_SIZE_MB: 10240
-            MB_WRITE_PER_SEC: 2
+

  post-benchmarks:
    steps:
@@ -273,12 +269,12 @@ jobs:
          ./sst_dump --help | grep -E -q 'Supported compression types: kNoCompression$' # Verify no compiled in compression
      - post-steps

-  build-linux-static_lib-alt_namespace-status_checked:
+  build-linux-shared_lib-alt_namespace-status_checked:
    executor: linux-docker
    resource_class: 2xlarge
    steps:
      - pre-steps
-      - run: ASSERT_STATUS_CHECKED=1 TEST_UINT128_COMPAT=1 ROCKSDB_MODIFY_NPHASH=1 LIB_MODE=static OPT="-DROCKSDB_NAMESPACE=alternative_rocksdb_ns" make V=1 -j24 check
+      - run: ASSERT_STATUS_CHECKED=1 TEST_UINT128_COMPAT=1 ROCKSDB_MODIFY_NPHASH=1 LIB_MODE=shared OPT="-DROCKSDB_NAMESPACE=alternative_rocksdb_ns" make V=1 -j32 check
      - post-steps

  build-linux-release:
@@ -286,21 +282,11 @@ jobs:
    resource_class: 2xlarge
    steps:
      - checkout # check out the code in the project directory
-      - run: make V=1 -j32 LIB_MODE=shared release
-      - run: ls librocksdb.so # ensure shared lib built
-      - run: ./db_stress --version # ensure with gflags
-      - run: make clean
      - run: make V=1 -j32 release
-      - run: ls librocksdb.a # ensure static lib built
      - run: ./db_stress --version # ensure with gflags
      - run: make clean
      - run: apt-get remove -y libgflags-dev
-      - run: make V=1 -j32 LIB_MODE=shared release
-      - run: ls librocksdb.so # ensure shared lib built
-      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
-      - run: make clean
      - run: make V=1 -j32 release
-      - run: ls librocksdb.a # ensure static lib built
      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
      - post-steps

@@ -316,6 +302,27 @@ jobs:
      - run: USE_RTTI=1 DEBUG_LEVEL=0 make V=1 -j16 static_lib tools db_bench
      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags

+  build-linux-lite:
+    executor: linux-docker
+    resource_class: large
+    steps:
+      - pre-steps
+      - run: LITE=1 make V=1 J=8 -j8 check
+      - post-steps
+
+  build-linux-lite-release:
+    executor: linux-docker
+    resource_class: large
+    steps:
+      - checkout # check out the code in the project directory
+      - run: LITE=1 make V=1 -j8 release
+      - run: ./db_stress --version # ensure with gflags
+      - run: make clean
+      - run: apt-get remove -y libgflags-dev
+      - run: LITE=1 make V=1 -j8 release
+      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
+      - post-steps
+
  build-linux-clang-no_test_run:
    executor: linux-docker
    resource_class: xlarge
@@ -420,10 +427,7 @@ jobs:
    steps:
      - checkout # check out the code in the project directory
      - run: apt-get update -y && apt-get install -y libgflags-dev
-      - run:
-          name: "Unity build"
-          command: make V=1 -j8 unity_test
-          no_output_timeout: 20m
+      - run: make V=1 -j8 unity_test
      - run: make V=1 -j8 -k check-headers # could be moved to a different build
      - post-steps

@@ -434,7 +438,7 @@ jobs:
      - pre-steps
      - setup-folly
      - build-folly
-      - run: USE_FOLLY=1 LIB_MODE=static CC=gcc-7 CXX=g++-7 V=1 make -j32 check # TODO: LIB_MODE only to work around unresolved linker failures
+      - run: USE_FOLLY=1 CC=gcc-7 CXX=g++-7 V=1 make -j32 check
      - post-steps

  build-linux-gcc-7-with-folly-lite-no-test:
@@ -480,7 +484,7 @@ jobs:
    resource_class: 2xlarge
    steps:
      - pre-steps
-      - run: LIB_MODE=static CC=gcc-11 CXX=g++-11 V=1 make -j32 all microbench # TODO: LIB_MODE only to work around unresolved linker failures
+      - run: CC=gcc-11 CXX=g++-11 V=1 make -j32 all microbench
      - post-steps

  build-linux-clang-13-no_test_run:
@@ -499,7 +503,7 @@ jobs:
      - pre-steps
      - setup-folly
      - build-folly
-      - run: CC=clang-13 CXX=clang++-13 LIB_MODE=static USE_CLANG=1 USE_FOLLY=1 COMPILE_WITH_UBSAN=1 COMPILE_WITH_ASAN=1 make -j32 check # TODO: LIB_MODE only to work around unresolved linker failures
+      - run: CC=clang-13 CXX=clang++-13 USE_CLANG=1 USE_FOLLY=1 COMPILE_WITH_UBSAN=1 COMPILE_WITH_ASAN=1 make -j32 check
      - post-steps

  # This job is only to make sure the microbench tests are able to run, the benchmark result is not meaningful as the CI host is changing.
@@ -516,7 +520,7 @@ jobs:
    resource_class: large
    steps:
      - pre-steps
-      - run: ulimit -S -n `ulimit -H -n` && make V=1 -j8 CRASH_TEST_EXT_ARGS='--duration=960 --max_key=2500000 --use_io_uring=0' blackbox_crash_test_with_atomic_flush
+      - run: ulimit -S -n `ulimit -H -n` && make V=1 -j8 CRASH_TEST_EXT_ARGS='--duration=960 --max_key=2500000' blackbox_crash_test_with_atomic_flush
      - post-steps

  build-linux-crashtest-tiered-storage-bb:
@@ -526,7 +530,7 @@ jobs:
      - pre-steps
      - run:
          name: "run crashtest"
-          command: ulimit -S -n `ulimit -H -n` && make V=1 -j32 CRASH_TEST_EXT_ARGS='--duration=10800 --use_io_uring=0' blackbox_crash_test_with_tiered_storage
+          command: ulimit -S -n `ulimit -H -n` && make V=1 -j32 CRASH_TEST_EXT_ARGS=--duration=10800 blackbox_crash_test_with_tiered_storage
          no_output_timeout: 100m
      - post-steps

@@ -537,7 +541,7 @@ jobs:
      - pre-steps
      - run:
          name: "run crashtest"
-          command: ulimit -S -n `ulimit -H -n` && make V=1 -j32 CRASH_TEST_EXT_ARGS='--duration=10800 --use_io_uring=0' whitebox_crash_test_with_tiered_storage
+          command: ulimit -S -n `ulimit -H -n` && make V=1 -j32 CRASH_TEST_EXT_ARGS=--duration=10800 whitebox_crash_test_with_tiered_storage
          no_output_timeout: 100m
      - post-steps

@@ -817,16 +821,18 @@ workflows:
      - build-linux-cmake-with-folly-coroutines
      - build-linux-cmake-with-benchmark
      - build-linux-encrypted_env-no_compression
+      - build-linux-lite
  jobs-linux-run-tests-san:
    jobs:
      - build-linux-clang10-asan
      - build-linux-clang10-ubsan
      - build-linux-clang10-mini-tsan
-      - build-linux-static_lib-alt_namespace-status_checked
+      - build-linux-shared_lib-alt_namespace-status_checked
  jobs-linux-no-test-run:
    jobs:
      - build-linux-release
      - build-linux-release-rtti
+      - build-linux-lite-release
      - build-examples
      - build-fuzzers
      - build-linux-clang-no_test_run
@@ -496,6 +496,12 @@ if(CMAKE_COMPILER_IS_GNUCXX)
  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-builtin-memcmp")
 endif()

+option(ROCKSDB_LITE "Build RocksDBLite version" OFF)
+if(ROCKSDB_LITE)
+  add_definitions(-DROCKSDB_LITE)
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions -Os")
+endif()
+
 if(CMAKE_SYSTEM_NAME MATCHES "Cygwin")
  add_definitions(-fno-builtin-memcmp -DCYGWIN)
 elseif(CMAKE_SYSTEM_NAME MATCHES "Darwin")
@@ -661,7 +667,6 @@ set(SOURCES
        cache/compressed_secondary_cache.cc
        cache/lru_cache.cc
        cache/secondary_cache.cc
-        cache/secondary_cache_adapter.cc
        cache/sharded_cache.cc
        db/arena_wrapped_db_iter.cc
        db/blob/blob_contents.cc
@@ -748,7 +753,6 @@ set(SOURCES
        db/write_batch.cc
        db/write_batch_base.cc
        db/write_controller.cc
-        db/write_stall_stats.cc
        db/write_thread.cc
        env/composite_env.cc
        env/env.cc
@@ -840,7 +844,6 @@ set(SOURCES
        table/get_context.cc
        table/iterator.cc
        table/merging_iterator.cc
-        table/compaction_merging_iterator.cc
        table/meta_blocks.cc
        table/persistent_cache_helper.cc
        table/plain/plain_table_bloom.cc
@@ -882,7 +885,6 @@ set(SOURCES
        util/compression_context_cache.cc
        util/concurrent_task_limiter_impl.cc
        util/crc32c.cc
-        util/data_structure.cc
        util/dynamic_bloom.cc
        util/hash.cc
        util/murmurhash.cc
@@ -1281,7 +1283,6 @@ if(WITH_TESTS OR WITH_BENCHMARK_TOOLS)
  add_subdirectory(third-party/gtest-1.8.1/fused-src/gtest)
  add_library(testharness STATIC
  test_util/mock_time_env.cc
-  test_util/secondary_cache_test_util.cc
  test_util/testharness.cc)
  target_link_libraries(testharness gtest)
 endif()
@@ -1,66 +1,14 @@
 # Rocksdb Change Log
-## 8.1.1 (04/06/2023)
+## 7.10.2 (02/10/2023)
 ### Bug Fixes
-* In the DB::VerifyFileChecksums API, ensure that file system reads of SST files are equal to the readahead_size in ReadOptions, if specified. Previously, each read was 2x the readahead_size.
-
-## 8.1.0 (03/18/2023)
-### Behavior changes
-* Compaction output file cutting logic now considers range tombstone start keys. For example, SST partitioner now may receive ParitionRequest for range tombstone start keys.
-* If the async_io ReadOption is specified for MultiGet or NewIterator on a platform that doesn't support IO uring, the option is ignored and synchronous IO is used.
-
-### Bug Fixes
-* Fixed an issue for backward iteration when user defined timestamp is enabled in combination with BlobDB.
-* Fixed a couple of cases where a Merge operand encountered during iteration wasn't reflected in the `internal_merge_count` PerfContext counter.
-* Fixed a bug in CreateColumnFamilyWithImport()/ExportColumnFamily() which did not support range tombstones (#11252).
-* Fixed a bug where an excluded column family from an atomic flush contains unflushed data that should've been included in this atomic flush (i.e, data of seqno less than the max seqno of this atomic flush), leading to potential data loss in this excluded column family when `WriteOptions::disableWAL == true` (#11148).
-
-### New Features
-* Add statistics rocksdb.secondary.cache.filter.hits, rocksdb.secondary.cache.index.hits, and rocksdb.secondary.cache.filter.hits
-* Added a new PerfContext counter `internal_merge_point_lookup_count` which tracks the number of Merge operands applied while serving point lookup queries.
-* Add new statistics rocksdb.table.open.prefetch.tail.read.bytes, rocksdb.table.open.prefetch.tail.{miss|hit}
-* Add support for SecondaryCache with HyperClockCache (`HyperClockCacheOptions` inherits `secondary_cache` option from `ShardedCacheOptions`)
-* Add new db properties `rocksdb.cf-write-stall-stats`, `rocksdb.db-write-stall-stats`and APIs to examine them in a structured way. In particular, users of `GetMapProperty()` with property `kCFWriteStallStats`/`kDBWriteStallStats` can now use the functions in `WriteStallStatsMapKeys` to find stats in the map.
-
-### Public API Changes
-* Changed various functions and features in `Cache` that are mostly relevant to custom implementations or wrappers. Especially, asychronous lookup functionality is moved from `Lookup()` to a new `StartAsyncLookup()` function.
-
-## 8.0.0 (02/19/2023)
-### Behavior changes
-* `ReadOptions::verify_checksums=false` disables checksum verification for more reads of non-`CacheEntryRole::kDataBlock` blocks.
-* In case of scan with async_io enabled, if posix doesn't support IOUring, Status::NotSupported error will be returned to the users. Initially that error was swallowed and reads were switched to synchronous reads.
+* Fixed a bug in DB open/recovery from a compressed WAL that was caused due to incorrect handling of certain record fragments with the same offset within a WAL block.

+## 7.10.1 (02/01/2023)
 ### Bug Fixes
 * Fixed a data race on `ColumnFamilyData::flush_reason` caused by concurrent flushes.
-* Fixed an issue in `Get` and `MultiGet` when user-defined timestamps is enabled in combination with BlobDB.
-* Fixed some atypical behaviors for `LockWAL()` such as allowing concurrent/recursive use and not expecting `UnlockWAL()` after non-OK result. See API comments.
-* Fixed a feature interaction bug where for blobs `GetEntity` would expose the blob reference instead of the blob value.
 * Fixed `DisableManualCompaction()` and `CompactRangeOptions::canceled` to cancel compactions even when they are waiting on conflicting compactions to finish
 * Fixed a bug in which a successful `GetMergeOperands()` could transiently return `Status::MergeInProgress()`
 * Return the correct error (Status::NotSupported()) to MultiGet caller when ReadOptions::async_io flag is true and IO uring is not enabled. Previously, Status::Corruption() was being returned when the actual failure was lack of async IO support.
-* Fixed a bug in DB open/recovery from a compressed WAL that was caused due to incorrect handling of certain record fragments with the same offset within a WAL block.
-
-### Feature Removal
-* Remove RocksDB Lite.
-* The feature block_cache_compressed is removed. Statistics related to it are removed too.
-* Remove deprecated Env::LoadEnv(). Use Env::CreateFromString() instead.
-* Remove deprecated FileSystem::Load(). Use FileSystem::CreateFromString() instead.
-* Removed the deprecated version of these utility functions and the corresponding Java bindings: `LoadOptionsFromFile`, `LoadLatestOptions`, `CheckOptionsCompatibility`.
-* Remove the FactoryFunc from the LoadObject method from the Customizable helper methods.
-
-### Public API Changes
-* Moved rarely-needed Cache class definition to new advanced_cache.h, and added a CacheWrapper class to advanced_cache.h. Minor changes to SimCache API definitions.
-* Completely removed the following deprecated/obsolete statistics: the tickers `BLOCK_CACHE_INDEX_BYTES_EVICT`, `BLOCK_CACHE_FILTER_BYTES_EVICT`, `BLOOM_FILTER_MICROS`, `NO_FILE_CLOSES`, `STALL_L0_SLOWDOWN_MICROS`, `STALL_MEMTABLE_COMPACTION_MICROS`, `STALL_L0_NUM_FILES_MICROS`, `RATE_LIMIT_DELAY_MILLIS`, `NO_ITERATORS`, `NUMBER_FILTERED_DELETES`, `WRITE_TIMEDOUT`, `BLOB_DB_GC_NUM_KEYS_OVERWRITTEN`, `BLOB_DB_GC_NUM_KEYS_EXPIRED`, `BLOB_DB_GC_BYTES_OVERWRITTEN`, `BLOB_DB_GC_BYTES_EXPIRED`, `BLOCK_CACHE_COMPRESSION_DICT_BYTES_EVICT` as well as the histograms `STALL_L0_SLOWDOWN_COUNT`, `STALL_MEMTABLE_COMPACTION_COUNT`, `STALL_L0_NUM_FILES_COUNT`, `HARD_RATE_LIMIT_DELAY_COUNT`, `SOFT_RATE_LIMIT_DELAY_COUNT`, `BLOB_DB_GC_MICROS`, and `NUM_DATA_BLOCKS_READ_PER_LEVEL`. Note that as a result, the C++ enum values of the still supported statistics have changed. Developers are advised to not rely on the actual numeric values.
-* Deprecated IngestExternalFileOptions::write_global_seqno and change default to false. This option only needs to be set to true to generate a DB compatible with RocksDB versions before 5.16.0.
-* Remove deprecated APIs `GetColumnFamilyOptionsFrom{Map|String}(const ColumnFamilyOptions&, ..)`, `GetDBOptionsFrom{Map|String}(const DBOptions&, ..)`, `GetBlockBasedTableOptionsFrom{Map|String}(const BlockBasedTableOptions& table_options, ..)` and ` GetPlainTableOptionsFrom{Map|String}(const PlainTableOptions& table_options,..)`.
-* Added a subcode of `Status::Corruption`, `Status::SubCode::kMergeOperatorFailed`, for users to identify corruption failures originating in the merge operator, as opposed to RocksDB's internally identified data corruptions
-
-### Build Changes
-* The `make` build now builds a shared library by default instead of a static library. Use `LIB_MODE=static` to override.
-
-### New Features
-* Compaction filters are now supported for wide-column entities by means of the `FilterV3` API. See the comment of the API for more details.
-* Added `do_not_compress_roles` to `CompressedSecondaryCacheOptions` to disable compression on certain kinds of block. Filter blocks are now not compressed by CompressedSecondaryCache by default.
-* Added a new `MultiGetEntity` API that enables batched wide-column point lookups. See the API comments for more details.

 ## 7.10.0 (01/23/2023)
 ### Behavior changes
@@ -48,11 +48,6 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
 * If you wish to build the RocksJava static target, then cmake is required for building Snappy.

 * If you wish to run microbench (e.g, `make microbench`, `make ribbon_bench` or `cmake -DWITH_BENCHMARK=1`), Google benchmark >= 1.6.0 is needed.
-* You can do the following to install Google benchmark. These commands are copied from `./build_tools/ubuntu20_image/Dockerfile`:
-
-`$ git clone --depth 1 --branch v1.7.0 https://github.com/google/benchmark.git ~/benchmark`
-
-`$ cd ~/benchmark && mkdir build && cd build && cmake .. -GNinja -DCMAKE_BUILD_TYPE=Release -DBENCHMARK_ENABLE_GTEST_TESTS=0 && ninja && ninja install`

 ## Supported platforms

@@ -183,7 +178,7 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
        gmake rocksdbjava

 * **iOS**:
-  * Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define an important pre-processing macros: `IOS_CROSS_COMPILE`.
+  * Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define two important pre-processing macros: `ROCKSDB_LITE` and `IOS_CROSS_COMPILE`.

 * **Windows** (Visual Studio 2017 to up):
  * Read and follow the instructions at CMakeLists.txt
@@ -44,6 +44,13 @@ quoted_perl_command = $(subst ','\'',$(perl_command))
 # Set the default DEBUG_LEVEL to 1
 DEBUG_LEVEL?=1

+# LIB_MODE says whether or not to use/build "shared" or "static" libraries.
+# Mode "static" means to link against static libraries (.a)
+# Mode "shared" means to link against shared libraries (.so, .sl, .dylib, etc)
+#
+# Set the default LIB_MODE to static
+LIB_MODE?=static
+
 # OBJ_DIR is where the object files reside.  Default to the current directory
 OBJ_DIR?=.

@@ -74,23 +81,29 @@ else ifneq ($(filter jtest rocksdbjava%, $(MAKECMDGOALS)),)
 	endif
 endif

-# LIB_MODE says whether or not to use/build "shared" or "static" libraries.
-# Mode "static" means to link against static libraries (.a)
-# Mode "shared" means to link against shared libraries (.so, .sl, .dylib, etc)
-#
-ifeq ($(DEBUG_LEVEL), 0)
-# For optimized, set the default LIB_MODE to static for code size/efficiency
-	LIB_MODE?=static
-else
-# For debug, set the default LIB_MODE to shared for efficient `make check` etc.
-	LIB_MODE?=shared
-endif
+$(info $$DEBUG_LEVEL is ${DEBUG_LEVEL})

-$(info $$DEBUG_LEVEL is $(DEBUG_LEVEL), $$LIB_MODE is $(LIB_MODE))
+# Lite build flag.
+LITE ?= 0
+ifeq ($(LITE), 0)
+ifneq ($(filter -DROCKSDB_LITE,$(OPT)),)
+  # Be backward compatible and support older format where OPT=-DROCKSDB_LITE is
+  # specified instead of LITE=1 on the command line.
+  LITE=1
+endif
+else ifeq ($(LITE), 1)
+ifeq ($(filter -DROCKSDB_LITE,$(OPT)),)
+	OPT += -DROCKSDB_LITE
+endif
+endif

 # Figure out optimize level.
 ifneq ($(DEBUG_LEVEL), 2)
+ifeq ($(LITE), 0)
 	OPTIMIZE_LEVEL ?= -O2
+else
+	OPTIMIZE_LEVEL ?= -Os
+endif
 endif
 # `OPTIMIZE_LEVEL` is empty when the user does not set it and `DEBUG_LEVEL=2`.
 # In that case, the compiler default (`-O0` for gcc and clang) will be used.
@@ -324,6 +337,13 @@ endif
 ifeq ($(PLATFORM), OS_SOLARIS)
 	PLATFORM_CXXFLAGS += -D _GLIBCXX_USE_C99
 endif
+ifneq ($(filter -DROCKSDB_LITE,$(OPT)),)
+	# found
+	CFLAGS += -fno-exceptions
+	CXXFLAGS += -fno-exceptions
+	# LUA is not supported under ROCKSDB_LITE
+	LUA_PATH =
+endif

 ifeq ($(LIB_MODE),shared)
 # So that binaries are executable from build location, in addition to install location
@@ -1062,11 +1082,13 @@ check: all
 	rm -rf $(TEST_TMPDIR)
 ifneq ($(PLATFORM), OS_AIX)
 	$(PYTHON) tools/check_all_python.py
+ifeq ($(filter -DROCKSDB_LITE,$(OPT)),)
 ifndef ASSERT_STATUS_CHECKED # not yet working with these tests
 	$(PYTHON) tools/ldb_test.py
 	sh tools/rocksdb_dump_test.sh
 endif
 endif
+endif
 ifndef SKIP_FORMAT_BUCK_CHECKS
 	$(MAKE) check-format
 	$(MAKE) check-buck-targets
@@ -1222,9 +1244,9 @@ clean: clean-ext-libraries-all clean-rocks clean-rocksjava
 clean-not-downloaded: clean-ext-libraries-bin clean-rocks clean-not-downloaded-rocksjava

 clean-rocks:
-# Not practical to exactly match all versions/variants in naming (e.g. debug or not)
-	rm -f ${LIBNAME}*.so* ${LIBNAME}*.a
-	rm -f $(BENCHMARKS) $(TOOLS) $(TESTS) $(PARALLEL_TEST) $(MICROBENCHS)
+	echo shared=$(ALL_SHARED_LIBS)
+	echo static=$(ALL_STATIC_LIBS)
+	rm -f $(BENCHMARKS) $(TOOLS) $(TESTS) $(PARALLEL_TEST) $(ALL_STATIC_LIBS) $(ALL_SHARED_LIBS) $(MICROBENCHS)
 	rm -rf $(CLEAN_FILES) ios-x86 ios-arm scan_build_report
 	$(FIND) . -name "*.[oda]" -exec rm -f {} \;
 	$(FIND) . -type f \( -name "*.gcda" -o -name "*.gcno" \) -exec rm -f {} \;
@@ -2467,6 +2489,18 @@ build_size:
 	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.shared_lib $$(stat --printf="%s" `readlink -f librocksdb.so`)
 	strip `readlink -f librocksdb.so`
 	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.shared_lib_stripped $$(stat --printf="%s" `readlink -f librocksdb.so`)
+	# === lite build, static ===
+	$(MAKE) clean
+	$(MAKE) LITE=1 static_lib
+	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.static_lib_lite $$(stat --printf="%s" librocksdb.a)
+	strip librocksdb.a
+	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.static_lib_lite_stripped $$(stat --printf="%s" librocksdb.a)
+	# === lite build, shared ===
+	$(MAKE) clean
+	$(MAKE) LITE=1 shared_lib
+	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.shared_lib_lite $$(stat --printf="%s" `readlink -f librocksdb.so`)
+	strip `readlink -f librocksdb.so`
+	$(REPORT_BUILD_STATISTIC) rocksdb.build_size.shared_lib_lite_stripped $$(stat --printf="%s" `readlink -f librocksdb.so`)

 # ---------------------------------------------------------------------------
 #  	Platform-specific compilation
@@ -0,0 +1,21 @@
+# RocksDBLite
+
+RocksDBLite is a project focused on mobile use cases, which don't need a lot of fancy things we've built for server workloads and they are very sensitive to binary size. For that reason, we added a compile flag ROCKSDB_LITE that comments out a lot of the nonessential code and keeps the binary lean.
+
+Some examples of the features disabled by ROCKSDB_LITE:
+* compiled-in support for LDB tool
+* No backup engine
+* No support for replication (which we provide in form of TransactionalIterator)
+* No advanced monitoring tools
+* No special-purpose memtables that are highly optimized for specific use cases
+* No Transactions
+
+When adding a new big feature to RocksDB, please add ROCKSDB_LITE compile guard if:
+* Nobody from mobile really needs your feature,
+* Your feature is adding a lot of weight to the binary.
+
+Don't add ROCKSDB_LITE compile guard if:
+* It would introduce a lot of code complexity. Compile guards make code harder to read. It's a trade-off.
+* Your feature is not adding a lot of weight.
+
+If unsure, ask. :)
@@ -19,7 +19,6 @@ cpp_library_wrapper(name="rocksdb_lib", srcs=[
        "cache/compressed_secondary_cache.cc",
        "cache/lru_cache.cc",
        "cache/secondary_cache.cc",
-        "cache/secondary_cache_adapter.cc",
        "cache/sharded_cache.cc",
        "db/arena_wrapped_db_iter.cc",
        "db/blob/blob_contents.cc",
@@ -106,7 +105,6 @@ cpp_library_wrapper(name="rocksdb_lib", srcs=[
        "db/write_batch.cc",
        "db/write_batch_base.cc",
        "db/write_controller.cc",
-        "db/write_stall_stats.cc",
        "db/write_thread.cc",
        "env/composite_env.cc",
        "env/env.cc",
@@ -202,7 +200,6 @@ cpp_library_wrapper(name="rocksdb_lib", srcs=[
        "table/block_based/reader_common.cc",
        "table/block_based/uncompression_dict_reader.cc",
        "table/block_fetcher.cc",
-        "table/compaction_merging_iterator.cc",
        "table/cuckoo/cuckoo_table_builder.cc",
        "table/cuckoo/cuckoo_table_factory.cc",
        "table/cuckoo/cuckoo_table_reader.cc",
@@ -250,7 +247,6 @@ cpp_library_wrapper(name="rocksdb_lib", srcs=[
        "util/concurrent_task_limiter_impl.cc",
        "util/crc32c.cc",
        "util/crc32c_arm64.cc",
-        "util/data_structure.cc",
        "util/dynamic_bloom.cc",
        "util/file_checksum_helper.cc",
        "util/hash.cc",
@@ -354,14 +350,353 @@ cpp_library_wrapper(name="rocksdb_lib", srcs=[
        "//folly/synchronization:distributed_mutex",
    ], headers=None, link_whole=False, extra_test_libs=False)

-cpp_library_wrapper(name="rocksdb_whole_archive_lib", srcs=[], deps=[":rocksdb_lib"], headers=None, link_whole=True, extra_test_libs=False)
+cpp_library_wrapper(name="rocksdb_whole_archive_lib", srcs=[
+        "cache/cache.cc",
+        "cache/cache_entry_roles.cc",
+        "cache/cache_helpers.cc",
+        "cache/cache_key.cc",
+        "cache/cache_reservation_manager.cc",
+        "cache/charged_cache.cc",
+        "cache/clock_cache.cc",
+        "cache/compressed_secondary_cache.cc",
+        "cache/lru_cache.cc",
+        "cache/secondary_cache.cc",
+        "cache/sharded_cache.cc",
+        "db/arena_wrapped_db_iter.cc",
+        "db/blob/blob_contents.cc",
+        "db/blob/blob_fetcher.cc",
+        "db/blob/blob_file_addition.cc",
+        "db/blob/blob_file_builder.cc",
+        "db/blob/blob_file_cache.cc",
+        "db/blob/blob_file_garbage.cc",
+        "db/blob/blob_file_meta.cc",
+        "db/blob/blob_file_reader.cc",
+        "db/blob/blob_garbage_meter.cc",
+        "db/blob/blob_log_format.cc",
+        "db/blob/blob_log_sequential_reader.cc",
+        "db/blob/blob_log_writer.cc",
+        "db/blob/blob_source.cc",
+        "db/blob/prefetch_buffer_collection.cc",
+        "db/builder.cc",
+        "db/c.cc",
+        "db/column_family.cc",
+        "db/compaction/compaction.cc",
+        "db/compaction/compaction_iterator.cc",
+        "db/compaction/compaction_job.cc",
+        "db/compaction/compaction_outputs.cc",
+        "db/compaction/compaction_picker.cc",
+        "db/compaction/compaction_picker_fifo.cc",
+        "db/compaction/compaction_picker_level.cc",
+        "db/compaction/compaction_picker_universal.cc",
+        "db/compaction/compaction_service_job.cc",
+        "db/compaction/compaction_state.cc",
+        "db/compaction/sst_partitioner.cc",
+        "db/compaction/subcompaction_state.cc",
+        "db/convenience.cc",
+        "db/db_filesnapshot.cc",
+        "db/db_impl/compacted_db_impl.cc",
+        "db/db_impl/db_impl.cc",
+        "db/db_impl/db_impl_compaction_flush.cc",
+        "db/db_impl/db_impl_debug.cc",
+        "db/db_impl/db_impl_experimental.cc",
+        "db/db_impl/db_impl_files.cc",
+        "db/db_impl/db_impl_open.cc",
+        "db/db_impl/db_impl_readonly.cc",
+        "db/db_impl/db_impl_secondary.cc",
+        "db/db_impl/db_impl_write.cc",
+        "db/db_info_dumper.cc",
+        "db/db_iter.cc",
+        "db/dbformat.cc",
+        "db/error_handler.cc",
+        "db/event_helpers.cc",
+        "db/experimental.cc",
+        "db/external_sst_file_ingestion_job.cc",
+        "db/file_indexer.cc",
+        "db/flush_job.cc",
+        "db/flush_scheduler.cc",
+        "db/forward_iterator.cc",
+        "db/import_column_family_job.cc",
+        "db/internal_stats.cc",
+        "db/log_reader.cc",
+        "db/log_writer.cc",
+        "db/logs_with_prep_tracker.cc",
+        "db/malloc_stats.cc",
+        "db/memtable.cc",
+        "db/memtable_list.cc",
+        "db/merge_helper.cc",
+        "db/merge_operator.cc",
+        "db/output_validator.cc",
+        "db/periodic_task_scheduler.cc",
+        "db/range_del_aggregator.cc",
+        "db/range_tombstone_fragmenter.cc",
+        "db/repair.cc",
+        "db/seqno_to_time_mapping.cc",
+        "db/snapshot_impl.cc",
+        "db/table_cache.cc",
+        "db/table_properties_collector.cc",
+        "db/transaction_log_impl.cc",
+        "db/trim_history_scheduler.cc",
+        "db/version_builder.cc",
+        "db/version_edit.cc",
+        "db/version_edit_handler.cc",
+        "db/version_set.cc",
+        "db/wal_edit.cc",
+        "db/wal_manager.cc",
+        "db/wide/wide_column_serialization.cc",
+        "db/wide/wide_columns.cc",
+        "db/write_batch.cc",
+        "db/write_batch_base.cc",
+        "db/write_controller.cc",
+        "db/write_thread.cc",
+        "env/composite_env.cc",
+        "env/env.cc",
+        "env/env_chroot.cc",
+        "env/env_encryption.cc",
+        "env/env_posix.cc",
+        "env/file_system.cc",
+        "env/file_system_tracer.cc",
+        "env/fs_posix.cc",
+        "env/fs_remap.cc",
+        "env/io_posix.cc",
+        "env/mock_env.cc",
+        "env/unique_id_gen.cc",
+        "file/delete_scheduler.cc",
+        "file/file_prefetch_buffer.cc",
+        "file/file_util.cc",
+        "file/filename.cc",
+        "file/line_file_reader.cc",
+        "file/random_access_file_reader.cc",
+        "file/read_write_util.cc",
+        "file/readahead_raf.cc",
+        "file/sequence_file_reader.cc",
+        "file/sst_file_manager_impl.cc",
+        "file/writable_file_writer.cc",
+        "logging/auto_roll_logger.cc",
+        "logging/event_logger.cc",
+        "logging/log_buffer.cc",
+        "memory/arena.cc",
+        "memory/concurrent_arena.cc",
+        "memory/jemalloc_nodump_allocator.cc",
+        "memory/memkind_kmem_allocator.cc",
+        "memory/memory_allocator.cc",
+        "memtable/alloc_tracker.cc",
+        "memtable/hash_linklist_rep.cc",
+        "memtable/hash_skiplist_rep.cc",
+        "memtable/skiplistrep.cc",
+        "memtable/vectorrep.cc",
+        "memtable/write_buffer_manager.cc",
+        "monitoring/histogram.cc",
+        "monitoring/histogram_windowing.cc",
+        "monitoring/in_memory_stats_history.cc",
+        "monitoring/instrumented_mutex.cc",
+        "monitoring/iostats_context.cc",
+        "monitoring/perf_context.cc",
+        "monitoring/perf_level.cc",
+        "monitoring/persistent_stats_history.cc",
+        "monitoring/statistics.cc",
+        "monitoring/thread_status_impl.cc",
+        "monitoring/thread_status_updater.cc",
+        "monitoring/thread_status_updater_debug.cc",
+        "monitoring/thread_status_util.cc",
+        "monitoring/thread_status_util_debug.cc",
+        "options/cf_options.cc",
+        "options/configurable.cc",
+        "options/customizable.cc",
+        "options/db_options.cc",
+        "options/options.cc",
+        "options/options_helper.cc",
+        "options/options_parser.cc",
+        "port/mmap.cc",
+        "port/port_posix.cc",
+        "port/stack_trace.cc",
+        "port/win/env_default.cc",
+        "port/win/env_win.cc",
+        "port/win/io_win.cc",
+        "port/win/port_win.cc",
+        "port/win/win_logger.cc",
+        "port/win/win_thread.cc",
+        "table/adaptive/adaptive_table_factory.cc",
+        "table/block_based/binary_search_index_reader.cc",
+        "table/block_based/block.cc",
+        "table/block_based/block_based_table_builder.cc",
+        "table/block_based/block_based_table_factory.cc",
+        "table/block_based/block_based_table_iterator.cc",
+        "table/block_based/block_based_table_reader.cc",
+        "table/block_based/block_builder.cc",
+        "table/block_based/block_cache.cc",
+        "table/block_based/block_prefetcher.cc",
+        "table/block_based/block_prefix_index.cc",
+        "table/block_based/data_block_footer.cc",
+        "table/block_based/data_block_hash_index.cc",
+        "table/block_based/filter_block_reader_common.cc",
+        "table/block_based/filter_policy.cc",
+        "table/block_based/flush_block_policy.cc",
+        "table/block_based/full_filter_block.cc",
+        "table/block_based/hash_index_reader.cc",
+        "table/block_based/index_builder.cc",
+        "table/block_based/index_reader_common.cc",
+        "table/block_based/parsed_full_filter_block.cc",
+        "table/block_based/partitioned_filter_block.cc",
+        "table/block_based/partitioned_index_iterator.cc",
+        "table/block_based/partitioned_index_reader.cc",
+        "table/block_based/reader_common.cc",
+        "table/block_based/uncompression_dict_reader.cc",
+        "table/block_fetcher.cc",
+        "table/cuckoo/cuckoo_table_builder.cc",
+        "table/cuckoo/cuckoo_table_factory.cc",
+        "table/cuckoo/cuckoo_table_reader.cc",
+        "table/format.cc",
+        "table/get_context.cc",
+        "table/iterator.cc",
+        "table/merging_iterator.cc",
+        "table/meta_blocks.cc",
+        "table/persistent_cache_helper.cc",
+        "table/plain/plain_table_bloom.cc",
+        "table/plain/plain_table_builder.cc",
+        "table/plain/plain_table_factory.cc",
+        "table/plain/plain_table_index.cc",
+        "table/plain/plain_table_key_coding.cc",
+        "table/plain/plain_table_reader.cc",
+        "table/sst_file_dumper.cc",
+        "table/sst_file_reader.cc",
+        "table/sst_file_writer.cc",
+        "table/table_factory.cc",
+        "table/table_properties.cc",
+        "table/two_level_iterator.cc",
+        "table/unique_id.cc",
+        "test_util/sync_point.cc",
+        "test_util/sync_point_impl.cc",
+        "test_util/transaction_test_util.cc",
+        "tools/dump/db_dump_tool.cc",
+        "tools/io_tracer_parser_tool.cc",
+        "tools/ldb_cmd.cc",
+        "tools/ldb_tool.cc",
+        "tools/sst_dump_tool.cc",
+        "trace_replay/block_cache_tracer.cc",
+        "trace_replay/io_tracer.cc",
+        "trace_replay/trace_record.cc",
+        "trace_replay/trace_record_handler.cc",
+        "trace_replay/trace_record_result.cc",
+        "trace_replay/trace_replay.cc",
+        "util/async_file_reader.cc",
+        "util/build_version.cc",
+        "util/cleanable.cc",
+        "util/coding.cc",
+        "util/compaction_job_stats_impl.cc",
+        "util/comparator.cc",
+        "util/compression.cc",
+        "util/compression_context_cache.cc",
+        "util/concurrent_task_limiter_impl.cc",
+        "util/crc32c.cc",
+        "util/crc32c_arm64.cc",
+        "util/dynamic_bloom.cc",
+        "util/file_checksum_helper.cc",
+        "util/hash.cc",
+        "util/murmurhash.cc",
+        "util/random.cc",
+        "util/rate_limiter.cc",
+        "util/ribbon_config.cc",
+        "util/slice.cc",
+        "util/status.cc",
+        "util/stderr_logger.cc",
+        "util/string_util.cc",
+        "util/thread_local.cc",
+        "util/threadpool_imp.cc",
+        "util/xxhash.cc",
+        "utilities/agg_merge/agg_merge.cc",
+        "utilities/backup/backup_engine.cc",
+        "utilities/blob_db/blob_compaction_filter.cc",
+        "utilities/blob_db/blob_db.cc",
+        "utilities/blob_db/blob_db_impl.cc",
+        "utilities/blob_db/blob_db_impl_filesnapshot.cc",
+        "utilities/blob_db/blob_dump_tool.cc",
+        "utilities/blob_db/blob_file.cc",
+        "utilities/cache_dump_load.cc",
+        "utilities/cache_dump_load_impl.cc",
+        "utilities/cassandra/cassandra_compaction_filter.cc",
+        "utilities/cassandra/format.cc",
+        "utilities/cassandra/merge_operator.cc",
+        "utilities/checkpoint/checkpoint_impl.cc",
+        "utilities/compaction_filters.cc",
+        "utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc",
+        "utilities/convenience/info_log_finder.cc",
+        "utilities/counted_fs.cc",
+        "utilities/debug.cc",
+        "utilities/env_mirror.cc",
+        "utilities/env_timed.cc",
+        "utilities/fault_injection_env.cc",
+        "utilities/fault_injection_fs.cc",
+        "utilities/fault_injection_secondary_cache.cc",
+        "utilities/leveldb_options/leveldb_options.cc",
+        "utilities/memory/memory_util.cc",
+        "utilities/merge_operators.cc",
+        "utilities/merge_operators/bytesxor.cc",
+        "utilities/merge_operators/max.cc",
+        "utilities/merge_operators/put.cc",
+        "utilities/merge_operators/sortlist.cc",
+        "utilities/merge_operators/string_append/stringappend.cc",
+        "utilities/merge_operators/string_append/stringappend2.cc",
+        "utilities/merge_operators/uint64add.cc",
+        "utilities/object_registry.cc",
+        "utilities/option_change_migration/option_change_migration.cc",
+        "utilities/options/options_util.cc",
+        "utilities/persistent_cache/block_cache_tier.cc",
+        "utilities/persistent_cache/block_cache_tier_file.cc",
+        "utilities/persistent_cache/block_cache_tier_metadata.cc",
+        "utilities/persistent_cache/persistent_cache_tier.cc",
+        "utilities/persistent_cache/volatile_tier_impl.cc",
+        "utilities/simulator_cache/cache_simulator.cc",
+        "utilities/simulator_cache/sim_cache.cc",
+        "utilities/table_properties_collectors/compact_on_deletion_collector.cc",
+        "utilities/trace/file_trace_reader_writer.cc",
+        "utilities/trace/replayer_impl.cc",
+        "utilities/transactions/lock/lock_manager.cc",
+        "utilities/transactions/lock/point/point_lock_manager.cc",
+        "utilities/transactions/lock/point/point_lock_tracker.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/concurrent_tree.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/keyrange.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/lock_request.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/locktree.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/manager.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/range_buffer.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/treenode.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/txnid_set.cc",
+        "utilities/transactions/lock/range/range_tree/lib/locktree/wfg.cc",
+        "utilities/transactions/lock/range/range_tree/lib/standalone_port.cc",
+        "utilities/transactions/lock/range/range_tree/lib/util/dbt.cc",
+        "utilities/transactions/lock/range/range_tree/lib/util/memarena.cc",
+        "utilities/transactions/lock/range/range_tree/range_tree_lock_manager.cc",
+        "utilities/transactions/lock/range/range_tree/range_tree_lock_tracker.cc",
+        "utilities/transactions/optimistic_transaction.cc",
+        "utilities/transactions/optimistic_transaction_db_impl.cc",
+        "utilities/transactions/pessimistic_transaction.cc",
+        "utilities/transactions/pessimistic_transaction_db.cc",
+        "utilities/transactions/snapshot_checker.cc",
+        "utilities/transactions/transaction_base.cc",
+        "utilities/transactions/transaction_db_mutex_impl.cc",
+        "utilities/transactions/transaction_util.cc",
+        "utilities/transactions/write_prepared_txn.cc",
+        "utilities/transactions/write_prepared_txn_db.cc",
+        "utilities/transactions/write_unprepared_txn.cc",
+        "utilities/transactions/write_unprepared_txn_db.cc",
+        "utilities/ttl/db_ttl_impl.cc",
+        "utilities/wal_filter.cc",
+        "utilities/write_batch_with_index/write_batch_with_index.cc",
+        "utilities/write_batch_with_index/write_batch_with_index_internal.cc",
+    ], deps=[
+        "//folly/container:f14_hash",
+        "//folly/experimental/coro:blocking_wait",
+        "//folly/experimental/coro:collect",
+        "//folly/experimental/coro:coroutine",
+        "//folly/experimental/coro:task",
+        "//folly/synchronization:distributed_mutex",
+    ], headers=None, link_whole=True, extra_test_libs=False)

 cpp_library_wrapper(name="rocksdb_test_lib", srcs=[
        "db/db_test_util.cc",
        "db/db_with_timestamp_test_util.cc",
        "table/mock_table.cc",
        "test_util/mock_time_env.cc",
-        "test_util/secondary_cache_test_util.cc",
        "test_util/testharness.cc",
        "test_util/testutil.cc",
        "tools/block_cache_analyzer/block_cache_trace_analyzer.cc",
@@ -15,9 +15,6 @@ At Facebook, we use RocksDB as storage engines in multiple data management servi

 [2] https://code.facebook.com/posts/357056558062811/logdevice-a-distributed-data-store-for-logs/

-## Microsoft
-The Bing search engine from Microsoft uses RocksDB as the storage engine for its web data platform: https://blogs.bing.com/Engineering-Blog/october-2021/RocksDB-in-Microsoft-Bing
-
 ## LinkedIn
 Two different use cases at Linkedin are using RocksDB as a storage engine:

@@ -26,7 +26,7 @@ from util import ColorString
 # $python3 buckifier/buckify_rocksdb.py \
 #        '{"fake": {
 #                      "extra_deps": [":test_dep", "//fakes/module:mock1"],
-#                      "extra_compiler_flags": ["-DFOO_BAR", "-Os"]
+#                      "extra_compiler_flags": ["-DROCKSDB_LITE", "-Os"]
 #                  }
 #         }'
 # (Generated TARGETS file has test_dep and mock1 as dependencies for RocksDB
@@ -154,9 +154,16 @@ def generate_targets(repo_path, deps_map):
    # rocksdb_whole_archive_lib
    TARGETS.add_library(
        "rocksdb_whole_archive_lib",
-        [],
+        src_mk["LIB_SOURCES"] +
+        # always add range_tree, it's only excluded on ppc64, which we don't use internally
+        src_mk["RANGE_TREE_SOURCES"] + src_mk["TOOL_LIB_SOURCES"],
        deps=[
-            ":rocksdb_lib",
+            "//folly/container:f14_hash",
+            "//folly/experimental/coro:blocking_wait",
+            "//folly/experimental/coro:collect",
+            "//folly/experimental/coro:coroutine",
+            "//folly/experimental/coro:task",
+            "//folly/synchronization:distributed_mutex",
        ],
        headers=None,
        extra_external_deps="",
@@ -63,7 +63,13 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
    if [ "$LIB_MODE" == "shared" ]; then
      PIC_BUILD=1
    fi
-    source "$PWD/build_tools/fbcode_config_platform010.sh"
+    if [ -n "$ROCKSDB_FBCODE_BUILD_WITH_PLATFORM010" ]; then
+      source "$PWD/build_tools/fbcode_config_platform010.sh"
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_PLATFORM009" ]; then
+      source "$PWD/build_tools/fbcode_config_platform009.sh"
+    else
+      source "$PWD/build_tools/fbcode_config_platform009.sh"
+    fi
 fi

 # Delete existing output, if it exists
@@ -148,7 +154,7 @@ case "$TARGET_OS" in
        ;;
    IOS)
        PLATFORM=IOS
-        COMMON_FLAGS="$COMMON_FLAGS -DOS_MACOSX -DIOS_CROSS_COMPILE "
+        COMMON_FLAGS="$COMMON_FLAGS -DOS_MACOSX -DIOS_CROSS_COMPILE -DROCKSDB_LITE"
        PLATFORM_SHARED_EXT=dylib
        PLATFORM_SHARED_LDFLAGS="-dynamiclib -install_name "
        CROSS_COMPILE=true
@@ -0,0 +1,22 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+GCC_BASE=/mnt/gvfs/third-party2/gcc/1795efe5f06778c15a92c8f9a2aba5dc496d9d4d/9.x/centos7-native/3bed279
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/7318eaac22659b6ff2fe43918e4b69fd0772a8a7/9.0.0/platform009/651ee30
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/4959b39cfbe5965a37c861c4c327fa7c5c759b87/9.x/platform009/9202ce7
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/45ce3375cdc77ecb2520bbf8f0ecddd3f98efd7a/2.30/platform009/f259413
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/be4de3205e029101b18aa8103daa696c2bef3b19/1.1.3/platform009/7f3b187
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/3c160ac5c67e257501e24c6c1d00ad5e01d73db6/1.2.8/platform009/7f3b187
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/73a237ac5bc0a5f5d67b39b8d253cfebaab88684/1.0.6/platform009/7f3b187
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/6ca38d3c390be2774d61a300f151464bbd632d62/1.9.1/platform009/7f3b187
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/64c58a207d2495e83abc57a500a956df09b79a7c/1.4.x/platform009/ba86d1f
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/824d0a8a5abb5b121afd1b35fc3896407ea50092/2.2.0/platform009/7f3b187
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/b62912d333ef33f9760efa6219dbe3fe6abb3b0e/master/platform009/c305944
+NUMA_BASE=/mnt/gvfs/third-party2/numa/0af65f71e23a67bf65dc91b11f95caa39325c432/2.0.11/platform009/7f3b187
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/02486dac347645d31dce116f44e1de3177315be2/1.4/platform009/5191652
+TBB_BASE=/mnt/gvfs/third-party2/tbb/2e0ec671e550bfca347300bf3f789d9c0fff24ad/2018_U5/platform009/7f3b187
+LIBURING_BASE=/mnt/gvfs/third-party2/liburing/70dbd9cfee63a25611417d09433a86d7711b3990/20200729/platform009/7f3b187
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/32b8a2407b634df3f8f948ba373fc4acc6a18296/fb/platform009/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/08634589372fa5f237bfd374e8c644a8364e78c1/2.32/platform009/ba86d1f/
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/6ae525939ad02e5e676855082fbbc7828dbafeac/3.15.0/platform009/7f3b187
+LUA_BASE=/mnt/gvfs/third-party2/lua/162efd9561a3d21f6869f4814011e9cf1b3ff4dc/5.3.4/platform009/a6271c4
+BENCHMARK_BASE=/mnt/gvfs/third-party2/benchmark/30bf49ad6414325e17f3425b0edcb64239427ae3/1.6.1/platform009/7f3b187
+GLOG_BASE=/mnt/gvfs/third-party2/glog/32d751bd5673375b438158717ab6a57c1cc57e3d/0.3.2_fb/platform009/10a364d
@@ -0,0 +1,170 @@
+#!/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_platform009.sh"
+
+CFLAGS=""
+
+# libgcc
+LIBGCC_INCLUDE="$LIBGCC_BASE/include/c++/9.3.0 -I $LIBGCC_BASE/include/c++/9.3.0/backward"
+LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
+
+# glibc
+GLIBC_INCLUDE="$GLIBC_BASE/include"
+GLIBC_LIBS=" -L $GLIBC_BASE/lib"
+
+if test -z $PIC_BUILD; then
+  MAYBE_PIC=
+else
+  MAYBE_PIC=_pic
+fi
+
+if ! test $ROCKSDB_DISABLE_SNAPPY; then
+  # snappy
+  SNAPPY_INCLUDE=" -I $SNAPPY_BASE/include/"
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy${MAYBE_PIC}.a"
+  CFLAGS+=" -DSNAPPY"
+fi
+
+if ! test $ROCKSDB_DISABLE_ZLIB; then
+  # location of zlib headers and libraries
+  ZLIB_INCLUDE=" -I $ZLIB_BASE/include/"
+  ZLIB_LIBS=" $ZLIB_BASE/lib/libz${MAYBE_PIC}.a"
+  CFLAGS+=" -DZLIB"
+fi
+
+if ! test $ROCKSDB_DISABLE_BZIP; then
+  # location of bzip headers and libraries
+  BZIP_INCLUDE=" -I $BZIP2_BASE/include/"
+  BZIP_LIBS=" $BZIP2_BASE/lib/libbz2${MAYBE_PIC}.a"
+  CFLAGS+=" -DBZIP2"
+fi
+
+if ! test $ROCKSDB_DISABLE_LZ4; then
+  LZ4_INCLUDE=" -I $LZ4_BASE/include/"
+  LZ4_LIBS=" $LZ4_BASE/lib/liblz4${MAYBE_PIC}.a"
+  CFLAGS+=" -DLZ4"
+fi
+
+if ! test $ROCKSDB_DISABLE_ZSTD; then
+  ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd${MAYBE_PIC}.a"
+  CFLAGS+=" -DZSTD"
+fi
+
+# location of gflags headers and libraries
+GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
+GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags${MAYBE_PIC}.a"
+CFLAGS+=" -DGFLAGS=gflags"
+
+BENCHMARK_INCLUDE=" -I $BENCHMARK_BASE/include/"
+BENCHMARK_LIBS=" $BENCHMARK_BASE/lib/libbenchmark${MAYBE_PIC}.a"
+
+GLOG_INCLUDE=" -I $GLOG_BASE/include/"
+GLOG_LIBS=" $GLOG_BASE/lib/libglog${MAYBE_PIC}.a"
+
+# location of jemalloc
+JEMALLOC_INCLUDE=" -I $JEMALLOC_BASE/include/"
+JEMALLOC_LIB=" $JEMALLOC_BASE/lib/libjemalloc${MAYBE_PIC}.a"
+
+# location of numa
+NUMA_INCLUDE=" -I $NUMA_BASE/include/"
+NUMA_LIB=" $NUMA_BASE/lib/libnuma${MAYBE_PIC}.a"
+CFLAGS+=" -DNUMA"
+
+# location of libunwind
+LIBUNWIND="$LIBUNWIND_BASE/lib/libunwind${MAYBE_PIC}.a"
+
+# location of TBB
+TBB_INCLUDE=" -isystem $TBB_BASE/include/"
+TBB_LIBS="$TBB_BASE/lib/libtbb${MAYBE_PIC}.a"
+CFLAGS+=" -DTBB"
+
+# location of LIBURING
+LIBURING_INCLUDE=" -isystem $LIBURING_BASE/include/"
+LIBURING_LIBS="$LIBURING_BASE/lib/liburing${MAYBE_PIC}.a"
+CFLAGS+=" -DLIBURING"
+
+test "$USE_SSE" || USE_SSE=1
+export USE_SSE
+test "$PORTABLE" || PORTABLE=1
+export PORTABLE
+
+BINUTILS="$BINUTILS_BASE/bin"
+AR="$BINUTILS/ar"
+AS="$BINUTILS/as"
+
+DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE $LIBURING_INCLUDE $BENCHMARK_INCLUDE $GLOG_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/clang/tools/scan-build/bin/scan-build"
+
+if [ -z "$USE_CLANG" ]; then
+  # gcc
+  CC="$GCC_BASE/bin/gcc"
+  CXX="$GCC_BASE/bin/g++"
+  AR="$GCC_BASE/bin/gcc-ar"
+
+  CFLAGS+=" -B$BINUTILS"
+  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++"
+  AR="$CLANG_BIN/llvm-ar"
+
+  KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
+
+  CFLAGS+=" -B$BINUTILS -nostdinc -nostdlib"
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/9.x "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/9.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 -DHAVE_SSE42 -DROCKSDB_IOURING_PRESENT"
+CXXFLAGS+=" $CFLAGS"
+
+EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS $LIBURING_LIBS $BENCHMARK_LIBS"
+EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/platform009/lib/ld.so"
+EXEC_LDFLAGS+=" $LIBUNWIND"
+EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/platform009/lib"
+EXEC_LDFLAGS+=" -Wl,-rpath=$GCC_BASE/lib64"
+# required by libtbb
+EXEC_LDFLAGS+=" -ldl"
+
+PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
+PLATFORM_LDFLAGS+=" -B$BINUTILS"
+
+EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS $LIBURING_LIBS $BENCHMARK_LIBS"
+
+VALGRIND_VER="$VALGRIND_BASE/bin/"
+
+# lua not supported because it's on track for deprecation, I think
+LUA_PATH=
+LUA_LIB=
+
+export CC CXX AR AS CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD LUA_PATH LUA_LIB
@@ -360,7 +360,7 @@ function send_to_ods {
    echo >&2 "ERROR: Key $key doesn't have a value."
    return
  fi
-  curl --silent "https://www.facebook.com/intern/agent/ods_set.php?entity=rocksdb_build&key=$key&value=$value" \
+  curl --silent "https://www.intern.facebook.com/intern/agent/ods_set.php?entity=rocksdb_build&key=$key&value=$value" \
    --connect-timeout 60
 }

@@ -104,3 +104,46 @@ get_lib_base valgrind   LATEST platform010
 get_lib_base lua        5.3.4  platform010

 git diff $OUTPUT
+
+
+###########################################################
+#                platform009 dependencies                 #
+###########################################################
+
+OUTPUT="$BASEDIR/dependencies_platform009.sh"
+
+rm -f "$OUTPUT"
+touch "$OUTPUT"
+
+echo "Writing dependencies to $OUTPUT"
+
+# Compilers locations
+GCC_BASE=`readlink -f $TP2_LATEST/gcc/9.x/centos7-native/*/`
+CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/9.0.0/platform009/*/`
+
+log_header
+log_variable GCC_BASE
+log_variable CLANG_BASE
+
+# Libraries locations
+get_lib_base libgcc     9.x     platform009
+get_lib_base glibc      2.30    platform009
+get_lib_base snappy     LATEST  platform009
+get_lib_base zlib       LATEST  platform009
+get_lib_base bzip2      LATEST  platform009
+get_lib_base lz4        LATEST  platform009
+get_lib_base zstd       LATEST  platform009
+get_lib_base gflags     LATEST  platform009
+get_lib_base jemalloc   LATEST  platform009
+get_lib_base numa       LATEST  platform009
+get_lib_base libunwind  LATEST  platform009
+get_lib_base tbb        2018_U5 platform009
+get_lib_base liburing   LATEST  platform009
+get_lib_base benchmark  LATEST  platform009
+
+get_lib_base kernel-headers fb platform009
+get_lib_base binutils   LATEST centos7-native
+get_lib_base valgrind   LATEST platform009
+get_lib_base lua        5.3.4  platform009
+
+git diff $OUTPUT
@@ -16,8 +16,7 @@
 #include "util/string_util.h"

 namespace ROCKSDB_NAMESPACE {
-const Cache::CacheItemHelper kNoopCacheItemHelper{};
-
+#ifndef ROCKSDB_LITE
 static std::unordered_map<std::string, OptionTypeInfo>
    lru_cache_options_type_info = {
        {"capacity",
@@ -65,6 +64,7 @@ static std::unordered_map<std::string, OptionTypeInfo>
          OptionType::kBoolean, OptionVerificationType::kNormal,
          OptionTypeFlags::kMutable}},
 };
+#endif  // ROCKSDB_LITE

 Status SecondaryCache::CreateFromString(
    const ConfigOptions& config_options, const std::string& value,
@@ -75,6 +75,7 @@ Status SecondaryCache::CreateFromString(
    Status status;
    std::shared_ptr<SecondaryCache> sec_cache;

+#ifndef ROCKSDB_LITE
    CompressedSecondaryCacheOptions sec_cache_opts;
    status = OptionTypeInfo::ParseStruct(config_options, "",
                                         &comp_sec_cache_options_type_info, "",
@@ -83,13 +84,19 @@ Status SecondaryCache::CreateFromString(
      sec_cache = NewCompressedSecondaryCache(sec_cache_opts);
    }

+#else
+    (void)config_options;
+    status = Status::NotSupported(
+        "Cannot load compressed secondary cache in LITE mode ", args);
+#endif  //! ROCKSDB_LITE

    if (status.ok()) {
      result->swap(sec_cache);
    }
    return status;
  } else {
-    return LoadSharedObject<SecondaryCache>(config_options, value, result);
+    return LoadSharedObject<SecondaryCache>(config_options, value, nullptr,
+                                            result);
  }
 }

@@ -101,6 +108,7 @@ Status Cache::CreateFromString(const ConfigOptions& config_options,
  if (value.find('=') == std::string::npos) {
    cache = NewLRUCache(ParseSizeT(value));
  } else {
+#ifndef ROCKSDB_LITE
    LRUCacheOptions cache_opts;
    status = OptionTypeInfo::ParseStruct(config_options, "",
                                         &lru_cache_options_type_info, "",
@@ -108,51 +116,14 @@ Status Cache::CreateFromString(const ConfigOptions& config_options,
    if (status.ok()) {
      cache = NewLRUCache(cache_opts);
    }
+#else
+    (void)config_options;
+    status = Status::NotSupported("Cannot load cache in LITE mode ", value);
+#endif  //! ROCKSDB_LITE
  }
  if (status.ok()) {
    result->swap(cache);
  }
  return status;
 }
-
-bool Cache::AsyncLookupHandle::IsReady() {
-  return pending_handle == nullptr || pending_handle->IsReady();
-}
-
-bool Cache::AsyncLookupHandle::IsPending() { return pending_handle != nullptr; }
-
-Cache::Handle* Cache::AsyncLookupHandle::Result() {
-  assert(!IsPending());
-  return result_handle;
-}
-
-void Cache::StartAsyncLookup(AsyncLookupHandle& async_handle) {
-  async_handle.found_dummy_entry = false;  // in case re-used
-  assert(!async_handle.IsPending());
-  async_handle.result_handle =
-      Lookup(async_handle.key, async_handle.helper, async_handle.create_context,
-             async_handle.priority, async_handle.stats);
-}
-
-Cache::Handle* Cache::Wait(AsyncLookupHandle& async_handle) {
-  WaitAll(&async_handle, 1);
-  return async_handle.Result();
-}
-
-void Cache::WaitAll(AsyncLookupHandle* async_handles, size_t count) {
-  for (size_t i = 0; i < count; ++i) {
-    if (async_handles[i].IsPending()) {
-      // If a pending handle gets here, it should be marked at "to be handled
-      // by a caller" by that caller erasing the pending_cache on it.
-      assert(async_handles[i].pending_cache == nullptr);
-    }
-  }
-}
-
-void Cache::SetEvictionCallback(EvictionCallback&& fn) {
-  // Overwriting non-empty with non-empty could indicate a bug
-  assert(!eviction_callback_ || !fn);
-  eviction_callback_ = std::move(fn);
-}
-
 }  // namespace ROCKSDB_NAMESPACE
@@ -16,7 +16,7 @@
 #include "db/db_impl/db_impl.h"
 #include "monitoring/histogram.h"
 #include "port/port.h"
-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "rocksdb/convenience.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
@@ -77,9 +77,11 @@ DEFINE_bool(lean, false,
            "If true, no additional computation is performed besides cache "
            "operations.");

+#ifndef ROCKSDB_LITE
 DEFINE_string(secondary_cache_uri, "",
              "Full URI for creating a custom secondary cache object");
 static class std::shared_ptr<ROCKSDB_NAMESPACE::SecondaryCache> secondary_cache;
+#endif  // ROCKSDB_LITE

 DEFINE_string(cache_type, "lru_cache", "Type of block cache.");

@@ -255,15 +257,12 @@ void DeleteFn(Cache::ObjectPtr value, MemoryAllocator* /*alloc*/) {
  delete[] static_cast<char*>(value);
 }

-Cache::CacheItemHelper helper1_wos(CacheEntryRole::kDataBlock, DeleteFn);
 Cache::CacheItemHelper helper1(CacheEntryRole::kDataBlock, DeleteFn, SizeFn,
-                               SaveToFn, CreateFn, &helper1_wos);
-Cache::CacheItemHelper helper2_wos(CacheEntryRole::kIndexBlock, DeleteFn);
+                               SaveToFn, CreateFn);
 Cache::CacheItemHelper helper2(CacheEntryRole::kIndexBlock, DeleteFn, SizeFn,
-                               SaveToFn, CreateFn, &helper2_wos);
-Cache::CacheItemHelper helper3_wos(CacheEntryRole::kFilterBlock, DeleteFn);
+                               SaveToFn, CreateFn);
 Cache::CacheItemHelper helper3(CacheEntryRole::kFilterBlock, DeleteFn, SizeFn,
-                               SaveToFn, CreateFn, &helper3_wos);
+                               SaveToFn, CreateFn);
 }  // namespace

 class CacheBench {
@@ -306,6 +305,7 @@ class CacheBench {
      LRUCacheOptions opts(FLAGS_cache_size, FLAGS_num_shard_bits,
                           false /* strict_capacity_limit */,
                           0.5 /* high_pri_pool_ratio */);
+#ifndef ROCKSDB_LITE
      if (!FLAGS_secondary_cache_uri.empty()) {
        Status s = SecondaryCache::CreateFromString(
            ConfigOptions(), FLAGS_secondary_cache_uri, &secondary_cache);
@@ -318,6 +318,7 @@ class CacheBench {
        }
        opts.secondary_cache = secondary_cache;
      }
+#endif  // ROCKSDB_LITE

      cache_ = NewLRUCache(opts);
    } else {
@@ -547,7 +548,7 @@ class CacheBench {
        }
        // do lookup
        handle = cache_->Lookup(key, &helper2, /*context*/ nullptr,
-                                Cache::Priority::LOW);
+                                Cache::Priority::LOW, true);
        if (handle) {
          if (!FLAGS_lean) {
            // do something with the data
@@ -576,7 +577,7 @@ class CacheBench {
        }
        // do lookup
        handle = cache_->Lookup(key, &helper2, /*context*/ nullptr,
-                                Cache::Priority::LOW);
+                                Cache::Priority::LOW, true);
        if (handle) {
          if (!FLAGS_lean) {
            // do something with the data
@@ -143,7 +143,7 @@ class CacheEntryStatsCollector {
      }
    }
    // If we reach here, shared entry is in cache with handle `h`.
-    assert(cache.get()->GetCacheItemHelper(h) == cache.GetBasicHelper());
+    assert(cache.get()->GetCacheItemHelper(h) == &cache.kBasicHelper);

    // Build an aliasing shared_ptr that keeps `ptr` in cache while there
    // are references.
@@ -7,7 +7,7 @@

 #include <cassert>

-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "rocksdb/rocksdb_namespace.h"

 namespace ROCKSDB_NAMESPACE {
@@ -8,7 +8,7 @@
 #include <algorithm>
 #include <atomic>

-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "table/unique_id_impl.h"
 #include "util/hash.h"
 #include "util/math.h"
@@ -169,7 +169,7 @@ Slice CacheReservationManagerImpl<R>::GetNextCacheKey() {
 template <CacheEntryRole R>
 const Cache::CacheItemHelper*
 CacheReservationManagerImpl<R>::TEST_GetCacheItemHelperForRole() {
-  return CacheInterface::GetHelper();
+  return &CacheInterface::kHelper;
 }

 template class CacheReservationManagerImpl<
@@ -18,7 +18,6 @@
 #include "cache/lru_cache.h"
 #include "cache/typed_cache.h"
 #include "port/stack_trace.h"
-#include "test_util/secondary_cache_test_util.h"
 #include "test_util/testharness.h"
 #include "util/coding.h"
 #include "util/string_util.h"
@@ -82,10 +81,13 @@ const Cache::CacheItemHelper kEraseOnDeleteHelper2{
      Cache* cache = static_cast<Cache*>(value);
      cache->Erase(EncodeKey16Bytes(1234));
    }};
+
+const std::string kLRU = "lru";
+const std::string kHyperClock = "hyper_clock";
+
 }  // anonymous namespace

-class CacheTest : public testing::Test,
-                  public secondary_cache_test_util::WithCacheTypeParam {
+class CacheTest : public testing::TestWithParam<std::string> {
 public:
  static CacheTest* current_;
  static std::string type_;
@@ -93,7 +95,8 @@ class CacheTest : public testing::Test,
  static void Deleter(Cache::ObjectPtr v, MemoryAllocator*) {
    current_->deleted_values_.push_back(DecodeValue(v));
  }
-  static const Cache::CacheItemHelper kHelper;
+  static constexpr Cache::CacheItemHelper kHelper{CacheEntryRole::kMisc,
+                                                  &Deleter};

  static const int kCacheSize = 1000;
  static const int kNumShardBits = 4;
@@ -105,6 +108,8 @@ class CacheTest : public testing::Test,
  std::shared_ptr<Cache> cache_;
  std::shared_ptr<Cache> cache2_;

+  size_t estimated_value_size_ = 1;
+
  CacheTest()
      : cache_(NewCache(kCacheSize, kNumShardBits, false)),
        cache2_(NewCache(kCacheSize2, kNumShardBits2, false)) {
@@ -114,6 +119,41 @@ class CacheTest : public testing::Test,

  ~CacheTest() override {}

+  std::shared_ptr<Cache> NewCache(size_t capacity) {
+    auto type = GetParam();
+    if (type == kLRU) {
+      return NewLRUCache(capacity);
+    }
+    if (type == kHyperClock) {
+      return HyperClockCacheOptions(
+                 capacity, estimated_value_size_ /*estimated_value_size*/)
+          .MakeSharedCache();
+    }
+    return nullptr;
+  }
+
+  std::shared_ptr<Cache> NewCache(
+      size_t capacity, int num_shard_bits, bool strict_capacity_limit,
+      CacheMetadataChargePolicy charge_policy = kDontChargeCacheMetadata) {
+    auto type = GetParam();
+    if (type == kLRU) {
+      LRUCacheOptions co;
+      co.capacity = capacity;
+      co.num_shard_bits = num_shard_bits;
+      co.strict_capacity_limit = strict_capacity_limit;
+      co.high_pri_pool_ratio = 0;
+      co.metadata_charge_policy = charge_policy;
+      return NewLRUCache(co);
+    }
+    if (type == kHyperClock) {
+      return HyperClockCacheOptions(capacity, 1 /*estimated_value_size*/,
+                                    num_shard_bits, strict_capacity_limit,
+                                    nullptr /*allocator*/, charge_policy)
+          .MakeSharedCache();
+    }
+    return nullptr;
+  }
+
  // These functions encode/decode keys in tests cases that use
  // int keys.
  // Currently, HyperClockCache requires keys to be 16B long, whereas
@@ -147,8 +187,8 @@ class CacheTest : public testing::Test,

  void Insert(std::shared_ptr<Cache> cache, int key, int value,
              int charge = 1) {
-    EXPECT_OK(cache->Insert(EncodeKey(key), EncodeValue(value), &kHelper,
-                            charge, /*handle*/ nullptr, Cache::Priority::HIGH));
+    EXPECT_OK(
+        cache->Insert(EncodeKey(key), EncodeValue(value), &kHelper, charge));
  }

  void Erase(std::shared_ptr<Cache> cache, int key) {
@@ -172,9 +212,6 @@ class CacheTest : public testing::Test,
  void Erase2(int key) { Erase(cache2_, key); }
 };

-const Cache::CacheItemHelper CacheTest::kHelper{CacheEntryRole::kMisc,
-                                                &CacheTest::Deleter};
-
 CacheTest* CacheTest::current_;
 std::string CacheTest::type_;

@@ -956,9 +993,8 @@ TEST_P(CacheTest, GetChargeAndDeleter) {
 }

 INSTANTIATE_TEST_CASE_P(CacheTestInstance, CacheTest,
-                        secondary_cache_test_util::GetTestingCacheTypes());
-INSTANTIATE_TEST_CASE_P(CacheTestInstance, LRUCacheTest,
-                        testing::Values(secondary_cache_test_util::kLRU));
+                        testing::Values(kLRU, kHyperClock));
+INSTANTIATE_TEST_CASE_P(CacheTestInstance, LRUCacheTest, testing::Values(kLRU));

 }  // namespace ROCKSDB_NAMESPACE

@@ -11,7 +11,7 @@ namespace ROCKSDB_NAMESPACE {

 ChargedCache::ChargedCache(std::shared_ptr<Cache> cache,
                           std::shared_ptr<Cache> block_cache)
-    : CacheWrapper(cache),
+    : cache_(cache),
      cache_res_mgr_(std::make_shared<ConcurrentCacheReservationManager>(
          std::make_shared<
              CacheReservationManagerImpl<CacheEntryRole::kBlobCache>>(
@@ -20,13 +20,13 @@ ChargedCache::ChargedCache(std::shared_ptr<Cache> cache,
 Status ChargedCache::Insert(const Slice& key, ObjectPtr obj,
                            const CacheItemHelper* helper, size_t charge,
                            Handle** handle, Priority priority) {
-  Status s = target_->Insert(key, obj, helper, charge, handle, priority);
+  Status s = cache_->Insert(key, obj, helper, charge, handle, priority);
  if (s.ok()) {
    // Insert may cause the cache entry eviction if the cache is full. So we
    // directly call the reservation manager to update the total memory used
    // in the cache.
    assert(cache_res_mgr_);
-    cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
+    cache_res_mgr_->UpdateCacheReservation(cache_->GetUsage())
        .PermitUncheckedError();
  }
  return s;
@@ -35,33 +35,25 @@ Status ChargedCache::Insert(const Slice& key, ObjectPtr obj,
 Cache::Handle* ChargedCache::Lookup(const Slice& key,
                                    const CacheItemHelper* helper,
                                    CreateContext* create_context,
-                                    Priority priority, Statistics* stats) {
-  auto handle = target_->Lookup(key, helper, create_context, priority, stats);
+                                    Priority priority, bool wait,
+                                    Statistics* stats) {
+  auto handle =
+      cache_->Lookup(key, helper, create_context, priority, wait, stats);
  // Lookup may promote the KV pair from the secondary cache to the primary
  // cache. So we directly call the reservation manager to update the total
  // memory used in the cache.
  if (helper && helper->create_cb) {
    assert(cache_res_mgr_);
-    cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
+    cache_res_mgr_->UpdateCacheReservation(cache_->GetUsage())
        .PermitUncheckedError();
  }
  return handle;
 }

-void ChargedCache::WaitAll(AsyncLookupHandle* async_handles, size_t count) {
-  target_->WaitAll(async_handles, count);
-  // In case of any promotions. Although some could finish by return of
-  // StartAsyncLookup, Wait/WaitAll will generally be used, so simpler to
-  // update here.
-  assert(cache_res_mgr_);
-  cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
-      .PermitUncheckedError();
-}
-
 bool ChargedCache::Release(Cache::Handle* handle, bool useful,
                           bool erase_if_last_ref) {
-  size_t memory_used_delta = target_->GetUsage(handle);
-  bool erased = target_->Release(handle, useful, erase_if_last_ref);
+  size_t memory_used_delta = cache_->GetUsage(handle);
+  bool erased = cache_->Release(handle, useful, erase_if_last_ref);
  if (erased) {
    assert(cache_res_mgr_);
    cache_res_mgr_
@@ -72,8 +64,8 @@ bool ChargedCache::Release(Cache::Handle* handle, bool useful,
 }

 bool ChargedCache::Release(Cache::Handle* handle, bool erase_if_last_ref) {
-  size_t memory_used_delta = target_->GetUsage(handle);
-  bool erased = target_->Release(handle, erase_if_last_ref);
+  size_t memory_used_delta = cache_->GetUsage(handle);
+  bool erased = cache_->Release(handle, erase_if_last_ref);
  if (erased) {
    assert(cache_res_mgr_);
    cache_res_mgr_
@@ -84,25 +76,25 @@ bool ChargedCache::Release(Cache::Handle* handle, bool erase_if_last_ref) {
 }

 void ChargedCache::Erase(const Slice& key) {
-  target_->Erase(key);
+  cache_->Erase(key);
  assert(cache_res_mgr_);
-  cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
+  cache_res_mgr_->UpdateCacheReservation(cache_->GetUsage())
      .PermitUncheckedError();
 }

 void ChargedCache::EraseUnRefEntries() {
-  target_->EraseUnRefEntries();
+  cache_->EraseUnRefEntries();
  assert(cache_res_mgr_);
-  cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
+  cache_res_mgr_->UpdateCacheReservation(cache_->GetUsage())
      .PermitUncheckedError();
 }

 void ChargedCache::SetCapacity(size_t capacity) {
-  target_->SetCapacity(capacity);
+  cache_->SetCapacity(capacity);
  // SetCapacity can result in evictions when the cache capacity is decreased,
  // so we would want to update the cache reservation here as well.
  assert(cache_res_mgr_);
-  cache_res_mgr_->UpdateCacheReservation(target_->GetUsage())
+  cache_res_mgr_->UpdateCacheReservation(cache_->GetUsage())
      .PermitUncheckedError();
 }

@@ -8,7 +8,7 @@
 #include <string>

 #include "port/port.h"
-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"

 namespace ROCKSDB_NAMESPACE {

@@ -17,10 +17,11 @@ class ConcurrentCacheReservationManager;
 // A cache interface which wraps around another cache and takes care of
 // reserving space in block cache towards a single global memory limit, and
 // forwards all the calls to the underlying cache.
-class ChargedCache : public CacheWrapper {
+class ChargedCache : public Cache {
 public:
  ChargedCache(std::shared_ptr<Cache> cache,
               std::shared_ptr<Cache> block_cache);
+  ~ChargedCache() override = default;

  Status Insert(const Slice& key, ObjectPtr obj, const CacheItemHelper* helper,
                size_t charge, Handle** handle = nullptr,
@@ -28,11 +29,9 @@ class ChargedCache : public CacheWrapper {

  Cache::Handle* Lookup(const Slice& key, const CacheItemHelper* helper,
                        CreateContext* create_context,
-                        Priority priority = Priority::LOW,
+                        Priority priority = Priority::LOW, bool wait = true,
                        Statistics* stats = nullptr) override;

-  void WaitAll(AsyncLookupHandle* async_handles, size_t count) override;
-
  bool Release(Cache::Handle* handle, bool useful,
               bool erase_if_last_ref = false) override;
  bool Release(Cache::Handle* handle, bool erase_if_last_ref = false) override;
@@ -43,9 +42,66 @@ class ChargedCache : public CacheWrapper {
  static const char* kClassName() { return "ChargedCache"; }
  const char* Name() const override { return kClassName(); }

+  uint64_t NewId() override { return cache_->NewId(); }
+
  void SetCapacity(size_t capacity) override;

-  inline Cache* GetCache() const { return target_.get(); }
+  void SetStrictCapacityLimit(bool strict_capacity_limit) override {
+    cache_->SetStrictCapacityLimit(strict_capacity_limit);
+  }
+
+  bool HasStrictCapacityLimit() const override {
+    return cache_->HasStrictCapacityLimit();
+  }
+
+  ObjectPtr Value(Cache::Handle* handle) override {
+    return cache_->Value(handle);
+  }
+
+  bool IsReady(Cache::Handle* handle) override {
+    return cache_->IsReady(handle);
+  }
+
+  void Wait(Cache::Handle* handle) override { cache_->Wait(handle); }
+
+  void WaitAll(std::vector<Handle*>& handles) override {
+    cache_->WaitAll(handles);
+  }
+
+  bool Ref(Cache::Handle* handle) override { return cache_->Ref(handle); }
+
+  size_t GetCapacity() const override { return cache_->GetCapacity(); }
+
+  size_t GetUsage() const override { return cache_->GetUsage(); }
+
+  size_t GetUsage(Cache::Handle* handle) const override {
+    return cache_->GetUsage(handle);
+  }
+
+  size_t GetPinnedUsage() const override { return cache_->GetPinnedUsage(); }
+
+  size_t GetCharge(Cache::Handle* handle) const override {
+    return cache_->GetCharge(handle);
+  }
+
+  const CacheItemHelper* GetCacheItemHelper(Handle* handle) const override {
+    return cache_->GetCacheItemHelper(handle);
+  }
+
+  void ApplyToAllEntries(
+      const std::function<void(const Slice& key, ObjectPtr value, size_t charge,
+                               const CacheItemHelper* helper)>& callback,
+      const Cache::ApplyToAllEntriesOptions& opts) override {
+    cache_->ApplyToAllEntries(callback, opts);
+  }
+
+  std::string GetPrintableOptions() const override {
+    return cache_->GetPrintableOptions();
+  }
+
+  void DisownData() override { return cache_->DisownData(); }
+
+  inline Cache* GetCache() const { return cache_.get(); }

  inline ConcurrentCacheReservationManager* TEST_GetCacheReservationManager()
      const {
@@ -53,6 +109,7 @@ class ChargedCache : public CacheWrapper {
  }

 private:
+  std::shared_ptr<Cache> cache_;
  std::shared_ptr<ConcurrentCacheReservationManager> cache_res_mgr_;
 };

@@ -9,16 +9,15 @@

 #include "cache/clock_cache.h"

+#include <cassert>
 #include <functional>
 #include <numeric>

 #include "cache/cache_key.h"
-#include "cache/secondary_cache_adapter.h"
 #include "logging/logging.h"
 #include "monitoring/perf_context_imp.h"
 #include "monitoring/statistics.h"
 #include "port/lang.h"
-#include "rocksdb/env.h"
 #include "util/hash.h"
 #include "util/math.h"
 #include "util/random.h"
@@ -51,7 +50,13 @@ inline uint64_t GetInitialCountdown(Cache::Priority priority) {
  }
 }

-inline void MarkEmpty(ClockHandle& h) {
+inline void FreeDataMarkEmpty(ClockHandle& h, MemoryAllocator* allocator) {
+  // NOTE: in theory there's more room for parallelism if we copy the handle
+  // data and delay actions like this until after marking the entry as empty,
+  // but performance tests only show a regression by copying the few words
+  // of data.
+  h.FreeData(allocator);
+
 #ifndef NDEBUG
  // Mark slot as empty, with assertion
  uint64_t meta = h.meta.exchange(0, std::memory_order_release);
@@ -62,16 +67,6 @@ inline void MarkEmpty(ClockHandle& h) {
 #endif
 }

-inline void FreeDataMarkEmpty(ClockHandle& h, MemoryAllocator* allocator) {
-  // NOTE: in theory there's more room for parallelism if we copy the handle
-  // data and delay actions like this until after marking the entry as empty,
-  // but performance tests only show a regression by copying the few words
-  // of data.
-  h.FreeData(allocator);
-
-  MarkEmpty(h);
-}
-
 inline bool ClockUpdate(ClockHandle& h) {
  uint64_t meta = h.meta.load(std::memory_order_relaxed);

@@ -129,16 +124,14 @@ void ClockHandleBasicData::FreeData(MemoryAllocator* allocator) const {
 HyperClockTable::HyperClockTable(
    size_t capacity, bool /*strict_capacity_limit*/,
    CacheMetadataChargePolicy metadata_charge_policy,
-    MemoryAllocator* allocator,
-    const Cache::EvictionCallback* eviction_callback, const Opts& opts)
+    MemoryAllocator* allocator, const Opts& opts)
    : length_bits_(CalcHashBits(capacity, opts.estimated_value_size,
                                metadata_charge_policy)),
      length_bits_mask_((size_t{1} << length_bits_) - 1),
      occupancy_limit_(static_cast<size_t>((uint64_t{1} << length_bits_) *
                                           kStrictLoadFactor)),
      array_(new HandleImpl[size_t{1} << length_bits_]),
-      allocator_(allocator),
-      eviction_callback_(*eviction_callback) {
+      allocator_(allocator) {
  if (metadata_charge_policy ==
      CacheMetadataChargePolicy::kFullChargeCacheMetadata) {
    usage_ += size_t{GetTableSize()} * sizeof(HandleImpl);
@@ -371,22 +364,21 @@ inline bool HyperClockTable::ChargeUsageMaybeEvictNonStrict(
  return true;
 }

-inline HyperClockTable::HandleImpl* HyperClockTable::StandaloneInsert(
+inline HyperClockTable::HandleImpl* HyperClockTable::DetachedInsert(
    const ClockHandleBasicData& proto) {
  // Heap allocated separate from table
  HandleImpl* h = new HandleImpl();
  ClockHandleBasicData* h_alias = h;
  *h_alias = proto;
-  h->SetStandalone();
-  // Single reference (standalone entries only created if returning a refed
+  h->SetDetached();
+  // Single reference (detached entries only created if returning a refed
  // Handle back to user)
  uint64_t meta = uint64_t{ClockHandle::kStateInvisible}
                  << ClockHandle::kStateShift;
  meta |= uint64_t{1} << ClockHandle::kAcquireCounterShift;
  h->meta.store(meta, std::memory_order_release);
-  // Keep track of how much of usage is standalone
-  standalone_usage_.fetch_add(proto.GetTotalCharge(),
-                              std::memory_order_relaxed);
+  // Keep track of how much of usage is detached
+  detached_usage_.fetch_add(proto.GetTotalCharge(), std::memory_order_relaxed);
  return h;
 }

@@ -404,7 +396,7 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,

  // Usage/capacity handling is somewhat different depending on
  // strict_capacity_limit, but mostly pessimistic.
-  bool use_standalone_insert = false;
+  bool use_detached_insert = false;
  const size_t total_charge = proto.GetTotalCharge();
  if (strict_capacity_limit) {
    Status s = ChargeUsageMaybeEvictStrict(total_charge, capacity,
@@ -425,9 +417,9 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
        proto.FreeData(allocator_);
        return Status::OK();
      } else {
-        // Need to track usage of fallback standalone insert
+        // Need to track usage of fallback detached insert
        usage_.fetch_add(total_charge, std::memory_order_relaxed);
-        use_standalone_insert = true;
+        use_detached_insert = true;
      }
    }
  }
@@ -437,7 +429,7 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
    assert(usage_.load(std::memory_order_relaxed) < SIZE_MAX / 2);
  };

-  if (!use_standalone_insert) {
+  if (!use_detached_insert) {
    // Attempt a table insert, but abort if we find an existing entry for the
    // key. If we were to overwrite old entries, we would either
    // * Have to gain ownership over an existing entry to overwrite it, which
@@ -508,8 +500,8 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
                  std::memory_order_acq_rel);
              // Correct for possible (but rare) overflow
              CorrectNearOverflow(old_meta, h->meta);
-              // Insert standalone instead (only if return handle needed)
-              use_standalone_insert = true;
+              // Insert detached instead (only if return handle needed)
+              use_detached_insert = true;
              return true;
            } else {
              // Mismatch. Pretend we never took the reference
@@ -547,9 +539,9 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
      // That should be infeasible for roughly n >= 256, so if this assertion
      // fails, that suggests something is going wrong.
      assert(GetTableSize() < 256);
-      use_standalone_insert = true;
+      use_detached_insert = true;
    }
-    if (!use_standalone_insert) {
+    if (!use_detached_insert) {
      // Successfully inserted
      if (handle) {
        *handle = e;
@@ -559,7 +551,7 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
    // Roll back table insertion
    Rollback(proto.hashed_key, e);
    revert_occupancy_fn();
-    // Maybe fall back on standalone insert
+    // Maybe fall back on detached insert
    if (handle == nullptr) {
      revert_usage_fn();
      // As if unrefed entry immdiately evicted
@@ -568,47 +560,19 @@ Status HyperClockTable::Insert(const ClockHandleBasicData& proto,
    }
  }

-  // Run standalone insert
-  assert(use_standalone_insert);
+  // Run detached insert
+  assert(use_detached_insert);

-  *handle = StandaloneInsert(proto);
+  *handle = DetachedInsert(proto);

  // The OkOverwritten status is used to count "redundant" insertions into
  // block cache. This implementation doesn't strictly check for redundant
  // insertions, but we instead are probably interested in how many insertions
-  // didn't go into the table (instead "standalone"), which could be redundant
-  // Insert or some other reason (use_standalone_insert reasons above).
+  // didn't go into the table (instead "detached"), which could be redundant
+  // Insert or some other reason (use_detached_insert reasons above).
  return Status::OkOverwritten();
 }

-HyperClockTable::HandleImpl* HyperClockTable::CreateStandalone(
-    ClockHandleBasicData& proto, size_t capacity, bool strict_capacity_limit,
-    bool allow_uncharged) {
-  const size_t total_charge = proto.GetTotalCharge();
-  if (strict_capacity_limit) {
-    Status s = ChargeUsageMaybeEvictStrict(total_charge, capacity,
-                                           /*need_evict_for_occupancy=*/false);
-    if (!s.ok()) {
-      if (allow_uncharged) {
-        proto.total_charge = 0;
-      } else {
-        return nullptr;
-      }
-    }
-  } else {
-    // Case strict_capacity_limit == false
-    bool success =
-        ChargeUsageMaybeEvictNonStrict(total_charge, capacity,
-                                       /*need_evict_for_occupancy=*/false);
-    if (!success) {
-      // Force the issue
-      usage_.fetch_add(total_charge, std::memory_order_relaxed);
-    }
-  }
-
-  return StandaloneInsert(proto);
-}
-
 HyperClockTable::HandleImpl* HyperClockTable::Lookup(
    const UniqueId64x2& hashed_key) {
  size_t probe = 0;
@@ -732,11 +696,11 @@ bool HyperClockTable::Release(HandleImpl* h, bool useful,
        std::memory_order_acquire));
    // Took ownership
    size_t total_charge = h->GetTotalCharge();
-    if (UNLIKELY(h->IsStandalone())) {
+    if (UNLIKELY(h->IsDetached())) {
      h->FreeData(allocator_);
-      // Delete standalone handle
+      // Delete detached handle
      delete h;
-      standalone_usage_.fetch_sub(total_charge, std::memory_order_relaxed);
+      detached_usage_.fetch_sub(total_charge, std::memory_order_relaxed);
      usage_.fetch_sub(total_charge, std::memory_order_relaxed);
    } else {
      Rollback(h->hashed_key, h);
@@ -995,9 +959,6 @@ inline void HyperClockTable::Evict(size_t requested_charge,
  uint64_t max_clock_pointer =
      old_clock_pointer + (ClockHandle::kMaxCountdown << length_bits_);

-  // For key reconstructed from hash
-  UniqueId64x2 unhashed;
-
  for (;;) {
    for (size_t i = 0; i < step_size; i++) {
      HandleImpl& h = array_[ModTableSize(Lower32of64(old_clock_pointer + i))];
@@ -1006,17 +967,7 @@ inline void HyperClockTable::Evict(size_t requested_charge,
        Rollback(h.hashed_key, &h);
        *freed_charge += h.GetTotalCharge();
        *freed_count += 1;
-        bool took_ownership = false;
-        if (eviction_callback_) {
-          took_ownership =
-              eviction_callback_(ClockCacheShard<HyperClockTable>::ReverseHash(
-                                     h.GetHash(), &unhashed),
-                                 reinterpret_cast<Cache::Handle*>(&h));
-        }
-        if (!took_ownership) {
-          h.FreeData(allocator_);
-        }
-        MarkEmpty(h);
+        FreeDataMarkEmpty(h, allocator_);
      }
    }

@@ -1038,12 +989,10 @@ template <class Table>
 ClockCacheShard<Table>::ClockCacheShard(
    size_t capacity, bool strict_capacity_limit,
    CacheMetadataChargePolicy metadata_charge_policy,
-    MemoryAllocator* allocator,
-    const Cache::EvictionCallback* eviction_callback,
-    const typename Table::Opts& opts)
+    MemoryAllocator* allocator, const typename Table::Opts& opts)
    : CacheShardBase(metadata_charge_policy),
      table_(capacity, strict_capacity_limit, metadata_charge_policy, allocator,
-             eviction_callback, opts),
+             opts),
      capacity_(capacity),
      strict_capacity_limit_(strict_capacity_limit) {
  // Initial charge metadata should not exceed capacity
@@ -1143,29 +1092,10 @@ Status ClockCacheShard<Table>::Insert(const Slice& key,
  proto.value = value;
  proto.helper = helper;
  proto.total_charge = charge;
-  return table_.Insert(proto, handle, priority,
-                       capacity_.load(std::memory_order_relaxed),
-                       strict_capacity_limit_.load(std::memory_order_relaxed));
-}
-
-template <class Table>
-typename ClockCacheShard<Table>::HandleImpl*
-ClockCacheShard<Table>::CreateStandalone(const Slice& key,
-                                         const UniqueId64x2& hashed_key,
-                                         Cache::ObjectPtr obj,
-                                         const Cache::CacheItemHelper* helper,
-                                         size_t charge, bool allow_uncharged) {
-  if (UNLIKELY(key.size() != kCacheKeySize)) {
-    return nullptr;
-  }
-  ClockHandleBasicData proto;
-  proto.hashed_key = hashed_key;
-  proto.value = obj;
-  proto.helper = helper;
-  proto.total_charge = charge;
-  return table_.CreateStandalone(
-      proto, capacity_.load(std::memory_order_relaxed),
-      strict_capacity_limit_.load(std::memory_order_relaxed), allow_uncharged);
+  Status s = table_.Insert(
+      proto, handle, priority, capacity_.load(std::memory_order_relaxed),
+      strict_capacity_limit_.load(std::memory_order_relaxed));
+  return s;
 }

 template <class Table>
@@ -1226,8 +1156,8 @@ size_t ClockCacheShard<Table>::GetUsage() const {
 }

 template <class Table>
-size_t ClockCacheShard<Table>::GetStandaloneUsage() const {
-  return table_.GetStandaloneUsage();
+size_t ClockCacheShard<Table>::GetDetachedUsage() const {
+  return table_.GetDetachedUsage();
 }

 template <class Table>
@@ -1261,7 +1191,7 @@ size_t ClockCacheShard<Table>::GetPinnedUsage() const {
      },
      0, table_.GetTableSize(), true);

-  return table_pinned_usage + table_.GetStandaloneUsage();
+  return table_pinned_usage + table_.GetDetachedUsage();
 }

 template <class Table>
@@ -1295,12 +1225,11 @@ HyperClockCache::HyperClockCache(
  // get to table entries
  size_t per_shard = GetPerShardCapacity();
  MemoryAllocator* alloc = this->memory_allocator();
-  const Cache::EvictionCallback* eviction_callback = &eviction_callback_;
  InitShards([=](Shard* cs) {
    HyperClockTable::Opts opts;
    opts.estimated_value_size = estimated_value_size;
    new (cs) Shard(per_shard, strict_capacity_limit, metadata_charge_policy,
-                   alloc, eviction_callback, opts);
+                   alloc, opts);
  });
 }

@@ -1330,7 +1259,7 @@ namespace {
 void AddShardEvaluation(const HyperClockCache::Shard& shard,
                        std::vector<double>& predicted_load_factors,
                        size_t& min_recommendation) {
-  size_t usage = shard.GetUsage() - shard.GetStandaloneUsage();
+  size_t usage = shard.GetUsage() - shard.GetDetachedUsage();
  size_t capacity = shard.GetCapacity();
  double usage_ratio = 1.0 * usage / capacity;

@@ -1470,13 +1399,9 @@ std::shared_ptr<Cache> HyperClockCacheOptions::MakeSharedCache() const {
    constexpr size_t min_shard_size = 32U * 1024U * 1024U;
    my_num_shard_bits = GetDefaultCacheShardBits(capacity, min_shard_size);
  }
-  std::shared_ptr<Cache> cache = std::make_shared<clock_cache::HyperClockCache>(
+  return std::make_shared<clock_cache::HyperClockCache>(
      capacity, estimated_entry_charge, my_num_shard_bits,
      strict_capacity_limit, metadata_charge_policy, memory_allocator);
-  if (secondary_cache) {
-    cache = std::make_shared<CacheWithSecondaryAdapter>(cache, secondary_cache);
-  }
-  return cache;
 }

 }  // namespace ROCKSDB_NAMESPACE
@@ -145,7 +145,7 @@ class ClockCacheTest;
 //     (erased by user) but can be read by existing references, and ref count
 //     changed by Ref and Release.
 //
-// A special case is "standalone" entries, which are heap-allocated handles
+// A special case is "detached" entries, which are heap-allocated handles
 // not in the table. They are always Invisible and freed on zero refs.
 //
 // State transitions:
@@ -200,8 +200,8 @@ class ClockCacheTest;
 // table occupancy limit has been reached. If strict_capacity_limit=false,
 // we must never fail Insert, and if a Handle* is provided, we have to return
 // a usable Cache handle on success. The solution to this (typically rare)
-// problem is "standalone" handles, which are usable by the caller but not
-// actually available for Lookup in the Cache. Standalone handles are allocated
+// problem is "detached" handles, which are usable by the caller but not
+// actually available for Lookup in the Cache. Detached handles are allocated
 // independently on the heap and specially marked so that they are freed on
 // the heap when their last reference is released.
 //
@@ -312,6 +312,12 @@ struct ClockHandleBasicData {
  UniqueId64x2 hashed_key = kNullUniqueId64x2;
  size_t total_charge = 0;

+  // For total_charge_and_flags
+  // "Detached" means the handle is allocated separately from hash table.
+  static constexpr uint64_t kFlagDetached = uint64_t{1} << 63;
+  // Extract just the total charge
+  static constexpr uint64_t kTotalChargeMask = kFlagDetached - 1;
+
  inline size_t GetTotalCharge() const { return total_charge; }

  // Calls deleter (if non-null) on cache key and value
@@ -392,11 +398,11 @@ class HyperClockTable {
    // TODO: ideally this would be packed into some other data field, such
    // as upper bits of total_charge, but that incurs a measurable performance
    // regression.
-    bool standalone = false;
+    bool detached = false;

-    inline bool IsStandalone() const { return standalone; }
+    inline bool IsDetached() const { return detached; }

-    inline void SetStandalone() { standalone = true; }
+    inline void SetDetached() { detached = true; }
  };  // struct HandleImpl

  struct Opts {
@@ -405,19 +411,13 @@ class HyperClockTable {

  HyperClockTable(size_t capacity, bool strict_capacity_limit,
                  CacheMetadataChargePolicy metadata_charge_policy,
-                  MemoryAllocator* allocator,
-                  const Cache::EvictionCallback* eviction_callback,
-                  const Opts& opts);
+                  MemoryAllocator* allocator, const Opts& opts);
  ~HyperClockTable();

  Status Insert(const ClockHandleBasicData& proto, HandleImpl** handle,
                Cache::Priority priority, size_t capacity,
                bool strict_capacity_limit);

-  HandleImpl* CreateStandalone(ClockHandleBasicData& proto, size_t capacity,
-                               bool strict_capacity_limit,
-                               bool allow_uncharged);
-
  HandleImpl* Lookup(const UniqueId64x2& hashed_key);

  bool Release(HandleImpl* handle, bool useful, bool erase_if_last_ref);
@@ -444,8 +444,8 @@ class HyperClockTable {

  size_t GetUsage() const { return usage_.load(std::memory_order_relaxed); }

-  size_t GetStandaloneUsage() const {
-    return standalone_usage_.load(std::memory_order_relaxed);
+  size_t GetDetachedUsage() const {
+    return detached_usage_.load(std::memory_order_relaxed);
  }

  // Acquire/release N references
@@ -514,10 +514,10 @@ class HyperClockTable {
                                             size_t capacity,
                                             bool need_evict_for_occupancy);

-  // Creates a "standalone" handle for returning from an Insert operation that
+  // Creates a "detached" handle for returning from an Insert operation that
  // cannot be completed by actually inserting into the table.
-  // Updates `standalone_usage_` but not `usage_` nor `occupancy_`.
-  inline HandleImpl* StandaloneInsert(const ClockHandleBasicData& proto);
+  // Updates `detached_usage_` but not `usage_` nor `occupancy_`.
+  inline HandleImpl* DetachedInsert(const ClockHandleBasicData& proto);

  MemoryAllocator* GetAllocator() const { return allocator_; }

@@ -543,9 +543,6 @@ class HyperClockTable {
  // From Cache, for deleter
  MemoryAllocator* const allocator_;

-  // A reference to Cache::eviction_callback_
-  const Cache::EvictionCallback& eviction_callback_;
-
  // We partition the following members into different cache lines
  // to avoid false sharing among Lookup, Release, Erase and Insert
  // operations in ClockCacheShard.
@@ -558,11 +555,11 @@ class HyperClockTable {
  // Number of elements in the table.
  std::atomic<size_t> occupancy_{};

-  // Memory usage by entries tracked by the cache (including standalone)
+  // Memory usage by entries tracked by the cache (including detached)
  std::atomic<size_t> usage_{};

-  // Part of usage by standalone entries (not in table)
-  std::atomic<size_t> standalone_usage_{};
+  // Part of usage by detached entries (not in table)
+  std::atomic<size_t> detached_usage_{};
 };  // class HyperClockTable

 // A single shard of sharded cache.
@@ -571,9 +568,7 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {
 public:
  ClockCacheShard(size_t capacity, bool strict_capacity_limit,
                  CacheMetadataChargePolicy metadata_charge_policy,
-                  MemoryAllocator* allocator,
-                  const Cache::EvictionCallback* eviction_callback,
-                  const typename Table::Opts& opts);
+                  MemoryAllocator* allocator, const typename Table::Opts& opts);

  // For CacheShard concept
  using HandleImpl = typename Table::HandleImpl;
@@ -614,11 +609,6 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {
                Cache::ObjectPtr value, const Cache::CacheItemHelper* helper,
                size_t charge, HandleImpl** handle, Cache::Priority priority);

-  HandleImpl* CreateStandalone(const Slice& key, const UniqueId64x2& hashed_key,
-                               Cache::ObjectPtr obj,
-                               const Cache::CacheItemHelper* helper,
-                               size_t charge, bool allow_uncharged);
-
  HandleImpl* Lookup(const Slice& key, const UniqueId64x2& hashed_key);

  bool Release(HandleImpl* handle, bool useful, bool erase_if_last_ref);
@@ -633,7 +623,7 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {

  size_t GetUsage() const;

-  size_t GetStandaloneUsage() const;
+  size_t GetDetachedUsage() const;

  size_t GetPinnedUsage() const;

@@ -656,10 +646,15 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {
  HandleImpl* Lookup(const Slice& key, const UniqueId64x2& hashed_key,
                     const Cache::CacheItemHelper* /*helper*/,
                     Cache::CreateContext* /*create_context*/,
-                     Cache::Priority /*priority*/, Statistics* /*stats*/) {
+                     Cache::Priority /*priority*/, bool /*wait*/,
+                     Statistics* /*stats*/) {
    return Lookup(key, hashed_key);
  }

+  bool IsReady(HandleImpl* /*handle*/) { return true; }
+
+  void Wait(HandleImpl* /*handle*/) {}
+
  // Acquire/release N references
  void TEST_RefN(HandleImpl* handle, size_t n);
  void TEST_ReleaseN(HandleImpl* handle, size_t n);
@@ -22,13 +22,12 @@ CompressedSecondaryCache::CompressedSecondaryCache(
    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
    CacheMetadataChargePolicy metadata_charge_policy,
    CompressionType compression_type, uint32_t compress_format_version,
-    bool enable_custom_split_merge,
-    const CacheEntryRoleSet& do_not_compress_roles)
+    bool enable_custom_split_merge)
    : cache_options_(capacity, num_shard_bits, strict_capacity_limit,
                     high_pri_pool_ratio, low_pri_pool_ratio, memory_allocator,
                     use_adaptive_mutex, metadata_charge_policy,
                     compression_type, compress_format_version,
-                     enable_custom_split_merge, do_not_compress_roles) {
+                     enable_custom_split_merge) {
  cache_ =
      NewLRUCache(capacity, num_shard_bits, strict_capacity_limit,
                  high_pri_pool_ratio, memory_allocator, use_adaptive_mutex,
@@ -40,10 +39,10 @@ CompressedSecondaryCache::~CompressedSecondaryCache() { cache_.reset(); }
 std::unique_ptr<SecondaryCacheResultHandle> CompressedSecondaryCache::Lookup(
    const Slice& key, const Cache::CacheItemHelper* helper,
    Cache::CreateContext* create_context, bool /*wait*/, bool advise_erase,
-    bool& kept_in_sec_cache) {
+    bool& is_in_sec_cache) {
  assert(helper);
  std::unique_ptr<SecondaryCacheResultHandle> handle;
-  kept_in_sec_cache = false;
+  is_in_sec_cache = false;
  Cache::Handle* lru_handle = cache_->Lookup(key);
  if (lru_handle == nullptr) {
    return nullptr;
@@ -72,8 +71,7 @@ std::unique_ptr<SecondaryCacheResultHandle> CompressedSecondaryCache::Lookup(
  Status s;
  Cache::ObjectPtr value{nullptr};
  size_t charge{0};
-  if (cache_options_.compression_type == kNoCompression ||
-      cache_options_.do_not_compress_roles.Contains(helper->role)) {
+  if (cache_options_.compression_type == kNoCompression) {
    s = helper->create_cb(Slice(ptr->get(), handle_value_charge),
                          create_context, allocator, &value, &charge);
  } else {
@@ -109,7 +107,7 @@ std::unique_ptr<SecondaryCacheResultHandle> CompressedSecondaryCache::Lookup(
                 /*charge=*/0)
        .PermitUncheckedError();
  } else {
-    kept_in_sec_cache = true;
+    is_in_sec_cache = true;
    cache_->Release(lru_handle, /*erase_if_last_ref=*/false);
  }
  handle.reset(new CompressedSecondaryCacheResultHandle(value, charge));
@@ -145,8 +143,7 @@ Status CompressedSecondaryCache::Insert(const Slice& key,
  Slice val(ptr.get(), size);

  std::string compressed_val;
-  if (cache_options_.compression_type != kNoCompression &&
-      !cache_options_.do_not_compress_roles.Contains(helper->role)) {
+  if (cache_options_.compression_type != kNoCompression) {
    PERF_COUNTER_ADD(compressed_sec_cache_uncompressed_bytes, size);
    CompressionOptions compression_opts;
    CompressionContext compression_context(cache_options_.compression_type);
@@ -317,13 +314,12 @@ std::shared_ptr<SecondaryCache> NewCompressedSecondaryCache(
    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
    CacheMetadataChargePolicy metadata_charge_policy,
    CompressionType compression_type, uint32_t compress_format_version,
-    bool enable_custom_split_merge,
-    const CacheEntryRoleSet& do_not_compress_roles) {
+    bool enable_custom_split_merge) {
  return std::make_shared<CompressedSecondaryCache>(
      capacity, num_shard_bits, strict_capacity_limit, high_pri_pool_ratio,
      low_pri_pool_ratio, memory_allocator, use_adaptive_mutex,
      metadata_charge_policy, compression_type, compress_format_version,
-      enable_custom_split_merge, do_not_compress_roles);
+      enable_custom_split_merge);
 }

 std::shared_ptr<SecondaryCache> NewCompressedSecondaryCache(
@@ -335,7 +331,7 @@ std::shared_ptr<SecondaryCache> NewCompressedSecondaryCache(
      opts.high_pri_pool_ratio, opts.low_pri_pool_ratio, opts.memory_allocator,
      opts.use_adaptive_mutex, opts.metadata_charge_policy,
      opts.compression_type, opts.compress_format_version,
-      opts.enable_custom_split_merge, opts.do_not_compress_roles);
+      opts.enable_custom_split_merge);
 }

 }  // namespace ROCKSDB_NAMESPACE
@@ -78,9 +78,7 @@ class CompressedSecondaryCache : public SecondaryCache {
          kDefaultCacheMetadataChargePolicy,
      CompressionType compression_type = CompressionType::kLZ4Compression,
      uint32_t compress_format_version = 2,
-      bool enable_custom_split_merge = false,
-      const CacheEntryRoleSet& do_not_compress_roles = {
-          CacheEntryRole::kFilterBlock});
+      bool enable_custom_split_merge = false);
  ~CompressedSecondaryCache() override;

  const char* Name() const override { return "CompressedSecondaryCache"; }
@@ -91,7 +89,7 @@ class CompressedSecondaryCache : public SecondaryCache {
  std::unique_ptr<SecondaryCacheResultHandle> Lookup(
      const Slice& key, const Cache::CacheItemHelper* helper,
      Cache::CreateContext* create_context, bool /*wait*/, bool advise_erase,
-      bool& kept_in_sec_cache) override;
+      bool& is_in_sec_cache) override;

  bool SupportForceErase() const override { return true; }

@@ -106,7 +104,7 @@ class CompressedSecondaryCache : public SecondaryCache {
  std::string GetPrintableOptions() const override;

 private:
-  friend class CompressedSecondaryCacheTestBase;
+  friend class CompressedSecondaryCacheTest;
  static constexpr std::array<uint16_t, 8> malloc_bin_sizes_{
      128, 256, 512, 1024, 2048, 4096, 8192, 16384};

@@ -5,43 +5,93 @@

 #include "cache/compressed_secondary_cache.h"

-#include <array>
 #include <iterator>
 #include <memory>
 #include <tuple>

 #include "memory/jemalloc_nodump_allocator.h"
 #include "rocksdb/convenience.h"
-#include "test_util/secondary_cache_test_util.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"

 namespace ROCKSDB_NAMESPACE {

-using secondary_cache_test_util::GetTestingCacheTypes;
-using secondary_cache_test_util::WithCacheType;
-
-// 16 bytes for HCC compatibility
-const std::string key0 = "____    ____key0";
-const std::string key1 = "____    ____key1";
-const std::string key2 = "____    ____key2";
-const std::string key3 = "____    ____key3";
-
-class CompressedSecondaryCacheTestBase : public testing::Test,
-                                         public WithCacheType {
+class CompressedSecondaryCacheTest : public testing::Test,
+                                     public Cache::CreateContext {
 public:
-  CompressedSecondaryCacheTestBase() {}
-  ~CompressedSecondaryCacheTestBase() override = default;
+  CompressedSecondaryCacheTest() : fail_create_(false) {}
+  ~CompressedSecondaryCacheTest() override = default;

 protected:
+  class TestItem {
+   public:
+    TestItem(const char* buf, size_t size) : buf_(new char[size]), size_(size) {
+      memcpy(buf_.get(), buf, size);
+    }
+    ~TestItem() = default;
+
+    char* Buf() { return buf_.get(); }
+    [[nodiscard]] size_t Size() const { return size_; }
+
+   private:
+    std::unique_ptr<char[]> buf_;
+    size_t size_;
+  };
+
+  static size_t SizeCallback(Cache::ObjectPtr obj) {
+    return static_cast<TestItem*>(obj)->Size();
+  }
+
+  static Status SaveToCallback(Cache::ObjectPtr from_obj, size_t from_offset,
+                               size_t length, char* out) {
+    auto item = static_cast<TestItem*>(from_obj);
+    const char* buf = item->Buf();
+    EXPECT_EQ(length, item->Size());
+    EXPECT_EQ(from_offset, 0);
+    memcpy(out, buf, length);
+    return Status::OK();
+  }
+
+  static void DeletionCallback(Cache::ObjectPtr obj,
+                               MemoryAllocator* /*alloc*/) {
+    delete static_cast<TestItem*>(obj);
+    obj = nullptr;
+  }
+
+  static Status SaveToCallbackFail(Cache::ObjectPtr /*obj*/, size_t /*offset*/,
+                                   size_t /*size*/, char* /*out*/) {
+    return Status::NotSupported();
+  }
+
+  static Status CreateCallback(const Slice& data, Cache::CreateContext* context,
+                               MemoryAllocator* /*allocator*/,
+                               Cache::ObjectPtr* out_obj, size_t* out_charge) {
+    auto t = static_cast<CompressedSecondaryCacheTest*>(context);
+    if (t->fail_create_) {
+      return Status::NotSupported();
+    }
+    *out_obj = new TestItem(data.data(), data.size());
+    *out_charge = data.size();
+    return Status::OK();
+  }
+
+  static constexpr Cache::CacheItemHelper kHelper{
+      CacheEntryRole::kMisc, &DeletionCallback, &SizeCallback, &SaveToCallback,
+      &CreateCallback};
+
+  static constexpr Cache::CacheItemHelper kHelperFail{
+      CacheEntryRole::kMisc, &DeletionCallback, &SizeCallback,
+      &SaveToCallbackFail, &CreateCallback};
+
+  void SetFailCreate(bool fail) { fail_create_ = fail; }
+
  void BasicTestHelper(std::shared_ptr<SecondaryCache> sec_cache,
                       bool sec_cache_is_compressed) {
    get_perf_context()->Reset();
-    bool kept_in_sec_cache{true};
+    bool is_in_sec_cache{true};
    // Lookup an non-existent key.
-    std::unique_ptr<SecondaryCacheResultHandle> handle0 =
-        sec_cache->Lookup(key0, GetHelper(), this, true, /*advise_erase=*/true,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle0 = sec_cache->Lookup(
+        "k0", &kHelper, this, true, /*advise_erase=*/true, is_in_sec_cache);
    ASSERT_EQ(handle0, nullptr);

    Random rnd(301);
@@ -49,25 +99,23 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::string str1(rnd.RandomString(1000));
    TestItem item1(str1.data(), str1.length());
    // A dummy handle is inserted if the item is inserted for the first time.
-    ASSERT_OK(sec_cache->Insert(key1, &item1, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k1", &item1, &kHelper));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 1);
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes, 0);
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_compressed_bytes, 0);

-    std::unique_ptr<SecondaryCacheResultHandle> handle1_1 =
-        sec_cache->Lookup(key1, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle1_1 = sec_cache->Lookup(
+        "k1", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_EQ(handle1_1, nullptr);

    // Insert and Lookup the item k1 for the second time and advise erasing it.
-    ASSERT_OK(sec_cache->Insert(key1, &item1, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k1", &item1, &kHelper));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 1);

-    std::unique_ptr<SecondaryCacheResultHandle> handle1_2 =
-        sec_cache->Lookup(key1, GetHelper(), this, true, /*advise_erase=*/true,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle1_2 = sec_cache->Lookup(
+        "k1", &kHelper, this, true, /*advise_erase=*/true, is_in_sec_cache);
    ASSERT_NE(handle1_2, nullptr);
-    ASSERT_FALSE(kept_in_sec_cache);
+    ASSERT_FALSE(is_in_sec_cache);
    if (sec_cache_is_compressed) {
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes,
                1000);
@@ -84,22 +132,20 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    ASSERT_EQ(memcmp(val1->Buf(), item1.Buf(), item1.Size()), 0);

    // Lookup the item k1 again.
-    std::unique_ptr<SecondaryCacheResultHandle> handle1_3 =
-        sec_cache->Lookup(key1, GetHelper(), this, true, /*advise_erase=*/true,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle1_3 = sec_cache->Lookup(
+        "k1", &kHelper, this, true, /*advise_erase=*/true, is_in_sec_cache);
    ASSERT_EQ(handle1_3, nullptr);

    // Insert and Lookup the item k2.
    std::string str2(rnd.RandomString(1000));
    TestItem item2(str2.data(), str2.length());
-    ASSERT_OK(sec_cache->Insert(key2, &item2, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k2", &item2, &kHelper));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 2);
-    std::unique_ptr<SecondaryCacheResultHandle> handle2_1 =
-        sec_cache->Lookup(key2, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle2_1 = sec_cache->Lookup(
+        "k2", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_EQ(handle2_1, nullptr);

-    ASSERT_OK(sec_cache->Insert(key2, &item2, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k2", &item2, &kHelper));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 2);
    if (sec_cache_is_compressed) {
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes,
@@ -110,9 +156,8 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes, 0);
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_compressed_bytes, 0);
    }
-    std::unique_ptr<SecondaryCacheResultHandle> handle2_2 =
-        sec_cache->Lookup(key2, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle2_2 = sec_cache->Lookup(
+        "k2", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_NE(handle2_2, nullptr);
    std::unique_ptr<TestItem> val2 =
        std::unique_ptr<TestItem>(static_cast<TestItem*>(handle2_2->Value()));
@@ -181,26 +226,24 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::string str1(rnd.RandomString(1000));
    TestItem item1(str1.data(), str1.length());
    // Insert a dummy handle.
-    ASSERT_OK(sec_cache->Insert(key1, &item1, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k1", &item1, &kHelper));
    // Insert k1.
-    ASSERT_OK(sec_cache->Insert(key1, &item1, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k1", &item1, &kHelper));

    // Insert and Lookup the second item.
    std::string str2(rnd.RandomString(200));
    TestItem item2(str2.data(), str2.length());
    // Insert a dummy handle, k1 is not evicted.
-    ASSERT_OK(sec_cache->Insert(key2, &item2, GetHelper()));
-    bool kept_in_sec_cache{false};
-    std::unique_ptr<SecondaryCacheResultHandle> handle1 =
-        sec_cache->Lookup(key1, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    ASSERT_OK(sec_cache->Insert("k2", &item2, &kHelper));
+    bool is_in_sec_cache{false};
+    std::unique_ptr<SecondaryCacheResultHandle> handle1 = sec_cache->Lookup(
+        "k1", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_EQ(handle1, nullptr);

    // Insert k2 and k1 is evicted.
-    ASSERT_OK(sec_cache->Insert(key2, &item2, GetHelper()));
-    std::unique_ptr<SecondaryCacheResultHandle> handle2 =
-        sec_cache->Lookup(key2, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    ASSERT_OK(sec_cache->Insert("k2", &item2, &kHelper));
+    std::unique_ptr<SecondaryCacheResultHandle> handle2 = sec_cache->Lookup(
+        "k2", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_NE(handle2, nullptr);
    std::unique_ptr<TestItem> val2 =
        std::unique_ptr<TestItem>(static_cast<TestItem*>(handle2->Value()));
@@ -208,26 +251,24 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    ASSERT_EQ(memcmp(val2->Buf(), item2.Buf(), item2.Size()), 0);

    // Insert k1 again and a dummy handle is inserted.
-    ASSERT_OK(sec_cache->Insert(key1, &item1, GetHelper()));
+    ASSERT_OK(sec_cache->Insert("k1", &item1, &kHelper));

-    std::unique_ptr<SecondaryCacheResultHandle> handle1_1 =
-        sec_cache->Lookup(key1, GetHelper(), this, true, /*advise_erase=*/false,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle1_1 = sec_cache->Lookup(
+        "k1", &kHelper, this, true, /*advise_erase=*/false, is_in_sec_cache);
    ASSERT_EQ(handle1_1, nullptr);

    // Create Fails.
    SetFailCreate(true);
-    std::unique_ptr<SecondaryCacheResultHandle> handle2_1 =
-        sec_cache->Lookup(key2, GetHelper(), this, true, /*advise_erase=*/true,
-                          kept_in_sec_cache);
+    std::unique_ptr<SecondaryCacheResultHandle> handle2_1 = sec_cache->Lookup(
+        "k2", &kHelper, this, true, /*advise_erase=*/true, is_in_sec_cache);
    ASSERT_EQ(handle2_1, nullptr);

    // Save Fails.
    std::string str3 = rnd.RandomString(10);
    TestItem item3(str3.data(), str3.length());
    // The Status is OK because a dummy handle is inserted.
-    ASSERT_OK(sec_cache->Insert(key3, &item3, GetHelperFail()));
-    ASSERT_NOK(sec_cache->Insert(key3, &item3, GetHelperFail()));
+    ASSERT_OK(sec_cache->Insert("k3", &item3, &kHelperFail));
+    ASSERT_NOK(sec_cache->Insert("k3", &item3, &kHelperFail));

    sec_cache.reset();
  }
@@ -251,22 +292,26 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    secondary_cache_opts.enable_custom_split_merge = enable_custom_split_merge;
    std::shared_ptr<SecondaryCache> secondary_cache =
        NewCompressedSecondaryCache(secondary_cache_opts);
-    std::shared_ptr<Cache> cache = NewCache(
+    LRUCacheOptions lru_cache_opts(
        /*_capacity =*/1300, /*_num_shard_bits =*/0,
-        /*_strict_capacity_limit =*/true, secondary_cache);
+        /*_strict_capacity_limit =*/false, /*_high_pri_pool_ratio =*/0.5,
+        /*_memory_allocator =*/nullptr, kDefaultToAdaptiveMutex,
+        kDefaultCacheMetadataChargePolicy, /*_low_pri_pool_ratio =*/0.0);
+    lru_cache_opts.secondary_cache = secondary_cache;
+    std::shared_ptr<Cache> cache = NewLRUCache(lru_cache_opts);
    std::shared_ptr<Statistics> stats = CreateDBStatistics();

    get_perf_context()->Reset();
    Random rnd(301);
    std::string str1 = rnd.RandomString(1001);
    auto item1_1 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1_1, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1_1, &kHelper, str1.length()));

    std::string str2 = rnd.RandomString(1012);
    auto item2_1 = new TestItem(str2.data(), str2.length());
    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's dummy item.
-    ASSERT_OK(cache->Insert(key2, item2_1, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2_1, &kHelper, str2.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 1);
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes, 0);
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_compressed_bytes, 0);
@@ -275,19 +320,19 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    auto item3_1 = new TestItem(str3.data(), str3.length());
    // After this Insert, primary cache contains k3 and secondary cache contains
    // k1's dummy item and k2's dummy item.
-    ASSERT_OK(cache->Insert(key3, item3_1, GetHelper(), str3.length()));
+    ASSERT_OK(cache->Insert("k3", item3_1, &kHelper, str3.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 2);

    // After this Insert, primary cache contains k1 and secondary cache contains
    // k1's dummy item, k2's dummy item, and k3's dummy item.
    auto item1_2 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1_2, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1_2, &kHelper, str1.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 3);

    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's item, k2's dummy item, and k3's dummy item.
    auto item2_2 = new TestItem(str2.data(), str2.length());
-    ASSERT_OK(cache->Insert(key2, item2_2, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2_2, &kHelper, str2.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 1);
    if (sec_cache_is_compressed) {
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes,
@@ -302,7 +347,7 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    // After this Insert, primary cache contains k3 and secondary cache contains
    // k1's item and k2's item.
    auto item3_2 = new TestItem(str3.data(), str3.length());
-    ASSERT_OK(cache->Insert(key3, item3_2, GetHelper(), str3.length()));
+    ASSERT_OK(cache->Insert("k3", item3_2, &kHelper, str3.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 2);
    if (sec_cache_is_compressed) {
      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes,
@@ -315,7 +360,7 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    }

    Cache::Handle* handle;
-    handle = cache->Lookup(key3, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k3", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_NE(handle, nullptr);
    auto val3 = static_cast<TestItem*>(cache->Value(handle));
@@ -324,13 +369,13 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    cache->Release(handle);

    // Lookup an non-existent key.
-    handle = cache->Lookup(key0, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k0", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_EQ(handle, nullptr);

    // This Lookup should just insert a dummy handle in the primary cache
    // and the k1 is still in the secondary cache.
-    handle = cache->Lookup(key1, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k1", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_NE(handle, nullptr);
    ASSERT_EQ(get_perf_context()->block_cache_standalone_handle_count, 1);
@@ -342,7 +387,7 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    // This Lookup should erase k1 from the secondary cache and insert
    // it into primary cache; then k3 is demoted.
    // k2 and k3 are in secondary cache.
-    handle = cache->Lookup(key1, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k1", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_NE(handle, nullptr);
    ASSERT_EQ(get_perf_context()->block_cache_standalone_handle_count, 1);
@@ -350,7 +395,7 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    cache->Release(handle);

    // k2 is still in secondary cache.
-    handle = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_NE(handle, nullptr);
    ASSERT_EQ(get_perf_context()->block_cache_standalone_handle_count, 2);
@@ -358,7 +403,7 @@ class CompressedSecondaryCacheTestBase : public testing::Test,

    // Testing SetCapacity().
    ASSERT_OK(secondary_cache->SetCapacity(0));
-    handle = cache->Lookup(key3, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k3", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());
    ASSERT_EQ(handle, nullptr);

@@ -368,30 +413,30 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    ASSERT_EQ(capacity, 7000);
    auto item1_3 = new TestItem(str1.data(), str1.length());
    // After this Insert, primary cache contains k1.
-    ASSERT_OK(cache->Insert(key1, item1_3, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k1", item1_3, &kHelper, str2.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 3);
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 4);

    auto item2_3 = new TestItem(str2.data(), str2.length());
    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's dummy item.
-    ASSERT_OK(cache->Insert(key2, item2_3, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k2", item2_3, &kHelper, str1.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 4);

    auto item1_4 = new TestItem(str1.data(), str1.length());
    // After this Insert, primary cache contains k1 and secondary cache contains
    // k1's dummy item and k2's dummy item.
-    ASSERT_OK(cache->Insert(key1, item1_4, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k1", item1_4, &kHelper, str2.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 5);

    auto item2_4 = new TestItem(str2.data(), str2.length());
    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's real item and k2's dummy item.
-    ASSERT_OK(cache->Insert(key2, item2_4, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2_4, &kHelper, str2.length()));
    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 5);
    // This Lookup should just insert a dummy handle in the primary cache
    // and the k1 is still in the secondary cache.
-    handle = cache->Lookup(key1, GetHelper(), this, Cache::Priority::LOW,
+    handle = cache->Lookup("k1", &kHelper, this, Cache::Priority::LOW, true,
                           stats.get());

    ASSERT_NE(handle, nullptr);
@@ -419,31 +464,26 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::shared_ptr<SecondaryCache> secondary_cache =
        NewCompressedSecondaryCache(secondary_cache_opts);

-    std::shared_ptr<Cache> cache = NewCache(
+    LRUCacheOptions opts(
        /*_capacity=*/1300, /*_num_shard_bits=*/0,
-        /*_strict_capacity_limit=*/false, secondary_cache);
+        /*_strict_capacity_limit=*/false, /*_high_pri_pool_ratio=*/0.5,
+        /*_memory_allocator=*/nullptr, kDefaultToAdaptiveMutex,
+        kDefaultCacheMetadataChargePolicy, /*_low_pri_pool_ratio=*/0.0);
+    opts.secondary_cache = secondary_cache;
+    std::shared_ptr<Cache> cache = NewLRUCache(opts);

    Random rnd(301);
    std::string str1 = rnd.RandomString(1001);
    auto item1 = std::make_unique<TestItem>(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1.get(), GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1.get(), &kHelper, str1.length()));
    item1.release();  // Appease clang-analyze "potential memory leak"

    Cache::Handle* handle;
-    handle = cache->Lookup(key2, nullptr, this, Cache::Priority::LOW);
+    handle = cache->Lookup("k2", nullptr, this, Cache::Priority::LOW, true);
    ASSERT_EQ(handle, nullptr);
-    handle = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW);
+    handle = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, false);
    ASSERT_EQ(handle, nullptr);

-    Cache::AsyncLookupHandle ah;
-    ah.key = key2;
-    ah.helper = GetHelper();
-    ah.create_context = this;
-    ah.priority = Cache::Priority::LOW;
-    cache->StartAsyncLookup(ah);
-    cache->Wait(ah);
-    ASSERT_EQ(ah.Result(), nullptr);
-
    cache.reset();
    secondary_cache.reset();
  }
@@ -466,29 +506,36 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::shared_ptr<SecondaryCache> secondary_cache =
        NewCompressedSecondaryCache(secondary_cache_opts);

-    std::shared_ptr<Cache> cache = NewCache(
+    LRUCacheOptions opts(
        /*_capacity=*/1300, /*_num_shard_bits=*/0,
-        /*_strict_capacity_limit=*/true, secondary_cache);
+        /*_strict_capacity_limit=*/false, /*_high_pri_pool_ratio=*/0.5,
+        /*_memory_allocator=*/nullptr, kDefaultToAdaptiveMutex,
+        kDefaultCacheMetadataChargePolicy, /*_low_pri_pool_ratio=*/0.0);
+    opts.secondary_cache = secondary_cache;
+    std::shared_ptr<Cache> cache = NewLRUCache(opts);

    Random rnd(301);
    std::string str1 = rnd.RandomString(1001);
    auto item1 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1, GetHelperFail(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1, &kHelperFail, str1.length()));

    std::string str2 = rnd.RandomString(1002);
    auto item2 = new TestItem(str2.data(), str2.length());
    // k1 should be demoted to the secondary cache.
-    ASSERT_OK(cache->Insert(key2, item2, GetHelperFail(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2, &kHelperFail, str2.length()));

    Cache::Handle* handle;
-    handle = cache->Lookup(key2, GetHelperFail(), this, Cache::Priority::LOW);
+    handle =
+        cache->Lookup("k2", &kHelperFail, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle, nullptr);
    cache->Release(handle);
    // This lookup should fail, since k1 demotion would have failed.
-    handle = cache->Lookup(key1, GetHelperFail(), this, Cache::Priority::LOW);
+    handle =
+        cache->Lookup("k1", &kHelperFail, this, Cache::Priority::LOW, true);
    ASSERT_EQ(handle, nullptr);
    // Since k1 was not promoted, k2 should still be in cache.
-    handle = cache->Lookup(key2, GetHelperFail(), this, Cache::Priority::LOW);
+    handle =
+        cache->Lookup("k2", &kHelperFail, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle, nullptr);
    cache->Release(handle);

@@ -514,30 +561,34 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::shared_ptr<SecondaryCache> secondary_cache =
        NewCompressedSecondaryCache(secondary_cache_opts);

-    std::shared_ptr<Cache> cache = NewCache(
+    LRUCacheOptions opts(
        /*_capacity=*/1300, /*_num_shard_bits=*/0,
-        /*_strict_capacity_limit=*/true, secondary_cache);
+        /*_strict_capacity_limit=*/false, /*_high_pri_pool_ratio=*/0.5,
+        /*_memory_allocator=*/nullptr, kDefaultToAdaptiveMutex,
+        kDefaultCacheMetadataChargePolicy, /*_low_pri_pool_ratio=*/0.0);
+    opts.secondary_cache = secondary_cache;
+    std::shared_ptr<Cache> cache = NewLRUCache(opts);

    Random rnd(301);
    std::string str1 = rnd.RandomString(1001);
    auto item1 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1, &kHelper, str1.length()));

    std::string str2 = rnd.RandomString(1002);
    auto item2 = new TestItem(str2.data(), str2.length());
    // k1 should be demoted to the secondary cache.
-    ASSERT_OK(cache->Insert(key2, item2, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2, &kHelper, str2.length()));

    Cache::Handle* handle;
    SetFailCreate(true);
-    handle = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW);
+    handle = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle, nullptr);
    cache->Release(handle);
    // This lookup should fail, since k1 creation would have failed
-    handle = cache->Lookup(key1, GetHelper(), this, Cache::Priority::LOW);
+    handle = cache->Lookup("k1", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_EQ(handle, nullptr);
    // Since k1 didn't get promoted, k2 should still be in cache
-    handle = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW);
+    handle = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle, nullptr);
    cache->Release(handle);

@@ -563,34 +614,38 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::shared_ptr<SecondaryCache> secondary_cache =
        NewCompressedSecondaryCache(secondary_cache_opts);

-    std::shared_ptr<Cache> cache = NewCache(
+    LRUCacheOptions opts(
        /*_capacity=*/1300, /*_num_shard_bits=*/0,
-        /*_strict_capacity_limit=*/false, secondary_cache);
+        /*_strict_capacity_limit=*/false, /*_high_pri_pool_ratio=*/0.5,
+        /*_memory_allocator=*/nullptr, kDefaultToAdaptiveMutex,
+        kDefaultCacheMetadataChargePolicy, /*_low_pri_pool_ratio=*/0.0);
+    opts.secondary_cache = secondary_cache;
+    std::shared_ptr<Cache> cache = NewLRUCache(opts);

    Random rnd(301);
    std::string str1 = rnd.RandomString(1001);
    auto item1_1 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1_1, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1_1, &kHelper, str1.length()));

    std::string str2 = rnd.RandomString(1002);
    std::string str2_clone{str2};
    auto item2 = new TestItem(str2.data(), str2.length());
    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's dummy item.
-    ASSERT_OK(cache->Insert(key2, item2, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2, &kHelper, str2.length()));

    // After this Insert, primary cache contains k1 and secondary cache contains
    // k1's dummy item and k2's dummy item.
    auto item1_2 = new TestItem(str1.data(), str1.length());
-    ASSERT_OK(cache->Insert(key1, item1_2, GetHelper(), str1.length()));
+    ASSERT_OK(cache->Insert("k1", item1_2, &kHelper, str1.length()));

    auto item2_2 = new TestItem(str2.data(), str2.length());
    // After this Insert, primary cache contains k2 and secondary cache contains
    // k1's item and k2's dummy item.
-    ASSERT_OK(cache->Insert(key2, item2_2, GetHelper(), str2.length()));
+    ASSERT_OK(cache->Insert("k2", item2_2, &kHelper, str2.length()));

    Cache::Handle* handle2;
-    handle2 = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW);
+    handle2 = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle2, nullptr);
    cache->Release(handle2);

@@ -598,12 +653,12 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    // strict_capacity_limit is true, but the lookup should still succeed.
    // A k1's dummy item is inserted into primary cache.
    Cache::Handle* handle1;
-    handle1 = cache->Lookup(key1, GetHelper(), this, Cache::Priority::LOW);
+    handle1 = cache->Lookup("k1", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle1, nullptr);
    cache->Release(handle1);

    // Since k1 didn't get inserted, k2 should still be in cache
-    handle2 = cache->Lookup(key2, GetHelper(), this, Cache::Priority::LOW);
+    handle2 = cache->Lookup("k2", &kHelper, this, Cache::Priority::LOW, true);
    ASSERT_NE(handle2, nullptr);
    cache->Release(handle2);

@@ -727,27 +782,19 @@ class CompressedSecondaryCacheTestBase : public testing::Test,

    sec_cache->GetHelper(true)->del_cb(chunks_head, /*alloc*/ nullptr);
  }
-};

-class CompressedSecondaryCacheTest
-    : public CompressedSecondaryCacheTestBase,
-      public testing::WithParamInterface<std::string> {
-  const std::string& Type() override { return GetParam(); }
+ private:
+  bool fail_create_;
 };

-INSTANTIATE_TEST_CASE_P(CompressedSecondaryCacheTest,
-                        CompressedSecondaryCacheTest, GetTestingCacheTypes());
-
 class CompressedSecCacheTestWithCompressAndAllocatorParam
-    : public CompressedSecondaryCacheTestBase,
-      public ::testing::WithParamInterface<
-          std::tuple<bool, bool, std::string>> {
+    : public CompressedSecondaryCacheTest,
+      public ::testing::WithParamInterface<std::tuple<bool, bool>> {
 public:
  CompressedSecCacheTestWithCompressAndAllocatorParam() {
    sec_cache_is_compressed_ = std::get<0>(GetParam());
    use_jemalloc_ = std::get<1>(GetParam());
  }
-  const std::string& Type() override { return std::get<2>(GetParam()); }
  bool sec_cache_is_compressed_;
  bool use_jemalloc_;
 };
@@ -758,20 +805,20 @@ TEST_P(CompressedSecCacheTestWithCompressAndAllocatorParam, BasicTes) {

 INSTANTIATE_TEST_CASE_P(CompressedSecCacheTests,
                        CompressedSecCacheTestWithCompressAndAllocatorParam,
-                        ::testing::Combine(testing::Bool(), testing::Bool(),
-                                           GetTestingCacheTypes()));
+                        ::testing::Combine(testing::Bool(), testing::Bool()));

 class CompressedSecondaryCacheTestWithCompressionParam
-    : public CompressedSecondaryCacheTestBase,
-      public ::testing::WithParamInterface<std::tuple<bool, std::string>> {
+    : public CompressedSecondaryCacheTest,
+      public ::testing::WithParamInterface<bool> {
 public:
  CompressedSecondaryCacheTestWithCompressionParam() {
-    sec_cache_is_compressed_ = std::get<0>(GetParam());
+    sec_cache_is_compressed_ = GetParam();
  }
-  const std::string& Type() override { return std::get<1>(GetParam()); }
  bool sec_cache_is_compressed_;
 };

+#ifndef ROCKSDB_LITE
+
 TEST_P(CompressedSecondaryCacheTestWithCompressionParam, BasicTestFromString) {
  std::shared_ptr<SecondaryCache> sec_cache{nullptr};
  std::string sec_cache_uri;
@@ -835,6 +882,7 @@ TEST_P(CompressedSecondaryCacheTestWithCompressionParam,
  BasicTestHelper(sec_cache, sec_cache_is_compressed_);
 }

+#endif  // ROCKSDB_LITE

 TEST_P(CompressedSecondaryCacheTestWithCompressionParam, FailsTest) {
  FailsTest(sec_cache_is_compressed_);
@@ -860,92 +908,18 @@ TEST_P(CompressedSecondaryCacheTestWithCompressionParam,
  IntegrationFullCapacityTest(sec_cache_is_compressed_);
 }

-TEST_P(CompressedSecondaryCacheTestWithCompressionParam, EntryRoles) {
-  CompressedSecondaryCacheOptions opts;
-  opts.capacity = 2048;
-  opts.num_shard_bits = 0;
-
-  if (sec_cache_is_compressed_) {
-    if (!LZ4_Supported()) {
-      ROCKSDB_GTEST_SKIP("This test requires LZ4 support.");
-      return;
-    }
-  } else {
-    opts.compression_type = CompressionType::kNoCompression;
-  }
-
-  // Select a random subset to include, for fast test
-  Random& r = *Random::GetTLSInstance();
-  CacheEntryRoleSet do_not_compress;
-  for (uint32_t i = 0; i < kNumCacheEntryRoles; ++i) {
-    // A few included on average, but decent chance of zero
-    if (r.OneIn(5)) {
-      do_not_compress.Add(static_cast<CacheEntryRole>(i));
-    }
-  }
-  opts.do_not_compress_roles = do_not_compress;
-
-  std::shared_ptr<SecondaryCache> sec_cache = NewCompressedSecondaryCache(opts);
-
-  // Fixed seed to ensure consistent compressibility (doesn't compress)
-  std::string junk(Random(301).RandomString(1000));
-
-  for (uint32_t i = 0; i < kNumCacheEntryRoles; ++i) {
-    CacheEntryRole role = static_cast<CacheEntryRole>(i);
-
-    // Uniquify `junk`
-    junk[0] = static_cast<char>(i);
-    TestItem item{junk.data(), junk.length()};
-    Slice ith_key = Slice(junk.data(), 16);
-
-    get_perf_context()->Reset();
-    ASSERT_OK(sec_cache->Insert(ith_key, &item, GetHelper(role)));
-    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_dummy_count, 1U);
-
-    ASSERT_OK(sec_cache->Insert(ith_key, &item, GetHelper(role)));
-    ASSERT_EQ(get_perf_context()->compressed_sec_cache_insert_real_count, 1U);
-
-    bool kept_in_sec_cache{true};
-    std::unique_ptr<SecondaryCacheResultHandle> handle =
-        sec_cache->Lookup(ith_key, GetHelper(role), this, true,
-                          /*advise_erase=*/true, kept_in_sec_cache);
-    ASSERT_NE(handle, nullptr);
-
-    // Lookup returns the right data
-    std::unique_ptr<TestItem> val =
-        std::unique_ptr<TestItem>(static_cast<TestItem*>(handle->Value()));
-    ASSERT_NE(val, nullptr);
-    ASSERT_EQ(memcmp(val->Buf(), item.Buf(), item.Size()), 0);
-
-    bool compressed =
-        sec_cache_is_compressed_ && !do_not_compress.Contains(role);
-    if (compressed) {
-      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes,
-                1000);
-      ASSERT_EQ(get_perf_context()->compressed_sec_cache_compressed_bytes,
-                1007);
-    } else {
-      ASSERT_EQ(get_perf_context()->compressed_sec_cache_uncompressed_bytes, 0);
-      ASSERT_EQ(get_perf_context()->compressed_sec_cache_compressed_bytes, 0);
-    }
-  }
-}
-
 INSTANTIATE_TEST_CASE_P(CompressedSecCacheTests,
                        CompressedSecondaryCacheTestWithCompressionParam,
-                        testing::Combine(testing::Bool(),
-                                         GetTestingCacheTypes()));
+                        testing::Bool());

 class CompressedSecCacheTestWithCompressAndSplitParam
-    : public CompressedSecondaryCacheTestBase,
-      public ::testing::WithParamInterface<
-          std::tuple<bool, bool, std::string>> {
+    : public CompressedSecondaryCacheTest,
+      public ::testing::WithParamInterface<std::tuple<bool, bool>> {
 public:
  CompressedSecCacheTestWithCompressAndSplitParam() {
    sec_cache_is_compressed_ = std::get<0>(GetParam());
    enable_custom_split_merge_ = std::get<1>(GetParam());
  }
-  const std::string& Type() override { return std::get<2>(GetParam()); }
  bool sec_cache_is_compressed_;
  bool enable_custom_split_merge_;
 };
@@ -956,18 +930,17 @@ TEST_P(CompressedSecCacheTestWithCompressAndSplitParam, BasicIntegrationTest) {

 INSTANTIATE_TEST_CASE_P(CompressedSecCacheTests,
                        CompressedSecCacheTestWithCompressAndSplitParam,
-                        ::testing::Combine(testing::Bool(), testing::Bool(),
-                                           GetTestingCacheTypes()));
+                        ::testing::Combine(testing::Bool(), testing::Bool()));

-TEST_P(CompressedSecondaryCacheTest, SplitValueIntoChunksTest) {
+TEST_F(CompressedSecondaryCacheTest, SplitValueIntoChunksTest) {
  SplitValueIntoChunksTest();
 }

-TEST_P(CompressedSecondaryCacheTest, MergeChunksIntoValueTest) {
+TEST_F(CompressedSecondaryCacheTest, MergeChunksIntoValueTest) {
  MergeChunksIntoValueTest();
 }

-TEST_P(CompressedSecondaryCacheTest, SplictValueAndMergeChunksTest) {
+TEST_F(CompressedSecondaryCacheTest, SplictValueAndMergeChunksTest) {
  SplictValueAndMergeChunksTest();
 }

@@ -14,7 +14,6 @@
 #include <cstdio>
 #include <cstdlib>

-#include "cache/secondary_cache_adapter.h"
 #include "monitoring/perf_context_imp.h"
 #include "monitoring/statistics.h"
 #include "port/lang.h"
@@ -23,6 +22,14 @@
 namespace ROCKSDB_NAMESPACE {
 namespace lru_cache {

+namespace {
+// A distinct pointer value for marking "dummy" cache entries
+struct DummyValue {
+  char val[12] = "kDummyValue";
+};
+DummyValue kDummyValue{};
+}  // namespace
+
 LRUHandleTable::LRUHandleTable(int max_upper_hash_bits,
                               MemoryAllocator* allocator)
    : length_bits_(/* historical starting size*/ 4),
@@ -96,7 +103,7 @@ void LRUHandleTable::Resize() {
  std::unique_ptr<LRUHandle* []> new_list {
    new LRUHandle* [size_t{1} << new_length_bits] {}
  };
-  [[maybe_unused]] uint32_t count = 0;
+  uint32_t count = 0;
  for (uint32_t i = 0; i < old_length; i++) {
    LRUHandle* h = list_[i];
    while (h != nullptr) {
@@ -120,7 +127,7 @@ LRUCacheShard::LRUCacheShard(size_t capacity, bool strict_capacity_limit,
                             CacheMetadataChargePolicy metadata_charge_policy,
                             int max_upper_hash_bits,
                             MemoryAllocator* allocator,
-                             const Cache::EvictionCallback* eviction_callback)
+                             SecondaryCache* secondary_cache)
    : CacheShardBase(metadata_charge_policy),
      capacity_(0),
      high_pri_pool_usage_(0),
@@ -134,7 +141,7 @@ LRUCacheShard::LRUCacheShard(size_t capacity, bool strict_capacity_limit,
      usage_(0),
      lru_usage_(0),
      mutex_(use_adaptive_mutex),
-      eviction_callback_(*eviction_callback) {
+      secondary_cache_(secondary_cache) {
  // Make empty circular linked list.
  lru_.next = &lru_;
  lru_.prev = &lru_;
@@ -334,19 +341,16 @@ void LRUCacheShard::EvictFromLRU(size_t charge,
  }
 }

-void LRUCacheShard::NotifyEvicted(
-    const autovector<LRUHandle*>& evicted_handles) {
-  MemoryAllocator* alloc = table_.GetAllocator();
-  for (LRUHandle* entry : evicted_handles) {
-    if (eviction_callback_ &&
-        eviction_callback_(entry->key(),
-                           reinterpret_cast<Cache::Handle*>(entry))) {
-      // Callback took ownership of obj; just free handle
-      free(entry);
-    } else {
-      // Free the entries here outside of mutex for performance reasons.
-      entry->Free(alloc);
+void LRUCacheShard::TryInsertIntoSecondaryCache(
+    autovector<LRUHandle*> evicted_handles) {
+  for (auto entry : evicted_handles) {
+    if (secondary_cache_ && entry->IsSecondaryCacheCompatible() &&
+        !entry->IsInSecondaryCache()) {
+      secondary_cache_->Insert(entry->key(), entry->value, entry->helper)
+          .PermitUncheckedError();
    }
+    // Free the entries here outside of mutex for performance reasons.
+    entry->Free(table_.GetAllocator());
  }
 }

@@ -360,7 +364,7 @@ void LRUCacheShard::SetCapacity(size_t capacity) {
    EvictFromLRU(0, &last_reference_list);
  }

-  NotifyEvicted(last_reference_list);
+  TryInsertIntoSecondaryCache(last_reference_list);
 }

 void LRUCacheShard::SetStrictCapacityLimit(bool strict_capacity_limit) {
@@ -368,7 +372,8 @@ void LRUCacheShard::SetStrictCapacityLimit(bool strict_capacity_limit) {
  strict_capacity_limit_ = strict_capacity_limit;
 }

-Status LRUCacheShard::InsertItem(LRUHandle* e, LRUHandle** handle) {
+Status LRUCacheShard::InsertItem(LRUHandle* e, LRUHandle** handle,
+                                 bool free_handle_on_fail) {
  Status s = Status::OK();
  autovector<LRUHandle*> last_reference_list;

@@ -387,9 +392,10 @@ Status LRUCacheShard::InsertItem(LRUHandle* e, LRUHandle** handle) {
        // into cache and get evicted immediately.
        last_reference_list.push_back(e);
      } else {
-        free(e);
-        e = nullptr;
-        *handle = nullptr;
+        if (free_handle_on_fail) {
+          free(e);
+          *handle = nullptr;
+        }
        s = Status::MemoryLimit("Insert failed due to LRU cache being full.");
      }
    } else {
@@ -421,27 +427,185 @@ Status LRUCacheShard::InsertItem(LRUHandle* e, LRUHandle** handle) {
    }
  }

-  NotifyEvicted(last_reference_list);
+  TryInsertIntoSecondaryCache(last_reference_list);

  return s;
 }

-LRUHandle* LRUCacheShard::Lookup(const Slice& key, uint32_t hash,
-                                 const Cache::CacheItemHelper* /*helper*/,
-                                 Cache::CreateContext* /*create_context*/,
-                                 Cache::Priority /*priority*/,
-                                 Statistics* /*stats*/) {
-  DMutexLock l(mutex_);
-  LRUHandle* e = table_.Lookup(key, hash);
-  if (e != nullptr) {
-    assert(e->InCache());
-    if (!e->HasRefs()) {
-      // The entry is in LRU since it's in hash and has no external
-      // references.
-      LRU_Remove(e);
+void LRUCacheShard::Promote(LRUHandle* e) {
+  SecondaryCacheResultHandle* secondary_handle = e->sec_handle;
+
+  assert(secondary_handle->IsReady());
+  // e is not thread-shared here; OK to modify "immutable" fields as well as
+  // "mutable" (normally requiring mutex)
+  e->SetIsPending(false);
+  e->value = secondary_handle->Value();
+  assert(e->total_charge == 0);
+  size_t value_size = secondary_handle->Size();
+  delete secondary_handle;
+
+  if (e->value) {
+    e->CalcTotalCharge(value_size, metadata_charge_policy_);
+    Status s;
+    if (e->IsStandalone()) {
+      assert(secondary_cache_ && secondary_cache_->SupportForceErase());
+
+      // Insert a dummy handle and return a standalone handle to caller.
+      // Charge the standalone handle.
+      autovector<LRUHandle*> last_reference_list;
+      bool free_standalone_handle{false};
+      {
+        DMutexLock l(mutex_);
+
+        // Free the space following strict LRU policy until enough space
+        // is freed or the lru list is empty.
+        EvictFromLRU(e->total_charge, &last_reference_list);
+
+        if ((usage_ + e->total_charge) > capacity_ && strict_capacity_limit_) {
+          free_standalone_handle = true;
+        } else {
+          usage_ += e->total_charge;
+        }
+      }
+
+      TryInsertIntoSecondaryCache(last_reference_list);
+      if (free_standalone_handle) {
+        e->Unref();
+        e->Free(table_.GetAllocator());
+        e = nullptr;
+      } else {
+        PERF_COUNTER_ADD(block_cache_standalone_handle_count, 1);
+      }
+
+      // Insert a dummy handle into the primary cache. This dummy handle is
+      // not IsSecondaryCacheCompatible().
+      // FIXME? This should not overwrite an existing non-dummy entry in the
+      // rare case that one exists
+      Cache::Priority priority =
+          e->IsHighPri() ? Cache::Priority::HIGH : Cache::Priority::LOW;
+      s = Insert(e->key(), e->hash, &kDummyValue, &kNoopCacheItemHelper,
+                 /*charge=*/0,
+                 /*handle=*/nullptr, priority);
+    } else {
+      e->SetInCache(true);
+      LRUHandle* handle = e;
+      // This InsertItem() could fail if the cache is over capacity and
+      // strict_capacity_limit_ is true. In such a case, we don't want
+      // InsertItem() to free the handle, since the item is already in memory
+      // and the caller will most likely just read it from disk if we erase it
+      // here.
+      s = InsertItem(e, &handle, /*free_handle_on_fail=*/false);
+      if (s.ok()) {
+        PERF_COUNTER_ADD(block_cache_real_handle_count, 1);
+      }
+    }
+
+    if (!s.ok()) {
+      // Item is in memory, but not accounted against the cache capacity.
+      // When the handle is released, the item should get deleted.
+      assert(!e->InCache());
+    }
+  } else {
+    // Secondary cache lookup failed. The caller will take care of detecting
+    // this and eventually releasing e.
+    assert(!e->value);
+    assert(!e->InCache());
+  }
+}
+
+LRUHandle* LRUCacheShard::Lookup(const Slice& key, uint32_t hash,
+                                 const Cache::CacheItemHelper* helper,
+                                 Cache::CreateContext* create_context,
+                                 Cache::Priority priority, bool wait,
+                                 Statistics* stats) {
+  LRUHandle* e = nullptr;
+  bool found_dummy_entry{false};
+  {
+    DMutexLock l(mutex_);
+    e = table_.Lookup(key, hash);
+    if (e != nullptr) {
+      assert(e->InCache());
+      if (e->value == &kDummyValue) {
+        // For a dummy handle, if it was retrieved from secondary cache,
+        // it may still exist in secondary cache.
+        // If the handle exists in secondary cache, the value should be
+        // erased from sec cache and be inserted into primary cache.
+        found_dummy_entry = true;
+        // Let the dummy entry be overwritten
+        e = nullptr;
+      } else {
+        if (!e->HasRefs()) {
+          // The entry is in LRU since it's in hash and has no external
+          // references.
+          LRU_Remove(e);
+        }
+        e->Ref();
+        e->SetHit();
+      }
+    }
+  }
+
+  // If handle table lookup failed or the handle is a dummy one, allocate
+  // a handle outside the mutex if we re going to lookup in the secondary cache.
+  //
+  // When a block is firstly Lookup from CompressedSecondaryCache, we just
+  // insert a dummy block into the primary cache (charging the actual size of
+  // the block) and don't erase the block from CompressedSecondaryCache. A
+  // standalone handle is returned to the caller. Only if the block is hit
+  // again, we erase it from CompressedSecondaryCache and add it into the
+  // primary cache.
+  if (!e && secondary_cache_ && helper && helper->create_cb) {
+    bool is_in_sec_cache{false};
+    std::unique_ptr<SecondaryCacheResultHandle> secondary_handle =
+        secondary_cache_->Lookup(key, helper, create_context, wait,
+                                 found_dummy_entry, is_in_sec_cache);
+    if (secondary_handle != nullptr) {
+      e = static_cast<LRUHandle*>(malloc(sizeof(LRUHandle) - 1 + key.size()));
+
+      e->m_flags = 0;
+      e->im_flags = 0;
+      e->helper = helper;
+      e->key_length = key.size();
+      e->hash = hash;
+      e->refs = 0;
+      e->next = e->prev = nullptr;
+      e->SetPriority(priority);
+      memcpy(e->key_data, key.data(), key.size());
+      e->value = nullptr;
+      e->sec_handle = secondary_handle.release();
+      e->total_charge = 0;
+      e->Ref();
+      e->SetIsInSecondaryCache(is_in_sec_cache);
+      e->SetIsStandalone(secondary_cache_->SupportForceErase() &&
+                         !found_dummy_entry);
+
+      if (wait) {
+        Promote(e);
+        if (e) {
+          if (!e->value) {
+            // The secondary cache returned a handle, but the lookup failed.
+            e->Unref();
+            e->Free(table_.GetAllocator());
+            e = nullptr;
+          } else {
+            PERF_COUNTER_ADD(secondary_cache_hit_count, 1);
+            RecordTick(stats, SECONDARY_CACHE_HITS);
+          }
+        }
+      } else {
+        // If wait is false, we always return a handle and let the caller
+        // release the handle after checking for success or failure.
+        e->SetIsPending(true);
+        // This may be slightly inaccurate, if the lookup eventually fails.
+        // But the probability is very low.
+        PERF_COUNTER_ADD(secondary_cache_hit_count, 1);
+        RecordTick(stats, SECONDARY_CACHE_HITS);
+      }
+    } else {
+      // Caller will most likely overwrite the dummy entry with an Insert
+      // after this Lookup fails
+      assert(e == nullptr);
    }
-    e->Ref();
-    e->SetHit();
  }
  return e;
 }
@@ -450,6 +614,8 @@ bool LRUCacheShard::Ref(LRUHandle* e) {
  DMutexLock l(mutex_);
  // To create another reference - entry must be already externally referenced.
  assert(e->HasRefs());
+  // Pending handles are not for sharing
+  assert(!e->IsPending());
  e->Ref();
  return true;
 }
@@ -473,13 +639,14 @@ bool LRUCacheShard::Release(LRUHandle* e, bool /*useful*/,
  if (e == nullptr) {
    return false;
  }
-  bool must_free;
-  bool was_in_cache;
+  bool last_reference = false;
+  // Must Wait or WaitAll first on pending handles. Otherwise, would leak
+  // a secondary cache handle.
+  assert(!e->IsPending());
  {
    DMutexLock l(mutex_);
-    must_free = e->Unref();
-    was_in_cache = e->InCache();
-    if (must_free && was_in_cache) {
+    last_reference = e->Unref();
+    if (last_reference && e->InCache()) {
      // The item is still in cache, and nobody else holds a reference to it.
      if (usage_ > capacity_ || erase_if_last_ref) {
        // The LRU list must be empty since the cache is full.
@@ -490,38 +657,29 @@ bool LRUCacheShard::Release(LRUHandle* e, bool /*useful*/,
      } else {
        // Put the item back on the LRU list, and don't free it.
        LRU_Insert(e);
-        must_free = false;
+        last_reference = false;
      }
    }
-    // If about to be freed, then decrement the cache usage.
-    if (must_free) {
+    // If it was the last reference, then decrement the cache usage.
+    if (last_reference) {
      assert(usage_ >= e->total_charge);
      usage_ -= e->total_charge;
    }
  }

  // Free the entry here outside of mutex for performance reasons.
-  if (must_free) {
-    // Only call eviction callback if we're sure no one requested erasure
-    // FIXME: disabled because of test churn
-    if (false && was_in_cache && !erase_if_last_ref && eviction_callback_ &&
-        eviction_callback_(e->key(), reinterpret_cast<Cache::Handle*>(e))) {
-      // Callback took ownership of obj; just free handle
-      free(e);
-    } else {
-      e->Free(table_.GetAllocator());
-    }
+  if (last_reference) {
+    e->Free(table_.GetAllocator());
  }
-  return must_free;
+  return last_reference;
 }

-LRUHandle* LRUCacheShard::CreateHandle(const Slice& key, uint32_t hash,
-                                       Cache::ObjectPtr value,
-                                       const Cache::CacheItemHelper* helper,
-                                       size_t charge) {
+Status LRUCacheShard::Insert(const Slice& key, uint32_t hash,
+                             Cache::ObjectPtr value,
+                             const Cache::CacheItemHelper* helper,
+                             size_t charge, LRUHandle** handle,
+                             Cache::Priority priority) {
  assert(helper);
-  // value == nullptr is reserved for indicating failure in SecondaryCache
-  assert(!(helper->IsSecondaryCacheCompatible() && value == nullptr));

  // Allocate the memory here outside of the mutex.
  // If the cache is full, we'll have to release it.
@@ -537,53 +695,16 @@ LRUHandle* LRUCacheShard::CreateHandle(const Slice& key, uint32_t hash,
  e->hash = hash;
  e->refs = 0;
  e->next = e->prev = nullptr;
+  e->SetInCache(true);
+  e->SetPriority(priority);
  memcpy(e->key_data, key.data(), key.size());
  e->CalcTotalCharge(charge, metadata_charge_policy_);

-  return e;
-}
+  // value == nullptr is reserved for indicating failure for when secondary
+  // cache compatible
+  assert(!(e->IsSecondaryCacheCompatible() && value == nullptr));

-Status LRUCacheShard::Insert(const Slice& key, uint32_t hash,
-                             Cache::ObjectPtr value,
-                             const Cache::CacheItemHelper* helper,
-                             size_t charge, LRUHandle** handle,
-                             Cache::Priority priority) {
-  LRUHandle* e = CreateHandle(key, hash, value, helper, charge);
-  e->SetPriority(priority);
-  e->SetInCache(true);
-  return InsertItem(e, handle);
-}
-
-LRUHandle* LRUCacheShard::CreateStandalone(const Slice& key, uint32_t hash,
-                                           Cache::ObjectPtr value,
-                                           const Cache::CacheItemHelper* helper,
-                                           size_t charge,
-                                           bool allow_uncharged) {
-  LRUHandle* e = CreateHandle(key, hash, value, helper, charge);
-  e->SetIsStandalone(true);
-  e->Ref();
-
-  autovector<LRUHandle*> last_reference_list;
-
-  {
-    DMutexLock l(mutex_);
-
-    EvictFromLRU(e->total_charge, &last_reference_list);
-
-    if (strict_capacity_limit_ && (usage_ + e->total_charge) > capacity_) {
-      if (allow_uncharged) {
-        e->total_charge = 0;
-      } else {
-        free(e);
-        e = nullptr;
-      }
-    } else {
-      usage_ += e->total_charge;
-    }
-  }
-
-  NotifyEvicted(last_reference_list);
-  return e;
+  return InsertItem(e, handle, /* free_handle_on_fail */ true);
 }

 void LRUCacheShard::Erase(const Slice& key, uint32_t hash) {
@@ -612,6 +733,16 @@ void LRUCacheShard::Erase(const Slice& key, uint32_t hash) {
  }
 }

+bool LRUCacheShard::IsReady(LRUHandle* e) {
+  bool ready = true;
+  if (e->IsPending()) {
+    assert(secondary_cache_);
+    assert(e->sec_handle);
+    ready = e->sec_handle->IsReady();
+  }
+  return ready;
+}
+
 size_t LRUCacheShard::GetUsage() const {
  DMutexLock l(mutex_);
  return usage_;
@@ -651,23 +782,26 @@ LRUCache::LRUCache(size_t capacity, int num_shard_bits,
                   double low_pri_pool_ratio,
                   std::shared_ptr<MemoryAllocator> allocator,
                   bool use_adaptive_mutex,
-                   CacheMetadataChargePolicy metadata_charge_policy)
+                   CacheMetadataChargePolicy metadata_charge_policy,
+                   std::shared_ptr<SecondaryCache> _secondary_cache)
    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit,
-                   std::move(allocator)) {
+                   std::move(allocator)),
+      secondary_cache_(std::move(_secondary_cache)) {
  size_t per_shard = GetPerShardCapacity();
+  SecondaryCache* secondary_cache = secondary_cache_.get();
  MemoryAllocator* alloc = memory_allocator();
-  const EvictionCallback* eviction_callback = &eviction_callback_;
  InitShards([=](LRUCacheShard* cs) {
-    new (cs) LRUCacheShard(per_shard, strict_capacity_limit,
-                           high_pri_pool_ratio, low_pri_pool_ratio,
-                           use_adaptive_mutex, metadata_charge_policy,
-                           /* max_upper_hash_bits */ 32 - num_shard_bits, alloc,
-                           eviction_callback);
+    new (cs) LRUCacheShard(
+        per_shard, strict_capacity_limit, high_pri_pool_ratio,
+        low_pri_pool_ratio, use_adaptive_mutex, metadata_charge_policy,
+        /* max_upper_hash_bits */ 32 - num_shard_bits, alloc, secondary_cache);
  });
 }

 Cache::ObjectPtr LRUCache::Value(Handle* handle) {
  auto h = reinterpret_cast<const LRUHandle*>(handle);
+  assert(!h->IsPending() || h->value == nullptr);
+  assert(h->value != &kDummyValue);
  return h->value;
 }

@@ -690,6 +824,42 @@ double LRUCache::GetHighPriPoolRatio() {
  return GetShard(0).GetHighPriPoolRatio();
 }

+void LRUCache::WaitAll(std::vector<Handle*>& handles) {
+  if (secondary_cache_) {
+    std::vector<SecondaryCacheResultHandle*> sec_handles;
+    sec_handles.reserve(handles.size());
+    for (Handle* handle : handles) {
+      if (!handle) {
+        continue;
+      }
+      LRUHandle* lru_handle = reinterpret_cast<LRUHandle*>(handle);
+      if (!lru_handle->IsPending()) {
+        continue;
+      }
+      sec_handles.emplace_back(lru_handle->sec_handle);
+    }
+    secondary_cache_->WaitAll(sec_handles);
+    for (Handle* handle : handles) {
+      if (!handle) {
+        continue;
+      }
+      LRUHandle* lru_handle = reinterpret_cast<LRUHandle*>(handle);
+      if (!lru_handle->IsPending()) {
+        continue;
+      }
+      GetShard(lru_handle->hash).Promote(lru_handle);
+    }
+  }
+}
+
+void LRUCache::AppendPrintableOptions(std::string& str) const {
+  ShardedCache::AppendPrintableOptions(str);  // options from shard
+  if (secondary_cache_) {
+    str.append("  secondary_cache:\n");
+    str.append(secondary_cache_->GetPrintableOptions());
+  }
+}
+
 }  // namespace lru_cache

 std::shared_ptr<Cache> NewLRUCache(
@@ -717,14 +887,10 @@ std::shared_ptr<Cache> NewLRUCache(
  if (num_shard_bits < 0) {
    num_shard_bits = GetDefaultCacheShardBits(capacity);
  }
-  std::shared_ptr<Cache> cache = std::make_shared<LRUCache>(
+  return std::make_shared<LRUCache>(
      capacity, num_shard_bits, strict_capacity_limit, high_pri_pool_ratio,
      low_pri_pool_ratio, std::move(memory_allocator), use_adaptive_mutex,
-      metadata_charge_policy);
-  if (secondary_cache) {
-    cache = std::make_shared<CacheWithSecondaryAdapter>(cache, secondary_cache);
-  }
-  return cache;
+      metadata_charge_policy, secondary_cache);
 }

 std::shared_ptr<Cache> NewLRUCache(const LRUCacheOptions& cache_opts) {
@@ -16,6 +16,7 @@
 #include "port/likely.h"
 #include "port/malloc.h"
 #include "port/port.h"
+#include "rocksdb/secondary_cache.h"
 #include "util/autovector.h"
 #include "util/distributed_mutex.h"

@@ -50,7 +51,12 @@ namespace lru_cache {
 struct LRUHandle {
  Cache::ObjectPtr value;
  const Cache::CacheItemHelper* helper;
-  LRUHandle* next_hash;
+  // An entry is not added to the LRUHandleTable until the secondary cache
+  // lookup is complete, so its safe to have this union.
+  union {
+    LRUHandle* next_hash;
+    SecondaryCacheResultHandle* sec_handle;
+  };
  LRUHandle* next;
  LRUHandle* prev;
  size_t total_charge;  // TODO(opt): Only allow uint32_t?
@@ -83,8 +89,12 @@ struct LRUHandle {
    IM_IS_HIGH_PRI = (1 << 0),
    // Whether this entry is low priority entry.
    IM_IS_LOW_PRI = (1 << 1),
+    // Is the handle still being read from a lower tier.
+    IM_IS_PENDING = (1 << 2),
+    // Whether this handle is still in a lower tier
+    IM_IS_IN_SECONDARY_CACHE = (1 << 3),
    // Marks result handles that should not be inserted into cache
-    IM_IS_STANDALONE = (1 << 2),
+    IM_IS_STANDALONE = (1 << 4),
  };

  // Beginning of the key (MUST BE THE LAST FIELD IN THIS STRUCT!)
@@ -114,6 +124,11 @@ struct LRUHandle {
  bool IsLowPri() const { return im_flags & IM_IS_LOW_PRI; }
  bool InLowPriPool() const { return m_flags & M_IN_LOW_PRI_POOL; }
  bool HasHit() const { return m_flags & M_HAS_HIT; }
+  bool IsSecondaryCacheCompatible() const { return helper->size_cb != nullptr; }
+  bool IsPending() const { return im_flags & IM_IS_PENDING; }
+  bool IsInSecondaryCache() const {
+    return im_flags & IM_IS_IN_SECONDARY_CACHE;
+  }
  bool IsStandalone() const { return im_flags & IM_IS_STANDALONE; }

  void SetInCache(bool in_cache) {
@@ -155,6 +170,22 @@ struct LRUHandle {

  void SetHit() { m_flags |= M_HAS_HIT; }

+  void SetIsPending(bool pending) {
+    if (pending) {
+      im_flags |= IM_IS_PENDING;
+    } else {
+      im_flags &= ~IM_IS_PENDING;
+    }
+  }
+
+  void SetIsInSecondaryCache(bool is_in_secondary_cache) {
+    if (is_in_secondary_cache) {
+      im_flags |= IM_IS_IN_SECONDARY_CACHE;
+    } else {
+      im_flags &= ~IM_IS_IN_SECONDARY_CACHE;
+    }
+  }
+
  void SetIsStandalone(bool is_standalone) {
    if (is_standalone) {
      im_flags |= IM_IS_STANDALONE;
@@ -165,6 +196,14 @@ struct LRUHandle {

  void Free(MemoryAllocator* allocator) {
    assert(refs == 0);
+
+    if (UNLIKELY(IsPending())) {
+      assert(sec_handle != nullptr);
+      SecondaryCacheResultHandle* tmp_sec_handle = sec_handle;
+      tmp_sec_handle->Wait();
+      value = tmp_sec_handle->Value();
+      delete tmp_sec_handle;
+    }
    assert(helper);
    if (helper->del_cb) {
      helper->del_cb(value, allocator);
@@ -264,14 +303,12 @@ class LRUHandleTable {
 // A single shard of sharded cache.
 class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
 public:
-  // NOTE: the eviction_callback ptr is saved, as is it assumed to be kept
-  // alive in Cache.
  LRUCacheShard(size_t capacity, bool strict_capacity_limit,
                double high_pri_pool_ratio, double low_pri_pool_ratio,
                bool use_adaptive_mutex,
                CacheMetadataChargePolicy metadata_charge_policy,
                int max_upper_hash_bits, MemoryAllocator* allocator,
-                const Cache::EvictionCallback* eviction_callback);
+                SecondaryCache* secondary_cache);

 public:  // Type definitions expected as parameter to ShardedCache
  using HandleImpl = LRUHandle;
@@ -302,17 +339,14 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
                const Cache::CacheItemHelper* helper, size_t charge,
                LRUHandle** handle, Cache::Priority priority);

-  LRUHandle* CreateStandalone(const Slice& key, uint32_t hash,
-                              Cache::ObjectPtr obj,
-                              const Cache::CacheItemHelper* helper,
-                              size_t charge, bool allow_uncharged);
-
  LRUHandle* Lookup(const Slice& key, uint32_t hash,
                    const Cache::CacheItemHelper* helper,
                    Cache::CreateContext* create_context,
-                    Cache::Priority priority, Statistics* stats);
+                    Cache::Priority priority, bool wait, Statistics* stats);

  bool Release(LRUHandle* handle, bool useful, bool erase_if_last_ref);
+  bool IsReady(LRUHandle* /*handle*/);
+  void Wait(LRUHandle* /*handle*/) {}
  bool Ref(LRUHandle* handle);
  void Erase(const Slice& key, uint32_t hash);

@@ -352,10 +386,20 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
 private:
  friend class LRUCache;
  // Insert an item into the hash table and, if handle is null, insert into
-  // the LRU list. Older items are evicted as necessary. Frees `item` on
-  // non-OK status.
-  Status InsertItem(LRUHandle* item, LRUHandle** handle);
-
+  // the LRU list. Older items are evicted as necessary. If the cache is full
+  // and free_handle_on_fail is true, the item is deleted and handle is set to
+  // nullptr.
+  Status InsertItem(LRUHandle* item, LRUHandle** handle,
+                    bool free_handle_on_fail);
+  // Promote an item looked up from the secondary cache to the LRU cache.
+  // The item may be still in the secondary cache.
+  // It is only inserted into the hash table and not the LRU list, and only
+  // if the cache is not at full capacity, as is the case during Insert.  The
+  // caller should hold a reference on the LRUHandle. When the caller releases
+  // the last reference, the item is added to the LRU list.
+  // The item is promoted to the high pri or low pri pool as specified by the
+  // caller in Lookup.
+  void Promote(LRUHandle* e);
  void LRU_Remove(LRUHandle* e);
  void LRU_Insert(LRUHandle* e);

@@ -369,11 +413,8 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
  // holding the mutex_.
  void EvictFromLRU(size_t charge, autovector<LRUHandle*>* deleted);

-  void NotifyEvicted(const autovector<LRUHandle*>& evicted_handles);
-
-  LRUHandle* CreateHandle(const Slice& key, uint32_t hash,
-                          Cache::ObjectPtr value,
-                          const Cache::CacheItemHelper* helper, size_t charge);
+  // Try to insert the evicted handles into the secondary cache.
+  void TryInsertIntoSecondaryCache(autovector<LRUHandle*> evicted_handles);

  // Initialized before use.
  size_t capacity_;
@@ -436,8 +477,8 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
  // don't mind mutex_ invoking the non-const actions.
  mutable DMutex mutex_;

-  // A reference to Cache::eviction_callback_
-  const Cache::EvictionCallback& eviction_callback_;
+  // Owned by LRUCache
+  SecondaryCache* secondary_cache_;
 };

 class LRUCache
@@ -451,16 +492,23 @@ class LRUCache
           std::shared_ptr<MemoryAllocator> memory_allocator = nullptr,
           bool use_adaptive_mutex = kDefaultToAdaptiveMutex,
           CacheMetadataChargePolicy metadata_charge_policy =
-               kDontChargeCacheMetadata);
+               kDontChargeCacheMetadata,
+           std::shared_ptr<SecondaryCache> secondary_cache = nullptr);
  const char* Name() const override { return "LRUCache"; }
  ObjectPtr Value(Handle* handle) override;
  size_t GetCharge(Handle* handle) const override;
  const CacheItemHelper* GetCacheItemHelper(Handle* handle) const override;
+  void WaitAll(std::vector<Handle*>& handles) override;

  // Retrieves number of elements in LRU, for unit test purpose only.
  size_t TEST_GetLRUSize();
  // Retrieves high pri pool ratio.
  double GetHighPriPoolRatio();
+
+  void AppendPrintableOptions(std::string& str) const override;
+
+ private:
+  std::shared_ptr<SecondaryCache> secondary_cache_;
 };

 }  // namespace lru_cache
@@ -32,11 +32,8 @@ Status FailCreate(const Slice&, Cache::CreateContext*, MemoryAllocator*,
 }  // namespace

 Status SecondaryCache::InsertSaved(const Slice& key, const Slice& saved) {
-  static Cache::CacheItemHelper helper_no_secondary{CacheEntryRole::kMisc,
-                                                    &NoopDelete};
-  static Cache::CacheItemHelper helper{
-      CacheEntryRole::kMisc, &NoopDelete, &SliceSize,
-      &SliceSaveTo,          &FailCreate, &helper_no_secondary};
+  static Cache::CacheItemHelper helper{CacheEntryRole::kMisc, &NoopDelete,
+                                       &SliceSize, &SliceSaveTo, &FailCreate};
  // NOTE: depends on Insert() being synchronous, not keeping pointer `&saved`
  return Insert(key, const_cast<Slice*>(&saved), &helper);
 }
@@ -1,296 +0,0 @@
-//  Copyright (c) Meta Platforms, Inc. and affiliates.
-//  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 "cache/secondary_cache_adapter.h"
-
-#include "monitoring/perf_context_imp.h"
-
-namespace ROCKSDB_NAMESPACE {
-
-namespace {
-// A distinct pointer value for marking "dummy" cache entries
-struct Dummy {
-  char val[7] = "kDummy";
-};
-const Dummy kDummy{};
-Cache::ObjectPtr const kDummyObj = const_cast<Dummy*>(&kDummy);
-}  // namespace
-
-CacheWithSecondaryAdapter::CacheWithSecondaryAdapter(
-    std::shared_ptr<Cache> target,
-    std::shared_ptr<SecondaryCache> secondary_cache)
-    : CacheWrapper(std::move(target)),
-      secondary_cache_(std::move(secondary_cache)) {
-  target_->SetEvictionCallback([this](const Slice& key, Handle* handle) {
-    return EvictionHandler(key, handle);
-  });
-}
-
-CacheWithSecondaryAdapter::~CacheWithSecondaryAdapter() {
-  // `*this` will be destroyed before `*target_`, so we have to prevent
-  // use after free
-  target_->SetEvictionCallback({});
-}
-
-bool CacheWithSecondaryAdapter::EvictionHandler(const Slice& key,
-                                                Handle* handle) {
-  auto helper = GetCacheItemHelper(handle);
-  if (helper->IsSecondaryCacheCompatible()) {
-    auto obj = target_->Value(handle);
-    // Ignore dummy entry
-    if (obj != kDummyObj) {
-      // Spill into secondary cache.
-      secondary_cache_->Insert(key, obj, helper).PermitUncheckedError();
-    }
-  }
-  // Never takes ownership of obj
-  return false;
-}
-
-bool CacheWithSecondaryAdapter::ProcessDummyResult(Cache::Handle** handle,
-                                                   bool erase) {
-  if (*handle && target_->Value(*handle) == kDummyObj) {
-    target_->Release(*handle, erase);
-    *handle = nullptr;
-    return true;
-  } else {
-    return false;
-  }
-}
-
-void CacheWithSecondaryAdapter::CleanupCacheObject(
-    ObjectPtr obj, const CacheItemHelper* helper) {
-  if (helper->del_cb) {
-    helper->del_cb(obj, memory_allocator());
-  }
-}
-
-Cache::Handle* CacheWithSecondaryAdapter::Promote(
-    std::unique_ptr<SecondaryCacheResultHandle>&& secondary_handle,
-    const Slice& key, const CacheItemHelper* helper, Priority priority,
-    Statistics* stats, bool found_dummy_entry, bool kept_in_sec_cache) {
-  assert(secondary_handle->IsReady());
-
-  ObjectPtr obj = secondary_handle->Value();
-  if (!obj) {
-    // Nothing found.
-    return nullptr;
-  }
-  // Found something.
-  switch (helper->role) {
-    case CacheEntryRole::kFilterBlock:
-      RecordTick(stats, SECONDARY_CACHE_FILTER_HITS);
-      break;
-    case CacheEntryRole::kIndexBlock:
-      RecordTick(stats, SECONDARY_CACHE_INDEX_HITS);
-      break;
-    case CacheEntryRole::kDataBlock:
-      RecordTick(stats, SECONDARY_CACHE_DATA_HITS);
-      break;
-    default:
-      break;
-  }
-  PERF_COUNTER_ADD(secondary_cache_hit_count, 1);
-  RecordTick(stats, SECONDARY_CACHE_HITS);
-
-  // FIXME? using charge=Size() is suspicious but inherited from older
-  // implementation
-  size_t charge = secondary_handle->Size();
-  Handle* result = nullptr;
-  // Insert into primary cache, possibly as a standalone+dummy entries.
-  if (secondary_cache_->SupportForceErase() && !found_dummy_entry) {
-    // Create standalone and insert dummy
-    // Allow standalone to be created even if cache is full, to avoid
-    // reading the entry from storage.
-    result =
-        CreateStandalone(key, obj, helper, charge, /*allow_uncharged*/ true);
-    assert(result);
-    PERF_COUNTER_ADD(block_cache_standalone_handle_count, 1);
-
-    // Insert dummy to record recent use
-    // TODO: try to avoid case where inserting this dummy could overwrite a
-    // regular entry
-    Status s = Insert(key, kDummyObj, &kNoopCacheItemHelper, /*charge=*/0,
-                      /*handle=*/nullptr, priority);
-    s.PermitUncheckedError();
-    // Nothing to do or clean up on dummy insertion failure
-  } else {
-    // Insert regular entry into primary cache.
-    // Don't allow it to spill into secondary cache again if it was kept there.
-    Status s = Insert(
-        key, obj, kept_in_sec_cache ? helper->without_secondary_compat : helper,
-        charge, &result, priority);
-    if (s.ok()) {
-      assert(result);
-      PERF_COUNTER_ADD(block_cache_real_handle_count, 1);
-    } else {
-      // Create standalone result instead, even if cache is full, to avoid
-      // reading the entry from storage.
-      result =
-          CreateStandalone(key, obj, helper, charge, /*allow_uncharged*/ true);
-      assert(result);
-      PERF_COUNTER_ADD(block_cache_standalone_handle_count, 1);
-    }
-  }
-  return result;
-}
-
-Cache::Handle* CacheWithSecondaryAdapter::Lookup(const Slice& key,
-                                                 const CacheItemHelper* helper,
-                                                 CreateContext* create_context,
-                                                 Priority priority,
-                                                 Statistics* stats) {
-  // NOTE: we could just StartAsyncLookup() and Wait(), but this should be a bit
-  // more efficient
-  Handle* result =
-      target_->Lookup(key, helper, create_context, priority, stats);
-  bool secondary_compatible = helper && helper->IsSecondaryCacheCompatible();
-  bool found_dummy_entry =
-      ProcessDummyResult(&result, /*erase=*/secondary_compatible);
-  if (!result && secondary_compatible) {
-    // Try our secondary cache
-    bool kept_in_sec_cache = false;
-    std::unique_ptr<SecondaryCacheResultHandle> secondary_handle =
-        secondary_cache_->Lookup(key, helper, create_context, /*wait*/ true,
-                                 found_dummy_entry, /*out*/ kept_in_sec_cache);
-    if (secondary_handle) {
-      result = Promote(std::move(secondary_handle), key, helper, priority,
-                       stats, found_dummy_entry, kept_in_sec_cache);
-    }
-  }
-  return result;
-}
-
-Cache::ObjectPtr CacheWithSecondaryAdapter::Value(Handle* handle) {
-  ObjectPtr v = target_->Value(handle);
-  // TODO with stacked secondaries: might fail in EvictionHandler
-  assert(v != kDummyObj);
-  return v;
-}
-
-void CacheWithSecondaryAdapter::StartAsyncLookupOnMySecondary(
-    AsyncLookupHandle& async_handle) {
-  assert(!async_handle.IsPending());
-  assert(async_handle.result_handle == nullptr);
-
-  std::unique_ptr<SecondaryCacheResultHandle> secondary_handle =
-      secondary_cache_->Lookup(async_handle.key, async_handle.helper,
-                               async_handle.create_context, /*wait*/ false,
-                               async_handle.found_dummy_entry,
-                               /*out*/ async_handle.kept_in_sec_cache);
-  if (secondary_handle) {
-    // TODO with stacked secondaries: Check & process if already ready?
-    async_handle.pending_handle = secondary_handle.release();
-    async_handle.pending_cache = secondary_cache_.get();
-  }
-}
-
-void CacheWithSecondaryAdapter::StartAsyncLookup(
-    AsyncLookupHandle& async_handle) {
-  target_->StartAsyncLookup(async_handle);
-  if (!async_handle.IsPending()) {
-    bool secondary_compatible =
-        async_handle.helper &&
-        async_handle.helper->IsSecondaryCacheCompatible();
-    async_handle.found_dummy_entry |= ProcessDummyResult(
-        &async_handle.result_handle, /*erase=*/secondary_compatible);
-
-    if (async_handle.Result() == nullptr && secondary_compatible) {
-      // Not found and not pending on another secondary cache
-      StartAsyncLookupOnMySecondary(async_handle);
-    }
-  }
-}
-
-void CacheWithSecondaryAdapter::WaitAll(AsyncLookupHandle* async_handles,
-                                        size_t count) {
-  if (count == 0) {
-    // Nothing to do
-    return;
-  }
-  // Requests that are pending on *my* secondary cache, at the start of this
-  // function
-  std::vector<AsyncLookupHandle*> my_pending;
-  // Requests that are pending on an "inner" secondary cache (managed somewhere
-  // under target_), as of the start of this function
-  std::vector<AsyncLookupHandle*> inner_pending;
-
-  // Initial accounting of pending handles, excluding those already handled
-  // by "outer" secondary caches. (See cur->pending_cache = nullptr.)
-  for (size_t i = 0; i < count; ++i) {
-    AsyncLookupHandle* cur = async_handles + i;
-    if (cur->pending_cache) {
-      assert(cur->IsPending());
-      assert(cur->helper);
-      assert(cur->helper->IsSecondaryCacheCompatible());
-      if (cur->pending_cache == secondary_cache_.get()) {
-        my_pending.push_back(cur);
-        // Mark as "to be handled by this caller"
-        cur->pending_cache = nullptr;
-      } else {
-        // Remember as potentially needing a lookup in my secondary
-        inner_pending.push_back(cur);
-      }
-    }
-  }
-
-  // Wait on inner-most cache lookups first
-  // TODO with stacked secondaries: because we are not using proper
-  // async/await constructs here yet, there is a false synchronization point
-  // here where all the results at one level are needed before initiating
-  // any lookups at the next level. Probably not a big deal, but worth noting.
-  if (!inner_pending.empty()) {
-    target_->WaitAll(async_handles, count);
-  }
-
-  // For those that failed to find something, convert to lookup in my
-  // secondary cache.
-  for (AsyncLookupHandle* cur : inner_pending) {
-    if (cur->Result() == nullptr) {
-      // Not found, try my secondary
-      StartAsyncLookupOnMySecondary(*cur);
-      if (cur->IsPending()) {
-        assert(cur->pending_cache == secondary_cache_.get());
-        my_pending.push_back(cur);
-        // Mark as "to be handled by this caller"
-        cur->pending_cache = nullptr;
-      }
-    }
-  }
-
-  // Wait on all lookups on my secondary cache
-  {
-    std::vector<SecondaryCacheResultHandle*> my_secondary_handles;
-    for (AsyncLookupHandle* cur : my_pending) {
-      my_secondary_handles.push_back(cur->pending_handle);
-    }
-    secondary_cache_->WaitAll(my_secondary_handles);
-  }
-
-  // Process results
-  for (AsyncLookupHandle* cur : my_pending) {
-    std::unique_ptr<SecondaryCacheResultHandle> secondary_handle(
-        cur->pending_handle);
-    cur->pending_handle = nullptr;
-    cur->result_handle = Promote(
-        std::move(secondary_handle), cur->key, cur->helper, cur->priority,
-        cur->stats, cur->found_dummy_entry, cur->kept_in_sec_cache);
-    assert(cur->pending_cache == nullptr);
-  }
-}
-
-std::string CacheWithSecondaryAdapter::GetPrintableOptions() const {
-  std::string str = target_->GetPrintableOptions();
-  str.append("  secondary_cache:\n");
-  str.append(secondary_cache_->GetPrintableOptions());
-  return str;
-}
-
-const char* CacheWithSecondaryAdapter::Name() const {
-  // To the user, at least for now, configure the underlying cache with
-  // a secondary cache. So we pretend to be that cache
-  return target_->Name();
-}
-}  // namespace ROCKSDB_NAMESPACE
@@ -1,52 +0,0 @@
-//  Copyright (c) Meta Platforms, Inc. and affiliates.
-//  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
-
-#include "rocksdb/secondary_cache.h"
-
-namespace ROCKSDB_NAMESPACE {
-
-class CacheWithSecondaryAdapter : public CacheWrapper {
- public:
-  explicit CacheWithSecondaryAdapter(
-      std::shared_ptr<Cache> target,
-      std::shared_ptr<SecondaryCache> secondary_cache);
-
-  ~CacheWithSecondaryAdapter() override;
-
-  Handle* Lookup(const Slice& key, const CacheItemHelper* helper,
-                 CreateContext* create_context,
-                 Priority priority = Priority::LOW,
-                 Statistics* stats = nullptr) override;
-
-  ObjectPtr Value(Handle* handle) override;
-
-  void StartAsyncLookup(AsyncLookupHandle& async_handle) override;
-
-  void WaitAll(AsyncLookupHandle* async_handles, size_t count) override;
-
-  std::string GetPrintableOptions() const override;
-
-  const char* Name() const override;
-
- private:
-  bool EvictionHandler(const Slice& key, Handle* handle);
-
-  void StartAsyncLookupOnMySecondary(AsyncLookupHandle& async_handle);
-
-  Handle* Promote(
-      std::unique_ptr<SecondaryCacheResultHandle>&& secondary_handle,
-      const Slice& key, const CacheItemHelper* helper, Priority priority,
-      Statistics* stats, bool found_dummy_entry, bool kept_in_sec_cache);
-
-  bool ProcessDummyResult(Cache::Handle** handle, bool erase);
-
-  void CleanupCacheObject(ObjectPtr obj, const CacheItemHelper* helper);
-
-  std::shared_ptr<SecondaryCache> secondary_cache_;
-};
-
-}  // namespace ROCKSDB_NAMESPACE
@@ -15,7 +15,7 @@

 #include "port/lang.h"
 #include "port/port.h"
-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "util/hash.h"
 #include "util/mutexlock.h"

@@ -51,17 +51,15 @@ class CacheShardBase {
  };
  Status Insert(const Slice& key, HashCref hash, Cache::ObjectPtr value,
                const Cache::CacheItemHelper* helper, size_t charge,
-                HandleImpl** handle, Cache::Priority priority,
-                bool standalone) = 0;
-  Handle* CreateStandalone(const Slice& key, HashCref hash, ObjectPtr obj,
-                           const CacheItemHelper* helper,
-                           size_t charge, bool allow_uncharged) = 0;
+                HandleImpl** handle, Cache::Priority priority) = 0;
  HandleImpl* Lookup(const Slice& key, HashCref hash,
                        const Cache::CacheItemHelper* helper,
                        Cache::CreateContext* create_context,
-                        Cache::Priority priority,
+                        Cache::Priority priority, bool wait,
                        Statistics* stats) = 0;
  bool Release(HandleImpl* handle, bool useful, bool erase_if_last_ref) = 0;
+  bool IsReady(HandleImpl* handle) = 0;
+  void Wait(HandleImpl* handle) = 0;
  bool Ref(HandleImpl* handle) = 0;
  void Erase(const Slice& key, HashCref hash) = 0;
  void SetCapacity(size_t capacity) = 0;
@@ -171,33 +169,24 @@ class ShardedCache : public ShardedCacheBase {
        [s_c_l](CacheShard* cs) { cs->SetStrictCapacityLimit(s_c_l); });
  }

-  Status Insert(const Slice& key, ObjectPtr obj, const CacheItemHelper* helper,
-                size_t charge, Handle** handle = nullptr,
+  Status Insert(const Slice& key, ObjectPtr value,
+                const CacheItemHelper* helper, size_t charge,
+                Handle** handle = nullptr,
                Priority priority = Priority::LOW) override {
    assert(helper);
    HashVal hash = CacheShard::ComputeHash(key);
    auto h_out = reinterpret_cast<HandleImpl**>(handle);
-    return GetShard(hash).Insert(key, hash, obj, helper, charge, h_out,
+    return GetShard(hash).Insert(key, hash, value, helper, charge, h_out,
                                 priority);
  }

-  Handle* CreateStandalone(const Slice& key, ObjectPtr obj,
-                           const CacheItemHelper* helper, size_t charge,
-                           bool allow_uncharged) override {
-    assert(helper);
-    HashVal hash = CacheShard::ComputeHash(key);
-    HandleImpl* result = GetShard(hash).CreateStandalone(
-        key, hash, obj, helper, charge, allow_uncharged);
-    return reinterpret_cast<Handle*>(result);
-  }
-
  Handle* Lookup(const Slice& key, const CacheItemHelper* helper = nullptr,
                 CreateContext* create_context = nullptr,
-                 Priority priority = Priority::LOW,
+                 Priority priority = Priority::LOW, bool wait = true,
                 Statistics* stats = nullptr) override {
    HashVal hash = CacheShard::ComputeHash(key);
-    HandleImpl* result = GetShard(hash).Lookup(key, hash, helper,
-                                               create_context, priority, stats);
+    HandleImpl* result = GetShard(hash).Lookup(
+        key, hash, helper, create_context, priority, wait, stats);
    return reinterpret_cast<Handle*>(result);
  }

@@ -211,6 +200,14 @@ class ShardedCache : public ShardedCacheBase {
    auto h = reinterpret_cast<HandleImpl*>(handle);
    return GetShard(h->GetHash()).Release(h, useful, erase_if_last_ref);
  }
+  bool IsReady(Handle* handle) override {
+    auto h = reinterpret_cast<HandleImpl*>(handle);
+    return GetShard(h->GetHash()).IsReady(h);
+  }
+  void Wait(Handle* handle) override {
+    auto h = reinterpret_cast<HandleImpl*>(handle);
+    GetShard(h->GetHash()).Wait(h);
+  }
  bool Ref(Handle* handle) override {
    auto h = reinterpret_cast<HandleImpl*>(handle);
    return GetShard(h->GetHash()).Ref(h);
@@ -29,8 +29,8 @@
 #include <type_traits>

 #include "cache/cache_helpers.h"
-#include "rocksdb/advanced_cache.h"
 #include "rocksdb/advanced_options.h"
+#include "rocksdb/cache.h"

 namespace ROCKSDB_NAMESPACE {

@@ -83,14 +83,11 @@ class PlaceholderCacheInterface : public BaseCacheInterface<CachePtr> {
  using BaseCacheInterface<CachePtr>::BaseCacheInterface;

  inline Status Insert(const Slice& key, size_t charge, Handle** handle) {
-    return this->cache_->Insert(key, /*value=*/nullptr, GetHelper(), charge,
+    return this->cache_->Insert(key, /*value=*/nullptr, &kHelper, charge,
                                handle);
  }

-  static const Cache::CacheItemHelper* GetHelper() {
-    static const Cache::CacheItemHelper kHelper{kRole};
-    return &kHelper;
-  }
+  static constexpr Cache::CacheItemHelper kHelper{kRole};
 };

 template <CacheEntryRole kRole>
@@ -131,11 +128,8 @@ class BasicTypedCacheHelperFns {
 template <class TValue, CacheEntryRole kRole>
 class BasicTypedCacheHelper : public BasicTypedCacheHelperFns<TValue> {
 public:
-  static const Cache::CacheItemHelper* GetBasicHelper() {
-    static const Cache::CacheItemHelper kHelper{kRole,
-                                                &BasicTypedCacheHelper::Delete};
-    return &kHelper;
-  }
+  static constexpr Cache::CacheItemHelper kBasicHelper{
+      kRole, &BasicTypedCacheHelper::Delete};
 };

 // BasicTypedCacheInterface - Used for primary cache storage of objects of
@@ -150,14 +144,9 @@ class BasicTypedCacheInterface : public BaseCacheInterface<CachePtr>,
  CACHE_TYPE_DEFS();
  using typename BasicTypedCacheHelperFns<TValue>::TValuePtr;
  struct TypedHandle : public Handle {};
-  using BasicTypedCacheHelper<TValue, kRole>::GetBasicHelper;
+  using BasicTypedCacheHelper<TValue, kRole>::kBasicHelper;
  // ctor
  using BaseCacheInterface<CachePtr>::BaseCacheInterface;
-  struct TypedAsyncLookupHandle : public Cache::AsyncLookupHandle {
-    TypedHandle* Result() {
-      return reinterpret_cast<TypedHandle*>(Cache::AsyncLookupHandle::Result());
-    }
-  };

  inline Status Insert(const Slice& key, TValuePtr value, size_t charge,
                       TypedHandle** handle = nullptr,
@@ -165,7 +154,7 @@ class BasicTypedCacheInterface : public BaseCacheInterface<CachePtr>,
    auto untyped_handle = reinterpret_cast<Handle**>(handle);
    return this->cache_->Insert(
        key, BasicTypedCacheHelperFns<TValue>::UpCastValue(value),
-        GetBasicHelper(), charge, untyped_handle, priority);
+        &kBasicHelper, charge, untyped_handle, priority);
  }

  inline TypedHandle* Lookup(const Slice& key, Statistics* stats = nullptr) {
@@ -173,11 +162,6 @@ class BasicTypedCacheInterface : public BaseCacheInterface<CachePtr>,
        this->cache_->BasicLookup(key, stats));
  }

-  inline void StartAsyncLookup(TypedAsyncLookupHandle& async_handle) {
-    assert(async_handle.helper == nullptr);
-    this->cache_->StartAsyncLookup(async_handle);
-  }
-
  inline CacheHandleGuard<TValue> Guard(TypedHandle* handle) {
    if (handle) {
      return CacheHandleGuard<TValue>(&*this->cache_, handle);
@@ -255,16 +239,9 @@ template <class TValue, class TCreateContext, CacheEntryRole kRole>
 class FullTypedCacheHelper
    : public FullTypedCacheHelperFns<TValue, TCreateContext> {
 public:
-  static const Cache::CacheItemHelper* GetFullHelper() {
-    static const Cache::CacheItemHelper kHelper{
-        kRole,
-        &FullTypedCacheHelper::Delete,
-        &FullTypedCacheHelper::Size,
-        &FullTypedCacheHelper::SaveTo,
-        &FullTypedCacheHelper::Create,
-        BasicTypedCacheHelper<TValue, kRole>::GetBasicHelper()};
-    return &kHelper;
-  }
+  static constexpr Cache::CacheItemHelper kFullHelper{
+      kRole, &FullTypedCacheHelper::Delete, &FullTypedCacheHelper::Size,
+      &FullTypedCacheHelper::SaveTo, &FullTypedCacheHelper::Create};
 };

 // FullTypedCacheHelper - Used for secondary cache compatible storage of
@@ -285,11 +262,9 @@ class FullTypedCacheInterface
 public:
  CACHE_TYPE_DEFS();
  using typename BasicTypedCacheInterface<TValue, kRole, CachePtr>::TypedHandle;
-  using typename BasicTypedCacheInterface<TValue, kRole,
-                                          CachePtr>::TypedAsyncLookupHandle;
  using typename BasicTypedCacheHelperFns<TValue>::TValuePtr;
-  using BasicTypedCacheHelper<TValue, kRole>::GetBasicHelper;
-  using FullTypedCacheHelper<TValue, TCreateContext, kRole>::GetFullHelper;
+  using BasicTypedCacheHelper<TValue, kRole>::kBasicHelper;
+  using FullTypedCacheHelper<TValue, TCreateContext, kRole>::kFullHelper;
  using BasicTypedCacheHelperFns<TValue>::UpCastValue;
  using BasicTypedCacheHelperFns<TValue>::DownCastValue;
  // ctor
@@ -304,8 +279,8 @@ class FullTypedCacheInterface
      CacheTier lowest_used_cache_tier = CacheTier::kNonVolatileBlockTier) {
    auto untyped_handle = reinterpret_cast<Handle**>(handle);
    auto helper = lowest_used_cache_tier == CacheTier::kNonVolatileBlockTier
-                      ? GetFullHelper()
-                      : GetBasicHelper();
+                      ? &kFullHelper
+                      : &kBasicHelper;
    return this->cache_->Insert(key, UpCastValue(value), helper, charge,
                                untyped_handle, priority);
  }
@@ -319,9 +294,9 @@ class FullTypedCacheInterface
      size_t* out_charge = nullptr) {
    ObjectPtr value;
    size_t charge;
-    Status st = GetFullHelper()->create_cb(data, create_context,
-                                           this->cache_->memory_allocator(),
-                                           &value, &charge);
+    Status st = kFullHelper.create_cb(data, create_context,
+                                      this->cache_->memory_allocator(), &value,
+                                      &charge);
    if (out_charge) {
      *out_charge = charge;
    }
@@ -329,7 +304,7 @@ class FullTypedCacheInterface
      st = InsertFull(key, DownCastValue(value), charge, nullptr /*handle*/,
                      priority, lowest_used_cache_tier);
    } else {
-      GetFullHelper()->del_cb(value, this->cache_->memory_allocator());
+      kFullHelper.del_cb(value, this->cache_->memory_allocator());
    }
    return st;
  }
@@ -338,28 +313,17 @@ class FullTypedCacheInterface
  // (Basic Lookup() also inherited.)
  inline TypedHandle* LookupFull(
      const Slice& key, TCreateContext* create_context = nullptr,
-      Priority priority = Priority::LOW, Statistics* stats = nullptr,
+      Priority priority = Priority::LOW, bool wait = true,
+      Statistics* stats = nullptr,
      CacheTier lowest_used_cache_tier = CacheTier::kNonVolatileBlockTier) {
    if (lowest_used_cache_tier == CacheTier::kNonVolatileBlockTier) {
      return reinterpret_cast<TypedHandle*>(this->cache_->Lookup(
-          key, GetFullHelper(), create_context, priority, stats));
+          key, &kFullHelper, create_context, priority, wait, stats));
    } else {
      return BasicTypedCacheInterface<TValue, kRole, CachePtr>::Lookup(key,
                                                                       stats);
    }
  }
-
-  inline void StartAsyncLookupFull(
-      TypedAsyncLookupHandle& async_handle,
-      CacheTier lowest_used_cache_tier = CacheTier::kNonVolatileBlockTier) {
-    if (lowest_used_cache_tier == CacheTier::kNonVolatileBlockTier) {
-      async_handle.helper = GetFullHelper();
-      this->cache_->StartAsyncLookup(async_handle);
-    } else {
-      BasicTypedCacheInterface<TValue, kRole, CachePtr>::StartAsyncLookup(
-          async_handle);
-    }
-  }
 };

 // FullTypedSharedCacheInterface - Like FullTypedCacheInterface but with a
@@ -12,7 +12,7 @@ fi
 ROOT=".."
 # Fetch right version of gcov
 if [ -d /mnt/gvfs/third-party -a -z "$CXX" ]; then
-  source $ROOT/build_tools/fbcode_config_platform010.sh
+  source $ROOT/build_tools/fbcode_config_platform009.sh
  GCOV=$GCC_BASE/bin/gcov
 else
  GCOV=$(which gcov)
@@ -47,9 +47,6 @@ void ArenaWrappedDBIter::Init(
  read_options_ = read_options;
  allow_refresh_ = allow_refresh;
  memtable_range_tombstone_iter_ = nullptr;
-  if (!env->GetFileSystem()->use_async_io()) {
-    read_options_.async_io = false;
-  }
 }

 Status ArenaWrappedDBIter::Refresh() {
@@ -8,7 +8,7 @@
 #include <memory>

 #include "memory/memory_allocator.h"
-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "rocksdb/rocksdb_namespace.h"
 #include "rocksdb/slice.h"
 #include "rocksdb/status.h"
@@ -123,10 +123,6 @@ class BlobCountingIterator : public InternalIterator {
    return iter_->GetProperty(prop_name, prop);
  }

-  bool IsDeleteRangeSentinelKey() const override {
-    return iter_->IsDeleteRangeSentinelKey();
-  }
-
 private:
  void UpdateAndCountBlobIfNeeded() {
    assert(!iter_->Valid() || iter_->status().ok());
@@ -23,19 +23,32 @@ class BlobFileCompletionCallback {
      const std::vector<std::shared_ptr<EventListener>>& listeners,
      const std::string& dbname)
      : event_logger_(event_logger), listeners_(listeners), dbname_(dbname) {
+#ifndef ROCKSDB_LITE
    sst_file_manager_ = sst_file_manager;
    mutex_ = mutex;
    error_handler_ = error_handler;
+#else
+    (void)sst_file_manager;
+    (void)mutex;
+    (void)error_handler;
+#endif  // ROCKSDB_LITE
  }

  void OnBlobFileCreationStarted(const std::string& file_name,
                                 const std::string& column_family_name,
                                 int job_id,
                                 BlobFileCreationReason creation_reason) {
+#ifndef ROCKSDB_LITE
    // Notify the listeners.
    EventHelpers::NotifyBlobFileCreationStarted(listeners_, dbname_,
                                                column_family_name, file_name,
                                                job_id, creation_reason);
+#else
+    (void)file_name;
+    (void)column_family_name;
+    (void)job_id;
+    (void)creation_reason;
+#endif
  }

  Status OnBlobFileCompleted(const std::string& file_name,
@@ -48,6 +61,7 @@ class BlobFileCompletionCallback {
                             uint64_t blob_count, uint64_t blob_bytes) {
    Status s;

+#ifndef ROCKSDB_LITE
    auto sfm = static_cast<SstFileManagerImpl*>(sst_file_manager_);
    if (sfm) {
      // Report new blob files to SstFileManagerImpl
@@ -60,6 +74,7 @@ class BlobFileCompletionCallback {
        error_handler_->SetBGError(s, BackgroundErrorReason::kFlush);
      }
    }
+#endif  // !ROCKSDB_LITE

    // Notify the listeners.
    EventHelpers::LogAndNotifyBlobFileCreationFinished(
@@ -74,9 +89,11 @@ class BlobFileCompletionCallback {
  }

 private:
+#ifndef ROCKSDB_LITE
  SstFileManager* sst_file_manager_;
  InstrumentedMutex* mutex_;
  ErrorHandler* error_handler_;
+#endif  // ROCKSDB_LITE
  EventLogger* event_logger_;
  std::vector<std::shared_ptr<EventListener>> listeners_;
  std::string dbname_;
@@ -30,6 +30,7 @@ BlobSource::BlobSource(const ImmutableOptions* immutable_options,
      blob_file_cache_(blob_file_cache),
      blob_cache_(immutable_options->blob_cache),
      lowest_used_cache_tier_(immutable_options->lowest_used_cache_tier) {
+#ifndef ROCKSDB_LITE
  auto bbto =
      immutable_options->table_factory->GetOptions<BlockBasedTableOptions>();
  if (bbto &&
@@ -38,6 +39,7 @@ BlobSource::BlobSource(const ImmutableOptions* immutable_options,
    blob_cache_ = SharedCacheInterface{std::make_shared<ChargedCache>(
        immutable_options->blob_cache, bbto->block_cache)};
  }
+#endif  // ROCKSDB_LITE
 }

 BlobSource::~BlobSource() = default;
@@ -104,9 +106,9 @@ Status BlobSource::PutBlobIntoCache(
 }

 BlobSource::TypedHandle* BlobSource::GetEntryFromCache(const Slice& key) const {
-  return blob_cache_.LookupFull(key, nullptr /* context */,
-                                Cache::Priority::BOTTOM, statistics_,
-                                lowest_used_cache_tier_);
+  return blob_cache_.LookupFull(
+      key, nullptr /* context */, Cache::Priority::BOTTOM,
+      true /* wait_for_cache */, statistics_, lowest_used_cache_tier_);
 }

 void BlobSource::PinCachedBlob(CacheHandleGuard<BlobContents>* cached_blob,
@@ -1214,12 +1214,12 @@ TEST_F(BlobSecondaryCacheTest, GetBlobsFromSecondaryCache) {
      ASSERT_EQ(handle0, nullptr);

      // key0's item should be in the secondary cache.
-      bool kept_in_sec_cache = false;
+      bool is_in_sec_cache = false;
      auto sec_handle0 = secondary_cache->Lookup(
-          key0, BlobSource::SharedCacheInterface::GetFullHelper(),
+          key0, &BlobSource::SharedCacheInterface::kFullHelper,
          /*context*/ nullptr, true,
-          /*advise_erase=*/true, kept_in_sec_cache);
-      ASSERT_FALSE(kept_in_sec_cache);
+          /*advise_erase=*/true, is_in_sec_cache);
+      ASSERT_FALSE(is_in_sec_cache);
      ASSERT_NE(sec_handle0, nullptr);
      ASSERT_TRUE(sec_handle0->IsReady());
      auto value = static_cast<BlobContents*>(sec_handle0->Value());
@@ -1242,12 +1242,12 @@ TEST_F(BlobSecondaryCacheTest, GetBlobsFromSecondaryCache) {
      ASSERT_NE(handle1, nullptr);
      blob_cache->Release(handle1);

-      bool kept_in_sec_cache = false;
+      bool is_in_sec_cache = false;
      auto sec_handle1 = secondary_cache->Lookup(
-          key1, BlobSource::SharedCacheInterface::GetFullHelper(),
+          key1, &BlobSource::SharedCacheInterface::kFullHelper,
          /*context*/ nullptr, true,
-          /*advise_erase=*/true, kept_in_sec_cache);
-      ASSERT_FALSE(kept_in_sec_cache);
+          /*advise_erase=*/true, is_in_sec_cache);
+      ASSERT_FALSE(is_in_sec_cache);
      ASSERT_EQ(sec_handle1, nullptr);

      ASSERT_TRUE(blob_source.TEST_BlobInCache(file_number, file_size,
@@ -1372,7 +1372,7 @@ class BlobSourceCacheReservationTest : public DBTestBase {

  static constexpr std::size_t kSizeDummyEntry = CacheReservationManagerImpl<
      CacheEntryRole::kBlobCache>::GetDummyEntrySize();
-  static constexpr std::size_t kCacheCapacity = 2 * kSizeDummyEntry;
+  static constexpr std::size_t kCacheCapacity = 1 * kSizeDummyEntry;
  static constexpr int kNumShardBits = 0;  // 2^0 shard

  static constexpr uint32_t kColumnFamilyId = 1;
@@ -1391,6 +1391,7 @@ class BlobSourceCacheReservationTest : public DBTestBase {
  std::string db_session_id_;
 };

+#ifndef ROCKSDB_LITE
 TEST_F(BlobSourceCacheReservationTest, SimpleCacheReservation) {
  options_.cf_paths.emplace_back(
      test::PerThreadDBPath(
@@ -1505,10 +1506,11 @@ TEST_F(BlobSourceCacheReservationTest, SimpleCacheReservation) {
  }
 }

-TEST_F(BlobSourceCacheReservationTest, IncreaseCacheReservation) {
+TEST_F(BlobSourceCacheReservationTest, IncreaseCacheReservationOnFullCache) {
  options_.cf_paths.emplace_back(
      test::PerThreadDBPath(
-          env_, "BlobSourceCacheReservationTest_IncreaseCacheReservation"),
+          env_,
+          "BlobSourceCacheReservationTest_IncreaseCacheReservationOnFullCache"),
      0);

  GenerateKeysAndBlobs();
@@ -1516,7 +1518,7 @@ TEST_F(BlobSourceCacheReservationTest, IncreaseCacheReservation) {
  DestroyAndReopen(options_);

  ImmutableOptions immutable_options(options_);
-  constexpr size_t blob_size = 24 << 10;  // 24KB
+  constexpr size_t blob_size = kSizeDummyEntry / (kNumBlobs / 2);
  for (size_t i = 0; i < kNumBlobs; ++i) {
    blob_file_size_ -= blobs_[i].size();  // old blob size
    blob_strs_[i].resize(blob_size, '@');
@@ -1574,6 +1576,11 @@ TEST_F(BlobSourceCacheReservationTest, IncreaseCacheReservation) {

    std::vector<PinnableSlice> values(keys_.size());

+    // Since we resized each blob to be kSizeDummyEntry / (num_blobs / 2), we
+    // can't fit all the blobs in the cache at the same time, which means we
+    // should observe cache evictions once we reach the cache's capacity.
+    // Due to the overhead of the cache and the BlobContents objects, as well as
+    // jemalloc bin sizes, this happens after inserting seven blobs.
    uint64_t blob_bytes = 0;
    for (size_t i = 0; i < kNumBlobs; ++i) {
      ASSERT_OK(blob_source.GetBlob(
@@ -1584,21 +1591,22 @@ TEST_F(BlobSourceCacheReservationTest, IncreaseCacheReservation) {
      // Release cache handle
      values[i].Reset();

-      size_t charge = 0;
-      ASSERT_TRUE(blob_source.TEST_BlobInCache(kBlobFileNumber, blob_file_size_,
-                                               blob_offsets[i], &charge));
+      if (i < kNumBlobs / 2 - 1) {
+        size_t charge = 0;
+        ASSERT_TRUE(blob_source.TEST_BlobInCache(
+            kBlobFileNumber, blob_file_size_, blob_offsets[i], &charge));

-      blob_bytes += charge;
+        blob_bytes += charge;
+      }

-      ASSERT_EQ(cache_res_mgr->GetTotalReservedCacheSize(),
-                (blob_bytes <= kSizeDummyEntry) ? kSizeDummyEntry
-                                                : (2 * kSizeDummyEntry));
+      ASSERT_EQ(cache_res_mgr->GetTotalReservedCacheSize(), kSizeDummyEntry);
      ASSERT_EQ(cache_res_mgr->GetTotalMemoryUsed(), blob_bytes);
      ASSERT_EQ(cache_res_mgr->GetTotalMemoryUsed(),
                options_.blob_cache->GetUsage());
    }
  }
 }
+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE

@@ -11,7 +11,6 @@
 #include "db/blob/blob_index.h"
 #include "db/blob/blob_log_format.h"
 #include "db/db_test_util.h"
-#include "db/db_with_timestamp_test_util.h"
 #include "port/stack_trace.h"
 #include "test_util/sync_point.h"
 #include "utilities/fault_injection_env.h"
@@ -585,6 +584,7 @@ TEST_F(DBBlobBasicTest, MultiGetBlobsFromCache) {
  }
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobBasicTest, MultiGetWithDirectIO) {
  Options options = GetDefaultOptions();

@@ -773,6 +773,7 @@ TEST_F(DBBlobBasicTest, MultiGetWithDirectIO) {
    ASSERT_EQ(values[2], second_blob);
  }
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBBlobBasicTest, MultiGetBlobsFromMultipleFiles) {
  Options options = GetDefaultOptions();
@@ -1061,6 +1062,7 @@ TEST_F(DBBlobBasicTest, GetBlob_IndexWithInvalidFileNumber) {
                  .IsCorruption());
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobBasicTest, GenerateIOTracing) {
  Options options = GetDefaultOptions();
  options.enable_blob_files = true;
@@ -1115,6 +1117,7 @@ TEST_F(DBBlobBasicTest, GenerateIOTracing) {
    ASSERT_GT(blob_files_op_count, 2);
  }
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBBlobBasicTest, BestEffortsRecovery_MissingNewestBlobFile) {
  Options options = GetDefaultOptions();
@@ -1216,6 +1219,7 @@ TEST_F(DBBlobBasicTest, MultiGetMergeBlobWithPut) {
  ASSERT_EQ(values[2], "v2_0");
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobBasicTest, Properties) {
  Options options = GetDefaultOptions();
  options.enable_blob_files = true;
@@ -1378,6 +1382,7 @@ TEST_F(DBBlobBasicTest, PropertiesMultiVersion) {
                 BlobLogRecord::CalculateAdjustmentForRecordHeader(key_size) +
                 blob_size + BlobLogFooter::kSize));
 }
+#endif  // !ROCKSDB_LITE

 class DBBlobBasicIOErrorTest : public DBBlobBasicTest,
                               public testing::WithParamInterface<std::string> {
@@ -1627,6 +1632,7 @@ TEST_F(DBBlobBasicTest, WarmCacheWithBlobsDuringFlush) {
            options.statistics->getTickerCount(BLOB_DB_CACHE_ADD));
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobBasicTest, DynamicallyWarmCacheDuringFlush) {
  Options options = GetDefaultOptions();

@@ -1694,6 +1700,7 @@ TEST_F(DBBlobBasicTest, DynamicallyWarmCacheDuringFlush) {
                              /*end=*/nullptr));
  EXPECT_EQ(0, options.statistics->getTickerCount(BLOB_DB_CACHE_ADD));
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBBlobBasicTest, WarmCacheWithBlobsSecondary) {
  CompressedSecondaryCacheOptions secondary_cache_opts;
@@ -1772,461 +1779,6 @@ TEST_F(DBBlobBasicTest, WarmCacheWithBlobsSecondary) {
            1);
 }

-TEST_F(DBBlobBasicTest, GetEntityBlob) {
-  Options options = GetDefaultOptions();
-  options.enable_blob_files = true;
-  options.min_blob_size = 0;
-
-  Reopen(options);
-
-  constexpr char key[] = "key";
-  constexpr char blob_value[] = "blob_value";
-
-  constexpr char other_key[] = "other_key";
-  constexpr char other_blob_value[] = "other_blob_value";
-
-  ASSERT_OK(Put(key, blob_value));
-  ASSERT_OK(Put(other_key, other_blob_value));
-
-  ASSERT_OK(Flush());
-
-  WideColumns expected_columns{{kDefaultWideColumnName, blob_value}};
-  WideColumns other_expected_columns{
-      {kDefaultWideColumnName, other_blob_value}};
-
-  {
-    PinnableWideColumns result;
-    ASSERT_OK(db_->GetEntity(ReadOptions(), db_->DefaultColumnFamily(), key,
-                             &result));
-    ASSERT_EQ(result.columns(), expected_columns);
-  }
-
-  {
-    PinnableWideColumns result;
-    ASSERT_OK(db_->GetEntity(ReadOptions(), db_->DefaultColumnFamily(),
-                             other_key, &result));
-
-    ASSERT_EQ(result.columns(), other_expected_columns);
-  }
-
-  {
-    constexpr size_t num_keys = 2;
-
-    std::array<Slice, num_keys> keys{{key, other_key}};
-    std::array<PinnableWideColumns, num_keys> results;
-    std::array<Status, num_keys> statuses;
-
-    db_->MultiGetEntity(ReadOptions(), db_->DefaultColumnFamily(), num_keys,
-                        &keys[0], &results[0], &statuses[0]);
-
-    ASSERT_OK(statuses[0]);
-    ASSERT_EQ(results[0].columns(), expected_columns);
-
-    ASSERT_OK(statuses[1]);
-    ASSERT_EQ(results[1].columns(), other_expected_columns);
-  }
-}
-
-class DBBlobWithTimestampTest : public DBBasicTestWithTimestampBase {
- protected:
-  DBBlobWithTimestampTest()
-      : DBBasicTestWithTimestampBase("db_blob_with_timestamp_test") {}
-};
-
-TEST_F(DBBlobWithTimestampTest, GetBlob) {
-  Options options = GetDefaultOptions();
-  options.create_if_missing = true;
-  options.enable_blob_files = true;
-  options.min_blob_size = 0;
-  const size_t kTimestampSize = Timestamp(0, 0).size();
-  TestComparator test_cmp(kTimestampSize);
-  options.comparator = &test_cmp;
-
-  DestroyAndReopen(options);
-  WriteOptions write_opts;
-  const std::string ts = Timestamp(1, 0);
-  constexpr char key[] = "key";
-  constexpr char blob_value[] = "blob_value";
-
-  ASSERT_OK(db_->Put(write_opts, key, ts, blob_value));
-
-  ASSERT_OK(Flush());
-
-  const std::string read_ts = Timestamp(2, 0);
-  Slice read_ts_slice(read_ts);
-  ReadOptions read_opts;
-  read_opts.timestamp = &read_ts_slice;
-  std::string value;
-  ASSERT_OK(db_->Get(read_opts, key, &value));
-  ASSERT_EQ(value, blob_value);
-}
-
-TEST_F(DBBlobWithTimestampTest, MultiGetBlobs) {
-  constexpr size_t min_blob_size = 6;
-
-  Options options = GetDefaultOptions();
-  options.enable_blob_files = true;
-  options.min_blob_size = min_blob_size;
-  options.create_if_missing = true;
-  const size_t kTimestampSize = Timestamp(0, 0).size();
-  TestComparator test_cmp(kTimestampSize);
-  options.comparator = &test_cmp;
-
-  DestroyAndReopen(options);
-
-  // Put then retrieve three key-values. The first value is below the size limit
-  // and is thus stored inline; the other two are stored separately as blobs.
-  constexpr size_t num_keys = 3;
-
-  constexpr char first_key[] = "first_key";
-  constexpr char first_value[] = "short";
-  static_assert(sizeof(first_value) - 1 < min_blob_size,
-                "first_value too long to be inlined");
-
-  DestroyAndReopen(options);
-  WriteOptions write_opts;
-  const std::string ts = Timestamp(1, 0);
-  ASSERT_OK(db_->Put(write_opts, first_key, ts, first_value));
-
-  constexpr char second_key[] = "second_key";
-  constexpr char second_value[] = "long_value";
-  static_assert(sizeof(second_value) - 1 >= min_blob_size,
-                "second_value too short to be stored as blob");
-
-  ASSERT_OK(db_->Put(write_opts, second_key, ts, second_value));
-
-  constexpr char third_key[] = "third_key";
-  constexpr char third_value[] = "other_long_value";
-  static_assert(sizeof(third_value) - 1 >= min_blob_size,
-                "third_value too short to be stored as blob");
-
-  ASSERT_OK(db_->Put(write_opts, third_key, ts, third_value));
-
-  ASSERT_OK(Flush());
-
-  ReadOptions read_options;
-  const std::string read_ts = Timestamp(2, 0);
-  Slice read_ts_slice(read_ts);
-  read_options.timestamp = &read_ts_slice;
-  std::array<Slice, num_keys> keys{{first_key, second_key, third_key}};
-
-  {
-    std::array<PinnableSlice, num_keys> values;
-    std::array<Status, num_keys> statuses;
-
-    db_->MultiGet(read_options, db_->DefaultColumnFamily(), num_keys, &keys[0],
-                  &values[0], &statuses[0]);
-
-    ASSERT_OK(statuses[0]);
-    ASSERT_EQ(values[0], first_value);
-
-    ASSERT_OK(statuses[1]);
-    ASSERT_EQ(values[1], second_value);
-
-    ASSERT_OK(statuses[2]);
-    ASSERT_EQ(values[2], third_value);
-  }
-}
-
-TEST_F(DBBlobWithTimestampTest, GetMergeBlobWithPut) {
-  Options options = GetDefaultOptions();
-  options.merge_operator = MergeOperators::CreateStringAppendOperator();
-  options.enable_blob_files = true;
-  options.min_blob_size = 0;
-  options.create_if_missing = true;
-  const size_t kTimestampSize = Timestamp(0, 0).size();
-  TestComparator test_cmp(kTimestampSize);
-  options.comparator = &test_cmp;
-
-  DestroyAndReopen(options);
-
-  WriteOptions write_opts;
-  const std::string ts = Timestamp(1, 0);
-  ASSERT_OK(db_->Put(write_opts, "Key1", ts, "v1"));
-  ASSERT_OK(Flush());
-  ASSERT_OK(
-      db_->Merge(write_opts, db_->DefaultColumnFamily(), "Key1", ts, "v2"));
-  ASSERT_OK(Flush());
-  ASSERT_OK(
-      db_->Merge(write_opts, db_->DefaultColumnFamily(), "Key1", ts, "v3"));
-  ASSERT_OK(Flush());
-
-  std::string value;
-  const std::string read_ts = Timestamp(2, 0);
-  Slice read_ts_slice(read_ts);
-  ReadOptions read_opts;
-  read_opts.timestamp = &read_ts_slice;
-  ASSERT_OK(db_->Get(read_opts, "Key1", &value));
-  ASSERT_EQ(value, "v1,v2,v3");
-}
-
-TEST_F(DBBlobWithTimestampTest, MultiGetMergeBlobWithPut) {
-  constexpr size_t num_keys = 3;
-
-  Options options = GetDefaultOptions();
-  options.merge_operator = MergeOperators::CreateStringAppendOperator();
-  options.enable_blob_files = true;
-  options.min_blob_size = 0;
-  options.create_if_missing = true;
-  const size_t kTimestampSize = Timestamp(0, 0).size();
-  TestComparator test_cmp(kTimestampSize);
-  options.comparator = &test_cmp;
-
-  DestroyAndReopen(options);
-
-  WriteOptions write_opts;
-  const std::string ts = Timestamp(1, 0);
-
-  ASSERT_OK(db_->Put(write_opts, "Key0", ts, "v0_0"));
-  ASSERT_OK(db_->Put(write_opts, "Key1", ts, "v1_0"));
-  ASSERT_OK(db_->Put(write_opts, "Key2", ts, "v2_0"));
-  ASSERT_OK(Flush());
-  ASSERT_OK(
-      db_->Merge(write_opts, db_->DefaultColumnFamily(), "Key0", ts, "v0_1"));
-  ASSERT_OK(
-      db_->Merge(write_opts, db_->DefaultColumnFamily(), "Key1", ts, "v1_1"));
-  ASSERT_OK(Flush());
-  ASSERT_OK(
-      db_->Merge(write_opts, db_->DefaultColumnFamily(), "Key0", ts, "v0_2"));
-  ASSERT_OK(Flush());
-
-  const std::string read_ts = Timestamp(2, 0);
-  Slice read_ts_slice(read_ts);
-  ReadOptions read_opts;
-  read_opts.timestamp = &read_ts_slice;
-  std::array<Slice, num_keys> keys{{"Key0", "Key1", "Key2"}};
-  std::array<PinnableSlice, num_keys> values;
-  std::array<Status, num_keys> statuses;
-
-  db_->MultiGet(read_opts, db_->DefaultColumnFamily(), num_keys, &keys[0],
-                &values[0], &statuses[0]);
-
-  ASSERT_OK(statuses[0]);
-  ASSERT_EQ(values[0], "v0_0,v0_1,v0_2");
-
-  ASSERT_OK(statuses[1]);
-  ASSERT_EQ(values[1], "v1_0,v1_1");
-
-  ASSERT_OK(statuses[2]);
-  ASSERT_EQ(values[2], "v2_0");
-}
-
-TEST_F(DBBlobWithTimestampTest, IterateBlobs) {
-  Options options = GetDefaultOptions();
-  options.enable_blob_files = true;
-  options.create_if_missing = true;
-  const size_t kTimestampSize = Timestamp(0, 0).size();
-  TestComparator test_cmp(kTimestampSize);
-  options.comparator = &test_cmp;
-
-  DestroyAndReopen(options);
-
-  int num_blobs = 5;
-  std::vector<std::string> keys;
-  std::vector<std::string> blobs;
-
-  WriteOptions write_opts;
-  std::vector<std::string> write_timestamps = {Timestamp(1, 0),
-                                               Timestamp(2, 0)};
-
-  // For each key in ["key0", ... "keyi", ...], write two versions:
-  // Timestamp(1, 0), "blobi0"
-  // Timestamp(2, 0), "blobi1"
-  for (int i = 0; i < num_blobs; i++) {
-    keys.push_back("key" + std::to_string(i));
-    blobs.push_back("blob" + std::to_string(i));
-    for (size_t j = 0; j < write_timestamps.size(); j++) {
-      ASSERT_OK(db_->Put(write_opts, keys[i], write_timestamps[j],
-                         blobs[i] + std::to_string(j)));
-    }
-  }
-  ASSERT_OK(Flush());
-
-  ReadOptions read_options;
-  std::vector<std::string> read_timestamps = {Timestamp(0, 0), Timestamp(3, 0)};
-  Slice ts_upper_bound(read_timestamps[1]);
-  read_options.timestamp = &ts_upper_bound;
-
-  auto check_iter_entry =
-      [](const Iterator* iter, const std::string& expected_key,
-         const std::string& expected_ts, const std::string& expected_value,
-         bool key_is_internal = true) {
-        ASSERT_OK(iter->status());
-        if (key_is_internal) {
-          std::string expected_ukey_and_ts;
-          expected_ukey_and_ts.assign(expected_key.data(), expected_key.size());
-          expected_ukey_and_ts.append(expected_ts.data(), expected_ts.size());
-
-          ParsedInternalKey parsed_ikey;
-          ASSERT_OK(ParseInternalKey(iter->key(), &parsed_ikey,
-                                     true /* log_err_key */));
-          ASSERT_EQ(parsed_ikey.user_key, expected_ukey_and_ts);
-        } else {
-          ASSERT_EQ(iter->key(), expected_key);
-        }
-        ASSERT_EQ(iter->timestamp(), expected_ts);
-        ASSERT_EQ(iter->value(), expected_value);
-      };
-
-  // Forward iterating one version of each key, get in this order:
-  // [("key0", Timestamp(2, 0), "blob01"),
-  //  ("key1", Timestamp(2, 0), "blob11")...]
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToFirst();
-    for (int i = 0; i < num_blobs; i++) {
-      check_iter_entry(iter.get(), keys[i], write_timestamps[1],
-                       blobs[i] + std::to_string(1), /*key_is_internal*/ false);
-      iter->Next();
-    }
-  }
-
-  // Forward iteration, then reverse to backward.
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToFirst();
-    for (int i = 0; i < num_blobs * 2 - 1; i++) {
-      if (i < num_blobs) {
-        check_iter_entry(iter.get(), keys[i], write_timestamps[1],
-                         blobs[i] + std::to_string(1),
-                         /*key_is_internal*/ false);
-        if (i != num_blobs - 1) {
-          iter->Next();
-        }
-      } else {
-        if (i != num_blobs) {
-          check_iter_entry(iter.get(), keys[num_blobs * 2 - 1 - i],
-                           write_timestamps[1],
-                           blobs[num_blobs * 2 - 1 - i] + std::to_string(1),
-                           /*key_is_internal*/ false);
-        }
-        iter->Prev();
-      }
-    }
-  }
-
-  // Backward iterating one versions of each key, get in this order:
-  // [("key4", Timestamp(2, 0), "blob41"),
-  //  ("key3", Timestamp(2, 0), "blob31")...]
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToLast();
-    for (int i = 0; i < num_blobs; i++) {
-      check_iter_entry(iter.get(), keys[num_blobs - 1 - i], write_timestamps[1],
-                       blobs[num_blobs - 1 - i] + std::to_string(1),
-                       /*key_is_internal*/ false);
-      iter->Prev();
-    }
-  }
-
-  // Backward iteration, then reverse to forward.
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToLast();
-    for (int i = 0; i < num_blobs * 2 - 1; i++) {
-      if (i < num_blobs) {
-        check_iter_entry(iter.get(), keys[num_blobs - 1 - i],
-                         write_timestamps[1],
-                         blobs[num_blobs - 1 - i] + std::to_string(1),
-                         /*key_is_internal*/ false);
-        if (i != num_blobs - 1) {
-          iter->Prev();
-        }
-      } else {
-        if (i != num_blobs) {
-          check_iter_entry(iter.get(), keys[i - num_blobs], write_timestamps[1],
-                           blobs[i - num_blobs] + std::to_string(1),
-                           /*key_is_internal*/ false);
-        }
-        iter->Next();
-      }
-    }
-  }
-
-  Slice ts_lower_bound(read_timestamps[0]);
-  read_options.iter_start_ts = &ts_lower_bound;
-  // Forward iterating multiple versions of the same key, get in this order:
-  // [("key0", Timestamp(2, 0), "blob01"),
-  //  ("key0", Timestamp(1, 0), "blob00"),
-  //  ("key1", Timestamp(2, 0), "blob11")...]
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToFirst();
-    for (int i = 0; i < num_blobs; i++) {
-      for (size_t j = write_timestamps.size(); j > 0; --j) {
-        check_iter_entry(iter.get(), keys[i], write_timestamps[j - 1],
-                         blobs[i] + std::to_string(j - 1));
-        iter->Next();
-      }
-    }
-  }
-
-  // Backward iterating multiple versions of the same key, get in this order:
-  // [("key4", Timestamp(1, 0), "blob00"),
-  //  ("key4", Timestamp(2, 0), "blob01"),
-  //  ("key3", Timestamp(1, 0), "blob10")...]
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToLast();
-    for (int i = num_blobs; i > 0; i--) {
-      for (size_t j = 0; j < write_timestamps.size(); j++) {
-        check_iter_entry(iter.get(), keys[i - 1], write_timestamps[j],
-                         blobs[i - 1] + std::to_string(j));
-        iter->Prev();
-      }
-    }
-  }
-
-  int upper_bound_idx = num_blobs - 2;
-  int lower_bound_idx = 1;
-  Slice upper_bound_slice(keys[upper_bound_idx]);
-  Slice lower_bound_slice(keys[lower_bound_idx]);
-  read_options.iterate_upper_bound = &upper_bound_slice;
-  read_options.iterate_lower_bound = &lower_bound_slice;
-
-  // Forward iteration with upper and lower bound.
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToFirst();
-    for (int i = lower_bound_idx; i < upper_bound_idx; i++) {
-      for (size_t j = write_timestamps.size(); j > 0; --j) {
-        check_iter_entry(iter.get(), keys[i], write_timestamps[j - 1],
-                         blobs[i] + std::to_string(j - 1));
-        iter->Next();
-      }
-    }
-  }
-
-  // Backward iteration with upper and lower bound.
-  {
-    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
-    ASSERT_OK(iter->status());
-
-    iter->SeekToLast();
-    for (int i = upper_bound_idx; i > lower_bound_idx; i--) {
-      for (size_t j = 0; j < write_timestamps.size(); j++) {
-        check_iter_entry(iter.get(), keys[i - 1], write_timestamps[j],
-                         blobs[i - 1] + std::to_string(j));
-        iter->Prev();
-      }
-    }
-  }
-}
-
 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
@@ -16,6 +16,7 @@ class DBBlobCompactionTest : public DBTestBase {
  explicit DBBlobCompactionTest()
      : DBTestBase("db_blob_compaction_test", /*env_do_fsync=*/false) {}

+#ifndef ROCKSDB_LITE
  const std::vector<InternalStats::CompactionStats>& GetCompactionStats() {
    VersionSet* const versions = dbfull()->GetVersionSet();
    assert(versions);
@@ -29,6 +30,7 @@ class DBBlobCompactionTest : public DBTestBase {

    return internal_stats->TEST_GetCompactionStats();
  }
+#endif  // ROCKSDB_LITE
 };

 namespace {
@@ -248,6 +250,7 @@ TEST_F(DBBlobCompactionTest, FilterByKeyLength) {
  ASSERT_OK(db_->Get(ReadOptions(), long_key, &value));
  ASSERT_EQ("value", value);

+#ifndef ROCKSDB_LITE
  const auto& compaction_stats = GetCompactionStats();
  ASSERT_GE(compaction_stats.size(), 2);

@@ -255,6 +258,7 @@ TEST_F(DBBlobCompactionTest, FilterByKeyLength) {
  // this involves neither reading nor writing blobs
  ASSERT_EQ(compaction_stats[1].bytes_read_blob, 0);
  ASSERT_EQ(compaction_stats[1].bytes_written_blob, 0);
+#endif  // ROCKSDB_LITE

  Close();
 }
@@ -295,6 +299,7 @@ TEST_F(DBBlobCompactionTest, FilterByValueLength) {
    ASSERT_EQ(long_value, value);
  }

+#ifndef ROCKSDB_LITE
  const auto& compaction_stats = GetCompactionStats();
  ASSERT_GE(compaction_stats.size(), 2);

@@ -302,10 +307,12 @@ TEST_F(DBBlobCompactionTest, FilterByValueLength) {
  // this involves reading but not writing blobs
  ASSERT_GT(compaction_stats[1].bytes_read_blob, 0);
  ASSERT_EQ(compaction_stats[1].bytes_written_blob, 0);
+#endif  // ROCKSDB_LITE

  Close();
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobCompactionTest, BlobCompactWithStartingLevel) {
  Options options = GetDefaultOptions();

@@ -381,6 +388,7 @@ TEST_F(DBBlobCompactionTest, BlobCompactWithStartingLevel) {

  Close();
 }
+#endif

 TEST_F(DBBlobCompactionTest, BlindWriteFilter) {
  Options options = GetDefaultOptions();
@@ -405,6 +413,7 @@ TEST_F(DBBlobCompactionTest, BlindWriteFilter) {
    ASSERT_EQ(new_blob_value, Get(key));
  }

+#ifndef ROCKSDB_LITE
  const auto& compaction_stats = GetCompactionStats();
  ASSERT_GE(compaction_stats.size(), 2);

@@ -412,6 +421,7 @@ TEST_F(DBBlobCompactionTest, BlindWriteFilter) {
  // this involves writing but not reading blobs
  ASSERT_EQ(compaction_stats[1].bytes_read_blob, 0);
  ASSERT_GT(compaction_stats[1].bytes_written_blob, 0);
+#endif  // ROCKSDB_LITE

  Close();
 }
@@ -530,6 +540,7 @@ TEST_F(DBBlobCompactionTest, CompactionFilter) {
    ASSERT_EQ(kv.second + std::string(padding), Get(kv.first));
  }

+#ifndef ROCKSDB_LITE
  const auto& compaction_stats = GetCompactionStats();
  ASSERT_GE(compaction_stats.size(), 2);

@@ -537,6 +548,7 @@ TEST_F(DBBlobCompactionTest, CompactionFilter) {
  // this involves reading and writing blobs
  ASSERT_GT(compaction_stats[1].bytes_read_blob, 0);
  ASSERT_GT(compaction_stats[1].bytes_written_blob, 0);
+#endif  // ROCKSDB_LITE

  Close();
 }
@@ -594,6 +606,7 @@ TEST_F(DBBlobCompactionTest, CompactionFilterReadBlobAndKeep) {
                              /*end=*/nullptr));
  ASSERT_EQ(blob_files, GetBlobFileNumbers());

+#ifndef ROCKSDB_LITE
  const auto& compaction_stats = GetCompactionStats();
  ASSERT_GE(compaction_stats.size(), 2);

@@ -601,6 +614,7 @@ TEST_F(DBBlobCompactionTest, CompactionFilterReadBlobAndKeep) {
  // this involves reading but not writing blobs
  ASSERT_GT(compaction_stats[1].bytes_read_blob, 0);
  ASSERT_EQ(compaction_stats[1].bytes_written_blob, 0);
+#endif  // ROCKSDB_LITE

  Close();
 }
@@ -34,6 +34,7 @@ class DBBlobCorruptionTest : public DBTestBase {
  }
 };

+#ifndef ROCKSDB_LITE
 TEST_F(DBBlobCorruptionTest, VerifyWholeBlobFileChecksum) {
  Options options = GetDefaultOptions();
  options.enable_blob_files = true;
@@ -70,6 +71,7 @@ TEST_F(DBBlobCorruptionTest, VerifyWholeBlobFileChecksum) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
 }
+#endif  // !ROCKSDB_LITE
 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
@@ -131,7 +131,9 @@ class DBBlobIndexTest : public DBTestBase {
        ASSERT_OK(Flush());
        ASSERT_OK(
            dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
+#ifndef ROCKSDB_LITE
        ASSERT_EQ("0,1", FilesPerLevel());
+#endif  // !ROCKSDB_LITE
        break;
    }
  }
@@ -457,6 +459,7 @@ TEST_F(DBBlobIndexTest, Iterate) {
    verify(15, Status::kOk, get_value(16, 0), get_value(14, 0),
           create_blob_iterator, check_is_blob(false));

+#ifndef ROCKSDB_LITE
    // Iterator with blob support and using seek.
    ASSERT_OK(dbfull()->SetOptions(
        cfh(), {{"max_sequential_skip_in_iterations", "0"}}));
@@ -481,6 +484,7 @@ TEST_F(DBBlobIndexTest, Iterate) {
           create_blob_iterator, check_is_blob(false));
    verify(15, Status::kOk, get_value(16, 0), get_value(14, 0),
           create_blob_iterator, check_is_blob(false));
+#endif  // !ROCKSDB_LITE

    for (auto* snapshot : snapshots) {
      dbfull()->ReleaseSnapshot(snapshot);
@@ -580,10 +584,12 @@ TEST_F(DBBlobIndexTest, IntegratedBlobIterate) {
  Status expected_status;
  verify(1, expected_status, expected_value);

+#ifndef ROCKSDB_LITE
  // Test DBIter::FindValueForCurrentKeyUsingSeek flow.
  ASSERT_OK(dbfull()->SetOptions(cfh(),
                                 {{"max_sequential_skip_in_iterations", "0"}}));
  verify(1, expected_status, expected_value);
+#endif  // !ROCKSDB_LITE
 }

 }  // namespace ROCKSDB_NAMESPACE
@@ -107,9 +107,11 @@ Status BuildTable(
  std::vector<std::string> blob_file_paths;
  std::string file_checksum = kUnknownFileChecksum;
  std::string file_checksum_func_name = kUnknownFileChecksumFuncName;
+#ifndef ROCKSDB_LITE
  EventHelpers::NotifyTableFileCreationStarted(ioptions.listeners, dbname,
                                               tboptions.column_family_name,
                                               fname, job_id, tboptions.reason);
+#endif  // !ROCKSDB_LITE
  Env* env = db_options.env;
  assert(env);
  FileSystem* fs = db_options.fs.get();
@@ -7,6 +7,8 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

+#ifndef ROCKSDB_LITE
+
 #include "rocksdb/c.h"

 #include <cstdlib>
@@ -15,7 +17,7 @@
 #include <vector>

 #include "port/port.h"
-#include "rocksdb/advanced_cache.h"
+#include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/comparator.h"
 #include "rocksdb/convenience.h"
@@ -67,7 +69,6 @@ using ROCKSDB_NAMESPACE::CompactionOptionsFIFO;
 using ROCKSDB_NAMESPACE::CompactRangeOptions;
 using ROCKSDB_NAMESPACE::Comparator;
 using ROCKSDB_NAMESPACE::CompressionType;
-using ROCKSDB_NAMESPACE::ConfigOptions;
 using ROCKSDB_NAMESPACE::CuckooTableOptions;
 using ROCKSDB_NAMESPACE::DB;
 using ROCKSDB_NAMESPACE::DBOptions;
@@ -77,7 +78,6 @@ using ROCKSDB_NAMESPACE::EnvOptions;
 using ROCKSDB_NAMESPACE::FileLock;
 using ROCKSDB_NAMESPACE::FilterPolicy;
 using ROCKSDB_NAMESPACE::FlushOptions;
-using ROCKSDB_NAMESPACE::HyperClockCacheOptions;
 using ROCKSDB_NAMESPACE::InfoLogLevel;
 using ROCKSDB_NAMESPACE::IngestExternalFileOptions;
 using ROCKSDB_NAMESPACE::Iterator;
@@ -208,9 +208,6 @@ struct rocksdb_logger_t {
 struct rocksdb_lru_cache_options_t {
  LRUCacheOptions rep;
 };
-struct rocksdb_hyper_clock_cache_options_t {
-  HyperClockCacheOptions rep;
-};
 struct rocksdb_memory_allocator_t {
  std::shared_ptr<MemoryAllocator> rep;
 };
@@ -1808,17 +1805,6 @@ void rocksdb_flush_cf(rocksdb_t* db, const rocksdb_flushoptions_t* options,
  SaveError(errptr, db->rep->Flush(options->rep, column_family->rep));
 }

-void rocksdb_flush_cfs(rocksdb_t* db, const rocksdb_flushoptions_t* options,
-                       rocksdb_column_family_handle_t** column_families,
-                       int num_column_families, char** errptr) {
-  std::vector<ColumnFamilyHandle*> column_family_handles;
-  for (int i = 0; i < num_column_families; i++) {
-    column_family_handles.push_back(column_families[i]->rep);
-  }
-
-  SaveError(errptr, db->rep->Flush(options->rep, column_family_handles));
-}
-
 void rocksdb_flush_wal(rocksdb_t* db, unsigned char sync, char** errptr) {
  SaveError(errptr, db->rep->FlushWAL(sync));
 }
@@ -2512,12 +2498,8 @@ void rocksdb_load_latest_options(
    rocksdb_options_t*** list_column_family_options, char** errptr) {
  DBOptions db_opt;
  std::vector<ColumnFamilyDescriptor> cf_descs;
-  ConfigOptions config_opts;
-  config_opts.ignore_unknown_options = ignore_unknown_options;
-  config_opts.input_strings_escaped = true;
-  config_opts.env = env->rep;
-  Status s = LoadLatestOptions(config_opts, std::string(db_path), &db_opt,
-                               &cf_descs, &cache->rep);
+  Status s = LoadLatestOptions(std::string(db_path), env->rep, &db_opt,
+                               &cf_descs, ignore_unknown_options, &cache->rep);
  if (s.ok()) {
    char** cf_names = (char**)malloc(cf_descs.size() * sizeof(char*));
    rocksdb_options_t** cf_options = (rocksdb_options_t**)malloc(
@@ -2638,6 +2620,14 @@ void rocksdb_block_based_options_set_block_cache(
  }
 }

+void rocksdb_block_based_options_set_block_cache_compressed(
+    rocksdb_block_based_table_options_t* options,
+    rocksdb_cache_t* block_cache_compressed) {
+  if (block_cache_compressed) {
+    options->rep.block_cache_compressed = block_cache_compressed->rep;
+  }
+}
+
 void rocksdb_block_based_options_set_whole_key_filtering(
    rocksdb_block_based_table_options_t* options, unsigned char v) {
  options->rep.whole_key_filtering = v;
@@ -4685,53 +4675,6 @@ rocksdb_cache_t* rocksdb_cache_create_lru_opts(
  return c;
 }

-rocksdb_hyper_clock_cache_options_t* rocksdb_hyper_clock_cache_options_create(
-    size_t capacity, size_t estimated_entry_charge) {
-  return new rocksdb_hyper_clock_cache_options_t{
-      HyperClockCacheOptions(capacity, estimated_entry_charge)};
-}
-
-void rocksdb_hyper_clock_cache_options_destroy(
-    rocksdb_hyper_clock_cache_options_t* opt) {
-  delete opt;
-}
-
-void rocksdb_hyper_clock_cache_options_set_capacity(
-    rocksdb_hyper_clock_cache_options_t* opts, size_t capacity) {
-  opts->rep.capacity = capacity;
-}
-
-void rocksdb_hyper_clock_cache_options_set_estimated_entry_charge(
-    rocksdb_hyper_clock_cache_options_t* opts, size_t estimated_entry_charge) {
-  opts->rep.estimated_entry_charge = estimated_entry_charge;
-}
-
-void rocksdb_hyper_clock_cache_options_set_num_shard_bits(
-    rocksdb_hyper_clock_cache_options_t* opts, int num_shard_bits) {
-  opts->rep.num_shard_bits = num_shard_bits;
-}
-
-void rocksdb_hyper_clock_cache_options_set_memory_allocator(
-    rocksdb_hyper_clock_cache_options_t* opts,
-    rocksdb_memory_allocator_t* memory_allocator) {
-  opts->rep.memory_allocator = memory_allocator->rep;
-}
-
-rocksdb_cache_t* rocksdb_cache_create_hyper_clock(
-    size_t capacity, size_t estimated_entry_charge) {
-  HyperClockCacheOptions opts(capacity, estimated_entry_charge);
-  rocksdb_cache_t* c = new rocksdb_cache_t;
-  c->rep = opts.MakeSharedCache();
-  return c;
-}
-
-rocksdb_cache_t* rocksdb_cache_create_hyper_clock_opts(
-    rocksdb_hyper_clock_cache_options_t* opts) {
-  rocksdb_cache_t* c = new rocksdb_cache_t;
-  c->rep = opts->rep.MakeSharedCache();
-  return c;
-}
-
 void rocksdb_cache_destroy(rocksdb_cache_t* cache) { delete cache; }

 void rocksdb_cache_disown_data(rocksdb_cache_t* cache) {
@@ -4972,12 +4915,6 @@ void rocksdb_ingestexternalfileoptions_set_ingest_behind(
  opt->rep.ingest_behind = ingest_behind;
 }

-void rocksdb_ingestexternalfileoptions_set_fail_if_not_bottommost_level(
-    rocksdb_ingestexternalfileoptions_t* opt,
-    unsigned char fail_if_not_bottommost_level) {
-  opt->rep.fail_if_not_bottommost_level = fail_if_not_bottommost_level;
-}
-
 void rocksdb_ingestexternalfileoptions_destroy(
    rocksdb_ingestexternalfileoptions_t* opt) {
  delete opt;
@@ -5139,16 +5076,6 @@ rocksdb_fifo_compaction_options_t* rocksdb_fifo_compaction_options_create() {
  return result;
 }

-void rocksdb_fifo_compaction_options_set_allow_compaction(
-    rocksdb_fifo_compaction_options_t* fifo_opts, unsigned char allow_compaction) {
-  fifo_opts->rep.allow_compaction = allow_compaction;
-}
-
-unsigned char rocksdb_fifo_compaction_options_get_allow_compaction(
-    rocksdb_fifo_compaction_options_t* fifo_opts) {
-  return fifo_opts->rep.allow_compaction;
-}
-
 void rocksdb_fifo_compaction_options_set_max_table_files_size(
    rocksdb_fifo_compaction_options_t* fifo_opts, uint64_t size) {
  fifo_opts->rep.max_table_files_size = size;
@@ -6186,18 +6113,6 @@ void rocksdb_transactiondb_flush_cf(
  SaveError(errptr, txn_db->rep->Flush(options->rep, column_family->rep));
 }

-void rocksdb_transactiondb_flush_cfs(
-    rocksdb_transactiondb_t* txn_db, const rocksdb_flushoptions_t* options,
-    rocksdb_column_family_handle_t** column_families, int num_column_families,
-    char** errptr) {
-  std::vector<ColumnFamilyHandle*> column_family_handles;
-  for (int i = 0; i < num_column_families; i++) {
-    column_family_handles.push_back(column_families[i]->rep);
-  }
-
-  SaveError(errptr, txn_db->rep->Flush(options->rep, column_family_handles));
-}
-
 rocksdb_checkpoint_t* rocksdb_transactiondb_checkpoint_object_create(
    rocksdb_transactiondb_t* txn_db, char** errptr) {
  Checkpoint* checkpoint;
@@ -6486,3 +6401,5 @@ void rocksdb_enable_manual_compaction(rocksdb_t* db) {
 }

 }  // end extern "C"
+
+#endif  // !ROCKSDB_LITE
@@ -5,6 +5,8 @@

 #include <stdio.h>

+#ifndef ROCKSDB_LITE  // Lite does not support C API
+
 #include <assert.h>
 #include <stddef.h>
 #include <stdlib.h>
@@ -3466,3 +3468,12 @@ int main(int argc, char** argv) {
  fprintf(stderr, "PASS\n");
  return 0;
 }
+
+#else
+
+int main(void) {
+  fprintf(stderr, "SKIPPED\n");
+  return 0;
+}
+
+#endif  // !ROCKSDB_LITE
@@ -53,9 +53,11 @@ ColumnFamilyHandleImpl::ColumnFamilyHandleImpl(

 ColumnFamilyHandleImpl::~ColumnFamilyHandleImpl() {
  if (cfd_ != nullptr) {
+#ifndef ROCKSDB_LITE
    for (auto& listener : cfd_->ioptions()->listeners) {
      listener->OnColumnFamilyHandleDeletionStarted(this);
    }
+#endif  // ROCKSDB_LITE
    // Job id == 0 means that this is not our background process, but rather
    // user thread
    // Need to hold some shared pointers owned by the initial_cf_options
@@ -86,10 +88,15 @@ const std::string& ColumnFamilyHandleImpl::GetName() const {
 }

 Status ColumnFamilyHandleImpl::GetDescriptor(ColumnFamilyDescriptor* desc) {
+#ifndef ROCKSDB_LITE
  // accessing mutable cf-options requires db mutex.
  InstrumentedMutexLock l(mutex_);
  *desc = ColumnFamilyDescriptor(cfd()->GetName(), cfd()->GetLatestCFOptions());
  return Status::OK();
+#else
+  (void)desc;
+  return Status::NotSupported();
+#endif  // !ROCKSDB_LITE
 }

 const Comparator* ColumnFamilyHandleImpl::GetComparator() const {
@@ -340,6 +347,7 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
        result.hard_pending_compaction_bytes_limit;
  }

+#ifndef ROCKSDB_LITE
  // When the DB is stopped, it's possible that there are some .trash files that
  // were not deleted yet, when we open the DB we will find these .trash files
  // and schedule them to be deleted (or delete immediately if SstFileManager
@@ -351,6 +359,7 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
                                      result.cf_paths[i].path)
        .PermitUncheckedError();
  }
+#endif

  if (result.cf_paths.empty()) {
    result.cf_paths = db_options.db_paths;
@@ -593,6 +602,7 @@ ColumnFamilyData::ColumnFamilyData(
    if (ioptions_.compaction_style == kCompactionStyleLevel) {
      compaction_picker_.reset(
          new LevelCompactionPicker(ioptions_, &internal_comparator_));
+#ifndef ROCKSDB_LITE
    } else if (ioptions_.compaction_style == kCompactionStyleUniversal) {
      compaction_picker_.reset(
          new UniversalCompactionPicker(ioptions_, &internal_comparator_));
@@ -606,6 +616,7 @@ ColumnFamilyData::ColumnFamilyData(
                     "Column family %s does not use any background compaction. "
                     "Compactions can only be done via CompactFiles\n",
                     GetName().c_str());
+#endif  // !ROCKSDB_LITE
    } else {
      ROCKS_LOG_ERROR(ioptions_.logger,
                      "Unable to recognize the specified compaction style %d. "
@@ -869,7 +880,7 @@ int GetL0ThresholdSpeedupCompaction(int level0_file_num_compaction_trigger,
 }
 }  // anonymous namespace

-std::pair<WriteStallCondition, WriteStallCause>
+std::pair<WriteStallCondition, ColumnFamilyData::WriteStallCause>
 ColumnFamilyData::GetWriteStallConditionAndCause(
    int num_unflushed_memtables, int num_l0_files,
    uint64_t num_compaction_needed_bytes,
@@ -942,8 +953,7 @@ WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
      internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_STOPS, 1);
      if (compaction_picker_->IsLevel0CompactionInProgress()) {
        internal_stats_->AddCFStats(
-            InternalStats::L0_FILE_COUNT_LIMIT_STOPS_WITH_ONGOING_COMPACTION,
-            1);
+            InternalStats::LOCKED_L0_FILE_COUNT_LIMIT_STOPS, 1);
      }
      ROCKS_LOG_WARN(ioptions_.logger,
                     "[%s] Stopping writes because we have %d level-0 files",
@@ -964,7 +974,7 @@ WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
          SetupDelay(write_controller, compaction_needed_bytes,
                     prev_compaction_needed_bytes_, was_stopped,
                     mutable_cf_options.disable_auto_compactions);
-      internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_DELAYS, 1);
+      internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_SLOWDOWNS, 1);
      ROCKS_LOG_WARN(
          ioptions_.logger,
          "[%s] Stalling writes because we have %d immutable memtables "
@@ -982,11 +992,11 @@ WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
          SetupDelay(write_controller, compaction_needed_bytes,
                     prev_compaction_needed_bytes_, was_stopped || near_stop,
                     mutable_cf_options.disable_auto_compactions);
-      internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_DELAYS, 1);
+      internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_SLOWDOWNS,
+                                  1);
      if (compaction_picker_->IsLevel0CompactionInProgress()) {
        internal_stats_->AddCFStats(
-            InternalStats::L0_FILE_COUNT_LIMIT_DELAYS_WITH_ONGOING_COMPACTION,
-            1);
+            InternalStats::LOCKED_L0_FILE_COUNT_LIMIT_SLOWDOWNS, 1);
      }
      ROCKS_LOG_WARN(ioptions_.logger,
                     "[%s] Stalling writes because we have %d level-0 files "
@@ -1012,7 +1022,7 @@ WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
                     prev_compaction_needed_bytes_, was_stopped || near_stop,
                     mutable_cf_options.disable_auto_compactions);
      internal_stats_->AddCFStats(
-          InternalStats::PENDING_COMPACTION_BYTES_LIMIT_DELAYS, 1);
+          InternalStats::PENDING_COMPACTION_BYTES_LIMIT_SLOWDOWNS, 1);
      ROCKS_LOG_WARN(
          ioptions_.logger,
          "[%s] Stalling writes because of estimated pending compaction "
@@ -1430,6 +1440,7 @@ Status ColumnFamilyData::ValidateOptions(
  return s;
 }

+#ifndef ROCKSDB_LITE
 Status ColumnFamilyData::SetOptions(
    const DBOptions& db_opts,
    const std::unordered_map<std::string, std::string>& options_map) {
@@ -1448,6 +1459,7 @@ Status ColumnFamilyData::SetOptions(
  }
  return s;
 }
+#endif  // ROCKSDB_LITE

 // REQUIRES: DB mutex held
 Env::WriteLifeTimeHint ColumnFamilyData::CalculateSSTWriteHint(int level) {
@@ -335,10 +335,12 @@ class ColumnFamilyData {
  // Validate CF options against DB options
  static Status ValidateOptions(const DBOptions& db_options,
                                const ColumnFamilyOptions& cf_options);
+#ifndef ROCKSDB_LITE
  // REQUIRES: DB mutex held
  Status SetOptions(
      const DBOptions& db_options,
      const std::unordered_map<std::string, std::string>& options_map);
+#endif  // ROCKSDB_LITE

  InternalStats* internal_stats() { return internal_stats_.get(); }

@@ -462,6 +464,12 @@ class ColumnFamilyData {
  bool queued_for_flush() { return queued_for_flush_; }
  bool queued_for_compaction() { return queued_for_compaction_; }

+  enum class WriteStallCause {
+    kNone,
+    kMemtableLimit,
+    kL0FileCountLimit,
+    kPendingCompactionBytes,
+  };
  static std::pair<WriteStallCondition, WriteStallCause>
  GetWriteStallConditionAndCause(
      int num_unflushed_memtables, int num_l0_files,
@@ -71,7 +71,11 @@ class ColumnFamilyTestBase : public testing::Test {
    for (auto h : handles_) {
      ColumnFamilyDescriptor cfdescriptor;
      Status s = h->GetDescriptor(&cfdescriptor);
+#ifdef ROCKSDB_LITE
+      EXPECT_TRUE(s.IsNotSupported());
+#else
      EXPECT_OK(s);
+#endif  // ROCKSDB_LITE
      column_families.push_back(cfdescriptor);
    }
    ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
@@ -193,10 +197,12 @@ class ColumnFamilyTestBase : public testing::Test {
                               &db_);
  }

+#ifndef ROCKSDB_LITE  // ReadOnlyDB is not supported
  void AssertOpenReadOnly(std::vector<std::string> cf,
                          std::vector<ColumnFamilyOptions> options = {}) {
    ASSERT_OK(OpenReadOnly(cf, options));
  }
+#endif  // !ROCKSDB_LITE

  void Open(std::vector<std::string> cf,
            std::vector<ColumnFamilyOptions> options = {}) {
@@ -218,16 +224,27 @@ class ColumnFamilyTestBase : public testing::Test {
  }

  bool IsDbWriteStopped() {
+#ifndef ROCKSDB_LITE
    uint64_t v;
    EXPECT_TRUE(dbfull()->GetIntProperty("rocksdb.is-write-stopped", &v));
    return (v == 1);
+#else
+    return dbfull()->TEST_write_controler().IsStopped();
+#endif  // !ROCKSDB_LITE
  }

  uint64_t GetDbDelayedWriteRate() {
+#ifndef ROCKSDB_LITE
    uint64_t v;
    EXPECT_TRUE(
        dbfull()->GetIntProperty("rocksdb.actual-delayed-write-rate", &v));
    return v;
+#else
+    if (!dbfull()->TEST_write_controler().NeedsDelay()) {
+      return 0;
+    }
+    return dbfull()->TEST_write_controler().delayed_write_rate();
+#endif  // !ROCKSDB_LITE
  }

  void Destroy(const std::vector<ColumnFamilyDescriptor>& column_families =
@@ -250,6 +267,7 @@ class ColumnFamilyTestBase : public testing::Test {
          db_->CreateColumnFamily(current_cf_opt, cfs[i], &handles_[cfi]));
      names_[cfi] = cfs[i];

+#ifndef ROCKSDB_LITE  // RocksDBLite does not support GetDescriptor
      // Verify the CF options of the returned CF handle.
      ColumnFamilyDescriptor desc;
      ASSERT_OK(handles_[cfi]->GetDescriptor(&desc));
@@ -258,6 +276,7 @@ class ColumnFamilyTestBase : public testing::Test {
      ASSERT_OK(RocksDBOptionsParser::VerifyCFOptions(
          ConfigOptions(), desc.options,
          SanitizeOptions(dbfull()->immutable_db_options(), current_cf_opt)));
+#endif  // !ROCKSDB_LITE
      cfi++;
    }
  }
@@ -306,6 +325,7 @@ class ColumnFamilyTestBase : public testing::Test {
    ASSERT_OK(db_->FlushWAL(/*sync=*/false));
  }

+#ifndef ROCKSDB_LITE  // TEST functions in DB are not supported in lite
  void WaitForFlush(int cf) {
    ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[cf]));
  }
@@ -319,6 +339,7 @@ class ColumnFamilyTestBase : public testing::Test {
  void AssertMaxTotalInMemoryState(uint64_t value) {
    ASSERT_EQ(value, MaxTotalInMemoryState());
  }
+#endif  // !ROCKSDB_LITE

  Status Put(int cf, const std::string& key, const std::string& value) {
    return db_->Put(WriteOptions(), handles_[cf], Slice(key), Slice(value));
@@ -356,6 +377,7 @@ class ColumnFamilyTestBase : public testing::Test {
                       "rocksdb.num-files-at-level" + std::to_string(level));
  }

+#ifndef ROCKSDB_LITE
  // Return spread of files per level
  std::string FilesPerLevel(int cf) {
    std::string result;
@@ -372,19 +394,31 @@ class ColumnFamilyTestBase : public testing::Test {
    result.resize(last_non_zero_offset);
    return result;
  }
+#endif

  void AssertFilesPerLevel(const std::string& value, int cf) {
+#ifndef ROCKSDB_LITE
    ASSERT_EQ(value, FilesPerLevel(cf));
+#else
+    (void)value;
+    (void)cf;
+#endif
  }

+#ifndef ROCKSDB_LITE  // GetLiveFilesMetaData is not supported
  int CountLiveFiles() {
    std::vector<LiveFileMetaData> metadata;
    db_->GetLiveFilesMetaData(&metadata);
    return static_cast<int>(metadata.size());
  }
+#endif  // !ROCKSDB_LITE

  void AssertCountLiveFiles(int expected_value) {
+#ifndef ROCKSDB_LITE
    ASSERT_EQ(expected_value, CountLiveFiles());
+#else
+    (void)expected_value;
+#endif
  }

  // Do n memtable flushes, each of which produces an sstable
@@ -398,6 +432,7 @@ class ColumnFamilyTestBase : public testing::Test {
    }
  }

+#ifndef ROCKSDB_LITE  // GetSortedWalFiles is not supported
  int CountLiveLogFiles() {
    int micros_wait_for_log_deletion = 20000;
    env_->SleepForMicroseconds(micros_wait_for_log_deletion);
@@ -426,18 +461,25 @@ class ColumnFamilyTestBase : public testing::Test {
    return ret;
    return 0;
  }
+#endif  // !ROCKSDB_LITE

  void AssertCountLiveLogFiles(int value) {
+#ifndef ROCKSDB_LITE  // GetSortedWalFiles is not supported
    ASSERT_EQ(value, CountLiveLogFiles());
+#else
+    (void)value;
+#endif  // !ROCKSDB_LITE
  }

  void AssertNumberOfImmutableMemtables(std::vector<int> num_per_cf) {
    assert(num_per_cf.size() == handles_.size());

+#ifndef ROCKSDB_LITE  // GetProperty is not supported in lite
    for (size_t i = 0; i < num_per_cf.size(); ++i) {
      ASSERT_EQ(num_per_cf[i], GetProperty(static_cast<int>(i),
                                           "rocksdb.num-immutable-mem-table"));
    }
+#endif  // !ROCKSDB_LITE
  }

  void CopyFile(const std::string& source, const std::string& destination,
@@ -533,6 +575,7 @@ TEST_P(ColumnFamilyTest, DontReuseColumnFamilyID) {
  }
 }

+#ifndef ROCKSDB_LITE
 TEST_P(ColumnFamilyTest, CreateCFRaceWithGetAggProperty) {
  Open();

@@ -555,6 +598,7 @@ TEST_P(ColumnFamilyTest, CreateCFRaceWithGetAggProperty) {

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
 }
+#endif  // !ROCKSDB_LITE

 class FlushEmptyCFTestWithParam
    : public ColumnFamilyTestBase,
@@ -898,6 +942,7 @@ TEST_P(ColumnFamilyTest, IgnoreRecoveredLog) {
  }
 }

+#ifndef ROCKSDB_LITE  // TEST functions used are not supported
 TEST_P(ColumnFamilyTest, FlushTest) {
  Open();
  CreateColumnFamiliesAndReopen({"one", "two"});
@@ -1012,6 +1057,7 @@ TEST_P(ColumnFamilyTest, LogDeletionTest) {
  AssertCountLiveLogFiles(4);
  Close();
 }
+#endif  // !ROCKSDB_LITE

 TEST_P(ColumnFamilyTest, CrashAfterFlush) {
  std::unique_ptr<FaultInjectionTestEnv> fault_env(
@@ -1051,6 +1097,7 @@ TEST_P(ColumnFamilyTest, OpenNonexistentColumnFamily) {
  ASSERT_TRUE(TryOpen({"default", "dne"}).IsInvalidArgument());
 }

+#ifndef ROCKSDB_LITE  // WaitForFlush() is not supported
 // Makes sure that obsolete log files get deleted
 TEST_P(ColumnFamilyTest, DifferentWriteBufferSizes) {
  // disable flushing stale column families
@@ -1158,12 +1205,14 @@ TEST_P(ColumnFamilyTest, DifferentWriteBufferSizes) {
  AssertCountLiveLogFiles(7);
  Close();
 }
+#endif  // !ROCKSDB_LITE

 // 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();
@@ -1183,6 +1232,7 @@ TEST_P(ColumnFamilyTest, DifferentWriteBufferSizes) {
 //   {second}); auto* s3 = dbfull()->GetSnapshot(); ASSERT_TRUE(s3 == nullptr);
 //   Close();
 // }
+// #endif  // !ROCKSDB_LITE

 class TestComparator : public Comparator {
  int Compare(const ROCKSDB_NAMESPACE::Slice& /*a*/,
@@ -1249,6 +1299,7 @@ TEST_P(ColumnFamilyTest, DifferentMergeOperators) {
  Close();
 }

+#ifndef ROCKSDB_LITE  // WaitForFlush() is not supported
 TEST_P(ColumnFamilyTest, DifferentCompactionStyles) {
  Open();
  CreateColumnFamilies({"one", "two"});
@@ -1316,7 +1367,9 @@ TEST_P(ColumnFamilyTest, DifferentCompactionStyles) {

  Close();
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE
 // Sync points not supported in RocksDB Lite

 TEST_P(ColumnFamilyTest, MultipleManualCompactions) {
@@ -1980,7 +2033,9 @@ TEST_P(ColumnFamilyTest, SameCFAutomaticManualCompactions) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  // Tailing iterator not supported
 namespace {
 std::string IterStatus(Iterator* iter) {
  std::string result;
@@ -2038,7 +2093,9 @@ TEST_P(ColumnFamilyTest, NewIteratorsTest) {
    Destroy();
  }
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  // ReadOnlyDB is not supported
 TEST_P(ColumnFamilyTest, ReadOnlyDBTest) {
  Open();
  CreateColumnFamiliesAndReopen({"one", "two", "three", "four"});
@@ -2087,7 +2144,9 @@ TEST_P(ColumnFamilyTest, ReadOnlyDBTest) {
  s = OpenReadOnly({"one", "four"});
  ASSERT_TRUE(!s.ok());
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  //  WaitForFlush() is not supported in lite
 TEST_P(ColumnFamilyTest, DontRollEmptyLogs) {
  Open();
  CreateColumnFamiliesAndReopen({"one", "two", "three", "four"});
@@ -2109,7 +2168,9 @@ TEST_P(ColumnFamilyTest, DontRollEmptyLogs) {
  ASSERT_EQ(static_cast<size_t>(total_new_writable_files), handles_.size() + 1);
  Close();
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  //  WaitForCompaction() is not supported in lite
 TEST_P(ColumnFamilyTest, FlushStaleColumnFamilies) {
  Open();
  CreateColumnFamilies({"one", "two"});
@@ -2156,6 +2217,7 @@ TEST_P(ColumnFamilyTest, FlushStaleColumnFamilies) {
  ASSERT_EQ(0, dbfull()->TEST_total_log_size());
  Close();
 }
+#endif  // !ROCKSDB_LITE

 TEST_P(ColumnFamilyTest, CreateMissingColumnFamilies) {
  Status s = TryOpen({"one", "two"});
@@ -2395,6 +2457,8 @@ TEST_P(ColumnFamilyTest, FlushAndDropRaceCondition) {
  Destroy();
 }

+#ifndef ROCKSDB_LITE
+// skipped as persisting options is not supported in ROCKSDB_LITE
 namespace {
 std::atomic<int> test_stage(0);
 std::atomic<bool> ordered_by_writethread(false);
@@ -2476,6 +2540,7 @@ TEST_P(ColumnFamilyTest, CreateAndDropRace) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
 }
+#endif  // !ROCKSDB_LITE

 TEST_P(ColumnFamilyTest, WriteStallSingleColumnFamily) {
  const uint64_t kBaseRate = 800000u;
@@ -2885,6 +2950,7 @@ TEST_P(ColumnFamilyTest, CreateDropAndDestroy) {
  ASSERT_OK(db_->DestroyColumnFamilyHandle(cfh));
 }

+#ifndef ROCKSDB_LITE
 TEST_P(ColumnFamilyTest, CreateDropAndDestroyWithoutFileDeletion) {
  ColumnFamilyHandle* cfh;
  Open();
@@ -2939,7 +3005,9 @@ TEST_P(ColumnFamilyTest, FlushCloseWALFiles) {
  db_options_.env = env_;
  Close();
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  // WaitForFlush() is not supported
 TEST_P(ColumnFamilyTest, IteratorCloseWALFile1) {
  SpecialEnv env(Env::Default());
  db_options_.env = &env;
@@ -3046,7 +3114,9 @@ TEST_P(ColumnFamilyTest, IteratorCloseWALFile2) {
  db_options_.env = env_;
  Close();
 }
+#endif  // !ROCKSDB_LITE

+#ifndef ROCKSDB_LITE  // TEST functions are not supported in lite
 TEST_P(ColumnFamilyTest, ForwardIteratorCloseWALFile) {
  SpecialEnv env(Env::Default());
  // Allow both of flush and purge job to schedule.
@@ -3122,6 +3192,7 @@ TEST_P(ColumnFamilyTest, ForwardIteratorCloseWALFile) {
  db_options_.env = env_;
  Close();
 }
+#endif  // !ROCKSDB_LITE

 // Disable on windows because SyncWAL requires env->IsSyncThreadSafe()
 // to return true which is not so in unbuffered mode.
@@ -3,6 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).

+#ifndef ROCKSDB_LITE

 #include <mutex>
 #include <string>
@@ -489,3 +490,13 @@ int main(int argc, char** argv) {
  return RUN_ALL_TESTS();
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr,
+          "SKIPPED as DBImpl::CompactFiles is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // !ROCKSDB_LITE
@@ -188,11 +188,6 @@ class ClippingIterator : public InternalIterator {
    return iter_->GetProperty(prop_name, prop);
  }

-  bool IsDeleteRangeSentinelKey() const override {
-    assert(valid_);
-    return iter_->IsDeleteRangeSentinelKey();
-  }
-
 private:
  void UpdateValid() {
    assert(!iter_->Valid() || iter_->status().ok());
@@ -13,7 +13,6 @@
 #include "db/blob/blob_index.h"
 #include "db/blob/prefetch_buffer_collection.h"
 #include "db/snapshot_checker.h"
-#include "db/wide/wide_column_serialization.h"
 #include "logging/logging.h"
 #include "port/likely.h"
 #include "rocksdb/listener.h"
@@ -222,47 +221,39 @@ void CompactionIterator::Next() {

 bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
                                              Slice* skip_until) {
-  if (!compaction_filter_) {
+  // TODO: support compaction filter for wide-column entities
+  if (!compaction_filter_ ||
+      (ikey_.type != kTypeValue && ikey_.type != kTypeBlobIndex)) {
    return true;
  }
-
-  if (ikey_.type != kTypeValue && ikey_.type != kTypeBlobIndex &&
-      ikey_.type != kTypeWideColumnEntity) {
-    return true;
-  }
-
-  CompactionFilter::Decision decision =
-      CompactionFilter::Decision::kUndetermined;
+  bool error = false;
+  // 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,
+  // replace the entry with a deletion marker.
+  CompactionFilter::Decision filter = CompactionFilter::Decision::kUndetermined;
+  compaction_filter_value_.clear();
+  compaction_filter_skip_until_.Clear();
  CompactionFilter::ValueType value_type =
      ikey_.type == kTypeValue ? CompactionFilter::ValueType::kValue
-      : ikey_.type == kTypeBlobIndex
-          ? CompactionFilter::ValueType::kBlobIndex
-          : CompactionFilter::ValueType::kWideColumnEntity;
-
+                               : CompactionFilter::ValueType::kBlobIndex;
  // Hack: pass internal key to BlobIndexCompactionFilter since it needs
  // to get sequence number.
  assert(compaction_filter_);
-  const Slice& filter_key =
-      (ikey_.type != kTypeBlobIndex ||
+  Slice& filter_key =
+      (ikey_.type == kTypeValue ||
       !compaction_filter_->IsStackedBlobDbInternalCompactionFilter())
          ? ikey_.user_key
          : key_;
-
-  compaction_filter_value_.clear();
-  compaction_filter_skip_until_.Clear();
-
-  std::vector<std::pair<std::string, std::string>> new_columns;
-
  {
    StopWatchNano timer(clock_, report_detailed_time_);
-
-    if (ikey_.type == kTypeBlobIndex) {
-      decision = compaction_filter_->FilterBlobByKey(
+    if (kTypeBlobIndex == ikey_.type) {
+      filter = compaction_filter_->FilterBlobByKey(
          level_, filter_key, &compaction_filter_value_,
          compaction_filter_skip_until_.rep());
-      if (decision == CompactionFilter::Decision::kUndetermined &&
+      if (CompactionFilter::Decision::kUndetermined == filter &&
          !compaction_filter_->IsStackedBlobDbInternalCompactionFilter()) {
-        if (!compaction_) {
+        if (compaction_ == nullptr) {
          status_ =
              Status::Corruption("Unexpected blob index outside of compaction");
          validity_info_.Invalidate();
@@ -308,61 +299,33 @@ bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
        value_type = CompactionFilter::ValueType::kValue;
      }
    }
-
-    if (decision == CompactionFilter::Decision::kUndetermined) {
-      const Slice* existing_val = nullptr;
-      const WideColumns* existing_col = nullptr;
-
-      WideColumns existing_columns;
-
-      if (ikey_.type != kTypeWideColumnEntity) {
-        if (!blob_value_.empty()) {
-          existing_val = &blob_value_;
-        } else {
-          existing_val = &value_;
-        }
-      } else {
-        Slice value_copy = value_;
-        const Status s =
-            WideColumnSerialization::Deserialize(value_copy, existing_columns);
-
-        if (!s.ok()) {
-          status_ = s;
-          validity_info_.Invalidate();
-          return false;
-        }
-
-        existing_col = &existing_columns;
-      }
-
-      decision = compaction_filter_->FilterV3(
-          level_, filter_key, value_type, existing_val, existing_col,
-          &compaction_filter_value_, &new_columns,
+    if (CompactionFilter::Decision::kUndetermined == filter) {
+      filter = compaction_filter_->FilterV2(
+          level_, filter_key, value_type,
+          blob_value_.empty() ? value_ : blob_value_, &compaction_filter_value_,
          compaction_filter_skip_until_.rep());
    }
-
    iter_stats_.total_filter_time +=
        env_ != nullptr && report_detailed_time_ ? timer.ElapsedNanos() : 0;
  }

-  if (decision == CompactionFilter::Decision::kUndetermined) {
-    // Should not reach here, since FilterV2/FilterV3 should never return
-    // kUndetermined.
-    status_ = Status::NotSupported(
-        "FilterV2/FilterV3 should never return kUndetermined");
+  if (CompactionFilter::Decision::kUndetermined == filter) {
+    // Should not reach here, since FilterV2 should never return kUndetermined.
+    status_ =
+        Status::NotSupported("FilterV2() should never return kUndetermined");
    validity_info_.Invalidate();
    return false;
  }

-  if (decision == CompactionFilter::Decision::kRemoveAndSkipUntil &&
+  if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil &&
      cmp_->Compare(*compaction_filter_skip_until_.rep(), ikey_.user_key) <=
          0) {
    // Can't skip to a key smaller than the current one.
-    // Keep the key as per FilterV2/FilterV3 documentation.
-    decision = CompactionFilter::Decision::kKeep;
+    // Keep the key as per FilterV2 documentation.
+    filter = CompactionFilter::Decision::kKeep;
  }

-  if (decision == CompactionFilter::Decision::kRemove) {
+  if (filter == CompactionFilter::Decision::kRemove) {
    // convert the current key to a delete; key_ is pointing into
    // current_key_ at this point, so updating current_key_ updates key()
    ikey_.type = kTypeDeletion;
@@ -370,7 +333,7 @@ bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
    // no value associated with delete
    value_.clear();
    iter_stats_.num_record_drop_user++;
-  } else if (decision == CompactionFilter::Decision::kPurge) {
+  } else if (filter == CompactionFilter::Decision::kPurge) {
    // convert the current key to a single delete; key_ is pointing into
    // current_key_ at this point, so updating current_key_ updates key()
    ikey_.type = kTypeSingleDeletion;
@@ -378,19 +341,19 @@ bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
    // no value associated with single delete
    value_.clear();
    iter_stats_.num_record_drop_user++;
-  } else if (decision == CompactionFilter::Decision::kChangeValue) {
-    if (ikey_.type != kTypeValue) {
+  } else if (filter == CompactionFilter::Decision::kChangeValue) {
+    if (ikey_.type == kTypeBlobIndex) {
+      // value transfer from blob file to inlined data
      ikey_.type = kTypeValue;
-      current_key_.UpdateInternalKey(ikey_.sequence, kTypeValue);
+      current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
    }
-
    value_ = compaction_filter_value_;
-  } else if (decision == CompactionFilter::Decision::kRemoveAndSkipUntil) {
+  } else if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil) {
    *need_skip = true;
    compaction_filter_skip_until_.ConvertFromUserKey(kMaxSequenceNumber,
                                                     kValueTypeForSeek);
    *skip_until = compaction_filter_skip_until_.Encode();
-  } else if (decision == CompactionFilter::Decision::kChangeBlobIndex) {
+  } else if (filter == CompactionFilter::Decision::kChangeBlobIndex) {
    // Only the StackableDB-based BlobDB impl's compaction filter should return
    // kChangeBlobIndex. Decision about rewriting blob and changing blob index
    // in the integrated BlobDB impl is made in subsequent call to
@@ -402,56 +365,23 @@ bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
      validity_info_.Invalidate();
      return false;
    }
-
-    if (ikey_.type != kTypeBlobIndex) {
+    if (ikey_.type == kTypeValue) {
+      // value transfer from inlined data to blob file
      ikey_.type = kTypeBlobIndex;
-      current_key_.UpdateInternalKey(ikey_.sequence, kTypeBlobIndex);
+      current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
    }
-
    value_ = compaction_filter_value_;
-  } else if (decision == CompactionFilter::Decision::kIOError) {
+  } else if (filter == CompactionFilter::Decision::kIOError) {
    if (!compaction_filter_->IsStackedBlobDbInternalCompactionFilter()) {
      status_ = Status::NotSupported(
          "CompactionFilter for integrated BlobDB should not return kIOError");
      validity_info_.Invalidate();
      return false;
    }
-
    status_ = Status::IOError("Failed to access blob during compaction filter");
-    validity_info_.Invalidate();
-    return false;
-  } else if (decision == CompactionFilter::Decision::kChangeWideColumnEntity) {
-    WideColumns sorted_columns;
-
-    sorted_columns.reserve(new_columns.size());
-    for (const auto& column : new_columns) {
-      sorted_columns.emplace_back(column.first, column.second);
-    }
-
-    std::sort(sorted_columns.begin(), sorted_columns.end(),
-              [](const WideColumn& lhs, const WideColumn& rhs) {
-                return lhs.name().compare(rhs.name()) < 0;
-              });
-
-    {
-      const Status s = WideColumnSerialization::Serialize(
-          sorted_columns, compaction_filter_value_);
-      if (!s.ok()) {
-        status_ = s;
-        validity_info_.Invalidate();
-        return false;
-      }
-    }
-
-    if (ikey_.type != kTypeWideColumnEntity) {
-      ikey_.type = kTypeWideColumnEntity;
-      current_key_.UpdateInternalKey(ikey_.sequence, kTypeWideColumnEntity);
-    }
-
-    value_ = compaction_filter_value_;
+    error = true;
  }
-
-  return true;
+  return !error;
 }

 void CompactionIterator::NextFromInput() {
@@ -464,7 +394,6 @@ void CompactionIterator::NextFromInput() {
    value_ = input_.value();
    blob_value_.Reset();
    iter_stats_.num_input_records++;
-    is_range_del_ = input_.IsDeleteRangeSentinelKey();

    Status pik_status = ParseInternalKey(key_, &ikey_, allow_data_in_errors_);
    if (!pik_status.ok()) {
@@ -484,10 +413,7 @@ void CompactionIterator::NextFromInput() {
      break;
    }
    TEST_SYNC_POINT_CALLBACK("CompactionIterator:ProcessKV", &ikey_);
-    if (is_range_del_) {
-      validity_info_.SetValid(kRangeDeletion);
-      break;
-    }
+
    // Update input statistics
    if (ikey_.type == kTypeDeletion || ikey_.type == kTypeSingleDeletion ||
        ikey_.type == kTypeDeletionWithTimestamp) {
@@ -709,14 +635,6 @@ void CompactionIterator::NextFromInput() {

      ParsedInternalKey next_ikey;
      AdvanceInputIter();
-      while (input_.Valid() && input_.IsDeleteRangeSentinelKey() &&
-             ParseInternalKey(input_.key(), &next_ikey, allow_data_in_errors_)
-                 .ok() &&
-             cmp_->EqualWithoutTimestamp(ikey_.user_key, next_ikey.user_key)) {
-        // skip range tombstone start keys with the same user key
-        // since they are not "real" point keys.
-        AdvanceInputIter();
-      }

      // Check whether the next key exists, is not corrupt, and is the same key
      // as the single delete.
@@ -724,7 +642,6 @@ void CompactionIterator::NextFromInput() {
          ParseInternalKey(input_.key(), &next_ikey, allow_data_in_errors_)
              .ok() &&
          cmp_->EqualWithoutTimestamp(ikey_.user_key, next_ikey.user_key)) {
-        assert(!input_.IsDeleteRangeSentinelKey());
 #ifndef NDEBUG
        const Compaction* c =
            compaction_ ? compaction_->real_compaction() : nullptr;
@@ -949,14 +866,12 @@ void CompactionIterator::NextFromInput() {
      // Note that a deletion marker of type kTypeDeletionWithTimestamp will be
      // considered to have a different user key unless the timestamp is older
      // than *full_history_ts_low_.
-      //
-      // Range tombstone start keys are skipped as they are not "real" keys.
      while (!IsPausingManualCompaction() && !IsShuttingDown() &&
             input_.Valid() &&
             (ParseInternalKey(input_.key(), &next_ikey, allow_data_in_errors_)
                  .ok()) &&
             cmp_->EqualWithoutTimestamp(ikey_.user_key, next_ikey.user_key) &&
-             (prev_snapshot == 0 || input_.IsDeleteRangeSentinelKey() ||
+             (prev_snapshot == 0 ||
              DefinitelyNotInSnapshot(next_ikey.sequence, prev_snapshot))) {
        AdvanceInputIter();
      }
@@ -1250,12 +1165,10 @@ void CompactionIterator::DecideOutputLevel() {

 void CompactionIterator::PrepareOutput() {
  if (Valid()) {
-    if (LIKELY(!is_range_del_)) {
-      if (ikey_.type == kTypeValue) {
-        ExtractLargeValueIfNeeded();
-      } else if (ikey_.type == kTypeBlobIndex) {
-        GarbageCollectBlobIfNeeded();
-      }
+    if (ikey_.type == kTypeValue) {
+      ExtractLargeValueIfNeeded();
+    } else if (ikey_.type == kTypeBlobIndex) {
+      GarbageCollectBlobIfNeeded();
    }

    if (compaction_ != nullptr && compaction_->SupportsPerKeyPlacement()) {
@@ -1278,7 +1191,7 @@ void CompactionIterator::PrepareOutput() {
        DefinitelyInSnapshot(ikey_.sequence, earliest_snapshot_) &&
        ikey_.type != kTypeMerge && current_key_committed_ &&
        !output_to_penultimate_level_ &&
-        ikey_.sequence < preserve_time_min_seqno_ && !is_range_del_) {
+        ikey_.sequence < preserve_time_min_seqno_) {
      if (ikey_.type == kTypeDeletion ||
          (ikey_.type == kTypeSingleDeletion && timestamp_size_ == 0)) {
        ROCKS_LOG_FATAL(
@@ -63,10 +63,6 @@ class SequenceIterWrapper : public InternalIterator {
  void SeekToLast() override { assert(false); }

  uint64_t num_itered() const { return num_itered_; }
-  bool IsDeleteRangeSentinelKey() const override {
-    assert(Valid());
-    return inner_iter_->IsDeleteRangeSentinelKey();
-  }

 private:
  InternalKeyComparator icmp_;
@@ -246,12 +242,7 @@ class CompactionIterator {
  const Status& status() const { return status_; }
  const ParsedInternalKey& ikey() const { return ikey_; }
  inline bool Valid() const { return validity_info_.IsValid(); }
-  const Slice& user_key() const {
-    if (UNLIKELY(is_range_del_)) {
-      return ikey_.user_key;
-    }
-    return current_user_key_;
-  }
+  const Slice& user_key() const { return current_user_key_; }
  const CompactionIterationStats& iter_stats() const { return iter_stats_; }
  uint64_t num_input_entry_scanned() const { return input_.num_itered(); }
  // If the current key should be placed on penultimate level, only valid if
@@ -261,8 +252,6 @@ class CompactionIterator {
  }
  Status InputStatus() const { return input_.status(); }

-  bool IsDeleteRangeSentinelKey() const { return is_range_del_; }
-
 private:
  // Processes the input stream to find the next output
  void NextFromInput();
@@ -396,7 +385,6 @@ class CompactionIterator {
    kKeepSD = 8,
    kKeepDel = 9,
    kNewUserKey = 10,
-    kRangeDeletion = 11,
  };

  struct ValidityInfo {
@@ -505,10 +493,6 @@ class CompactionIterator {
    // This is a best-effort facility, so memory_order_relaxed is sufficient.
    return manual_compaction_canceled_.load(std::memory_order_relaxed);
  }
-
-  // Stores whether the current compaction iterator output
-  // is a range tombstone start key.
-  bool is_range_del_{false};
 };

 inline bool CompactionIterator::DefinitelyInSnapshot(SequenceNumber seq,
@@ -978,6 +978,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {

 void CompactionJob::NotifyOnSubcompactionBegin(
    SubcompactionState* sub_compact) {
+#ifndef ROCKSDB_LITE
  Compaction* c = compact_->compaction;

  if (db_options_.listeners.empty()) {
@@ -1003,10 +1004,14 @@ void CompactionJob::NotifyOnSubcompactionBegin(
  }
  info.status.PermitUncheckedError();

+#else
+  (void)sub_compact;
+#endif  // ROCKSDB_LITE
 }

 void CompactionJob::NotifyOnSubcompactionCompleted(
    SubcompactionState* sub_compact) {
+#ifndef ROCKSDB_LITE

  if (db_options_.listeners.empty()) {
    return;
@@ -1027,12 +1032,16 @@ void CompactionJob::NotifyOnSubcompactionCompleted(
  for (const auto& listener : db_options_.listeners) {
    listener->OnSubcompactionCompleted(info);
  }
+#else
+  (void)sub_compact;
+#endif  // ROCKSDB_LITE
 }

 void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  assert(sub_compact);
  assert(sub_compact->compaction);

+#ifndef ROCKSDB_LITE
  if (db_options_.compaction_service) {
    CompactionServiceJobStatus comp_status =
        ProcessKeyValueCompactionWithCompactionService(sub_compact);
@@ -1043,6 +1052,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
    // fallback to local compaction
    assert(comp_status == CompactionServiceJobStatus::kUseLocal);
  }
+#endif  // !ROCKSDB_LITE

  uint64_t prev_cpu_micros = db_options_.clock->CPUMicros();

@@ -1118,8 +1128,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  IterKey end_ikey;
  Slice start_slice;
  Slice end_slice;
-  Slice start_user_key{};
-  Slice end_user_key{};

  static constexpr char kMaxTs[] =
      "\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff";
@@ -1142,7 +1150,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
                                   &ts_slice);
    }
    start_slice = start_ikey.GetInternalKey();
-    start_user_key = start_ikey.GetUserKey();
  }
  if (end.has_value()) {
    end_ikey.SetInternalKey(end.value(), kMaxSequenceNumber, kValueTypeForSeek);
@@ -1151,7 +1158,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
                                 &ts_slice);
    }
    end_slice = end_ikey.GetInternalKey();
-    end_user_key = end_ikey.GetUserKey();
  }

  std::unique_ptr<InternalIterator> clip;
@@ -1267,15 +1273,11 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
      [this, sub_compact](CompactionOutputs& outputs) {
        return this->OpenCompactionOutputFile(sub_compact, outputs);
      };
-
  const CompactionFileCloseFunc close_file_func =
-      [this, sub_compact, start_user_key, end_user_key](
-          CompactionOutputs& outputs, const Status& status,
-          const Slice& next_table_min_key) {
-        return this->FinishCompactionOutputFile(
-            status, sub_compact, outputs, next_table_min_key,
-            sub_compact->start.has_value() ? &start_user_key : nullptr,
-            sub_compact->end.has_value() ? &end_user_key : nullptr);
+      [this, sub_compact](CompactionOutputs& outputs, const Status& status,
+                          const Slice& next_table_min_key) {
+        return this->FinishCompactionOutputFile(status, sub_compact, outputs,
+                                                next_table_min_key);
      };

  Status status;
@@ -1286,6 +1288,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
  while (status.ok() && !cfd->IsDropped() && c_iter->Valid()) {
    // Invariant: c_iter.status() is guaranteed to be OK if c_iter->Valid()
    // returns true.
+
    assert(!end.has_value() || cfd->user_comparator()->Compare(
                                   c_iter->user_key(), end.value()) < 0);

@@ -1465,8 +1468,7 @@ void CompactionJob::RecordDroppedKeys(

 Status CompactionJob::FinishCompactionOutputFile(
    const Status& input_status, SubcompactionState* sub_compact,
-    CompactionOutputs& outputs, const Slice& next_table_min_key,
-    const Slice* comp_start_user_key, const Slice* comp_end_user_key) {
+    CompactionOutputs& outputs, const Slice& next_table_min_key) {
  AutoThreadOperationStageUpdater stage_updater(
      ThreadStatus::STAGE_COMPACTION_SYNC_FILE);
  assert(sub_compact != nullptr);
@@ -1496,10 +1498,12 @@ Status CompactionJob::FinishCompactionOutputFile(
    // output_to_penultimate_level compaction here, as it's only used to decide
    // if range dels could be dropped.
    if (outputs.HasRangeDel()) {
-      s = outputs.AddRangeDels(comp_start_user_key, comp_end_user_key,
-                               range_del_out_stats, bottommost_level_,
-                               cfd->internal_comparator(), earliest_snapshot,
-                               next_table_min_key, full_history_ts_low_);
+      s = outputs.AddRangeDels(
+          sub_compact->start.has_value() ? &(sub_compact->start.value())
+                                         : nullptr,
+          sub_compact->end.has_value() ? &(sub_compact->end.value()) : nullptr,
+          range_del_out_stats, bottommost_level_, cfd->internal_comparator(),
+          earliest_snapshot, next_table_min_key, full_history_ts_low_);
    }
    RecordDroppedKeys(range_del_out_stats, &sub_compact->compaction_job_stats);
    TEST_SYNC_POINT("CompactionJob::FinishCompactionOutputFile1");
@@ -1611,6 +1615,7 @@ Status CompactionJob::FinishCompactionOutputFile(
      TableFileCreationReason::kCompaction, status_for_listener, file_checksum,
      file_checksum_func_name);

+#ifndef ROCKSDB_LITE
  // Report new file to SstFileManagerImpl
  auto sfm =
      static_cast<SstFileManagerImpl*>(db_options_.sst_file_manager.get());
@@ -1629,6 +1634,7 @@ Status CompactionJob::FinishCompactionOutputFile(
      db_error_handler_->SetBGError(s, BackgroundErrorReason::kCompaction);
    }
  }
+#endif

  outputs.ResetBuilder();
  return s;
@@ -1753,9 +1759,11 @@ Status CompactionJob::OpenCompactionOutputFile(SubcompactionState* sub_compact,
  std::string fname = GetTableFileName(file_number);
  // Fire events.
  ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
+#ifndef ROCKSDB_LITE
  EventHelpers::NotifyTableFileCreationStarted(
      cfd->ioptions()->listeners, dbname_, cfd->GetName(), fname, job_id_,
      TableFileCreationReason::kCompaction);
+#endif  // !ROCKSDB_LITE
  // Make the output file
  std::unique_ptr<FSWritableFile> writable_file;
 #ifndef NDEBUG
@@ -1892,6 +1900,7 @@ void CompactionJob::CleanupCompaction() {
  compact_ = nullptr;
 }

+#ifndef ROCKSDB_LITE
 namespace {
 void CopyPrefix(const Slice& src, size_t prefix_length, std::string* dst) {
  assert(prefix_length > 0);
@@ -1900,6 +1909,7 @@ void CopyPrefix(const Slice& src, size_t prefix_length, std::string* dst) {
 }
 }  // namespace

+#endif  // !ROCKSDB_LITE

 void CompactionJob::UpdateCompactionStats() {
  assert(compact_);
@@ -1946,6 +1956,7 @@ void CompactionJob::UpdateCompactionInputStatsHelper(int* num_files,

 void CompactionJob::UpdateCompactionJobStats(
    const InternalStats::CompactionStats& stats) const {
+#ifndef ROCKSDB_LITE
  compaction_job_stats_->elapsed_micros = stats.micros;

  // input information
@@ -1972,6 +1983,9 @@ void CompactionJob::UpdateCompactionJobStats(
    CopyPrefix(compact_->LargestUserKey(), CompactionJobStats::kMaxPrefixLength,
               &compaction_job_stats_->largest_output_key_prefix);
  }
+#else
+  (void)stats;
+#endif  // !ROCKSDB_LITE
 }

 void CompactionJob::LogCompaction() {
@@ -256,9 +256,7 @@ class CompactionJob {
  Status FinishCompactionOutputFile(const Status& input_status,
                                    SubcompactionState* sub_compact,
                                    CompactionOutputs& outputs,
-                                    const Slice& next_table_min_key,
-                                    const Slice* comp_start_user_key,
-                                    const Slice* comp_end_user_key);
+                                    const Slice& next_table_min_key);
  Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options);
  Status OpenCompactionOutputFile(SubcompactionState* sub_compact,
                                  CompactionOutputs& outputs);
@@ -59,6 +59,7 @@
 #include "utilities/merge_operators.h"

 #if !defined(IOS_CROSS_COMPILE)
+#ifndef ROCKSDB_LITE
 namespace ROCKSDB_NAMESPACE {

 static std::string RandomString(Random* rnd, int len, double ratio) {
@@ -958,6 +959,15 @@ int main(int argc, char** argv) {
  return RUN_ALL_TESTS();
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr, "SKIPPED, not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // !ROCKSDB_LITE

 #else

@@ -3,6 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).

+#ifndef ROCKSDB_LITE

 #include "db/compaction/compaction_job.h"

@@ -2441,3 +2442,13 @@ int main(int argc, char** argv) {
  return RUN_ALL_TESTS();
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr,
+          "SKIPPED as CompactionJobStats is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // ROCKSDB_LITE
@@ -226,15 +226,6 @@ uint64_t CompactionOutputs::GetCurrentKeyGrandparentOverlappedBytes(
 bool CompactionOutputs::ShouldStopBefore(const CompactionIterator& c_iter) {
  assert(c_iter.Valid());
  const Slice& internal_key = c_iter.key();
-#ifndef NDEBUG
-  bool should_stop = false;
-  std::pair<bool*, const Slice> p{&should_stop, internal_key};
-  TEST_SYNC_POINT_CALLBACK(
-      "CompactionOutputs::ShouldStopBefore::manual_decision", (void*)&p);
-  if (should_stop) {
-    return true;
-  }
-#endif  // NDEBUG
  const uint64_t previous_overlapped_bytes = grandparent_overlapped_bytes_;
  const InternalKeyComparator* icmp =
      &compaction_->column_family_data()->internal_comparator();
@@ -356,14 +347,8 @@ Status CompactionOutputs::AddToOutput(
    const CompactionFileOpenFunc& open_file_func,
    const CompactionFileCloseFunc& close_file_func) {
  Status s;
-  bool is_range_del = c_iter.IsDeleteRangeSentinelKey();
-  if (is_range_del && compaction_->bottommost_level()) {
-    // We don't consider range tombstone for bottommost level since:
-    // 1. there is no grandparent and hence no overlap to consider
-    // 2. range tombstone may be dropped at bottommost level.
-    return s;
-  }
  const Slice& key = c_iter.key();
+
  if (ShouldStopBefore(c_iter) && HasBuilder()) {
    s = close_file_func(*this, c_iter.InputStatus(), key);
    if (!s.ok()) {
@@ -373,13 +358,6 @@ Status CompactionOutputs::AddToOutput(
    grandparent_boundary_switched_num_ = 0;
    grandparent_overlapped_bytes_ =
        GetCurrentKeyGrandparentOverlappedBytes(key);
-    if (UNLIKELY(is_range_del)) {
-      // lower bound for this new output file, this is needed as the lower bound
-      // does not come from the smallest point key in this case.
-      range_tombstone_lower_bound_.DecodeFrom(key);
-    } else {
-      range_tombstone_lower_bound_.Clear();
-    }
  }

  // Open output file if necessary
@@ -390,17 +368,6 @@ Status CompactionOutputs::AddToOutput(
    }
  }

-  // c_iter may emit range deletion keys, so update `last_key_for_partitioner_`
-  // here before returning below when `is_range_del` is true
-  if (partitioner_) {
-    last_key_for_partitioner_.assign(c_iter.user_key().data_,
-                                     c_iter.user_key().size_);
-  }
-
-  if (UNLIKELY(is_range_del)) {
-    return s;
-  }
-
  assert(builder_ != nullptr);
  const Slice& value = c_iter.value();
  s = current_output().validator.Add(key, value);
@@ -424,33 +391,28 @@ Status CompactionOutputs::AddToOutput(
  s = current_output().meta.UpdateBoundaries(key, value, ikey.sequence,
                                             ikey.type);

+  if (partitioner_) {
+    last_key_for_partitioner_.assign(c_iter.user_key().data_,
+                                     c_iter.user_key().size_);
+  }
+
  return s;
 }

-namespace {
-void SetMaxSeqAndTs(InternalKey& internal_key, const Slice& user_key,
-                    const size_t ts_sz) {
-  if (ts_sz) {
-    static constexpr char kTsMax[] = "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
-    if (ts_sz <= strlen(kTsMax)) {
-      internal_key = InternalKey(user_key, kMaxSequenceNumber,
-                                 kTypeRangeDeletion, Slice(kTsMax, ts_sz));
-    } else {
-      internal_key =
-          InternalKey(user_key, kMaxSequenceNumber, kTypeRangeDeletion,
-                      std::string(ts_sz, '\xff'));
-    }
-  } else {
-    internal_key.Set(user_key, kMaxSequenceNumber, kTypeRangeDeletion);
-  }
-}
-}  // namespace
-
 Status CompactionOutputs::AddRangeDels(
    const Slice* comp_start_user_key, const Slice* comp_end_user_key,
    CompactionIterationStats& range_del_out_stats, bool bottommost_level,
    const InternalKeyComparator& icmp, SequenceNumber earliest_snapshot,
    const Slice& next_table_min_key, const std::string& full_history_ts_low) {
+  assert(HasRangeDel());
+  FileMetaData& meta = current_output().meta;
+  const Comparator* ucmp = icmp.user_comparator();
+
+  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;
+
  // The following example does not happen since
  // CompactionOutput::ShouldStopBefore() always return false for the first
  // point key. But we should consider removing this dependency. Suppose for the
@@ -462,134 +424,98 @@ Status CompactionOutputs::AddRangeDels(
  // Then meta.smallest will be set to comp_start_user_key@seqno
  // and meta.largest will be set to comp_start_user_key@kMaxSequenceNumber
  // which violates the assumption that meta.smallest should be <= meta.largest.
-  assert(HasRangeDel());
-  FileMetaData& meta = current_output().meta;
-  const Comparator* ucmp = icmp.user_comparator();
-  InternalKey lower_bound_buf, upper_bound_buf;
-  Slice lower_bound_guard, upper_bound_guard;
-  std::string smallest_user_key;
-  const Slice *lower_bound, *upper_bound;
-
-  // We first determine the internal key lower_bound and upper_bound for
-  // this output file. All and only range tombstones that overlap with
-  // [lower_bound, upper_bound] should be added to this file. File
-  // boundaries (meta.smallest/largest) should be updated accordingly when
-  // extended by range tombstones.
  size_t output_size = outputs_.size();
  if (output_size == 1) {
-    // This is the first file in the subcompaction.
-    //
-    // When outputting a range tombstone that spans a subcompaction boundary,
-    // the files on either side of that boundary need to include that
-    // boundary's user key. Otherwise, the spanning range tombstone would lose
-    // coverage.
-    //
-    // To achieve this while preventing files from overlapping in internal key
-    // (an LSM invariant violation), we allow the earlier file to include the
-    // boundary user key up to `kMaxSequenceNumber,kTypeRangeDeletion`. The
-    // later file can begin at the boundary user key at the newest key version
-    // it contains. At this point that version number is unknown since we have
-    // not processed the range tombstones yet, so permit any version. Same story
-    // applies to timestamp, and a non-nullptr `comp_start_user_key` should have
-    // `kMaxTs` here, which similarly permits any timestamp.
-    if (comp_start_user_key) {
-      lower_bound_buf.Set(*comp_start_user_key, kMaxSequenceNumber,
-                          kTypeRangeDeletion);
-      lower_bound_guard = lower_bound_buf.Encode();
-      lower_bound = &lower_bound_guard;
-    } else {
-      lower_bound = nullptr;
-    }
-  } else {
+    // For the first output table, include range tombstones before the min
+    // key but after the subcompaction boundary.
+    lower_bound = comp_start_user_key;
+    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
    // tombstones falling before min key.
-    if (range_tombstone_lower_bound_.size() > 0) {
-      assert(meta.smallest.size() == 0 ||
-             icmp.Compare(range_tombstone_lower_bound_, meta.smallest) < 0);
-      lower_bound_guard = range_tombstone_lower_bound_.Encode();
-    } else {
-      assert(meta.smallest.size() > 0);
-      lower_bound_guard = meta.smallest.Encode();
-    }
+    smallest_user_key = meta.smallest.user_key().ToString(false /*hex*/);
+    lower_bound_guard = Slice(smallest_user_key);
    lower_bound = &lower_bound_guard;
+  } else {
+    lower_bound = nullptr;
  }
-
-  const size_t ts_sz = ucmp->timestamp_size();
-  if (next_table_min_key.empty()) {
-    // Last file of the subcompaction.
-    if (comp_end_user_key) {
-      upper_bound_buf.Set(*comp_end_user_key, kMaxSequenceNumber,
-                          kTypeRangeDeletion);
-      upper_bound_guard = upper_bound_buf.Encode();
-      upper_bound = &upper_bound_guard;
+  if (!next_table_min_key.empty()) {
+    // 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);
+    if (comp_end_user_key != nullptr &&
+        ucmp->CompareWithoutTimestamp(upper_bound_guard, *comp_end_user_key) >=
+            0) {
+      upper_bound = comp_end_user_key;
    } else {
-      upper_bound = nullptr;
+      upper_bound = &upper_bound_guard;
    }
  } else {
-    // There is another file coming whose coverage will begin at
-    // `next_table_min_key`. The current file needs to extend range tombstone
-    // coverage through its own keys (through `meta.largest`) and through user
-    // keys preceding `next_table_min_key`'s user key.
-    ParsedInternalKey next_table_min_key_parsed;
-    ParseInternalKey(next_table_min_key, &next_table_min_key_parsed,
-                     false /* log_err_key */)
-        .PermitUncheckedError();
-    assert(next_table_min_key_parsed.sequence < kMaxSequenceNumber);
-    assert(meta.largest.size() == 0 ||
-           icmp.Compare(meta.largest.Encode(), next_table_min_key) < 0);
-    assert(!lower_bound || icmp.Compare(*lower_bound, next_table_min_key) <= 0);
-    if (meta.largest.size() > 0 &&
-        ucmp->EqualWithoutTimestamp(meta.largest.user_key(),
-                                    next_table_min_key_parsed.user_key)) {
-      // Caution: this assumes meta.largest.Encode() lives longer than
-      // upper_bound, which is only true if meta.largest is never updated.
-      // This just happens to be the case here since meta.largest serves
-      // as the upper_bound.
-      upper_bound_guard = meta.largest.Encode();
-    } else {
-      SetMaxSeqAndTs(upper_bound_buf, next_table_min_key_parsed.user_key,
-                     ts_sz);
-      upper_bound_guard = upper_bound_buf.Encode();
-    }
-    upper_bound = &upper_bound_guard;
+    // This is the last file in the subcompaction, so extend until the
+    // subcompaction ends.
+    upper_bound = comp_end_user_key;
  }
-  if (lower_bound && upper_bound &&
-      icmp.Compare(*lower_bound, *upper_bound) > 0) {
-    assert(meta.smallest.size() == 0 &&
-           ucmp->EqualWithoutTimestamp(ExtractUserKey(*lower_bound),
-                                       ExtractUserKey(*upper_bound)));
-    // This can only happen when lower_bound have the same user key as
-    // next_table_min_key and that there is no point key in the current
-    // compaction output file.
-    return Status::OK();
+  bool has_overlapping_endpoints;
+  if (upper_bound != nullptr && meta.largest.size() > 0) {
+    has_overlapping_endpoints = ucmp->CompareWithoutTimestamp(
+                                    meta.largest.user_key(), *upper_bound) == 0;
+  } 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(comp_end_user_key == nullptr || upper_bound == nullptr ||
-         ucmp->CompareWithoutTimestamp(ExtractUserKey(*upper_bound),
-                                       *comp_end_user_key) <= 0);
-  auto it = range_del_agg_->NewIterator(lower_bound, upper_bound);
+         ucmp->CompareWithoutTimestamp(*upper_bound, *comp_end_user_key) <= 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();
+  }
  Slice last_tombstone_start_user_key{};
-  bool reached_lower_bound = false;
-  for (it->SeekToFirst(); it->Valid(); it->Next()) {
+  for (; it->Valid(); it->Next()) {
    auto tombstone = it->Tombstone();
-    auto kv = tombstone.Serialize();
-    InternalKey tombstone_end = tombstone.SerializeEndKey();
-    // TODO: the underlying iterator should support clamping the bounds.
-    // tombstone_end.Encode is of form user_key@kMaxSeqno
-    // if it is equal to lower_bound, there is no need to include
-    // such range tombstone.
-    if (!reached_lower_bound && lower_bound &&
-        icmp.Compare(tombstone_end.Encode(), *lower_bound) <= 0) {
-      continue;
+    if (upper_bound != nullptr) {
+      int cmp =
+          ucmp->CompareWithoutTimestamp(*upper_bound, tombstone.start_key_);
+      // Tombstones starting after upper_bound only need to be included in
+      // the next table.
+      // If the current SST ends before upper_bound, i.e.,
+      // `has_overlapping_endpoints == false`, we can also skip over range
+      // tombstones that start exactly at upper_bound. Such range
+      // tombstones will be included in the next file and are not relevant
+      // to the point keys or endpoints of the current file.
+      // If the current SST ends at the same user key at upper_bound,
+      // i.e., `has_overlapping_endpoints == true`, AND the tombstone has
+      // the same start key as upper_bound, i.e., cmp == 0, then
+      // the tombstone is relevant only if the tombstone's sequence number
+      // is no larger than this file's largest key's sequence number. This
+      // is because the upper bound to truncate this file's range tombstone
+      // will be meta.largest in this case, and any tombstone that starts after
+      // it will not be relevant.
+      if (cmp < 0) {
+        break;
+      } else if (cmp == 0) {
+        if (!has_overlapping_endpoints ||
+            tombstone.seq_ < GetInternalKeySeqno(meta.largest.Encode())) {
+          break;
+        }
+      }
    }
-    assert(!lower_bound ||
-           icmp.Compare(*lower_bound, tombstone_end.Encode()) <= 0);
-    reached_lower_bound = true;

+    const size_t ts_sz = ucmp->timestamp_size();
    // Garbage collection for range tombstones.
    // If user-defined timestamp is enabled, range tombstones are dropped if
    // they are at bottommost_level, below full_history_ts_low and not visible
@@ -608,93 +534,83 @@ Status CompactionOutputs::AddRangeDels(
      continue;
    }

+    auto kv = tombstone.Serialize();
    assert(lower_bound == nullptr ||
-           ucmp->CompareWithoutTimestamp(ExtractUserKey(*lower_bound),
-                                         kv.second) < 0);
-    InternalKey tombstone_start = kv.first;
-    if (lower_bound &&
-        ucmp->CompareWithoutTimestamp(tombstone_start.user_key(),
-                                      ExtractUserKey(*lower_bound)) < 0) {
-      // This just updates the non-timestamp portion of `tombstone_start`'s user
-      // key. Ideally there would be a simpler API usage
-      ParsedInternalKey tombstone_start_parsed;
-      ParseInternalKey(tombstone_start.Encode(), &tombstone_start_parsed,
-                       false /* log_err_key */)
-          .PermitUncheckedError();
-      // timestamp should be from where sequence number is from, which is from
-      // tombstone in this case
-      std::string ts =
-          tombstone_start_parsed.GetTimestamp(ucmp->timestamp_size())
-              .ToString();
-      tombstone_start_parsed.user_key = ExtractUserKey(*lower_bound);
-      tombstone_start.SetFrom(tombstone_start_parsed, ts);
-    }
-    if (upper_bound != nullptr &&
-        icmp.Compare(*upper_bound, tombstone_start.Encode()) < 0) {
-      break;
-    }
-    // Here we show that *only* range tombstones that overlap with
-    // [lower_bound, upper_bound] are added to the current file, and
-    // sanity checking invariants that should hold:
-    // - [tombstone_start, tombstone_end] overlaps with [lower_bound,
-    // upper_bound]
-    // - meta.smallest <= meta.largest
-    // Corresponding assertions are made, the proof is broken is any of them
-    // fails.
-    // TODO: show that *all* range tombstones that overlap with
-    //  [lower_bound, upper_bound] are added.
-    // TODO: some invariant about boundaries are correctly updated.
-    //
-    // Note that `tombstone_start` is updated in the if condition above, we use
-    // tombstone_start to refer to its initial value, i.e.,
-    // it->Tombstone().first, and use tombstone_start* to refer to its value
-    // after the update.
-    //
-    // To show [lower_bound, upper_bound] overlaps with [tombstone_start,
-    // tombstone_end]:
-    // lower_bound <= upper_bound from the if condition right after all
-    // bounds are initialized. We assume each tombstone fragment has
-    // start_key.user_key < end_key.user_key, so
-    // tombstone_start < tombstone_end by
-    // FragmentedTombstoneIterator::Tombstone(). So these two ranges are both
-    // non-emtpy. The flag `reached_lower_bound` and the if logic before it
-    // ensures lower_bound <= tombstone_end. tombstone_start is only updated
-    // if it has a smaller user_key than lower_bound user_key, so
-    // tombstone_start <= tombstone_start*. The above if condition implies
-    // tombstone_start* <= upper_bound. So we have
-    // tombstone_start <= upper_bound and lower_bound <= tombstone_end
-    // and the two ranges overlap.
-    //
-    // To show meta.smallest <= meta.largest:
-    // From the implementation of UpdateBoundariesForRange(), it suffices to
-    // prove that when it is first called in this function, its parameters
-    // satisfy `start <= end`, where start = max(tombstone_start*, lower_bound)
-    // and end = min(tombstone_end, upper_bound). From the above proof we have
-    // lower_bound <= tombstone_end and lower_bound <= upper_bound. We only need
-    // to show that tombstone_start* <= min(tombstone_end, upper_bound).
-    // Note that tombstone_start*.user_key = max(tombstone_start.user_key,
-    // lower_bound.user_key). Assuming tombstone_end always has
-    // kMaxSequenceNumber and lower_bound.seqno < kMaxSequenceNumber.
-    // Since lower_bound <= tombstone_end and lower_bound.seqno <
-    // tombstone_end.seqno (in absolute number order, not internal key order),
-    // lower_bound.user_key < tombstone_end.user_key.
-    // Since lower_bound.user_key < tombstone_end.user_key and
-    // tombstone_start.user_key < tombstone_end.user_key, tombstone_start* <
-    // tombstone_end. Since tombstone_start* <= upper_bound from the above proof
-    // and tombstone_start* < tombstone_end, tombstone_start* <=
-    // min(tombstone_end, upper_bound), so the two ranges overlap.
-
+           ucmp->CompareWithoutTimestamp(*lower_bound, kv.second) < 0);
    // Range tombstone is not supported by output validator yet.
    builder_->Add(kv.first.Encode(), kv.second);
-    if (lower_bound &&
-        icmp.Compare(tombstone_start.Encode(), *lower_bound) < 0) {
-      tombstone_start.DecodeFrom(*lower_bound);
+    InternalKey tombstone_start = std::move(kv.first);
+    InternalKey smallest_candidate{tombstone_start};
+    if (lower_bound != nullptr &&
+        ucmp->CompareWithoutTimestamp(smallest_candidate.user_key(),
+                                      *lower_bound) <= 0) {
+      // Pretend the smallest key has the same user key as lower_bound
+      // (the max key in the previous table or subcompaction) in order for
+      // files to appear key-space partitioned.
+      if (lower_bound_from_sub_compact) {
+        // When lower_bound is chosen by a subcompaction
+        // (lower_bound_from_sub_compact), 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 (ts_sz) {
+          assert(tombstone.ts_.size() == ts_sz);
+          smallest_candidate = InternalKey(*lower_bound, tombstone.seq_,
+                                           kTypeRangeDeletion, tombstone.ts_);
+        } else {
+          smallest_candidate =
+              InternalKey(*lower_bound, tombstone.seq_, kTypeRangeDeletion);
+        }
+      } else {
+        // 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, 0, kTypeRangeDeletion);
+      }
    }
-    if (upper_bound && icmp.Compare(*upper_bound, tombstone_end.Encode()) < 0) {
-      tombstone_end.DecodeFrom(*upper_bound);
+    InternalKey tombstone_end = tombstone.SerializeEndKey();
+    InternalKey largest_candidate{tombstone_end};
+    if (upper_bound != nullptr &&
+        ucmp->CompareWithoutTimestamp(*upper_bound,
+                                      largest_candidate.user_key()) <= 0) {
+      // Pretend the largest key has the same user key as upper_bound (the
+      // min key in the following table or subcompaction) in order for files
+      // to appear key-space partitioned.
+      //
+      // Choose highest seqnum so this file's largest internal key comes
+      // before the next file's/subcompaction's smallest. The fake seqnum is
+      // OK because the read path's file-picking code only considers the
+      // user key portion.
+      //
+      // Note Seek() also creates InternalKey with (user_key,
+      // kMaxSequenceNumber), but with kTypeDeletion (0x7) instead of
+      // kTypeRangeDeletion (0xF), so the range tombstone comes before the
+      // Seek() key in InternalKey's ordering. So Seek() will look in the
+      // next file for the user key
+      if (ts_sz) {
+        static constexpr char kTsMax[] = "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
+        if (ts_sz <= strlen(kTsMax)) {
+          largest_candidate =
+              InternalKey(*upper_bound, kMaxSequenceNumber, kTypeRangeDeletion,
+                          Slice(kTsMax, ts_sz));
+        } else {
+          largest_candidate =
+              InternalKey(*upper_bound, kMaxSequenceNumber, kTypeRangeDeletion,
+                          std::string(ts_sz, '\xff'));
+        }
+      } else {
+        largest_candidate =
+            InternalKey(*upper_bound, kMaxSequenceNumber, kTypeRangeDeletion);
+      }
    }
-    assert(icmp.Compare(tombstone_start, tombstone_end) <= 0);
-    meta.UpdateBoundariesForRange(tombstone_start, tombstone_end,
+    meta.UpdateBoundariesForRange(smallest_candidate, largest_candidate,
                                  tombstone.seq_, icmp);
    if (!bottommost_level) {
      bool start_user_key_changed =
@@ -702,8 +618,17 @@ Status CompactionOutputs::AddRangeDels(
          ucmp->CompareWithoutTimestamp(last_tombstone_start_user_key,
                                        it->start_key()) < 0;
      last_tombstone_start_user_key = it->start_key();
+      // Range tombstones are truncated at file boundaries
+      if (icmp.Compare(tombstone_start, meta.smallest) < 0) {
+        tombstone_start = meta.smallest;
+      }
+      if (icmp.Compare(tombstone_end, meta.largest) > 0) {
+        tombstone_end = meta.largest;
+      }
+      // this assertion validates invariant (2) in the comment below.
+      assert(icmp.Compare(tombstone_start, tombstone_end) <= 0);
      if (start_user_key_changed) {
-        // If tombstone_start >= tombstone_end, then either no key range is
+        // if tombstone_start >= tombstone_end, then either no key range is
        // covered, or that they have the same user key. If they have the same
        // user key, then the internal key range should only be within this
        // level, and no keys from older levels is covered.
@@ -721,6 +646,138 @@ Status CompactionOutputs::AddRangeDels(
        }
      }
    }
+    // TODO: show invariants that ensure all necessary range tombstones are
+    // added
+    //  and that file boundaries ensure no coverage is lost.
+    // Each range tombstone with internal key range [tombstone_start,
+    // tombstone_end] is being added to the current compaction output file here.
+    // The range tombstone is going to be truncated at range [meta.smallest,
+    // meta.largest] during reading/scanning. We should maintain invariants
+    // (1) meta.smallest <= meta.largest and,
+    // (2) [tombstone_start, tombstone_end] and [meta.smallest, meta.largest]
+    // overlaps, as there is no point adding range tombstone with a range
+    // outside the file's range.
+    // Since `tombstone_end` is always some user_key@kMaxSeqno, it is okay to
+    // use either open or closed range. Using closed range here to make
+    // reasoning easier, and it is more consistent with an ongoing work that
+    // tries to simplify this method.
+    //
+    // There are two cases:
+    // Case 1. Output file has no point key:
+    //   First we show this case only happens when the entire compaction output
+    //   is range tombstone only. This is true if CompactionIterator does not
+    //   emit any point key. Suppose CompactionIterator emits some point key.
+    //   Based on the assumption that CompactionOutputs::ShouldStopBefore()
+    //   always return false for the first point key, the first compaction
+    //   output file always contains a point key. Each new compaction output
+    //   file is created if there is a point key for which ShouldStopBefore()
+    //   returns true, and the point key would be added to the new compaction
+    //   output file. So each new compaction file always contains a point key.
+    //   So Case 1 only happens when CompactionIterator does not emit any
+    //   point key.
+    //
+    //   To show (1) meta.smallest <= meta.largest:
+    //   Since the compaction output is range tombstone only, `lower_bound` and
+    //   `upper_bound` are either null or comp_start/end_user_key respectively.
+    //   According to how UpdateBoundariesForRange() is implemented, it blindly
+    //   updates meta.smallest and meta.largest to smallest_candidate and
+    //   largest_candidate the first time it is called. Subsequently, it
+    //   compares input parameter with meta.smallest and meta.largest and only
+    //   updates them when input is smaller/larger. So we only need to show
+    //   smallest_candidate <= largest_candidate the first time
+    //   UpdateBoundariesForRange() is called. Here we show something stronger
+    //   that smallest_candidate.user_key < largest_candidate.user_key always
+    //   hold for Case 1.
+    //   We assume comp_start_user_key < comp_end_user_key, if provided. We
+    //   assume that tombstone_start < tombstone_end. This assumption is based
+    //   on that each fragment in FragmentedTombstoneList has
+    //   start_key < end_key (user_key) and that
+    //   FragmentedTombstoneIterator::Tombstone() returns the pair
+    //   (start_key@tombstone_seqno with op_type kTypeRangeDeletion, end_key).
+    //   The logic in this loop sets smallest_candidate to
+    //   max(tombstone_start.user_key, comp_start_user_key)@tombstone.seq_ with
+    //   op_type kTypeRangeDeletion, largest_candidate to
+    //   min(tombstone_end.user_key, comp_end_user_key)@kMaxSequenceNumber with
+    //   op_type kTypeRangeDeletion. When a bound is null, there is no
+    //   truncation on that end. To show that smallest_candidate.user_key <
+    //   largest_candidate.user_key, it suffices to show
+    //   tombstone_start.user_key < comp_end_user_key (if not null) AND
+    //   comp_start_user_key (if not null) < tombstone_end.user_key.
+    //   Since the file has no point key, `has_overlapping_endpoints` is false.
+    //   In the first sanity check of this for-loop, we compare
+    //   tombstone_start.user_key against upper_bound = comp_end_user_key,
+    //   and only proceed if tombstone_start.user_key < comp_end_user_key.
+    //   We assume FragmentedTombstoneIterator::Seek(k) lands
+    //   on a tombstone with end_key > k. So the call it->Seek(*lower_bound)
+    //   above implies compact_start_user_key < tombstone_end.user_key.
+    //
+    //   To show (2) [tombstone_start, tombstone_end] and [meta.smallest,
+    //   meta.largest] overlaps (after the call to UpdateBoundariesForRange()):
+    //   In the proof for (1) we have shown that
+    //   smallest_candidate <= largest_candidate. Since tombstone_start <=
+    //   smallest_candidate <= largest_candidate <= tombstone_end, for (2) to
+    //   hold, it suffices to show that [smallest_candidate, largest_candidate]
+    //   overlaps with [meta.smallest, meta.largest]. too.
+    //   Given meta.smallest <= meta.largest shown above, we need to show
+    //   that it is impossible to have largest_candidate < meta.smallest or
+    //   meta.largest < smallest_candidate. If the above
+    //   meta.UpdateBoundariesForRange(smallest_candidate, largest_candidate)
+    //   updates meta.largest or meta.smallest, then the two ranges overlap.
+    //   So we assume meta.UpdateBoundariesForRange(smallest_candidate,
+    //   largest_candidate) did not update meta.smallest nor meta.largest, which
+    //   means meta.smallest < smallest_candidate and largest_candidate <
+    //   meta.largest.
+    //
+    // Case 2. Output file has >= 1 point key. This means meta.smallest and
+    // meta.largest are not empty when AddRangeDels() is called.
+    //   To show (1) meta.smallest <= meta.largest:
+    //   Assume meta.smallest <= meta.largest when AddRangeDels() is called,
+    //   this follow from how UpdateBoundariesForRange() is implemented where it
+    //   takes min or max to update meta.smallest or meta.largest.
+    //
+    //   To show (2) [tombstone_start, tombstone_end] and [meta.smallest,
+    //   meta.largest] overlaps (after the call to UpdateBoundariesForRange()):
+    //   When smallest_candidate <= largest_candidate, the proof in Case 1
+    //   applies, so we only need to show (2) holds when smallest_candidate >
+    //   largest_candidate. When both bounds are either null or from
+    //   subcompaction boundary, the proof in Case 1 applies, so we only need to
+    //   show (2) holds when at least one bound is from a point key (either
+    //   meta.smallest for lower bound or next_table_min_key for upper bound).
+    //
+    //   Suppose lower bound is meta.smallest.user_key. The call
+    //   it->Seek(*lower_bound) implies tombstone_end.user_key >
+    //   meta.smallest.user_key. We have smallest_candidate.user_key =
+    //   max(tombstone_start.user_key, meta.smallest.user_key). For
+    //   smallest_candidate to be > largest_candidate, we need
+    //   largest_candidate.user_key = upper_bound = smallest_candidate.user_key,
+    //   where tombstone_end is truncated to largest_candidate.
+    //   Subcase 1:
+    //   Suppose largest_candidate.user_key = comp_end_user_key (there is no
+    //   next point key). Subcompaction ensures any point key from this
+    //   subcompaction has a user_key < comp_end_user_key, so 1)
+    //   meta.smallest.user_key < comp_end_user_key, 2)
+    //   `has_overlapping_endpoints` is false, and the first if condition in
+    //   this for-loop ensures tombstone_start.user_key < comp_end_user_key. So
+    //   smallest_candidate.user_key < largest_candidate.user_key. This case
+    //   cannot happen when smallest > largest_candidate.
+    //   Subcase 2:
+    //   Suppose largest_candidate.user_key = next_table_min_key.user_key.
+    //   The first if condition in this for-loop together with
+    //   smallest_candidate.user_key = next_table_min_key.user_key =
+    //   upper_bound implies `has_overlapping_endpoints` is true (so meta
+    //   largest.user_key = upper_bound) and
+    //   tombstone.seq_ < meta.largest.seqno. So
+    //   tombstone_start < meta.largest < tombstone_end.
+    //
+    //   Suppose lower bound is comp_start_user_key and upper_bound is
+    //   next_table_min_key. The call it->Seek(*lower_bound) implies we have
+    //   tombstone_end_key.user_key > comp_start_user_key. So
+    //   tombstone_end_key.user_key > smallest_candidate.user_key. For
+    //   smallest_candidate to be > largest_candidate, we need
+    //   tombstone_start.user_key = largest_candidate.user_key = upper_bound =
+    //   next_table_min_key.user_key. This means `has_overlapping_endpoints` is
+    //   true (so meta.largest.user_key = upper_bound) and tombstone.seq_ <
+    //   meta.largest.seqno. So tombstone_start < meta.largest < tombstone_end.
  }
  return Status::OK();
 }
@@ -167,15 +167,9 @@ class CompactionOutputs {
    current_output_file_size_ = 0;
  }

-  // Add range deletions from the range_del_agg_ to the current output file.
-  // Input parameters, `range_tombstone_lower_bound_` and current output's
-  // metadata determine the bounds on range deletions to add. Updates output
-  // file metadata boundary if extended by range tombstones.
-  //
+  // Add range-dels from the aggregator to the current output file
  // @param comp_start_user_key and comp_end_user_key include timestamp if
-  // user-defined timestamp is enabled. Their timestamp should be max timestamp.
-  // @param next_table_min_key internal key lower bound for the next compaction
-  // output.
+  // user-defined timestamp is enabled.
  // @param full_history_ts_low used for range tombstone garbage collection.
  Status AddRangeDels(const Slice* comp_start_user_key,
                      const Slice* comp_end_user_key,
@@ -320,7 +314,6 @@ class CompactionOutputs {
  std::unique_ptr<SstPartitioner> partitioner_;

  // A flag determines if this subcompaction has been split by the cursor
-  // for RoundRobin compaction
  bool is_split_ = false;

  // We also maintain the output split key for each subcompaction to avoid
@@ -352,10 +345,6 @@ class CompactionOutputs {
  // for the current output file, how many file boundaries has it crossed,
  // basically number of files overlapped * 2
  size_t grandparent_boundary_switched_num_ = 0;
-
-  // The smallest key of the current output file, this is set when current
-  // output file's smallest key is a range tombstone start key.
-  InternalKey range_tombstone_lower_bound_;
 };

 // helper struct to concatenate the last level and penultimate level outputs
@@ -877,6 +877,7 @@ Compaction* CompactionPicker::CompactRange(
  return compaction;
 }

+#ifndef ROCKSDB_LITE
 namespace {
 // Test whether two files have overlapping key-ranges.
 bool HaveOverlappingKeyRanges(const Comparator* c, const SstFileMetaData& a,
@@ -1115,6 +1116,7 @@ Status CompactionPicker::SanitizeCompactionInputFiles(

  return Status::OK();
 }
+#endif  // !ROCKSDB_LITE

 void CompactionPicker::RegisterCompaction(Compaction* c) {
  if (c == nullptr) {
@@ -93,9 +93,11 @@ class CompactionPicker {
 // into a valid one by adding more files, the function will return a
 // non-ok status with specific reason.
 //
+#ifndef ROCKSDB_LITE
  Status SanitizeCompactionInputFiles(std::unordered_set<uint64_t>* input_files,
                                      const ColumnFamilyMetaData& cf_meta,
                                      const int output_level) const;
+#endif  // ROCKSDB_LITE

  // Free up the files that participated in a compaction
  //
@@ -227,9 +229,11 @@ class CompactionPicker {

 // A helper function to SanitizeCompactionInputFiles() that
 // sanitizes "input_files" by adding necessary files.
+#ifndef ROCKSDB_LITE
  virtual Status SanitizeCompactionInputFilesForAllLevels(
      std::unordered_set<uint64_t>* input_files,
      const ColumnFamilyMetaData& cf_meta, const int output_level) const;
+#endif  // ROCKSDB_LITE

  // Keeps track of all compactions that are running on Level0.
  // Protected by DB mutex
@@ -242,6 +246,7 @@ class CompactionPicker {
  const InternalKeyComparator* const icmp_;
 };

+#ifndef ROCKSDB_LITE
 // A dummy compaction that never triggers any automatic
 // compaction.
 class NullCompactionPicker : public CompactionPicker {
@@ -282,6 +287,7 @@ class NullCompactionPicker : public CompactionPicker {
    return false;
  }
 };
+#endif  // !ROCKSDB_LITE

 // Attempts to find an intra L0 compaction conforming to the given parameters.
 //
@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

 #include "db/compaction/compaction_picker_fifo.h"
+#ifndef ROCKSDB_LITE

 #include <cinttypes>
 #include <string>
@@ -442,3 +443,4 @@ Compaction* FIFOCompactionPicker::CompactRange(
 }

 }  // namespace ROCKSDB_NAMESPACE
+#endif  // !ROCKSDB_LITE
@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

 #pragma once
+#ifndef ROCKSDB_LITE

 #include "db/compaction/compaction_picker.h"

@@ -59,3 +60,4 @@ class FIFOCompactionPicker : public CompactionPicker {
                                   LogBuffer* log_buffer);
 };
 }  // namespace ROCKSDB_NAMESPACE
+#endif  // !ROCKSDB_LITE
@@ -482,6 +482,8 @@ TEST_F(CompactionPickerTest, LevelTriggerDynamic4) {
  ASSERT_EQ(num_levels - 1, compaction->output_level());
 }

+// Universal and FIFO Compactions are not supported in ROCKSDB_LITE
+#ifndef ROCKSDB_LITE
 TEST_F(CompactionPickerTest, NeedsCompactionUniversal) {
  NewVersionStorage(1, kCompactionStyleUniversal);
  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
@@ -1243,6 +1245,7 @@ TEST_F(CompactionPickerTest, FIFOToWarmWithHotBetweenWarms) {
  ASSERT_EQ(2U, compaction->input(0, 0)->fd.GetNumber());
 }

+#endif  // ROCKSDB_LITE

 TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
  NewVersionStorage(6, kCompactionStyleLevel);
@@ -2870,6 +2873,7 @@ TEST_F(CompactionPickerTest, IntraL0MaxCompactionBytesHit) {
  ASSERT_EQ(0, compaction->output_level());
 }

+#ifndef ROCKSDB_LITE
 TEST_F(CompactionPickerTest, UniversalMarkedCompactionFullOverlap) {
  const uint64_t kFileSize = 100000;

@@ -3978,6 +3982,7 @@ TEST_P(PerKeyPlacementCompactionPickerTest,
 INSTANTIATE_TEST_CASE_P(PerKeyPlacementCompactionPickerTest,
                        PerKeyPlacementCompactionPickerTest, ::testing::Bool());

+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE

@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

 #include "db/compaction/compaction_picker_universal.h"
+#ifndef ROCKSDB_LITE

 #include <cinttypes>
 #include <limits>
@@ -1450,3 +1451,4 @@ uint64_t UniversalCompactionBuilder::GetMaxOverlappingBytes() const {
 }
 }  // namespace ROCKSDB_NAMESPACE

+#endif  // !ROCKSDB_LITE
@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

 #pragma once
+#ifndef ROCKSDB_LITE

 #include "db/compaction/compaction_picker.h"

@@ -28,3 +29,4 @@ class UniversalCompactionPicker : public CompactionPicker {
      const VersionStorageInfo* vstorage) const override;
 };
 }  // namespace ROCKSDB_NAMESPACE
+#endif  // !ROCKSDB_LITE
@@ -16,6 +16,7 @@
 #include "options/options_helper.h"
 #include "rocksdb/utilities/options_type.h"

+#ifndef ROCKSDB_LITE
 namespace ROCKSDB_NAMESPACE {
 class SubcompactionState;

@@ -831,3 +832,4 @@ bool CompactionServiceInput::TEST_Equals(CompactionServiceInput* other,
 #endif  // NDEBUG
 }  // namespace ROCKSDB_NAMESPACE

+#endif  // !ROCKSDB_LITE
@@ -3,6 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).

+#ifndef ROCKSDB_LITE

 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
@@ -953,3 +954,13 @@ int main(int argc, char** argv) {
  return RUN_ALL_TESTS();
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr,
+          "SKIPPED as CompactionService is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // ROCKSDB_LITE
@@ -15,9 +15,11 @@
 namespace ROCKSDB_NAMESPACE {
 static std::unordered_map<std::string, OptionTypeInfo>
    sst_fixed_prefix_type_info = {
+#ifndef ROCKSDB_LITE
        {"length",
         {0, OptionType::kSizeT, OptionVerificationType::kNormal,
          OptionTypeFlags::kNone}},
+#endif  // ROCKSDB_LITE
 };

 SstPartitionerFixedPrefixFactory::SstPartitionerFixedPrefixFactory(size_t len)
@@ -56,6 +58,7 @@ std::shared_ptr<SstPartitionerFactory> NewSstPartitionerFixedPrefixFactory(
  return std::make_shared<SstPartitionerFixedPrefixFactory>(prefix_len);
 }

+#ifndef ROCKSDB_LITE
 namespace {
 static int RegisterSstPartitionerFactories(ObjectLibrary& library,
                                           const std::string& /*arg*/) {
@@ -70,14 +73,18 @@ static int RegisterSstPartitionerFactories(ObjectLibrary& library,
  return 1;
 }
 }  // namespace
+#endif  // ROCKSDB_LITE

 Status SstPartitionerFactory::CreateFromString(
    const ConfigOptions& options, const std::string& value,
    std::shared_ptr<SstPartitionerFactory>* result) {
+#ifndef ROCKSDB_LITE
  static std::once_flag once;
  std::call_once(once, [&]() {
    RegisterSstPartitionerFactories(*(ObjectLibrary::Default().get()), "");
  });
-  return LoadSharedObject<SstPartitionerFactory>(options, value, result);
+#endif  // ROCKSDB_LITE
+  return LoadSharedObject<SstPartitionerFactory>(options, value, nullptr,
+                                                 result);
 }
 }  // namespace ROCKSDB_NAMESPACE
@@ -84,11 +84,6 @@ class SubcompactionState {
  // Assign range dels aggregator, for each range_del, it can only be assigned
  // to one output level, for per_key_placement, it's going to be the
  // penultimate level.
-  // TODO: This does not work for per_key_placement + user-defined timestamp +
-  //  DeleteRange() combo. If user-defined timestamp is enabled,
-  //  it is possible for a range tombstone to belong to bottommost level (
-  //  seqno < earliest snapshot) without being dropped (garbage collection
-  //  for user-defined timestamp).
  void AssignRangeDelAggregator(
      std::unique_ptr<CompactionRangeDelAggregator>&& range_del_agg) {
    if (compaction->SupportsPerKeyPlacement()) {
@@ -104,6 +99,7 @@ class SubcompactionState {
    penultimate_level_outputs_.RemoveLastEmptyOutput();
  }

+#ifndef ROCKSDB_LITE
  void BuildSubcompactionJobInfo(
      SubcompactionJobInfo& subcompaction_job_info) const {
    const Compaction* c = compaction;
@@ -117,6 +113,7 @@ class SubcompactionState {
    subcompaction_job_info.output_level = c->output_level();
    subcompaction_job_info.stats = compaction_job_stats;
  }
+#endif  // !ROCKSDB_LITE

  SubcompactionState() = delete;
  SubcompactionState(const SubcompactionState&) = delete;
@@ -17,6 +17,7 @@

 namespace ROCKSDB_NAMESPACE {

+#if !defined(ROCKSDB_LITE)

 class TieredCompactionTest : public DBTestBase,
                             public testing::WithParamInterface<bool> {
@@ -2138,11 +2139,18 @@ TEST_F(PrecludeLastLevelTest, RangeDelsCauseFileEndpointsToOverlap) {
  Close();
 }

+#endif  // !defined(ROCKSDB_LITE)

 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
+#if !defined(ROCKSDB_LITE)
  ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
  ::testing::InitGoogleTest(&argc, argv);
  return RUN_ALL_TESTS();
+#else
+  (void)argc;
+  (void)argv;
+  return 0;
+#endif
 }
@@ -4,6 +4,7 @@
 //  (found in the LICENSE.Apache file in the root directory).
 //

+#ifndef ROCKSDB_LITE

 #include "rocksdb/convenience.h"

@@ -77,3 +78,4 @@ Status VerifySstFileChecksum(const Options& options,

 }  // namespace ROCKSDB_NAMESPACE

+#endif  // ROCKSDB_LITE
@@ -8,6 +8,7 @@
 // found in the LICENSE file. See the AUTHORS file for names of contributors.

 #include "rocksdb/options.h"
+#ifndef ROCKSDB_LITE

 #include <fcntl.h>
 #include <sys/stat.h>
@@ -42,36 +43,33 @@ namespace ROCKSDB_NAMESPACE {
 static constexpr int kValueSize = 1000;
 namespace {
 // A wrapper that allows injection of errors.
-class ErrorFS : public FileSystemWrapper {
+class ErrorEnv : public EnvWrapper {
 public:
  bool writable_file_error_;
  int num_writable_file_errors_;

-  explicit ErrorFS(const std::shared_ptr<FileSystem>& _target)
-      : FileSystemWrapper(_target),
+  explicit ErrorEnv(Env* _target)
+      : EnvWrapper(_target),
        writable_file_error_(false),
        num_writable_file_errors_(0) {}
  const char* Name() const override { return "ErrorEnv"; }

-  virtual IOStatus NewWritableFile(const std::string& fname,
-                                   const FileOptions& opts,
-                                   std::unique_ptr<FSWritableFile>* result,
-                                   IODebugContext* dbg) override {
+  virtual Status NewWritableFile(const std::string& fname,
+                                 std::unique_ptr<WritableFile>* result,
+                                 const EnvOptions& soptions) override {
    result->reset();
    if (writable_file_error_) {
      ++num_writable_file_errors_;
-      return IOStatus::IOError(fname, "fake error");
+      return Status::IOError(fname, "fake error");
    }
-    return target()->NewWritableFile(fname, opts, result, dbg);
+    return target()->NewWritableFile(fname, result, soptions);
  }
 };
 }  // anonymous namespace
 class CorruptionTest : public testing::Test {
 public:
  std::shared_ptr<Env> env_guard_;
-  std::shared_ptr<ErrorFS> fs_;
-  std::unique_ptr<Env> env_;
-  Env* base_env_;
+  ErrorEnv* env_;
  std::string dbname_;
  std::shared_ptr<Cache> tiny_cache_;
  Options options_;
@@ -82,15 +80,14 @@ class CorruptionTest : public testing::Test {
    // set it to 0), test SequenceNumberRecovery will fail, likely because of a
    // bug in recovery code. Keep it 4 for now to make the test passes.
    tiny_cache_ = NewLRUCache(100, 4);
-    base_env_ = Env::Default();
+    Env* base_env = Env::Default();
    EXPECT_OK(
-        test::CreateEnvFromSystem(ConfigOptions(), &base_env_, &env_guard_));
-    EXPECT_NE(base_env_, nullptr);
-    fs_.reset(new ErrorFS(base_env_->GetFileSystem()));
-    env_ = NewCompositeEnv(fs_);
+        test::CreateEnvFromSystem(ConfigOptions(), &base_env, &env_guard_));
+    EXPECT_NE(base_env, nullptr);
+    env_ = new ErrorEnv(base_env);
    options_.wal_recovery_mode = WALRecoveryMode::kTolerateCorruptedTailRecords;
-    options_.env = env_.get();
-    dbname_ = test::PerThreadDBPath(env_.get(), "corruption_test");
+    options_.env = env_;
+    dbname_ = test::PerThreadDBPath(env_, "corruption_test");
    Status s = DestroyDB(dbname_, options_);
    EXPECT_OK(s);

@@ -113,9 +110,10 @@ class CorruptionTest : public testing::Test {
      fprintf(stdout, "db is still at %s\n", dbname_.c_str());
    } else {
      Options opts;
-      opts.env = base_env_;
+      opts.env = env_->target();
      EXPECT_OK(DestroyDB(dbname_, opts));
    }
+    delete env_;
  }

  void CloseDb() {
@@ -130,7 +128,7 @@ class CorruptionTest : public testing::Test {
    if (opt.env == Options().env) {
      // If env is not overridden, replace it with ErrorEnv.
      // Otherwise, the test already uses a non-default Env.
-      opt.env = env_.get();
+      opt.env = env_;
    }
    opt.arena_block_size = 4096;
    BlockBasedTableOptions table_options;
@@ -167,10 +165,6 @@ class CorruptionTest : public testing::Test {
  void Build(int n, int flush_every = 0) { Build(n, 0, flush_every); }

  void Check(int min_expected, int max_expected) {
-    Check(min_expected, max_expected, ReadOptions(false, true));
-  }
-
-  void Check(int min_expected, int max_expected, ReadOptions read_options) {
    uint64_t next_expected = 0;
    uint64_t missed = 0;
    int bad_keys = 0;
@@ -182,7 +176,7 @@ class CorruptionTest : public testing::Test {
    // Instead, we want the reads to be successful and this test
    // will detect whether the appropriate corruptions have
    // occurred.
-    Iterator* iter = db_->NewIterator(read_options);
+    Iterator* iter = db_->NewIterator(ReadOptions(false, true));
    for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
      ASSERT_OK(iter->status());
      uint64_t key;
@@ -229,7 +223,7 @@ class CorruptionTest : public testing::Test {
    }
    ASSERT_TRUE(!fname.empty()) << filetype;

-    ASSERT_OK(test::CorruptFile(env_.get(), fname, offset, bytes_to_corrupt,
+    ASSERT_OK(test::CorruptFile(env_, fname, offset, bytes_to_corrupt,
                                /*verify_checksum*/ filetype == kTableFile));
  }

@@ -240,7 +234,7 @@ class CorruptionTest : public testing::Test {
    db_->GetLiveFilesMetaData(&metadata);
    for (const auto& m : metadata) {
      if (m.level == level) {
-        ASSERT_OK(test::CorruptFile(env_.get(), dbname_ + "/" + m.name, offset,
+        ASSERT_OK(test::CorruptFile(env_, dbname_ + "/" + m.name, offset,
                                    bytes_to_corrupt));
        return;
      }
@@ -314,7 +308,7 @@ class CorruptionTest : public testing::Test {
    if (bytes_to_truncate == 0) {
      new_size = 0;
    }
-    ASSERT_OK(test::TruncateFile(env_.get(), path, new_size));
+    ASSERT_OK(test::TruncateFile(env_, path, new_size));
  }
 };

@@ -408,14 +402,14 @@ TEST_F(CorruptionTest, PostPITRCorruptionWALsRetained) {
 }

 TEST_F(CorruptionTest, RecoverWriteError) {
-  fs_->writable_file_error_ = true;
+  env_->writable_file_error_ = true;
  Status s = TryReopen();
  ASSERT_TRUE(!s.ok());
 }

 TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
  // Do enough writing to force minor compaction
-  fs_->writable_file_error_ = true;
+  env_->writable_file_error_ = true;
  const int num =
      static_cast<int>(3 + (Options().write_buffer_size / kValueSize));
  std::string value_storage;
@@ -431,8 +425,8 @@ TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
    ASSERT_TRUE(!failed || !s.ok());
  }
  ASSERT_TRUE(!s.ok());
-  ASSERT_GE(fs_->num_writable_file_errors_, 1);
-  fs_->writable_file_error_ = false;
+  ASSERT_GE(env_->num_writable_file_errors_, 1);
+  env_->writable_file_error_ = false;
  Reopen();
 }

@@ -450,7 +444,7 @@ TEST_F(CorruptionTest, TableFile) {

 TEST_F(CorruptionTest, VerifyChecksumReadahead) {
  Options options;
-  SpecialEnv senv(base_env_);
+  SpecialEnv senv(env_->target());
  options.env = &senv;
  // Disable block cache as we are going to check checksum for
  // the same file twice and measure number of reads.
@@ -518,7 +512,7 @@ TEST_F(CorruptionTest, TableFileIndexData) {
  dbi = static_cast_with_check<DBImpl>(db_);
  // one full file may be readable, since only one was corrupted
  // the other file should be fully non-readable, since index was corrupted
-  Check(0, 5000, ReadOptions(true, true));
+  Check(0, 5000);
  ASSERT_NOK(dbi->VerifyChecksum());

  // In paranoid mode, the db cannot be opened due to the corrupted file.
@@ -593,8 +587,8 @@ TEST_F(CorruptionTest, TableFileWrongSize) {
  // Make the file smaller with truncation.
  // First leaving a partial footer, and then completely removing footer.
  for (size_t bytes_lost : {8, 100}) {
-    ASSERT_OK(test::TruncateFile(env_.get(), filename,
-                                 metadata[0].size - bytes_lost));
+    ASSERT_OK(
+        test::TruncateFile(env_, filename, metadata[0].size - bytes_lost));

    // Reported well with paranoid checks
    options_.paranoid_checks = true;
@@ -659,7 +653,7 @@ TEST_F(CorruptionTest, CorruptedDescriptor) {

 TEST_F(CorruptionTest, CompactionInputError) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  Reopen(&options);
  Build(10);
  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
@@ -680,7 +674,7 @@ TEST_F(CorruptionTest, CompactionInputError) {

 TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.paranoid_checks = true;
  options.write_buffer_size = 131072;
  options.max_write_buffer_number = 2;
@@ -767,7 +761,7 @@ TEST_F(CorruptionTest, RangeDeletionCorrupted) {
      ImmutableOptions(options_), kRangeDelBlockName, &range_del_handle));

  ASSERT_OK(TryReopen());
-  ASSERT_OK(test::CorruptFile(env_.get(), filename,
+  ASSERT_OK(test::CorruptFile(env_, filename,
                              static_cast<int>(range_del_handle.offset()), 1));
  ASSERT_TRUE(TryReopen().IsCorruption());
 }
@@ -775,7 +769,7 @@ TEST_F(CorruptionTest, RangeDeletionCorrupted) {
 TEST_F(CorruptionTest, FileSystemStateCorrupted) {
  for (int iter = 0; iter < 2; ++iter) {
    Options options;
-    options.env = env_.get();
+    options.env = env_;
    options.paranoid_checks = true;
    options.create_if_missing = true;
    Reopen(&options);
@@ -814,7 +808,7 @@ static const auto& corruption_modes = {

 TEST_F(CorruptionTest, ParanoidFileChecksOnFlush) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.check_flush_compaction_key_order = false;
  options.paranoid_file_checks = true;
  options.create_if_missing = true;
@@ -842,7 +836,7 @@ TEST_F(CorruptionTest, ParanoidFileChecksOnFlush) {

 TEST_F(CorruptionTest, ParanoidFileChecksOnCompact) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.paranoid_file_checks = true;
  options.create_if_missing = true;
  options.check_flush_compaction_key_order = false;
@@ -875,7 +869,7 @@ TEST_F(CorruptionTest, ParanoidFileChecksOnCompact) {

 TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeFirst) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.check_flush_compaction_key_order = false;
  options.paranoid_file_checks = true;
  options.create_if_missing = true;
@@ -911,7 +905,7 @@ TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeFirst) {

 TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRange) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.check_flush_compaction_key_order = false;
  options.paranoid_file_checks = true;
  options.create_if_missing = true;
@@ -950,7 +944,7 @@ TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRange) {

 TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeLast) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.check_flush_compaction_key_order = false;
  options.paranoid_file_checks = true;
  options.create_if_missing = true;
@@ -986,7 +980,7 @@ TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeLast) {

 TEST_F(CorruptionTest, LogCorruptionErrorsInCompactionIterator) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.create_if_missing = true;
  options.allow_data_in_errors = true;
  auto mode = mock::MockTableFactory::kCorruptKey;
@@ -1015,7 +1009,7 @@ TEST_F(CorruptionTest, LogCorruptionErrorsInCompactionIterator) {

 TEST_F(CorruptionTest, CompactionKeyOrderCheck) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.paranoid_file_checks = false;
  options.create_if_missing = true;
  options.check_flush_compaction_key_order = false;
@@ -1042,7 +1036,7 @@ TEST_F(CorruptionTest, CompactionKeyOrderCheck) {

 TEST_F(CorruptionTest, FlushKeyOrderCheck) {
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  options.paranoid_file_checks = false;
  options.create_if_missing = true;
  ASSERT_OK(db_->SetOptions({{"check_flush_compaction_key_order", "true"}}));
@@ -1095,7 +1089,7 @@ TEST_F(CorruptionTest, DisableKeyOrderCheck) {
 TEST_F(CorruptionTest, VerifyWholeTableChecksum) {
  CloseDb();
  Options options;
-  options.env = env_.get();
+  options.env = env_;
  ASSERT_OK(DestroyDB(dbname_, options));
  options.create_if_missing = true;
  options.file_checksum_gen_factory =
@@ -1184,7 +1178,7 @@ TEST_P(CrashDuringRecoveryWithCorruptionTest, CrashDuringRecovery) {
      track_and_verify_wals_in_manifest_;
  options.wal_recovery_mode = WALRecoveryMode::kPointInTimeRecovery;
  options.avoid_flush_during_recovery = false;
-  options.env = env_.get();
+  options.env = env_;
  ASSERT_OK(DestroyDB(dbname_, options));
  options.create_if_missing = true;
  options.max_write_buffer_number = 8;
@@ -1356,7 +1350,7 @@ TEST_P(CrashDuringRecoveryWithCorruptionTest, TxnDbCrashDuringRecovery) {
  options.track_and_verify_wals_in_manifest =
      track_and_verify_wals_in_manifest_;
  options.avoid_flush_during_recovery = false;
-  options.env = env_.get();
+  options.env = env_;
  ASSERT_OK(DestroyDB(dbname_, options));
  options.create_if_missing = true;
  options.max_write_buffer_number = 3;
@@ -1551,7 +1545,7 @@ TEST_P(CrashDuringRecoveryWithCorruptionTest, CrashDuringRecoveryWithFlush) {
  Options options;
  options.wal_recovery_mode = WALRecoveryMode::kPointInTimeRecovery;
  options.avoid_flush_during_recovery = false;
-  options.env = env_.get();
+  options.env = env_;
  options.create_if_missing = true;

  ASSERT_OK(DestroyDB(dbname_, options));
@@ -1667,3 +1661,12 @@ int main(int argc, char** argv) {
  return RUN_ALL_TESTS();
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr, "SKIPPED as RepairDB() is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // !ROCKSDB_LITE
@@ -3,6 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).

+#ifndef ROCKSDB_LITE

 #include "db/db_impl/db_impl.h"
 #include "db/db_test_util.h"
@@ -349,3 +350,12 @@ int main(int argc, char** argv) {
  }
 }

+#else
+#include <stdio.h>
+
+int main(int /*argc*/, char** /*argv*/) {
+  fprintf(stderr, "SKIPPED as Cuckoo table is not supported in ROCKSDB_LITE\n");
+  return 0;
+}
+
+#endif  // ROCKSDB_LITE
@@ -20,7 +20,9 @@
 #include "rocksdb/utilities/debug.h"
 #include "table/block_based/block_based_table_reader.h"
 #include "table/block_based/block_builder.h"
+#if !defined(ROCKSDB_LITE)
 #include "test_util/sync_point.h"
+#endif
 #include "util/file_checksum_helper.h"
 #include "util/random.h"
 #include "utilities/counted_fs.h"
@@ -96,6 +98,7 @@ TEST_F(DBBasicTest, UniqueSession) {
  EXPECT_MATCHES_REGEX(sid2, expected);
  EXPECT_MATCHES_REGEX(sid3, expected);

+#ifndef ROCKSDB_LITE
  Close();
  ASSERT_OK(ReadOnlyReopen(options));
  ASSERT_OK(db_->GetDbSessionId(sid1));
@@ -110,6 +113,7 @@ TEST_F(DBBasicTest, UniqueSession) {
  ASSERT_NE(sid1, sid2);

  ASSERT_EQ(sid2, sid3);
+#endif  // ROCKSDB_LITE

  CreateAndReopenWithCF({"goku"}, options);
  ASSERT_OK(db_->GetDbSessionId(sid1));
@@ -126,6 +130,7 @@ TEST_F(DBBasicTest, UniqueSession) {
  ASSERT_NE(sid1, sid4);
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBasicTest, ReadOnlyDB) {
  ASSERT_OK(Put("foo", "v1"));
  ASSERT_OK(Put("bar", "v2"));
@@ -362,6 +367,7 @@ TEST_F(DBBasicTest, LevelLimitReopen) {
  options.max_bytes_for_level_multiplier_additional.resize(10, 1);
  ASSERT_OK(TryReopenWithColumnFamilies({"default", "pikachu"}, options));
 }
+#endif  // ROCKSDB_LITE

 TEST_F(DBBasicTest, PutDeleteGet) {
  do {
@@ -423,6 +429,7 @@ TEST_F(DBBasicTest, GetFromVersions) {
  } while (ChangeOptions());
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBasicTest, GetSnapshot) {
  anon::OptionsOverride options_override;
  options_override.skip_policy = kSkipNoSnapshot;
@@ -443,6 +450,7 @@ TEST_F(DBBasicTest, GetSnapshot) {
    }
  } while (ChangeOptions());
 }
+#endif  // ROCKSDB_LITE

 TEST_F(DBBasicTest, CheckLock) {
  do {
@@ -674,27 +682,7 @@ TEST_F(DBBasicTest, IdentityAcrossRestarts) {
  } while (ChangeCompactOptions());
 }

-TEST_F(DBBasicTest, LockFileRecovery) {
-  Options options = CurrentOptions();
-  // Regardless of best_efforts_recovery
-  for (bool ber : {false, true}) {
-    options.best_efforts_recovery = ber;
-    DestroyAndReopen(options);
-    std::string id1, id2;
-    ASSERT_OK(db_->GetDbIdentity(id1));
-    Close();
-
-    // Should be OK to re-open DB after lock file deleted
-    std::string lockfilename = LockFileName(dbname_);
-    ASSERT_OK(env_->DeleteFile(lockfilename));
-    Reopen(options);
-
-    // Should be same DB as before
-    ASSERT_OK(db_->GetDbIdentity(id2));
-    ASSERT_EQ(id1, id2);
-  }
-}
-
+#ifndef ROCKSDB_LITE
 TEST_F(DBBasicTest, Snapshot) {
  env_->SetMockSleep();
  anon::OptionsOverride options_override;
@@ -766,6 +754,7 @@ TEST_F(DBBasicTest, Snapshot) {
  } while (ChangeOptions());
 }

+#endif  // ROCKSDB_LITE

 class DBBasicMultiConfigs : public DBBasicTest,
                            public ::testing::WithParamInterface<int> {
@@ -2302,7 +2291,9 @@ TEST_P(DBMultiGetAsyncIOTest, GetFromL0) {
    ASSERT_EQ(multiget_io_batch_size.count, 3);
  }
 #else   // ROCKSDB_IOURING_PRESENT
-  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 0);
+  if (GetParam()) {
+    ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 3);
+  }
 #endif  // ROCKSDB_IOURING_PRESENT
 }

@@ -2336,18 +2327,16 @@ TEST_P(DBMultiGetAsyncIOTest, GetFromL1) {
  ASSERT_EQ(values[1], "val_l1_" + std::to_string(54));
  ASSERT_EQ(values[2], "val_l1_" + std::to_string(102));

+#ifdef ROCKSDB_IOURING_PRESENT
  HistogramData multiget_io_batch_size;

  statistics()->histogramData(MULTIGET_IO_BATCH_SIZE, &multiget_io_batch_size);

-#ifdef ROCKSDB_IOURING_PRESENT
  // A batch of 3 async IOs is expected, one for each overlapping file in L1
  ASSERT_EQ(multiget_io_batch_size.count, 1);
  ASSERT_EQ(multiget_io_batch_size.max, 3);
-  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 3);
-#else   // ROCKSDB_IOURING_PRESENT
-  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 0);
 #endif  // ROCKSDB_IOURING_PRESENT
+  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 3);
 }

 #ifdef ROCKSDB_IOURING_PRESENT
@@ -2531,12 +2520,8 @@ TEST_P(DBMultiGetAsyncIOTest, GetFromL2WithRangeOverlapL0L1) {
  ASSERT_EQ(values[0], "val_l2_" + std::to_string(19));
  ASSERT_EQ(values[1], "val_l2_" + std::to_string(26));

-#ifdef ROCKSDB_IOURING_PRESENT
  // Bloom filters in L0/L1 will avoid the coroutine calls in those levels
  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 2);
-#else   // ROCKSDB_IOURING_PRESENT
-  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 0);
-#endif  // ROCKSDB_IOURING_PRESENT
 }

 #ifdef ROCKSDB_IOURING_PRESENT
@@ -2627,17 +2612,18 @@ TEST_P(DBMultiGetAsyncIOTest, GetNoIOUring) {
  dbfull()->MultiGet(ro, dbfull()->DefaultColumnFamily(), keys.size(),
                     keys.data(), values.data(), statuses.data());
  ASSERT_EQ(values.size(), 3);
-  ASSERT_EQ(statuses[0], Status::OK());
-  ASSERT_EQ(statuses[1], Status::OK());
-  ASSERT_EQ(statuses[2], Status::OK());
+  ASSERT_EQ(statuses[0], Status::NotSupported());
+  ASSERT_EQ(statuses[1], Status::NotSupported());
+  ASSERT_EQ(statuses[2], Status::NotSupported());

-  HistogramData async_read_bytes;
+  HistogramData multiget_io_batch_size;

-  statistics()->histogramData(ASYNC_READ_BYTES, &async_read_bytes);
+  statistics()->histogramData(MULTIGET_IO_BATCH_SIZE, &multiget_io_batch_size);

  // A batch of 3 async IOs is expected, one for each overlapping file in L1
-  ASSERT_EQ(async_read_bytes.count, 0);
-  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 0);
+  ASSERT_EQ(multiget_io_batch_size.count, 1);
+  ASSERT_EQ(multiget_io_batch_size.max, 3);
+  ASSERT_EQ(statistics()->getTickerCount(MULTIGET_COROUTINE_COUNT), 3);
 }

 INSTANTIATE_TEST_CASE_P(DBMultiGetAsyncIOTest, DBMultiGetAsyncIOTest,
@@ -2808,6 +2794,7 @@ TEST_P(MultiGetPrefixExtractorTest, Batched) {
 INSTANTIATE_TEST_CASE_P(MultiGetPrefix, MultiGetPrefixExtractorTest,
                        ::testing::Bool());

+#ifndef ROCKSDB_LITE
 class DBMultiGetRowCacheTest : public DBBasicTest,
                               public ::testing::WithParamInterface<bool> {};

@@ -2962,6 +2949,7 @@ TEST_F(DBBasicTest, ValueTypeString) {
    ASSERT_TRUE(key_version.GetTypeName() != "Invalid");
  }
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBBasicTest, MultiGetIOBufferOverrun) {
  Options options = CurrentOptions();
@@ -3062,6 +3050,7 @@ TEST_F(DBBasicTest, BestEffortsRecoveryWithVersionBuildingFailure) {
  SyncPoint::GetInstance()->ClearAllCallBacks();
 }

+#ifndef ROCKSDB_LITE
 namespace {
 class TableFileListener : public EventListener {
 public:
@@ -3336,6 +3325,7 @@ TEST_F(DBBasicTest, DisableTrackWal) {
  ASSERT_TRUE(dbfull()->GetVersionSet()->GetWalSet().GetWals().empty());
  Close();
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBBasicTest, ManifestChecksumMismatch) {
  Options options = CurrentOptions();
@@ -3379,6 +3369,7 @@ TEST_F(DBBasicTest, ConcurrentlyCloseDB) {
  }
 }

+#ifndef ROCKSDB_LITE
 class DBBasicTestTrackWal : public DBTestBase,
                            public testing::WithParamInterface<bool> {
 public:
@@ -3434,16 +3425,21 @@ TEST_P(DBBasicTestTrackWal, DoNotTrackObsoleteWal) {

 INSTANTIATE_TEST_CASE_P(DBBasicTestTrackWal, DBBasicTestTrackWal,
                        testing::Bool());
+#endif  // ROCKSDB_LITE

 class DBBasicTestMultiGet : public DBTestBase {
 public:
-  DBBasicTestMultiGet(std::string test_dir, int num_cfs,
+  DBBasicTestMultiGet(std::string test_dir, int num_cfs, bool compressed_cache,
                      bool uncompressed_cache, bool _compression_enabled,
                      bool _fill_cache, uint32_t compression_parallel_threads)
      : DBTestBase(test_dir, /*env_do_fsync=*/false) {
    compression_enabled_ = _compression_enabled;
    fill_cache_ = _fill_cache;

+    if (compressed_cache) {
+      std::shared_ptr<Cache> cache = NewLRUCache(1048576);
+      compressed_cache_ = std::make_shared<MyBlockCache>(cache);
+    }
    if (uncompressed_cache) {
      std::shared_ptr<Cache> cache = NewLRUCache(1048576);
      uncompressed_cache_ = std::make_shared<MyBlockCache>(cache);
@@ -3455,6 +3451,7 @@ class DBBasicTestMultiGet : public DBTestBase {
    Random rnd(301);
    BlockBasedTableOptions table_options;

+#ifndef ROCKSDB_LITE
    if (compression_enabled_) {
      std::vector<CompressionType> compression_types;
      compression_types = GetSupportedCompressions();
@@ -3473,6 +3470,12 @@ class DBBasicTestMultiGet : public DBTestBase {
        compression_enabled_ = false;
      }
    }
+#else
+    // GetSupportedCompressions() is not available in LITE build
+    if (!Snappy_Supported()) {
+      compression_enabled_ = false;
+    }
+#endif  // ROCKSDB_LITE

    table_options.block_cache = uncompressed_cache_;
    if (table_options.block_cache == nullptr) {
@@ -3480,6 +3483,7 @@ class DBBasicTestMultiGet : public DBTestBase {
    } else {
      table_options.pin_l0_filter_and_index_blocks_in_cache = true;
    }
+    table_options.block_cache_compressed = compressed_cache_;
    table_options.flush_block_policy_factory.reset(
        new MyFlushBlockPolicyFactory());
    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
@@ -3629,11 +3633,11 @@ class DBBasicTestMultiGet : public DBTestBase {

    Handle* Lookup(const Slice& key, const CacheItemHelper* helper,
                   CreateContext* create_context,
-                   Priority priority = Priority::LOW,
+                   Priority priority = Priority::LOW, bool wait = true,
                   Statistics* stats = nullptr) override {
      num_lookups_++;
      Handle* handle =
-          target_->Lookup(key, helper, create_context, priority, stats);
+          target_->Lookup(key, helper, create_context, priority, wait, stats);
      if (handle != nullptr) {
        num_found_++;
      }
@@ -3662,14 +3666,16 @@ class DBBasicTestMultiGet : public DBTestBase {
  std::vector<std::string> cf_names_;
 };

-class DBBasicTestWithParallelIO : public DBBasicTestMultiGet,
-                                  public testing::WithParamInterface<
-                                      std::tuple<bool, bool, bool, uint32_t>> {
+class DBBasicTestWithParallelIO
+    : public DBBasicTestMultiGet,
+      public testing::WithParamInterface<
+          std::tuple<bool, bool, bool, bool, uint32_t>> {
 public:
  DBBasicTestWithParallelIO()
      : DBBasicTestMultiGet("/db_basic_test_with_parallel_io", 1,
                            std::get<0>(GetParam()), std::get<1>(GetParam()),
-                            std::get<2>(GetParam()), std::get<3>(GetParam())) {}
+                            std::get<2>(GetParam()), std::get<3>(GetParam()),
+                            std::get<4>(GetParam())) {}
 };

 TEST_P(DBBasicTestWithParallelIO, MultiGet) {
@@ -3795,6 +3801,7 @@ TEST_P(DBBasicTestWithParallelIO, MultiGet) {
  }
 }

+#ifndef ROCKSDB_LITE
 TEST_P(DBBasicTestWithParallelIO, MultiGetDirectIO) {
  class FakeDirectIOEnv : public EnvWrapper {
    class FakeDirectIOSequentialFile;
@@ -3911,6 +3918,7 @@ TEST_P(DBBasicTestWithParallelIO, MultiGetDirectIO) {
  }
  Close();
 }
+#endif  // ROCKSDB_LITE

 TEST_P(DBBasicTestWithParallelIO, MultiGetWithChecksumMismatch) {
  std::vector<std::string> key_data(10);
@@ -3998,12 +4006,13 @@ TEST_P(DBBasicTestWithParallelIO, MultiGetWithMissingFile) {

 INSTANTIATE_TEST_CASE_P(ParallelIO, DBBasicTestWithParallelIO,
                        // Params are as follows -
-                        // Param 0 - Uncompressed cache enabled
-                        // Param 1 - Data compression enabled
-                        // Param 2 - ReadOptions::fill_cache
-                        // Param 3 - CompressionOptions::parallel_threads
+                        // Param 0 - Compressed cache enabled
+                        // Param 1 - Uncompressed cache enabled
+                        // Param 2 - Data compression enabled
+                        // Param 3 - ReadOptions::fill_cache
+                        // Param 4 - CompressionOptions::parallel_threads
                        ::testing::Combine(::testing::Bool(), ::testing::Bool(),
-                                           ::testing::Bool(),
+                                           ::testing::Bool(), ::testing::Bool(),
                                           ::testing::Values(1, 4)));

 // Forward declaration
@@ -4214,8 +4223,9 @@ class DBBasicTestMultiGetDeadline : public DBBasicTestMultiGet,
  DBBasicTestMultiGetDeadline()
      : DBBasicTestMultiGet(
            "db_basic_test_multiget_deadline" /*Test dir*/,
-            10 /*# of column families*/, true /*uncompressed cache enabled*/,
-            true /*compression enabled*/, true /*ReadOptions.fill_cache*/,
+            10 /*# of column families*/, false /*compressed cache enabled*/,
+            true /*uncompressed cache enabled*/, true /*compression enabled*/,
+            true /*ReadOptions.fill_cache*/,
            1 /*# of parallel compression threads*/) {}

  inline void CheckStatus(std::vector<Status>& statuses, size_t num_ok) {
@@ -4451,6 +4461,7 @@ TEST_F(DBBasicTest, FailOpenIfLoggerCreationFail) {
  SyncPoint::GetInstance()->ClearAllCallBacks();
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBBasicTest, VerifyFileChecksums) {
  Options options = GetDefaultOptions();
  options.create_if_missing = true;
@@ -4494,63 +4505,6 @@ TEST_F(DBBasicTest, VerifyFileChecksums) {
  ASSERT_TRUE(db_->VerifyFileChecksums(ReadOptions()).IsInvalidArgument());
 }

-TEST_F(DBBasicTest, VerifyFileChecksumsReadahead) {
-  Options options = GetDefaultOptions();
-  options.create_if_missing = true;
-  options.env = env_;
-  options.file_checksum_gen_factory = GetFileChecksumGenCrc32cFactory();
-  DestroyAndReopen(options);
-
-  Random rnd(301);
-  int alignment = 256 * 1024;
-  for (int i = 0; i < 16; ++i) {
-    ASSERT_OK(Put("key" + std::to_string(i), rnd.RandomString(alignment)));
-  }
-  ASSERT_OK(Flush());
-
-  std::vector<std::string> filenames;
-  int sst_cnt = 0;
-  std::string sst_name;
-  uint64_t sst_size;
-  uint64_t number;
-  FileType type;
-  ASSERT_OK(env_->GetChildren(dbname_, &filenames));
-  for (auto name : filenames) {
-    if (ParseFileName(name, &number, &type)) {
-      if (type == kTableFile) {
-        sst_cnt++;
-        sst_name = name;
-      }
-    }
-  }
-  ASSERT_EQ(sst_cnt, 1);
-  ASSERT_OK(env_->GetFileSize(dbname_ + '/' + sst_name, &sst_size));
-
-  bool last_read = false;
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
-      "GenerateOneFileChecksum::Chunk:0", [&](void* /*arg*/) {
-        if (env_->random_read_bytes_counter_.load() == sst_size) {
-          EXPECT_FALSE(last_read);
-          last_read = true;
-        } else {
-          ASSERT_EQ(env_->random_read_bytes_counter_.load() & (alignment - 1),
-                    0);
-        }
-      });
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
-  env_->count_random_reads_ = true;
-  env_->random_read_bytes_counter_ = 0;
-  env_->random_read_counter_.Reset();
-
-  ReadOptions ro;
-  ro.readahead_size = alignment;
-  ASSERT_OK(db_->VerifyFileChecksums(ro));
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
-  ASSERT_TRUE(last_read);
-  ASSERT_EQ(env_->random_read_counter_.Read(),
-            (sst_size + alignment - 1) / (alignment));
-}
-
 // TODO: re-enable after we provide finer-grained control for WAL tracking to
 // meet the needs of different use cases, durability levels and recovery modes.
 TEST_F(DBBasicTest, DISABLED_ManualWalSync) {
@@ -4573,6 +4527,7 @@ TEST_F(DBBasicTest, DISABLED_ManualWalSync) {

  ASSERT_TRUE(TryReopen(options).IsCorruption());
 }
+#endif  // !ROCKSDB_LITE

 // A test class for intercepting random reads and injecting artificial
 // delays. Used for testing the deadline/timeout feature
@@ -44,6 +44,10 @@ class DBBlockCacheTest : public DBTestBase {
  size_t compression_dict_miss_count_ = 0;
  size_t compression_dict_hit_count_ = 0;
  size_t compression_dict_insert_count_ = 0;
+  size_t compressed_miss_count_ = 0;
+  size_t compressed_hit_count_ = 0;
+  size_t compressed_insert_count_ = 0;
+  size_t compressed_failure_count_ = 0;

 public:
  const size_t kNumBlocks = 10;
@@ -81,6 +85,14 @@ class DBBlockCacheTest : public DBTestBase {
    hit_count_ = TestGetTickerCount(options, BLOCK_CACHE_HIT);
    insert_count_ = TestGetTickerCount(options, BLOCK_CACHE_ADD);
    failure_count_ = TestGetTickerCount(options, BLOCK_CACHE_ADD_FAILURES);
+    compressed_miss_count_ =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS);
+    compressed_hit_count_ =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT);
+    compressed_insert_count_ =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD);
+    compressed_failure_count_ =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD_FAILURES);
  }

  void RecordCacheCountersForCompressionDict(const Options& options) {
@@ -132,6 +144,30 @@ class DBBlockCacheTest : public DBTestBase {
    compression_dict_insert_count_ = new_compression_dict_insert_count;
  }

+  void CheckCompressedCacheCounters(const Options& options,
+                                    size_t expected_misses,
+                                    size_t expected_hits,
+                                    size_t expected_inserts,
+                                    size_t expected_failures) {
+    size_t new_miss_count =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS);
+    size_t new_hit_count =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT);
+    size_t new_insert_count =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD);
+    size_t new_failure_count =
+        TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_ADD_FAILURES);
+    ASSERT_EQ(compressed_miss_count_ + expected_misses, new_miss_count);
+    ASSERT_EQ(compressed_hit_count_ + expected_hits, new_hit_count);
+    ASSERT_EQ(compressed_insert_count_ + expected_inserts, new_insert_count);
+    ASSERT_EQ(compressed_failure_count_ + expected_failures, new_failure_count);
+    compressed_miss_count_ = new_miss_count;
+    compressed_hit_count_ = new_hit_count;
+    compressed_insert_count_ = new_insert_count;
+    compressed_failure_count_ = new_failure_count;
+  }
+
+#ifndef ROCKSDB_LITE
  const std::array<size_t, kNumCacheEntryRoles> GetCacheEntryRoleCountsBg() {
    // Verify in cache entry role stats
    std::array<size_t, kNumCacheEntryRoles> cache_entry_role_counts;
@@ -145,6 +181,7 @@ class DBBlockCacheTest : public DBTestBase {
    }
    return cache_entry_role_counts;
  }
+#endif  // ROCKSDB_LITE
 };

 TEST_F(DBBlockCacheTest, IteratorBlockCacheUsage) {
@@ -237,6 +274,84 @@ TEST_F(DBBlockCacheTest, TestWithoutCompressedBlockCache) {
 }

 #ifdef SNAPPY
+TEST_F(DBBlockCacheTest, TestWithCompressedBlockCache) {
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
+
+  BlockBasedTableOptions table_options;
+  table_options.no_block_cache = true;
+  table_options.block_cache_compressed = nullptr;
+  table_options.block_size = 1;
+  table_options.filter_policy.reset(NewBloomFilterPolicy(20));
+  table_options.cache_index_and_filter_blocks = false;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  options.compression = CompressionType::kSnappyCompression;
+
+  DestroyAndReopen(options);
+
+  std::string value(kValueSize, 'a');
+  for (size_t i = 0; i < kNumBlocks; i++) {
+    ASSERT_OK(Put(std::to_string(i), value));
+    ASSERT_OK(Flush());
+  }
+
+  ReadOptions read_options;
+  std::shared_ptr<Cache> compressed_cache = NewLRUCache(1 << 25, 0, false);
+  LRUCacheOptions co;
+  co.capacity = 0;
+  co.num_shard_bits = 0;
+  co.strict_capacity_limit = false;
+  // Needed not to count entry stats collector
+  co.metadata_charge_policy = kDontChargeCacheMetadata;
+  std::shared_ptr<Cache> cache = NewLRUCache(co);
+  table_options.block_cache = cache;
+  table_options.no_block_cache = false;
+  table_options.block_cache_compressed = compressed_cache;
+  table_options.max_auto_readahead_size = 0;
+  table_options.cache_index_and_filter_blocks = false;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  Reopen(options);
+  RecordCacheCounters(options);
+
+  // Load blocks into cache.
+  for (size_t i = 0; i < kNumBlocks - 1; i++) {
+    ASSERT_EQ(value, Get(std::to_string(i)));
+    CheckCacheCounters(options, 1, 0, 1, 0);
+    CheckCompressedCacheCounters(options, 1, 0, 1, 0);
+  }
+
+  size_t usage = cache->GetUsage();
+  ASSERT_EQ(0, usage);
+  ASSERT_EQ(usage, cache->GetPinnedUsage());
+  size_t compressed_usage = compressed_cache->GetUsage();
+  ASSERT_LT(0, compressed_usage);
+  // Compressed block cache cannot be pinned.
+  ASSERT_EQ(0, compressed_cache->GetPinnedUsage());
+
+  // Set strict capacity limit flag. Now block will only load into compressed
+  // block cache.
+  cache->SetCapacity(usage);
+  cache->SetStrictCapacityLimit(true);
+  ASSERT_EQ(usage, cache->GetPinnedUsage());
+
+  // Load last key block.
+  ASSERT_EQ(
+      "Operation aborted: Memory limit reached: Insert failed due to LRU cache "
+      "being full.",
+      Get(std::to_string(kNumBlocks - 1)));
+  // Failure will also record the miss counter.
+  CheckCacheCounters(options, 1, 0, 0, 1);
+  CheckCompressedCacheCounters(options, 1, 0, 1, 0);
+
+  // Clear strict capacity limit flag. This time we shall hit compressed block
+  // cache and load into block cache.
+  cache->SetStrictCapacityLimit(false);
+  // Load last key block.
+  ASSERT_EQ(value, Get(std::to_string(kNumBlocks - 1)));
+  CheckCacheCounters(options, 1, 0, 1, 0);
+  CheckCompressedCacheCounters(options, 0, 1, 0, 0);
+}

 namespace {
 class PersistentCacheFromCache : public PersistentCache {
@@ -287,11 +402,9 @@ class PersistentCacheFromCache : public PersistentCache {
 };

 class ReadOnlyCacheWrapper : public CacheWrapper {
- public:
  using CacheWrapper::CacheWrapper;

-  const char* Name() const override { return "ReadOnlyCacheWrapper"; }
-
+  using Cache::Insert;
  Status Insert(const Slice& /*key*/, Cache::ObjectPtr /*value*/,
                const CacheItemHelper* /*helper*/, size_t /*charge*/,
                Handle** /*handle*/, Priority /*priority*/) override {
@@ -300,8 +413,83 @@ class ReadOnlyCacheWrapper : public CacheWrapper {
 };

 }  // anonymous namespace
+
+TEST_F(DBBlockCacheTest, TestWithSameCompressed) {
+  auto table_options = GetTableOptions();
+  auto options = GetOptions(table_options);
+  InitTable(options);
+
+  std::shared_ptr<Cache> rw_cache{NewLRUCache(1000000)};
+  std::shared_ptr<PersistentCacheFromCache> rw_pcache{
+      new PersistentCacheFromCache(rw_cache, /*read_only*/ false)};
+  // Exercise some obscure behavior with read-only wrappers
+  std::shared_ptr<Cache> ro_cache{new ReadOnlyCacheWrapper(rw_cache)};
+  std::shared_ptr<PersistentCacheFromCache> ro_pcache{
+      new PersistentCacheFromCache(rw_cache, /*read_only*/ true)};
+
+  // Simple same pointer
+  table_options.block_cache = rw_cache;
+  table_options.block_cache_compressed = rw_cache;
+  table_options.persistent_cache.reset();
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: block_cache same as block_cache_compressed not "
+            "currently supported, and would be bad for performance anyway");
+
+  // Other cases
+  table_options.block_cache = ro_cache;
+  table_options.block_cache_compressed = rw_cache;
+  table_options.persistent_cache.reset();
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: block_cache and block_cache_compressed share "
+            "the same key space, which is not supported");
+
+  table_options.block_cache = rw_cache;
+  table_options.block_cache_compressed = ro_cache;
+  table_options.persistent_cache.reset();
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: block_cache_compressed and block_cache share "
+            "the same key space, which is not supported");
+
+  table_options.block_cache = ro_cache;
+  table_options.block_cache_compressed.reset();
+  table_options.persistent_cache = rw_pcache;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: block_cache and persistent_cache share the same "
+            "key space, which is not supported");
+
+  table_options.block_cache = rw_cache;
+  table_options.block_cache_compressed.reset();
+  table_options.persistent_cache = ro_pcache;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: persistent_cache and block_cache share the same "
+            "key space, which is not supported");
+
+  table_options.block_cache.reset();
+  table_options.no_block_cache = true;
+  table_options.block_cache_compressed = ro_cache;
+  table_options.persistent_cache = rw_pcache;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: block_cache_compressed and persistent_cache "
+            "share the same key space, which is not supported");
+
+  table_options.block_cache.reset();
+  table_options.no_block_cache = true;
+  table_options.block_cache_compressed = rw_cache;
+  table_options.persistent_cache = ro_pcache;
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  ASSERT_EQ(TryReopen(options).ToString(),
+            "Invalid argument: persistent_cache and block_cache_compressed "
+            "share the same key space, which is not supported");
+}
 #endif  // SNAPPY

+#ifndef ROCKSDB_LITE

 // Make sure that when options.block_cache is set, after a new table is
 // created its index/filter blocks are added to block cache.
@@ -421,6 +609,11 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
  ASSERT_EQ(cache->GetUsage(), index_bytes_insert + filter_bytes_insert);
  // set the cache capacity to the current usage
  cache->SetCapacity(index_bytes_insert + filter_bytes_insert);
+  // The index and filter eviction statistics were broken by the refactoring
+  // that moved the readers out of the block cache. Disabling these until we can
+  // bring the stats back.
+  // ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_EVICT), 0);
+  // ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_EVICT), 0);
  // 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"));
@@ -431,6 +624,13 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
            index_bytes_insert);
  ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_INSERT),
            filter_bytes_insert);
+  // The index and filter eviction statistics were broken by the refactoring
+  // that moved the readers out of the block cache. Disabling these until we can
+  // bring the stats back.
+  // ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_INDEX_BYTES_EVICT),
+  //           index_bytes_insert);
+  // ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_FILTER_BYTES_EVICT),
+  //           filter_bytes_insert);
 }

 #if (defined OS_LINUX || defined OS_WIN)
@@ -713,11 +913,10 @@ class LookupLiarCache : public CacheWrapper {
  explicit LookupLiarCache(std::shared_ptr<Cache> target)
      : CacheWrapper(std::move(target)) {}

-  const char* Name() const override { return "LookupLiarCache"; }
-
+  using Cache::Lookup;
  Handle* Lookup(const Slice& key, const CacheItemHelper* helper = nullptr,
                 CreateContext* create_context = nullptr,
-                 Priority priority = Priority::LOW,
+                 Priority priority = Priority::LOW, bool wait = true,
                 Statistics* stats = nullptr) override {
    if (nth_lookup_not_found_ == 1) {
      nth_lookup_not_found_ = 0;
@@ -726,7 +925,8 @@ class LookupLiarCache : public CacheWrapper {
    if (nth_lookup_not_found_ > 1) {
      --nth_lookup_not_found_;
    }
-    return CacheWrapper::Lookup(key, helper, create_context, priority, stats);
+    return CacheWrapper::Lookup(key, helper, create_context, priority, wait,
+                                stats);
  }

  // 1 == next lookup, 2 == after next, etc.
@@ -886,6 +1086,124 @@ TEST_F(DBBlockCacheTest, ParanoidFileChecks) {
            TestGetTickerCount(options, BLOCK_CACHE_ADD));
 }

+TEST_F(DBBlockCacheTest, CompressedCache) {
+  if (!Snappy_Supported()) {
+    return;
+  }
+  int num_iter = 80;
+
+  // Run this test three iterations.
+  // Iteration 1: only a uncompressed block cache
+  // Iteration 2: only a compressed block cache
+  // Iteration 3: both block cache and compressed cache
+  // Iteration 4: both block cache and compressed cache, but DB is not
+  // compressed
+  for (int iter = 0; iter < 4; iter++) {
+    Options options = CurrentOptions();
+    options.write_buffer_size = 64 * 1024;  // small write buffer
+    options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
+
+    BlockBasedTableOptions table_options;
+    switch (iter) {
+      case 0:
+        // only uncompressed block cache
+        table_options.block_cache = NewLRUCache(8 * 1024);
+        table_options.block_cache_compressed = nullptr;
+        options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+        break;
+      case 1:
+        // no block cache, only compressed cache
+        table_options.no_block_cache = true;
+        table_options.block_cache = nullptr;
+        table_options.block_cache_compressed = NewLRUCache(8 * 1024);
+        options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+        break;
+      case 2:
+        // both compressed and uncompressed block cache
+        table_options.block_cache = NewLRUCache(1024);
+        table_options.block_cache_compressed = NewLRUCache(8 * 1024);
+        options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+        break;
+      case 3:
+        // both block cache and compressed cache, but DB is not compressed
+        // also, make block cache sizes bigger, to trigger block cache hits
+        table_options.block_cache = NewLRUCache(1024 * 1024);
+        table_options.block_cache_compressed = NewLRUCache(8 * 1024 * 1024);
+        options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+        options.compression = kNoCompression;
+        break;
+      default:
+        FAIL();
+    }
+    CreateAndReopenWithCF({"pikachu"}, options);
+    // default column family doesn't have block cache
+    Options no_block_cache_opts;
+    no_block_cache_opts.statistics = options.statistics;
+    no_block_cache_opts = CurrentOptions(no_block_cache_opts);
+    BlockBasedTableOptions table_options_no_bc;
+    table_options_no_bc.no_block_cache = true;
+    no_block_cache_opts.table_factory.reset(
+        NewBlockBasedTableFactory(table_options_no_bc));
+    ReopenWithColumnFamilies(
+        {"default", "pikachu"},
+        std::vector<Options>({no_block_cache_opts, options}));
+
+    Random rnd(301);
+
+    // Write 8MB (80 values, each 100K)
+    ASSERT_EQ(NumTableFilesAtLevel(0, 1), 0);
+    std::vector<std::string> values;
+    std::string str;
+    for (int i = 0; i < num_iter; i++) {
+      if (i % 4 == 0) {  // high compression ratio
+        str = rnd.RandomString(1000);
+      }
+      values.push_back(str);
+      ASSERT_OK(Put(1, Key(i), values[i]));
+    }
+
+    // flush all data from memtable so that reads are from block cache
+    ASSERT_OK(Flush(1));
+
+    for (int i = 0; i < num_iter; i++) {
+      ASSERT_EQ(Get(1, Key(i)), values[i]);
+    }
+
+    // check that we triggered the appropriate code paths in the cache
+    switch (iter) {
+      case 0:
+        // only uncompressed block cache
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
+        ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
+        break;
+      case 1:
+        // no block cache, only compressed cache
+        ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
+        break;
+      case 2:
+        // both compressed and uncompressed block cache
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
+        break;
+      case 3:
+        // both compressed and uncompressed block cache
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_MISS), 0);
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_HIT), 0);
+        ASSERT_GT(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_MISS), 0);
+        // compressed doesn't have any hits since blocks are not compressed on
+        // storage
+        ASSERT_EQ(TestGetTickerCount(options, BLOCK_CACHE_COMPRESSED_HIT), 0);
+        break;
+      default:
+        FAIL();
+    }
+
+    options.create_if_missing = true;
+    DestroyAndReopen(options);
+  }
+}
+
 TEST_F(DBBlockCacheTest, CacheCompressionDict) {
  const int kNumFiles = 4;
  const int kNumEntriesPerFile = 128;
@@ -1334,6 +1652,7 @@ TEST_F(DBBlockCacheTest, HyperClockCacheReportProblems) {
  EXPECT_EQ(logger->PopCounts(), (std::array<int, 3>{{0, 1, 0}}));
 }

+#endif  // ROCKSDB_LITE

 class DBBlockCacheKeyTest
    : public DBTestBase,
@@ -1389,16 +1708,31 @@ TEST_P(DBBlockCacheKeyTest, StableCacheKeys) {
  uint64_t expected_stat = 0;

  std::function<void()> verify_stats;
-  table_options.cache_index_and_filter_blocks = true;
-  table_options.block_cache = NewLRUCache(1 << 25, 0, false);
-  verify_stats = [&options, &expected_stat] {
-    ASSERT_EQ(expected_stat,
-              options.statistics->getTickerCount(BLOCK_CACHE_DATA_ADD));
-    ASSERT_EQ(expected_stat,
-              options.statistics->getTickerCount(BLOCK_CACHE_INDEX_ADD));
-    ASSERT_EQ(expected_stat,
-              options.statistics->getTickerCount(BLOCK_CACHE_FILTER_ADD));
-  };
+  if (use_compressed_cache_) {
+    if (!Snappy_Supported()) {
+      ROCKSDB_GTEST_SKIP("Compressed cache test requires snappy support");
+      return;
+    }
+    options.compression = CompressionType::kSnappyCompression;
+    table_options.no_block_cache = true;
+    table_options.block_cache_compressed = NewLRUCache(1 << 25, 0, false);
+    verify_stats = [&options, &expected_stat] {
+      // One for ordinary SST file and one for external SST file
+      ASSERT_EQ(expected_stat,
+                options.statistics->getTickerCount(BLOCK_CACHE_COMPRESSED_ADD));
+    };
+  } else {
+    table_options.cache_index_and_filter_blocks = true;
+    table_options.block_cache = NewLRUCache(1 << 25, 0, false);
+    verify_stats = [&options, &expected_stat] {
+      ASSERT_EQ(expected_stat,
+                options.statistics->getTickerCount(BLOCK_CACHE_DATA_ADD));
+      ASSERT_EQ(expected_stat,
+                options.statistics->getTickerCount(BLOCK_CACHE_INDEX_ADD));
+      ASSERT_EQ(expected_stat,
+                options.statistics->getTickerCount(BLOCK_CACHE_FILTER_ADD));
+    };
+  }

  table_options.filter_policy.reset(NewBloomFilterPolicy(10, false));
  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
@@ -1440,6 +1774,7 @@ TEST_P(DBBlockCacheKeyTest, StableCacheKeys) {
    ++key_count;
  }

+#ifndef ROCKSDB_LITE
  // Save an export of those ordinary SST files for later
  std::string export_files_dir = dbname_ + "/exported";
  ExportImportFilesMetaData* metadata_ptr_ = nullptr;
@@ -1468,6 +1803,13 @@ TEST_P(DBBlockCacheKeyTest, StableCacheKeys) {
    ASSERT_OK(db_->IngestExternalFile(handles_[1], {f}, ingest_opts));
  }

+  if (exclude_file_numbers_) {
+    // FIXME(peterd): figure out where these extra ADDs are coming from
+    options.statistics->recordTick(BLOCK_CACHE_COMPRESSED_ADD,
+                                   uint64_t{0} - uint64_t{2});
+  }
+#endif
+
  perform_gets();
  verify_stats();

@@ -1480,6 +1822,7 @@ TEST_P(DBBlockCacheKeyTest, StableCacheKeys) {
  // Make sure we can cache hit even on a full copy of the DB. Using
  // StableCacheKeyTestFS, Checkpoint will resort to full copy not hard link.
  // (Checkpoint  not available in LITE mode to test this.)
+#ifndef ROCKSDB_LITE
  auto db_copy_name = dbname_ + "-copy";
  ASSERT_OK(Checkpoint::Create(db_, &checkpoint));
  ASSERT_OK(checkpoint->CreateCheckpoint(db_copy_name));
@@ -1518,6 +1861,7 @@ TEST_P(DBBlockCacheKeyTest, StableCacheKeys) {

  perform_gets();
  verify_stats();
+#endif  // !ROCKSDB_LITE

  Close();
  Destroy(options);
@@ -245,12 +245,14 @@ TEST_F(DBBloomFilterTest, GetFilterByPrefixBloomCustomPrefixExtractor) {
        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);

    // No bloom on extractor changed
+#ifndef ROCKSDB_LITE
    ASSERT_OK(db_->SetOptions({{"prefix_extractor", "capped:10"}}));
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
    ASSERT_EQ(
        3,
        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+#endif  // ROCKSDB_LITE

    // No bloom on extractor changed, after re-open
    options.prefix_extractor.reset(NewCappedPrefixTransform(10));
@@ -315,12 +317,14 @@ TEST_F(DBBloomFilterTest, GetFilterByPrefixBloom) {
        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);

    // No bloom on extractor changed
+#ifndef ROCKSDB_LITE
    ASSERT_OK(db_->SetOptions({{"prefix_extractor", "capped:10"}}));
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
    ASSERT_EQ(
        3,
        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+#endif  // ROCKSDB_LITE

    get_perf_context()->Reset();
  }
@@ -560,6 +564,7 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {
      ASSERT_LE(reads, 3 * N / 100);
    }

+#ifndef ROCKSDB_LITE
    // Sanity check some table properties
    std::map<std::string, std::string> props;
    ASSERT_TRUE(db_->GetMapProperty(
@@ -578,6 +583,7 @@ TEST_P(DBBloomFilterTestWithParam, BloomFilter) {

    uint64_t num_filter_entries = ParseUint64(props["num_filter_entries"]);
    EXPECT_EQ(num_filter_entries, nkeys);
+#endif  // ROCKSDB_LITE

    env_->delay_sstable_sync_.store(false, std::memory_order_release);
    Close();
@@ -643,8 +649,10 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
    PutFn();
    GetFn();
  };
+#ifndef ROCKSDB_LITE
  std::map<std::string, std::string> props;
  const auto& kAggTableProps = DB::Properties::kAggregatedTableProperties;
+#endif  // ROCKSDB_LITE

  Options options = CurrentOptions();
  options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
@@ -667,9 +675,11 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE), 0);
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE), 0);

+#ifndef ROCKSDB_LITE
  props.clear();
  ASSERT_TRUE(db_->GetMapProperty(kAggTableProps, &props));
  EXPECT_EQ(props["filter_size"], "0");
+#endif  // ROCKSDB_LITE

  // Test 2: use custom API to skip filters -> no filter constructed
  // or read.
@@ -683,9 +693,11 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE), 0);
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE), 0);

+#ifndef ROCKSDB_LITE
  props.clear();
  ASSERT_TRUE(db_->GetMapProperty(kAggTableProps, &props));
  EXPECT_EQ(props["filter_size"], "0");
+#endif  // ROCKSDB_LITE

  // Control test: using an actual filter with 100% FP rate -> the filter
  // is constructed and checked on read.
@@ -701,9 +713,11 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  EXPECT_EQ(
      TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE),
      maxKey);
+#ifndef ROCKSDB_LITE
  props.clear();
  ASSERT_TRUE(db_->GetMapProperty(kAggTableProps, &props));
  EXPECT_NE(props["filter_size"], "0");
+#endif  // ROCKSDB_LITE

  // Test 3 (options test): Able to read existing filters with longstanding
  // generated options file entry `filter_policy=rocksdb.BuiltinBloomFilter`
@@ -729,9 +743,11 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE), 0);
  EXPECT_EQ(TestGetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE), 0);

+#ifndef ROCKSDB_LITE
  props.clear();
  ASSERT_TRUE(db_->GetMapProperty(kAggTableProps, &props));
  EXPECT_EQ(props["filter_size"], "0");
+#endif  // ROCKSDB_LITE
 }

 #if !defined(ROCKSDB_VALGRIND_RUN) || defined(ROCKSDB_FULL_VALGRIND_RUN)
@@ -1487,6 +1503,7 @@ TEST_P(DBFilterConstructionCorruptionTestWithParam, DetectCorruption) {
 }

 // RocksDB lite does not support dynamic options
+#ifndef ROCKSDB_LITE
 TEST_P(DBFilterConstructionCorruptionTestWithParam,
       DynamicallyTurnOnAndOffDetectConstructCorruption) {
  Options options = CurrentOptions();
@@ -1570,6 +1587,7 @@ TEST_P(DBFilterConstructionCorruptionTestWithParam,
      db_->GetOptions().table_factory->GetOptions<BlockBasedTableOptions>();
  EXPECT_FALSE(updated_table_options->detect_filter_construct_corruption);
 }
+#endif  // ROCKSDB_LITE

 namespace {
 // NOTE: This class is referenced by HISTORY.md as a model for a wrapper
@@ -1738,6 +1756,7 @@ TEST_F(DBBloomFilterTest, ContextCustomFilterPolicy) {
        EXPECT_LE(useful_count, maxKey * 0.91);
      }
    } else {
+#ifndef ROCKSDB_LITE
      // Also try external SST file
      {
        std::string file_path = dbname_ + "/external.sst";
@@ -1749,6 +1768,7 @@ TEST_F(DBBloomFilterTest, ContextCustomFilterPolicy) {
      // Note: kCompactionStyleLevel is default, ignored if num_levels == -1
      EXPECT_EQ(policy->DumpTestReport(),
                "cf=abe,s=kCompactionStyleLevel,n=-1,l=-1,b=0,r=kMisc\n");
+#endif
    }

    // Destroy
@@ -2016,7 +2036,9 @@ TEST_P(DBBloomFilterTestVaryPrefixAndFormatVer, PartitionedMultiGet) {
    ASSERT_OK(Put(UKey(i), UKey(i)));
  }
  ASSERT_OK(Flush());
+#ifndef ROCKSDB_LITE
  ASSERT_EQ(TotalTableFiles(), 1);
+#endif

  constexpr uint32_t Q = 29;
  // MultiGet In
@@ -2168,6 +2190,7 @@ INSTANTIATE_TEST_CASE_P(DBBloomFilterTestVaryPrefixAndFormatVer,
                            std::make_tuple(true, 3), std::make_tuple(true, 4),
                            std::make_tuple(true, 5)));

+#ifndef ROCKSDB_LITE
 namespace {
 static const std::string kPlainTable = "test_PlainTableBloom";
 }  // anonymous namespace
@@ -3463,6 +3486,7 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter3) {
  }
 }

+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE

@@ -285,6 +285,7 @@ class ChangeFilterFactory : public CompactionFilterFactory {
  const char* Name() const override { return "ChangeFilterFactory"; }
 };

+#ifndef ROCKSDB_LITE
 TEST_F(DBTestCompactionFilter, CompactionFilter) {
  Options options = CurrentOptions();
  options.max_open_files = -1;
@@ -468,6 +469,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterDeletesAll) {

  delete itr;
 }
+#endif  // ROCKSDB_LITE

 TEST_F(DBTestCompactionFilter, CompactionFilterFlush) {
  // Tests a `CompactionFilterFactory` that filters when table file is created
@@ -653,6 +655,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterWithMergeOperator) {
  ASSERT_EQ(newvalue, four);
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
  KeepFilterFactory* filter = new KeepFilterFactory(true, true);

@@ -712,6 +715,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
    ASSERT_EQ(count, 0);
  }
 }
+#endif  // ROCKSDB_LITE

 TEST_F(DBTestCompactionFilter, CompactionFilterContextCfId) {
  KeepFilterFactory* filter = new KeepFilterFactory(false, true);
@@ -742,6 +746,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextCfId) {
  ASSERT_TRUE(filter->compaction_filter_created());
 }

+#ifndef ROCKSDB_LITE
 // Compaction filters aplies to all records, regardless snapshots.
 TEST_F(DBTestCompactionFilter, CompactionFilterIgnoreSnapshot) {
  std::string five = std::to_string(5);
@@ -802,6 +807,7 @@ TEST_F(DBTestCompactionFilter, CompactionFilterIgnoreSnapshot) {
  // removed.
  db_->ReleaseSnapshot(snapshot);
 }
+#endif  // ROCKSDB_LITE

 TEST_F(DBTestCompactionFilter, SkipUntil) {
  Options options = CurrentOptions();
@@ -30,6 +30,7 @@
 namespace ROCKSDB_NAMESPACE {

 // SYNC_POINT is not supported in released Windows mode.
+#if !defined(ROCKSDB_LITE)

 class CompactionStatsCollector : public EventListener {
 public:
@@ -9108,11 +9109,18 @@ TEST_F(DBCompactionTest, BottommostFileCompactionAllowIngestBehind) {
  // ASSERT_OK(dbfull()->TEST_WaitForCompact(true /* wait_unscheduled */));
 }

+#endif  // !defined(ROCKSDB_LITE)

 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
+#if !defined(ROCKSDB_LITE)
  ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
  ::testing::InitGoogleTest(&argc, argv);
  return RUN_ALL_TESTS();
+#else
+  (void)argc;
+  (void)argv;
+  return 0;
+#endif
 }
@@ -10,6 +10,7 @@
 // Introduction of SyncPoint effectively disabled building and running this test
 // in Release build.
 // which is a pity, it is a good test
+#if !defined(ROCKSDB_LITE)

 #include "db/db_test_util.h"
 #include "port/port.h"
@@ -491,9 +492,16 @@ TEST_F(DBTestDynamicLevel, DISABLED_MigrateToDynamicLevelMaxBytesBase) {
 }
 }  // namespace ROCKSDB_NAMESPACE

+#endif  // !defined(ROCKSDB_LITE)

 int main(int argc, char** argv) {
+#if !defined(ROCKSDB_LITE)
  ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
  ::testing::InitGoogleTest(&argc, argv);
  return RUN_ALL_TESTS();
+#else
+  (void)argc;
+  (void)argv;
+  return 0;
+#endif
 }
@@ -6,7 +6,9 @@
 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
 #include "rocksdb/perf_context.h"
+#if !defined(ROCKSDB_LITE)
 #include "test_util/sync_point.h"
+#endif
 #include <iostream>
 #include <string>

@@ -25,6 +27,7 @@ class DBEncryptionTest : public DBTestBase {
  }
 };

+#ifndef ROCKSDB_LITE

 TEST_F(DBEncryptionTest, CheckEncrypted) {
  ASSERT_OK(Put("foo567", "v1.fetdq"));
@@ -116,6 +119,7 @@ TEST_F(DBEncryptionTest, ReadEmptyFile) {
  ASSERT_TRUE(data.empty());
 }

+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE

@@ -4,6 +4,7 @@
 //  (found in the LICENSE.Apache file in the root directory).
 //

+#ifndef ROCKSDB_LITE

 #include <algorithm>
 #include <cstdint>
@@ -34,8 +35,11 @@ Status DBImpl::FlushForGetLiveFiles() {
  // flush all dirty data to disk.
  Status status;
  if (immutable_db_options_.atomic_flush) {
+    autovector<ColumnFamilyData*> cfds;
+    SelectColumnFamiliesForAtomicFlush(&cfds);
    mutex_.Unlock();
-    status = AtomicFlushMemTables(FlushOptions(), FlushReason::kGetLiveFiles);
+    status =
+        AtomicFlushMemTables(cfds, FlushOptions(), FlushReason::kGetLiveFiles);
    if (status.IsColumnFamilyDropped()) {
      status = Status::OK();
    }
@@ -434,3 +438,5 @@ Status DBImpl::GetLiveFilesStorageInfo(
 }

 }  // namespace ROCKSDB_NAMESPACE
+
+#endif  // ROCKSDB_LITE
@@ -73,7 +73,9 @@ TEST_F(DBFlushTest, FlushWhileWritingManifest) {
  ASSERT_OK(dbfull()->Flush(no_wait));
  // If the issue is hit we will wait here forever.
  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
+#ifndef ROCKSDB_LITE
  ASSERT_EQ(2, TotalTableFiles());
+#endif  // ROCKSDB_LITE
 }

 // Disable this test temporarily on Travis as it fails intermittently.
@@ -103,7 +105,9 @@ TEST_F(DBFlushTest, SyncFail) {
  // Now the background job will do the flush; wait for it.
  // Returns the IO error happend during flush.
  ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
+#ifndef ROCKSDB_LITE
  ASSERT_EQ("", FilesPerLevel());  // flush failed.
+#endif                             // ROCKSDB_LITE
  Destroy(options);
 }

@@ -660,6 +664,7 @@ TEST_F(DBFlushTest, StatisticsGarbageRangeDeletes) {
  Close();
 }

+#ifndef ROCKSDB_LITE
 // This simple Listener can only handle one flush at a time.
 class TestFlushListener : public EventListener {
 public:
@@ -739,98 +744,10 @@ class TestFlushListener : public EventListener {
  Env* env_;
  DBFlushTest* test_;
 };
+#endif  // !ROCKSDB_LITE

-TEST_F(
-    DBFlushTest,
-    FixUnrecoverableWriteDuringAtomicFlushWaitUntilFlushWouldNotStallWrites) {
-  Options options = CurrentOptions();
-  options.atomic_flush = true;
-
-  // To simulate a real-life crash where we can't flush during db's shutdown
-  options.avoid_flush_during_shutdown = true;
-
-  // Set 3 low thresholds (while `disable_auto_compactions=false`) here so flush
-  // adding one more L0 file during `GetLiveFiles()` will have to wait till such
-  // flush will not stall writes
-  options.level0_stop_writes_trigger = 2;
-  options.level0_slowdown_writes_trigger = 2;
-  // Disable level-0 compaction triggered by number of files to avoid
-  // stalling check being skipped (resulting in the flush mentioned above didn't
-  // wait)
-  options.level0_file_num_compaction_trigger = -1;
-
-  CreateAndReopenWithCF({"cf1"}, options);
-
-  // Manually pause compaction thread to ensure enough L0 files as
-  // `disable_auto_compactions=false`is needed, in order to meet the 3 low
-  // thresholds above
-  std::unique_ptr<test::SleepingBackgroundTask> sleeping_task_;
-  sleeping_task_.reset(new test::SleepingBackgroundTask());
-  env_->SetBackgroundThreads(1, Env::LOW);
-  env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask,
-                 sleeping_task_.get(), Env::Priority::LOW);
-  sleeping_task_->WaitUntilSleeping();
-
-  // Create some initial file to help meet the 3 low thresholds above
-  ASSERT_OK(Put(1, "dontcare", "dontcare"));
-  ASSERT_OK(Flush(1));
-
-  // Insert some initial data so we have something to atomic-flush later
-  // triggered by `GetLiveFiles()`
-  WriteOptions write_opts;
-  write_opts.disableWAL = true;
-  ASSERT_OK(Put(1, "k1", "v1", write_opts));
-
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency({{
-      "DBImpl::WaitUntilFlushWouldNotStallWrites:StallWait",
-      "DBFlushTest::"
-      "UnrecoverableWriteInAtomicFlushWaitUntilFlushWouldNotStallWrites::Write",
-  }});
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
-
-  // Write to db when atomic flush releases the lock to wait on write stall
-  // condition to be gone in `WaitUntilFlushWouldNotStallWrites()`
-  port::Thread write_thread([&] {
-    TEST_SYNC_POINT(
-        "DBFlushTest::"
-        "UnrecoverableWriteInAtomicFlushWaitUntilFlushWouldNotStallWrites::"
-        "Write");
-    // Before the fix, the empty default CF would've been prematurely excluded
-    // from this atomic flush. The following two writes together make default CF
-    // later contain data that should've been included in the atomic flush.
-    ASSERT_OK(Put(0, "k2", "v2", write_opts));
-    // The following write increases the max seqno of this atomic flush to be 3,
-    // which is greater than the seqno of default CF's data. This then violates
-    // the invariant that all entries of seqno less than the max seqno
-    // of this atomic flush should've been flushed by the time of this atomic
-    // flush finishes.
-    ASSERT_OK(Put(1, "k3", "v3", write_opts));
-
-    // Resume compaction threads and reduce L0 files so `GetLiveFiles()` can
-    // resume from the wait
-    sleeping_task_->WakeUp();
-    sleeping_task_->WaitUntilDone();
-    MoveFilesToLevel(1, 1);
-  });
-
-  // Trigger an atomic flush by `GetLiveFiles()`
-  std::vector<std::string> files;
-  uint64_t manifest_file_size;
-  ASSERT_OK(db_->GetLiveFiles(files, &manifest_file_size, /*flush*/ true));
-
-  write_thread.join();
-
-  ReopenWithColumnFamilies({"default", "cf1"}, options);
-
-  ASSERT_EQ(Get(1, "k3"), "v3");
-  // Prior to the fix, `Get()` will return `NotFound as "k2" entry in default CF
-  // can't be recovered from a crash right after the atomic flush finishes,
-  // resulting in a "recovery hole" as "k3" can be recovered. It's due to the
-  // invariant violation described above.
-  ASSERT_EQ(Get(0, "k2"), "v2");
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
-}
-
+// RocksDB lite does not support GetLiveFiles()
+#ifndef ROCKSDB_LITE
 TEST_F(DBFlushTest, FixFlushReasonRaceFromConcurrentFlushes) {
  Options options = CurrentOptions();
  options.atomic_flush = true;
@@ -885,6 +802,7 @@ TEST_F(DBFlushTest, FixFlushReasonRaceFromConcurrentFlushes) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBFlushTest, MemPurgeBasic) {
  Options options = CurrentOptions();
@@ -917,16 +835,24 @@ TEST_F(DBFlushTest, MemPurgeBasic) {

  // Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
  options.write_buffer_size = 1 << 20;
+#ifndef ROCKSDB_LITE
  // Initially deactivate the MemPurge prototype.
  options.experimental_mempurge_threshold = 0.0;
  TestFlushListener* listener = new TestFlushListener(options.env, this);
  options.listeners.emplace_back(listener);
+#else
+  // Activate directly the MemPurge prototype.
+  // (RocksDB lite does not support dynamic options)
+  options.experimental_mempurge_threshold = 1.0;
+#endif  // !ROCKSDB_LITE
  ASSERT_OK(TryReopen(options));

  // RocksDB lite does not support dynamic options
+#ifndef ROCKSDB_LITE
  // Dynamically activate the MemPurge prototype without restarting the DB.
  ColumnFamilyHandle* cfh = db_->DefaultColumnFamily();
  ASSERT_OK(db_->SetOptions(cfh, {{"experimental_mempurge_threshold", "1.0"}}));
+#endif

  std::atomic<uint32_t> mempurge_count{0};
  std::atomic<uint32_t> sst_count{0};
@@ -1059,6 +985,7 @@ TEST_F(DBFlushTest, MemPurgeBasic) {
 }

 // RocksDB lite does not support dynamic options
+#ifndef ROCKSDB_LITE
 TEST_F(DBFlushTest, MemPurgeBasicToggle) {
  Options options = CurrentOptions();

@@ -1171,10 +1098,12 @@ TEST_F(DBFlushTest, MemPurgeBasicToggle) {

  Close();
 }
+// Closes the "#ifndef ROCKSDB_LITE"
 // End of MemPurgeBasicToggle, which is not
 // supported with RocksDB LITE because it
 // relies on dynamically changing the option
 // flag experimental_mempurge_threshold.
+#endif

 // At the moment, MemPurge feature is deactivated
 // when atomic_flush is enabled. This is because the level
@@ -1292,8 +1221,10 @@ TEST_F(DBFlushTest, MemPurgeDeleteAndDeleteRange) {
  options.compression = kNoCompression;
  options.inplace_update_support = false;
  options.allow_concurrent_memtable_write = true;
+#ifndef ROCKSDB_LITE
  TestFlushListener* listener = new TestFlushListener(options.env, this);
  options.listeners.emplace_back(listener);
+#endif  // !ROCKSDB_LITE
  // Enforce size of a single MemTable to 64MB (64MB = 67108864 bytes).
  options.write_buffer_size = 1 << 20;
  // Activate the MemPurge prototype.
@@ -1491,8 +1422,10 @@ TEST_F(DBFlushTest, MemPurgeAndCompactionFilter) {
  options.compression = kNoCompression;
  options.inplace_update_support = false;
  options.allow_concurrent_memtable_write = true;
+#ifndef ROCKSDB_LITE
  TestFlushListener* listener = new TestFlushListener(options.env, this);
  options.listeners.emplace_back(listener);
+#endif  // !ROCKSDB_LITE
  // Create a ConditionalUpdate compaction filter
  // that will update all the values of the KV pairs
  // where the keys are "lower" than KEY4.
@@ -1945,10 +1878,12 @@ TEST_F(DBFlushTest, ManualFlushFailsInReadOnlyMode) {
  ASSERT_OK(db_->ContinueBackgroundWork());
  // We ingested the error to env, so the returned status is not OK.
  ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
+#ifndef ROCKSDB_LITE
  uint64_t num_bg_errors;
  ASSERT_TRUE(
      db_->GetIntProperty(DB::Properties::kBackgroundErrors, &num_bg_errors));
  ASSERT_GT(num_bg_errors, 0);
+#endif  // ROCKSDB_LITE

  // In the bug scenario, triggering another flush would cause the second flush
  // to hang forever. After the fix we expect it to return an error.
@@ -1990,6 +1925,7 @@ TEST_F(DBFlushTest, CFDropRaceWithWaitForFlushMemTables) {
  SyncPoint::GetInstance()->DisableProcessing();
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
  class TestListener : public EventListener {
   public:
@@ -2080,6 +2016,7 @@ TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
  SyncPoint::GetInstance()->DisableProcessing();
  SyncPoint::GetInstance()->ClearAllCallBacks();
 }
+#endif  // !ROCKSDB_LITE

 TEST_F(DBFlushTest, FlushWithBlob) {
  constexpr uint64_t min_blob_size = 10;
@@ -2141,6 +2078,7 @@ TEST_F(DBFlushTest, FlushWithBlob) {

  ASSERT_EQ(blob_file->GetTotalBlobCount(), 1);

+#ifndef ROCKSDB_LITE
  const InternalStats* const internal_stats = cfd->internal_stats();
  assert(internal_stats);

@@ -2156,6 +2094,7 @@ TEST_F(DBFlushTest, FlushWithBlob) {
  ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED],
            compaction_stats[0].bytes_written +
                compaction_stats[0].bytes_written_blob);
+#endif  // ROCKSDB_LITE
 }

 TEST_F(DBFlushTest, FlushWithChecksumHandoff1) {
@@ -2469,6 +2408,7 @@ TEST_P(DBFlushTestBlobError, FlushError) {
    ASSERT_NE(type, kBlobFile);
  }

+#ifndef ROCKSDB_LITE
  const InternalStats* const internal_stats = cfd->internal_stats();
  assert(internal_stats);

@@ -2492,8 +2432,10 @@ TEST_P(DBFlushTestBlobError, FlushError) {
  ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED],
            compaction_stats[0].bytes_written +
                compaction_stats[0].bytes_written_blob);
+#endif  // ROCKSDB_LITE
 }

+#ifndef ROCKSDB_LITE
 TEST_F(DBFlushTest, TombstoneVisibleInSnapshot) {
  class SimpleTestFlushListener : public EventListener {
   public:
@@ -2674,6 +2616,7 @@ TEST_P(DBAtomicFlushTest, ManualAtomicFlush) {
    ASSERT_TRUE(cfh->cfd()->mem()->IsEmpty());
  }
 }
+#endif  // ROCKSDB_LITE

 TEST_P(DBAtomicFlushTest, PrecomputeMinLogNumberToKeepNon2PC) {
  Options options = CurrentOptions();
@@ -3,6 +3,7 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).

+#ifndef ROCKSDB_LITE
 #include "db/db_impl/compacted_db_impl.h"

 #include "db/db_impl/db_impl.h"
@@ -253,3 +254,4 @@ Status CompactedDBImpl::Open(const Options& options, const std::string& dbname,
 }

 }  // namespace ROCKSDB_NAMESPACE
+#endif  // ROCKSDB_LITE
@@ -4,6 +4,7 @@
 //  (found in the LICENSE.Apache file in the root directory).

 #pragma once
+#ifndef ROCKSDB_LITE
 #include <string>
 #include <vector>

@@ -132,10 +133,12 @@ class CompactedDBImpl : public DBImpl {
  // Share with DBImplReadOnly?

 protected:
+#ifndef ROCKSDB_LITE
  Status FlushForGetLiveFiles() override {
    // No-op for read-only DB
    return Status::OK();
  }
+#endif  // !ROCKSDB_LITE

 private:
  friend class DB;
@@ -148,3 +151,4 @@ class CompactedDBImpl : public DBImpl {
  LevelFilesBrief files_;
 };
 }  // namespace ROCKSDB_NAMESPACE
+#endif  // ROCKSDB_LITE
@@ -213,13 +213,17 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
      has_unpersisted_data_(false),
      unable_to_release_oldest_log_(false),
      num_running_ingest_file_(0),
+#ifndef ROCKSDB_LITE
      wal_manager_(immutable_db_options_, file_options_, io_tracer_,
                   seq_per_batch),
+#endif  // ROCKSDB_LITE
      bg_work_paused_(0),
      bg_compaction_paused_(0),
      refitting_level_(false),
      opened_successfully_(false),
+#ifndef ROCKSDB_LITE
      periodic_task_scheduler_(),
+#endif  // ROCKSDB_LITE
      two_write_queues_(options.two_write_queues),
      manual_wal_flush_(options.manual_wal_flush),
      // last_sequencee_ is always maintained by the main queue that also writes
@@ -243,8 +247,7 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
      atomic_flush_install_cv_(&mutex_),
      blob_callback_(immutable_db_options_.sst_file_manager.get(), &mutex_,
                     &error_handler_, &event_logger_,
-                     immutable_db_options_.listeners, dbname_),
-      lock_wal_count_(0) {
+                     immutable_db_options_.listeners, dbname_) {
  // !batch_per_trx_ implies seq_per_batch_ because it is only unset for
  // WriteUnprepared, which should use seq_per_batch_.
  assert(batch_per_txn_ || seq_per_batch_);
@@ -262,6 +265,7 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
  SetDbSessionId();
  assert(!db_session_id_.empty());

+#ifndef ROCKSDB_LITE
  periodic_task_functions_.emplace(PeriodicTaskType::kDumpStats,
                                   [this]() { this->DumpStats(); });
  periodic_task_functions_.emplace(PeriodicTaskType::kPersistStats,
@@ -271,6 +275,7 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
  periodic_task_functions_.emplace(
      PeriodicTaskType::kRecordSeqnoTime,
      [this]() { this->RecordSeqnoToTimeMapping(); });
+#endif  // ROCKSDB_LITE

  versions_.reset(new VersionSet(dbname_, &immutable_db_options_, file_options_,
                                 table_cache_.get(), write_buffer_manager_,
@@ -387,8 +392,10 @@ Status DBImpl::ResumeImpl(DBRecoverContext context) {
    // We allow flush to stall write since we are trying to resume from error.
    flush_opts.allow_write_stall = true;
    if (immutable_db_options_.atomic_flush) {
+      autovector<ColumnFamilyData*> cfds;
+      SelectColumnFamiliesForAtomicFlush(&cfds);
      mutex_.Unlock();
-      s = AtomicFlushMemTables(flush_opts, context.flush_reason);
+      s = AtomicFlushMemTables(cfds, flush_opts, context.flush_reason);
      mutex_.Lock();
    } else {
      for (auto cfd : versions_->GetRefedColumnFamilySet()) {
@@ -489,6 +496,7 @@ void DBImpl::CancelAllBackgroundWork(bool wait) {
  ROCKS_LOG_INFO(immutable_db_options_.info_log,
                 "Shutdown: canceling all background work");

+#ifndef ROCKSDB_LITE
  for (uint8_t task_type = 0;
       task_type < static_cast<uint8_t>(PeriodicTaskType::kMax); task_type++) {
    Status s = periodic_task_scheduler_.Unregister(
@@ -499,14 +507,18 @@ void DBImpl::CancelAllBackgroundWork(bool wait) {
                     task_type, s.ToString().c_str());
    }
  }
+#endif  // !ROCKSDB_LITE

  InstrumentedMutexLock l(&mutex_);
  if (!shutting_down_.load(std::memory_order_acquire) &&
      has_unpersisted_data_.load(std::memory_order_relaxed) &&
      !mutable_db_options_.avoid_flush_during_shutdown) {
    if (immutable_db_options_.atomic_flush) {
+      autovector<ColumnFamilyData*> cfds;
+      SelectColumnFamiliesForAtomicFlush(&cfds);
      mutex_.Unlock();
-      Status s = AtomicFlushMemTables(FlushOptions(), FlushReason::kShutDown);
+      Status s =
+          AtomicFlushMemTables(cfds, FlushOptions(), FlushReason::kShutDown);
      s.PermitUncheckedError();  //**TODO: What to do on error?
      mutex_.Lock();
    } else {
@@ -692,6 +704,7 @@ Status DBImpl::CloseHelper() {
  ROCKS_LOG_INFO(immutable_db_options_.info_log, "Shutdown complete");
  LogFlush(immutable_db_options_.info_log);

+#ifndef ROCKSDB_LITE
  // If the sst_file_manager was allocated by us during DB::Open(), ccall
  // Close() on it before closing the info_log. Otherwise, background thread
  // in SstFileManagerImpl might try to log something
@@ -700,6 +713,7 @@ Status DBImpl::CloseHelper() {
        immutable_db_options_.sst_file_manager.get());
    sfm->Close();
  }
+#endif  // ROCKSDB_LITE

  if (immutable_db_options_.info_log && own_info_log_) {
    Status s = immutable_db_options_.info_log->Close();
@@ -777,6 +791,7 @@ void DBImpl::PrintStatistics() {
 }

 Status DBImpl::StartPeriodicTaskScheduler() {
+#ifndef ROCKSDB_LITE

 #ifndef NDEBUG
  // It only used by test to disable scheduler
@@ -819,9 +834,13 @@ Status DBImpl::StartPeriodicTaskScheduler() {
      periodic_task_functions_.at(PeriodicTaskType::kFlushInfoLog));

  return s;
+#else
+  return Status::OK();
+#endif  // !ROCKSDB_LITE
 }

 Status DBImpl::RegisterRecordSeqnoTimeWorker() {
+#ifndef ROCKSDB_LITE
  uint64_t min_time_duration = std::numeric_limits<uint64_t>::max();
  uint64_t max_time_duration = std::numeric_limits<uint64_t>::min();
  {
@@ -864,6 +883,9 @@ Status DBImpl::RegisterRecordSeqnoTimeWorker() {
  }

  return s;
+#else
+  return Status::OK();
+#endif  // !ROCKSDB_LITE
 }

 // esitmate the total size of stats_history_
@@ -884,6 +906,7 @@ size_t DBImpl::EstimateInMemoryStatsHistorySize() const {

 void DBImpl::PersistStats() {
  TEST_SYNC_POINT("DBImpl::PersistStats:Entry");
+#ifndef ROCKSDB_LITE
  if (shutdown_initiated_) {
    return;
  }
@@ -988,6 +1011,7 @@ void DBImpl::PersistStats() {
                   stats_history_size, stats_history_.size());
  }
  TEST_SYNC_POINT("DBImpl::PersistStats:End");
+#endif  // !ROCKSDB_LITE
 }

 bool DBImpl::FindStatsByTime(uint64_t start_time, uint64_t end_time,
@@ -1029,6 +1053,7 @@ Status DBImpl::GetStatsHistory(

 void DBImpl::DumpStats() {
  TEST_SYNC_POINT("DBImpl::DumpStats:1");
+#ifndef ROCKSDB_LITE
  std::string stats;
  if (shutdown_initiated_) {
    return;
@@ -1093,6 +1118,7 @@ void DBImpl::DumpStats() {
      ROCKS_LOG_INFO(immutable_db_options_.info_log, "%s", stats.c_str());
    }
  }
+#endif  // !ROCKSDB_LITE

  PrintStatistics();
 }
@@ -1147,6 +1173,11 @@ FSDirectory* DBImpl::GetDataDir(ColumnFamilyData* cfd, size_t path_id) const {
 Status DBImpl::SetOptions(
    ColumnFamilyHandle* column_family,
    const std::unordered_map<std::string, std::string>& options_map) {
+#ifdef ROCKSDB_LITE
+  (void)column_family;
+  (void)options_map;
+  return Status::NotSupported("Not supported in ROCKSDB LITE");
+#else
  auto* cfd =
      static_cast_with_check<ColumnFamilyHandleImpl>(column_family)->cfd();
  if (options_map.empty()) {
@@ -1204,10 +1235,15 @@ Status DBImpl::SetOptions(
  }
  LogFlush(immutable_db_options_.info_log);
  return s;
+#endif  // ROCKSDB_LITE
 }

 Status DBImpl::SetDBOptions(
    const std::unordered_map<std::string, std::string>& options_map) {
+#ifdef ROCKSDB_LITE
+  (void)options_map;
+  return Status::NotSupported("Not supported in ROCKSDB LITE");
+#else
  if (options_map.empty()) {
    ROCKS_LOG_WARN(immutable_db_options_.info_log,
                   "SetDBOptions(), empty input.");
@@ -1372,6 +1408,7 @@ Status DBImpl::SetDBOptions(
  }
  LogFlush(immutable_db_options_.info_log);
  return s;
+#endif  // ROCKSDB_LITE
 }

 // return the same level if it cannot be moved
@@ -1425,10 +1462,15 @@ Status DBImpl::FlushWAL(bool sync) {
  return SyncWAL();
 }

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

@@ -1530,83 +1572,29 @@ Status DBImpl::ApplyWALToManifest(VersionEdit* synced_wals) {
 Status DBImpl::LockWAL() {
  {
    InstrumentedMutexLock lock(&mutex_);
-    if (lock_wal_count_ > 0) {
-      assert(lock_wal_write_token_);
-      ++lock_wal_count_;
-    } else {
-      // NOTE: this will "unnecessarily" wait for other non-LockWAL() write
-      // stalls to clear before LockWAL returns, however fixing that would
-      // not be simple because if we notice the primary queue is already
-      // stalled, that stall might clear while we release DB mutex in
-      // EnterUnbatched() for the nonmem queue. And if we work around that in
-      // the naive way, we could deadlock by locking the two queues in different
-      // orders.
-
-      WriteThread::Writer w;
-      write_thread_.EnterUnbatched(&w, &mutex_);
-      WriteThread::Writer nonmem_w;
-      if (two_write_queues_) {
-        nonmem_write_thread_.EnterUnbatched(&nonmem_w, &mutex_);
-      }
-
-      // NOTE: releasing mutex in EnterUnbatched might mean we are actually
-      // now lock_wal_count > 0
-      if (lock_wal_count_ == 0) {
-        assert(!lock_wal_write_token_);
-        lock_wal_write_token_ = write_controller_.GetStopToken();
-      }
-      ++lock_wal_count_;
-
-      if (two_write_queues_) {
-        nonmem_write_thread_.ExitUnbatched(&nonmem_w);
-      }
-      write_thread_.ExitUnbatched(&w);
+    WriteThread::Writer w;
+    write_thread_.EnterUnbatched(&w, &mutex_);
+    WriteThread::Writer nonmem_w;
+    if (two_write_queues_) {
+      nonmem_write_thread_.EnterUnbatched(&nonmem_w, &mutex_);
    }
+
+    lock_wal_write_token_ = write_controller_.GetStopToken();
+
+    if (two_write_queues_) {
+      nonmem_write_thread_.ExitUnbatched(&nonmem_w);
+    }
+    write_thread_.ExitUnbatched(&w);
  }
-  // NOTE: avoid I/O holding DB mutex
-  Status s = FlushWAL(/*sync=*/false);
-  if (!s.ok()) {
-    // Non-OK return should not be in locked state
-    UnlockWAL().PermitUncheckedError();
-  }
-  return s;
+  return FlushWAL(/*sync=*/false);
 }

 Status DBImpl::UnlockWAL() {
-  bool signal = false;
-  uint64_t maybe_stall_begun_count = 0;
-  uint64_t nonmem_maybe_stall_begun_count = 0;
  {
    InstrumentedMutexLock lock(&mutex_);
-    if (lock_wal_count_ == 0) {
-      return Status::Aborted("No LockWAL() in effect");
-    }
-    --lock_wal_count_;
-    if (lock_wal_count_ == 0) {
-      lock_wal_write_token_.reset();
-      signal = true;
-      // For the last UnlockWAL, we don't want to return from UnlockWAL()
-      // until the thread(s) that called BeginWriteStall() have had a chance to
-      // call EndWriteStall(), so that no_slowdown writes after UnlockWAL() are
-      // guaranteed to succeed if there's no other source of stall.
-      maybe_stall_begun_count = write_thread_.GetBegunCountOfOutstandingStall();
-      if (two_write_queues_) {
-        nonmem_maybe_stall_begun_count =
-            nonmem_write_thread_.GetBegunCountOfOutstandingStall();
-      }
-    }
-  }
-  if (signal) {
-    // SignalAll outside of mutex for efficiency
-    bg_cv_.SignalAll();
-  }
-  // Ensure stalls have cleared
-  if (maybe_stall_begun_count) {
-    write_thread_.WaitForStallEndedCount(maybe_stall_begun_count);
-  }
-  if (nonmem_maybe_stall_begun_count) {
-    nonmem_write_thread_.WaitForStallEndedCount(nonmem_maybe_stall_begun_count);
+    lock_wal_write_token_.reset();
  }
+  bg_cv_.SignalAll();
  return Status::OK();
 }

@@ -2600,25 +2588,14 @@ void DBImpl::MultiGet(const ReadOptions& read_options, const size_t num_keys,
                      ColumnFamilyHandle** column_families, const Slice* keys,
                      PinnableSlice* values, Status* statuses,
                      const bool sorted_input) {
-  MultiGet(read_options, num_keys, column_families, keys, values,
-           /* timestamps */ nullptr, statuses, sorted_input);
+  return MultiGet(read_options, num_keys, column_families, keys, values,
+                  /*timestamps=*/nullptr, statuses, sorted_input);
 }

 void DBImpl::MultiGet(const ReadOptions& read_options, const size_t num_keys,
                      ColumnFamilyHandle** column_families, const Slice* keys,
                      PinnableSlice* values, std::string* timestamps,
                      Status* statuses, const bool sorted_input) {
-  MultiGetCommon(read_options, num_keys, column_families, keys, values,
-                 /* columns */ nullptr, timestamps, statuses, sorted_input);
-}
-
-void DBImpl::MultiGetCommon(const ReadOptions& read_options,
-                            const size_t num_keys,
-                            ColumnFamilyHandle** column_families,
-                            const Slice* keys, PinnableSlice* values,
-                            PinnableWideColumns* columns,
-                            std::string* timestamps, Status* statuses,
-                            const bool sorted_input) {
  if (num_keys == 0) {
    return;
  }
@@ -2664,20 +2641,8 @@ void DBImpl::MultiGetCommon(const ReadOptions& read_options,
  autovector<KeyContext*, MultiGetContext::MAX_BATCH_SIZE> sorted_keys;
  sorted_keys.resize(num_keys);
  for (size_t i = 0; i < num_keys; ++i) {
-    PinnableSlice* val = nullptr;
-    PinnableWideColumns* col = nullptr;
-
-    if (values) {
-      val = &values[i];
-      val->Reset();
-    } else {
-      assert(columns);
-
-      col = &columns[i];
-      col->Reset();
-    }
-
-    key_context.emplace_back(column_families[i], keys[i], val, col,
+    values[i].Reset();
+    key_context.emplace_back(column_families[i], keys[i], &values[i],
                             timestamps ? &timestamps[i] : nullptr,
                             &statuses[i]);
  }
@@ -2801,8 +2766,8 @@ void DBImpl::MultiGet(const ReadOptions& read_options,
                      ColumnFamilyHandle* column_family, const size_t num_keys,
                      const Slice* keys, PinnableSlice* values,
                      Status* statuses, const bool sorted_input) {
-  MultiGet(read_options, column_family, num_keys, keys, values,
-           /* timestamps */ nullptr, statuses, sorted_input);
+  return MultiGet(read_options, column_family, num_keys, keys, values,
+                  /*timestamp=*/nullptr, statuses, sorted_input);
 }

 void DBImpl::MultiGet(const ReadOptions& read_options,
@@ -2810,16 +2775,6 @@ void DBImpl::MultiGet(const ReadOptions& read_options,
                      const Slice* keys, PinnableSlice* values,
                      std::string* timestamps, Status* statuses,
                      const bool sorted_input) {
-  MultiGetCommon(read_options, column_family, num_keys, keys, values,
-                 /* columns */ nullptr, timestamps, statuses, sorted_input);
-}
-
-void DBImpl::MultiGetCommon(const ReadOptions& read_options,
-                            ColumnFamilyHandle* column_family,
-                            const size_t num_keys, const Slice* keys,
-                            PinnableSlice* values, PinnableWideColumns* columns,
-                            std::string* timestamps, Status* statuses,
-                            bool sorted_input) {
  if (tracer_) {
    // TODO: This mutex should be removed later, to improve performance when
    // tracing is enabled.
@@ -2833,20 +2788,8 @@ void DBImpl::MultiGetCommon(const ReadOptions& read_options,
  autovector<KeyContext*, MultiGetContext::MAX_BATCH_SIZE> sorted_keys;
  sorted_keys.resize(num_keys);
  for (size_t i = 0; i < num_keys; ++i) {
-    PinnableSlice* val = nullptr;
-    PinnableWideColumns* col = nullptr;
-
-    if (values) {
-      val = &values[i];
-      val->Reset();
-    } else {
-      assert(columns);
-
-      col = &columns[i];
-      col->Reset();
-    }
-
-    key_context.emplace_back(column_family, keys[i], val, col,
+    values[i].Reset();
+    key_context.emplace_back(column_family, keys[i], &values[i],
                             timestamps ? &timestamps[i] : nullptr,
                             &statuses[i]);
  }
@@ -3008,17 +2951,8 @@ Status DBImpl::MultiGetImpl(
  uint64_t bytes_read = 0;
  for (size_t i = start_key; i < start_key + num_keys - keys_left; ++i) {
    KeyContext* key = (*sorted_keys)[i];
-    assert(key);
-    assert(key->s);
-
    if (key->s->ok()) {
-      if (key->value) {
-        bytes_read += key->value->size();
-      } else {
-        assert(key->columns);
-        bytes_read += key->columns->serialized_size();
-      }
-
+      bytes_read += key->value->size();
      num_found++;
    }
  }
@@ -3042,22 +2976,6 @@ Status DBImpl::MultiGetImpl(
  return s;
 }

-void DBImpl::MultiGetEntity(const ReadOptions& options, size_t num_keys,
-                            ColumnFamilyHandle** column_families,
-                            const Slice* keys, PinnableWideColumns* results,
-                            Status* statuses, bool sorted_input) {
-  MultiGetCommon(options, num_keys, column_families, keys, /* values */ nullptr,
-                 results, /* timestamps */ nullptr, statuses, sorted_input);
-}
-
-void DBImpl::MultiGetEntity(const ReadOptions& options,
-                            ColumnFamilyHandle* column_family, size_t num_keys,
-                            const Slice* keys, PinnableWideColumns* results,
-                            Status* statuses, bool sorted_input) {
-  MultiGetCommon(options, column_family, num_keys, keys, /* values */ nullptr,
-                 results, /* timestamps */ nullptr, statuses, sorted_input);
-}
-
 Status DBImpl::CreateColumnFamily(const ColumnFamilyOptions& cf_options,
                                  const std::string& column_family,
                                  ColumnFamilyHandle** handle) {
@@ -3377,6 +3295,11 @@ Iterator* DBImpl::NewIterator(const ReadOptions& read_options,
  assert(cfd != nullptr);
  ReadCallback* read_callback = nullptr;  // No read callback provided.
  if (read_options.tailing) {
+#ifdef ROCKSDB_LITE
+    // not supported in lite version
+    result = nullptr;
+
+#else
    SuperVersion* sv = cfd->GetReferencedSuperVersion(this);
    auto iter = new ForwardIterator(this, read_options, cfd, sv,
                                    /* allow_unprepared_value */ true);
@@ -3385,6 +3308,7 @@ Iterator* DBImpl::NewIterator(const ReadOptions& read_options,
        cfd->user_comparator(), iter, sv->current, kMaxSequenceNumber,
        sv->mutable_cf_options.max_sequential_skip_in_iterations, read_callback,
        this, cfd);
+#endif
  } else {
    // Note: no need to consider the special case of
    // last_seq_same_as_publish_seq_==false since NewIterator is overridden in
@@ -3515,6 +3439,10 @@ Status DBImpl::NewIterators(
  iterators->clear();
  iterators->reserve(column_families.size());
  if (read_options.tailing) {
+#ifdef ROCKSDB_LITE
+    return Status::InvalidArgument(
+        "Tailing iterator not supported in RocksDB lite");
+#else
    for (auto cfh : column_families) {
      auto cfd = static_cast_with_check<ColumnFamilyHandleImpl>(cfh)->cfd();
      SuperVersion* sv = cfd->GetReferencedSuperVersion(this);
@@ -3526,6 +3454,7 @@ Status DBImpl::NewIterators(
          sv->mutable_cf_options.max_sequential_skip_in_iterations,
          read_callback, this, cfd));
    }
+#endif
  } else {
    // Note: no need to consider the special case of
    // last_seq_same_as_publish_seq_==false since NewIterators is overridden in
@@ -3547,9 +3476,11 @@ Status DBImpl::NewIterators(

 const Snapshot* DBImpl::GetSnapshot() { return GetSnapshotImpl(false); }

+#ifndef ROCKSDB_LITE
 const Snapshot* DBImpl::GetSnapshotForWriteConflictBoundary() {
  return GetSnapshotImpl(true);
 }
+#endif  // ROCKSDB_LITE

 std::pair<Status, std::shared_ptr<const Snapshot>>
 DBImpl::CreateTimestampedSnapshot(SequenceNumber snapshot_seq, uint64_t ts) {
@@ -3797,6 +3728,7 @@ void DBImpl::ReleaseSnapshot(const Snapshot* s) {
  delete casted_s;
 }

+#ifndef ROCKSDB_LITE
 Status DBImpl::GetPropertiesOfAllTables(ColumnFamilyHandle* column_family,
                                        TablePropertiesCollection* props) {
  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
@@ -3840,6 +3772,7 @@ Status DBImpl::GetPropertiesOfTablesInRange(ColumnFamilyHandle* column_family,
  return s;
 }

+#endif  // ROCKSDB_LITE

 const std::string& DBImpl::GetName() const { return dbname_; }

@@ -3858,6 +3791,7 @@ SystemClock* DBImpl::GetSystemClock() const {
  return immutable_db_options_.clock;
 }

+#ifndef ROCKSDB_LITE

 Status DBImpl::StartIOTrace(const TraceOptions& trace_options,
                            std::unique_ptr<TraceWriter>&& trace_writer) {
@@ -3871,6 +3805,7 @@ Status DBImpl::EndIOTrace() {
  return Status::OK();
 }

+#endif  // ROCKSDB_LITE

 Options DBImpl::GetOptions(ColumnFamilyHandle* column_family) const {
  InstrumentedMutexLock l(&mutex_);
@@ -3999,6 +3934,7 @@ bool DBImpl::GetPropertyHandleOptionsStatistics(std::string* value) {
  return true;
 }

+#ifndef ROCKSDB_LITE
 Status DBImpl::ResetStats() {
  InstrumentedMutexLock l(&mutex_);
  for (auto* cfd : *versions_->GetColumnFamilySet()) {
@@ -4008,6 +3944,7 @@ Status DBImpl::ResetStats() {
  }
  return Status::OK();
 }
+#endif  // ROCKSDB_LITE

 bool DBImpl::GetAggregatedIntProperty(const Slice& property,
                                      uint64_t* aggregated_value) {
@@ -4216,6 +4153,7 @@ void DBImpl::ReleaseFileNumberFromPendingOutputs(
  }
 }

+#ifndef ROCKSDB_LITE
 Status DBImpl::GetUpdatesSince(
    SequenceNumber seq, std::unique_ptr<TransactionLogIterator>* iter,
    const TransactionLogIterator::ReadOptions& read_options) {
@@ -4464,6 +4402,7 @@ void DBImpl::GetAllColumnFamilyMetaData(
  }
 }

+#endif  // ROCKSDB_LITE

 Status DBImpl::CheckConsistency() {
  mutex_.AssertHeld();
@@ -4814,6 +4753,7 @@ Status DestroyDB(const std::string& dbname, const Options& options,

 Status DBImpl::WriteOptionsFile(bool need_mutex_lock,
                                bool need_enter_write_thread) {
+#ifndef ROCKSDB_LITE
  WriteThread::Writer w;
  if (need_mutex_lock) {
    mutex_.Lock();
@@ -4870,9 +4810,14 @@ Status DBImpl::WriteOptionsFile(bool need_mutex_lock,
                             s.ToString().c_str());
    }
  }
+#else
+  (void)need_mutex_lock;
+  (void)need_enter_write_thread;
+#endif  // !ROCKSDB_LITE
  return Status::OK();
 }

+#ifndef ROCKSDB_LITE
 namespace {
 void DeleteOptionsFilesHelper(const std::map<uint64_t, std::string>& filenames,
                              const size_t num_files_to_keep,
@@ -4890,8 +4835,10 @@ void DeleteOptionsFilesHelper(const std::map<uint64_t, std::string>& filenames,
  }
 }
 }  // namespace
+#endif  // !ROCKSDB_LITE

 Status DBImpl::DeleteObsoleteOptionsFiles() {
+#ifndef ROCKSDB_LITE
  std::vector<std::string> filenames;
  // use ordered map to store keep the filenames sorted from the newest
  // to the oldest.
@@ -4919,9 +4866,13 @@ Status DBImpl::DeleteObsoleteOptionsFiles() {
  DeleteOptionsFilesHelper(options_filenames, kNumOptionsFilesKept,
                           immutable_db_options_.info_log, GetEnv());
  return Status::OK();
+#else
+  return Status::OK();
+#endif  // !ROCKSDB_LITE
 }

 Status DBImpl::RenameTempFileToOptionsFile(const std::string& file_name) {
+#ifndef ROCKSDB_LITE
  Status s;

  uint64_t options_file_number = versions_->NewFileNumber();
@@ -4965,6 +4916,10 @@ Status DBImpl::RenameTempFileToOptionsFile(const std::string& file_name) {
    DeleteObsoleteOptionsFiles().PermitUncheckedError();
  }
  return s;
+#else
+  (void)file_name;
+  return Status::OK();
+#endif  // !ROCKSDB_LITE
 }

 #ifdef ROCKSDB_USING_THREAD_STATUS
@@ -5012,6 +4967,7 @@ void DumpRocksDBBuildVersion(Logger* log) {
  }
 }

+#ifndef ROCKSDB_LITE
 SequenceNumber DBImpl::GetEarliestMemTableSequenceNumber(SuperVersion* sv,
                                                         bool include_history) {
  // Find the earliest sequence number that we know we can rely on reading
@@ -5345,10 +5301,12 @@ Status DBImpl::IngestExternalFiles(
      FlushOptions flush_opts;
      flush_opts.allow_write_stall = true;
      if (immutable_db_options_.atomic_flush) {
+        autovector<ColumnFamilyData*> cfds_to_flush;
+        SelectColumnFamiliesForAtomicFlush(&cfds_to_flush);
        mutex_.Unlock();
-        status = AtomicFlushMemTables(
-            flush_opts, FlushReason::kExternalFileIngestion,
-            {} /* provided_candidate_cfds */, true /* entered_write_thread */);
+        status = AtomicFlushMemTables(cfds_to_flush, flush_opts,
+                                      FlushReason::kExternalFileIngestion,
+                                      true /* entered_write_thread */);
        mutex_.Lock();
      } else {
        for (size_t i = 0; i != num_cfs; ++i) {
@@ -5966,5 +5924,6 @@ void DBImpl::RecordSeqnoToTimeMapping() {
                   seqno, unix_time);
  }
 }
+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE
@@ -54,9 +54,13 @@
 #include "rocksdb/env.h"
 #include "rocksdb/memtablerep.h"
 #include "rocksdb/status.h"
+#ifndef ROCKSDB_LITE
 #include "rocksdb/trace_reader_writer.h"
+#endif  // ROCKSDB_LITE
 #include "rocksdb/transaction_log.h"
+#ifndef ROCKSDB_LITE
 #include "rocksdb/utilities/replayer.h"
+#endif  // ROCKSDB_LITE
 #include "rocksdb/write_buffer_manager.h"
 #include "table/merging_iterator.h"
 #include "table/scoped_arena_iterator.h"
@@ -277,40 +281,33 @@ class DBImpl : public DB {
  // 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
-  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 MultiGet(const ReadOptions& options, ColumnFamilyHandle* column_family,
-                const size_t num_keys, const Slice* keys, PinnableSlice* values,
-                std::string* timestamps, Status* statuses,
-                const bool sorted_input = false) override;
+  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;
+  virtual void MultiGet(const ReadOptions& options,
+                        ColumnFamilyHandle* column_family,
+                        const size_t num_keys, const Slice* keys,
+                        PinnableSlice* values, std::string* timestamps,
+                        Status* statuses,
+                        const bool sorted_input = false) override;

-  void MultiGet(const ReadOptions& options, const size_t num_keys,
-                ColumnFamilyHandle** column_families, const Slice* keys,
-                PinnableSlice* values, Status* statuses,
-                const bool sorted_input = false) override;
-  void MultiGet(const ReadOptions& options, const size_t num_keys,
-                ColumnFamilyHandle** column_families, const Slice* keys,
-                PinnableSlice* values, std::string* timestamps,
-                Status* statuses, const bool sorted_input = false) override;
+  virtual void MultiGet(const ReadOptions& options, const size_t num_keys,
+                        ColumnFamilyHandle** column_families, const Slice* keys,
+                        PinnableSlice* values, Status* statuses,
+                        const bool sorted_input = false) override;
+  virtual void MultiGet(const ReadOptions& options, const size_t num_keys,
+                        ColumnFamilyHandle** column_families, const Slice* keys,
+                        PinnableSlice* values, std::string* timestamps,
+                        Status* statuses,
+                        const bool sorted_input = false) override;

-  void MultiGetWithCallback(
+  virtual void MultiGetWithCallback(
      const ReadOptions& options, ColumnFamilyHandle* column_family,
      ReadCallback* callback,
      autovector<KeyContext*, MultiGetContext::MAX_BATCH_SIZE>* sorted_keys);

-  using DB::MultiGetEntity;
-
-  void MultiGetEntity(const ReadOptions& options,
-                      ColumnFamilyHandle* column_family, size_t num_keys,
-                      const Slice* keys, PinnableWideColumns* results,
-                      Status* statuses, bool sorted_input) override;
-
-  void MultiGetEntity(const ReadOptions& options, size_t num_keys,
-                      ColumnFamilyHandle** column_families, const Slice* keys,
-                      PinnableWideColumns* results, Status* statuses,
-                      bool sorted_input) override;
-
  virtual Status CreateColumnFamily(const ColumnFamilyOptions& cf_options,
                                    const std::string& column_family,
                                    ColumnFamilyHandle** handle) override;
@@ -433,7 +430,7 @@ class DBImpl : public DB {
      const FlushOptions& options,
      const std::vector<ColumnFamilyHandle*>& column_families) override;
  virtual Status FlushWAL(bool sync) override;
-  bool WALBufferIsEmpty();
+  bool WALBufferIsEmpty(bool lock = true);
  virtual Status SyncWAL() override;
  virtual Status LockWAL() override;
  virtual Status UnlockWAL() override;
@@ -472,6 +469,7 @@ class DBImpl : public DB {
      uint64_t start_time, uint64_t end_time,
      std::unique_ptr<StatsHistoryIterator>* stats_iterator) override;

+#ifndef ROCKSDB_LITE
  using DB::ResetStats;
  virtual Status ResetStats() override;
  // All the returned filenames start with "/"
@@ -599,6 +597,7 @@ class DBImpl : public DB {
      ColumnFamilyHandle* column_family, const Range* range, std::size_t n,
      TablePropertiesCollection* props) override;

+#endif  // ROCKSDB_LITE

  // ---- End of implementations of the DB interface ----
  SystemClock* GetSystemClock() const;
@@ -656,6 +655,7 @@ class DBImpl : public DB {
  // depends also on data written to the WAL but not to the memtable.
  SequenceNumber TEST_GetLastVisibleSequence() const;

+#ifndef ROCKSDB_LITE
  // Similar to Write() but will call the callback once on the single write
  // thread to determine whether it is safe to perform the write.
  virtual Status WriteWithCallback(const WriteOptions& write_options,
@@ -720,6 +720,7 @@ class DBImpl : public DB {
  Status TraceIteratorSeekForPrev(const uint32_t& cf_id, const Slice& key,
                                  const Slice& lower_bound,
                                  const Slice upper_bound);
+#endif  // ROCKSDB_LITE

  // Similar to GetSnapshot(), but also lets the db know that this snapshot
  // will be used for transaction write-conflict checking.  The DB can then
@@ -1081,9 +1082,8 @@ class DBImpl : public DB {
  // is because in certain cases, we can flush column families, wait for the
  // flush to complete, but delete the column family handle before the wait
  // finishes. For example in CompactRange.
-  Status TEST_AtomicFlushMemTables(
-      const autovector<ColumnFamilyData*>& provided_candidate_cfds,
-      const FlushOptions& flush_opts);
+  Status TEST_AtomicFlushMemTables(const autovector<ColumnFamilyData*>& cfds,
+                                   const FlushOptions& flush_opts);

  // Wait for background threads to complete scheduled work.
  Status TEST_WaitForBackgroundWork();
@@ -1125,8 +1125,6 @@ class DBImpl : public DB {

  void TEST_UnlockMutex();

-  void TEST_SignalAllBgCv();
-
  // REQUIRES: mutex locked
  void* TEST_BeginWrite();

@@ -1178,7 +1176,9 @@ class DBImpl : public DB {
    return files_grabbed_for_purge_;
  }

+#ifndef ROCKSDB_LITE
  const PeriodicTaskScheduler& TEST_GetPeriodicTaskScheduler() const;
+#endif  // !ROCKSDB_LITE

 #endif  // NDEBUG

@@ -1401,10 +1401,12 @@ class DBImpl : public DB {
  void NotifyOnMemTableSealed(ColumnFamilyData* cfd,
                              const MemTableInfo& mem_table_info);

+#ifndef ROCKSDB_LITE
  void NotifyOnExternalFileIngested(
      ColumnFamilyData* cfd, const ExternalSstFileIngestionJob& ingestion_job);

  virtual Status FlushForGetLiveFiles();
+#endif  // !ROCKSDB_LITE

  void NewThreadStatusCfInfo(ColumnFamilyData* cfd) const;

@@ -1552,7 +1554,9 @@ class DBImpl : public DB {
  friend class WriteUnpreparedTxnDB;
  friend class WriteUnpreparedTxn;

+#ifndef ROCKSDB_LITE
  friend class ForwardIterator;
+#endif
  friend struct SuperVersion;
  friend class CompactedDBImpl;
  friend class DBTest_ConcurrentFlushWAL_Test;
@@ -1887,27 +1891,16 @@ class DBImpl : public DB {

  Status SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context);

-  // Select and output column families qualified for atomic flush in
-  // `selected_cfds`. If `provided_candidate_cfds` is non-empty, it will be used
-  // as candidate CFs to select qualified ones from. Otherwise, all column
-  // families are used as candidate to select from.
-  //
-  // REQUIRES: mutex held
-  void SelectColumnFamiliesForAtomicFlush(
-      autovector<ColumnFamilyData*>* selected_cfds,
-      const autovector<ColumnFamilyData*>& provided_candidate_cfds = {});
+  void SelectColumnFamiliesForAtomicFlush(autovector<ColumnFamilyData*>* cfds);

  // Force current memtable contents to be flushed.
  Status FlushMemTable(ColumnFamilyData* cfd, const FlushOptions& options,
                       FlushReason flush_reason,
                       bool entered_write_thread = false);

-  // Atomic-flush memtables from quanlified CFs among `provided_candidate_cfds`
-  // (if non-empty) or amomg all column families and atomically record the
-  // result to the MANIFEST.
  Status AtomicFlushMemTables(
+      const autovector<ColumnFamilyData*>& column_family_datas,
      const FlushOptions& options, FlushReason flush_reason,
-      const autovector<ColumnFamilyData*>& provided_candidate_cfds = {},
      bool entered_write_thread = false);

  // Wait until flushing this column family won't stall writes
@@ -2030,6 +2023,7 @@ class DBImpl : public DB {
  // Used by WriteImpl to update bg_error_ in case of memtable insert error.
  void MemTableInsertStatusCheck(const Status& memtable_insert_status);

+#ifndef ROCKSDB_LITE
  Status CompactFilesImpl(const CompactionOptions& compact_options,
                          ColumnFamilyData* cfd, Version* version,
                          const std::vector<std::string>& input_file_names,
@@ -2037,6 +2031,7 @@ class DBImpl : public DB {
                          const int output_level, int output_path_id,
                          JobContext* job_context, LogBuffer* log_buffer,
                          CompactionJobInfo* compaction_job_info);
+#endif  // ROCKSDB_LITE

  ColumnFamilyData* GetColumnFamilyDataByName(const std::string& cf_name);

@@ -2175,6 +2170,7 @@ 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,
@@ -2187,6 +2183,7 @@ class DBImpl : public DB {
      ColumnFamilyData* cfd, uint64_t num,
      std::unique_ptr<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);
@@ -2210,18 +2207,6 @@ class DBImpl : public DB {
      const size_t num_keys, bool sorted,
      autovector<KeyContext*, MultiGetContext::MAX_BATCH_SIZE>* key_ptrs);

-  void MultiGetCommon(const ReadOptions& options,
-                      ColumnFamilyHandle* column_family, const size_t num_keys,
-                      const Slice* keys, PinnableSlice* values,
-                      PinnableWideColumns* columns, std::string* timestamps,
-                      Status* statuses, bool sorted_input);
-
-  void MultiGetCommon(const ReadOptions& options, const size_t num_keys,
-                      ColumnFamilyHandle** column_families, const Slice* keys,
-                      PinnableSlice* values, PinnableWideColumns* columns,
-                      std::string* timestamps, Status* statuses,
-                      bool sorted_input);
-
  // A structure to hold the information required to process MultiGet of keys
  // belonging to one column family. For a multi column family MultiGet, there
  // will be a container of these objects.
@@ -2608,7 +2593,9 @@ class DBImpl : public DB {
  // REQUIRES: mutex held
  int num_running_ingest_file_;

+#ifndef ROCKSDB_LITE
  WalManager wal_manager_;
+#endif  // ROCKSDB_LITE

  // A value of > 0 temporarily disables scheduling of background work
  int bg_work_paused_;
@@ -2636,12 +2623,14 @@ class DBImpl : public DB {
  // Only to be set during initialization
  std::unique_ptr<PreReleaseCallback> recoverable_state_pre_release_callback_;

+#ifndef ROCKSDB_LITE
  // Scheduler to run DumpStats(), PersistStats(), and FlushInfoLog().
  // Currently, internally it has a global timer instance for running the tasks.
  PeriodicTaskScheduler periodic_task_scheduler_;

  // It contains the implementations for each periodic task.
  std::map<PeriodicTaskType, const PeriodicTaskFunc> periodic_task_functions_;
+#endif

  // When set, we use a separate queue for writes that don't write to memtable.
  // In 2PC these are the writes at Prepare phase.
@@ -2696,14 +2685,9 @@ class DBImpl : public DB {
  // thread safe, both read and write need db mutex hold.
  SeqnoToTimeMapping seqno_time_mapping_;

-  // Stop write token that is acquired when first LockWAL() is called.
-  // Destroyed when last UnlockWAL() is called. Controlled by DB mutex.
-  // See lock_wal_count_
+  // stop write token that is acquired when LockWal() is called. Destructed
+  // when UnlockWal() is called.
  std::unique_ptr<WriteControllerToken> lock_wal_write_token_;
-
-  // The number of LockWAL called without matching UnlockWAL call.
-  // See also lock_wal_write_token_
-  uint32_t lock_wal_count_;
 };

 class GetWithTimestampReadCallback : public ReadCallback {
@@ -30,6 +30,7 @@ bool DBImpl::EnoughRoomForCompaction(
    bool* sfm_reserved_compact_space, LogBuffer* log_buffer) {
  // Check if we have enough room to do the compaction
  bool enough_room = true;
+#ifndef ROCKSDB_LITE
  auto sfm = static_cast<SstFileManagerImpl*>(
      immutable_db_options_.sst_file_manager.get());
  if (sfm) {
@@ -44,6 +45,11 @@ bool DBImpl::EnoughRoomForCompaction(
      *sfm_reserved_compact_space = true;
    }
  }
+#else
+  (void)cfd;
+  (void)inputs;
+  (void)sfm_reserved_compact_space;
+#endif  // ROCKSDB_LITE
  if (!enough_room) {
    // Just in case tests want to change the value of enough_room
    TEST_SYNC_POINT_CALLBACK(
@@ -253,9 +259,11 @@ Status DBImpl::FlushMemTableToOutputFile(
  TEST_SYNC_POINT_CALLBACK(
      "DBImpl::FlushMemTableToOutputFile:AfterPickMemtables", &flush_job);

+#ifndef ROCKSDB_LITE
  // may temporarily unlock and lock the mutex.
  NotifyOnFlushBegin(cfd, &file_meta, mutable_cf_options, job_context->job_id,
                     flush_reason);
+#endif  // ROCKSDB_LITE

  bool switched_to_mempurge = false;
  // Within flush_job.Run, rocksdb may call event listener to notify
@@ -336,6 +344,7 @@ Status DBImpl::FlushMemTableToOutputFile(
  // If flush ran smoothly and no mempurge happened
  // install new SST file path.
  if (s.ok() && (!switched_to_mempurge)) {
+#ifndef ROCKSDB_LITE
    // may temporarily unlock and lock the mutex.
    NotifyOnFlushCompleted(cfd, mutable_cf_options,
                           flush_job.GetCommittedFlushJobsInfo());
@@ -358,6 +367,7 @@ Status DBImpl::FlushMemTableToOutputFile(
        error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush);
      }
    }
+#endif  // ROCKSDB_LITE
  }
  TEST_SYNC_POINT("DBImpl::FlushMemTableToOutputFile:Finish");
  return s;
@@ -414,7 +424,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
    assert(cfd->imm()->NumNotFlushed() != 0);
    assert(cfd->imm()->IsFlushPending());
  }
-  for (const auto& bg_flush_arg : bg_flush_args) {
+  for (const auto bg_flush_arg : bg_flush_args) {
    assert(bg_flush_arg.flush_reason_ == bg_flush_args[0].flush_reason_);
  }
 #endif /* !NDEBUG */
@@ -476,6 +486,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
  IOStatus log_io_s = IOStatus::OK();
  assert(num_cfs == static_cast<int>(jobs.size()));

+#ifndef ROCKSDB_LITE
  for (int i = 0; i != num_cfs; ++i) {
    const MutableCFOptions& mutable_cf_options = all_mutable_cf_options.at(i);
    // may temporarily unlock and lock the mutex.
@@ -483,6 +494,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
    NotifyOnFlushBegin(cfds[i], &file_meta[i], mutable_cf_options,
                       job_context->job_id, flush_reason);
  }
+#endif /* !ROCKSDB_LITE */

  if (logfile_number_ > 0) {
    // TODO (yanqin) investigate whether we should sync the closed logs for
@@ -691,8 +703,10 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
        mems_list.emplace_back(&mems);
        mutable_cf_options_list.emplace_back(&all_mutable_cf_options[i]);
        tmp_file_meta.emplace_back(&file_meta[i]);
+#ifndef ROCKSDB_LITE
        committed_flush_jobs_info.emplace_back(
            jobs[i]->GetCommittedFlushJobsInfo());
+#endif  //! ROCKSDB_LITE
      }
    }

@@ -744,6 +758,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
    if (made_progress) {
      *made_progress = true;
    }
+#ifndef ROCKSDB_LITE
    auto sfm = static_cast<SstFileManagerImpl*>(
        immutable_db_options_.sst_file_manager.get());
    assert(all_mutable_cf_options.size() == static_cast<size_t>(num_cfs));
@@ -774,6 +789,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
        }
      }
    }
+#endif  // ROCKSDB_LITE
  }

  // Need to undo atomic flush if something went wrong, i.e. s is not OK and
@@ -811,6 +827,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
 void DBImpl::NotifyOnFlushBegin(ColumnFamilyData* cfd, FileMetaData* file_meta,
                                const MutableCFOptions& mutable_cf_options,
                                int job_id, FlushReason flush_reason) {
+#ifndef ROCKSDB_LITE
  if (immutable_db_options_.listeners.size() == 0U) {
    return;
  }
@@ -850,11 +867,19 @@ void DBImpl::NotifyOnFlushBegin(ColumnFamilyData* cfd, FileMetaData* file_meta,
  mutex_.Lock();
 // no need to signal bg_cv_ as it will be signaled at the end of the
 // flush process.
+#else
+  (void)cfd;
+  (void)file_meta;
+  (void)mutable_cf_options;
+  (void)job_id;
+  (void)flush_reason;
+#endif  // ROCKSDB_LITE
 }

 void DBImpl::NotifyOnFlushCompleted(
    ColumnFamilyData* cfd, const MutableCFOptions& mutable_cf_options,
    std::list<std::unique_ptr<FlushJobInfo>>* flush_jobs_info) {
+#ifndef ROCKSDB_LITE
  assert(flush_jobs_info != nullptr);
  if (immutable_db_options_.listeners.size() == 0U) {
    return;
@@ -886,6 +911,11 @@ void DBImpl::NotifyOnFlushCompleted(
  mutex_.Lock();
  // no need to signal bg_cv_ as it will be signaled at the end of the
  // flush process.
+#else
+  (void)cfd;
+  (void)mutable_cf_options;
+  (void)flush_jobs_info;
+#endif  // ROCKSDB_LITE
 }

 Status DBImpl::CompactRange(const CompactRangeOptions& options,
@@ -1031,9 +1061,15 @@ Status DBImpl::CompactRangeInternal(const CompactRangeOptions& options,
    FlushOptions fo;
    fo.allow_write_stall = options.allow_write_stall;
    if (immutable_db_options_.atomic_flush) {
-      s = AtomicFlushMemTables(fo, FlushReason::kManualCompaction);
+      autovector<ColumnFamilyData*> cfds;
+      mutex_.Lock();
+      SelectColumnFamiliesForAtomicFlush(&cfds);
+      mutex_.Unlock();
+      s = AtomicFlushMemTables(cfds, fo, FlushReason::kManualCompaction,
+                               false /* entered_write_thread */);
    } else {
-      s = FlushMemTable(cfd, fo, FlushReason::kManualCompaction);
+      s = FlushMemTable(cfd, fo, FlushReason::kManualCompaction,
+                        false /* entered_write_thread */);
    }
    if (!s.ok()) {
      LogFlush(immutable_db_options_.info_log);
@@ -1262,6 +1298,17 @@ Status DBImpl::CompactFiles(const CompactionOptions& compact_options,
                            const int output_level, const int output_path_id,
                            std::vector<std::string>* const output_file_names,
                            CompactionJobInfo* compaction_job_info) {
+#ifdef ROCKSDB_LITE
+  (void)compact_options;
+  (void)column_family;
+  (void)input_file_names;
+  (void)output_level;
+  (void)output_path_id;
+  (void)output_file_names;
+  (void)compaction_job_info;
+  // not supported in lite version
+  return Status::NotSupported("Not supported in ROCKSDB LITE");
+#else
  if (column_family == nullptr) {
    return Status::InvalidArgument("ColumnFamilyHandle must be non-null.");
  }
@@ -1320,8 +1367,10 @@ Status DBImpl::CompactFiles(const CompactionOptions& compact_options,
  }

  return s;
+#endif  // ROCKSDB_LITE
 }

+#ifndef ROCKSDB_LITE
 Status DBImpl::CompactFilesImpl(
    const CompactionOptions& compact_options, ColumnFamilyData* cfd,
    Version* version, const std::vector<std::string>& input_file_names,
@@ -1463,12 +1512,14 @@ Status DBImpl::CompactFilesImpl(
  // SetBGError
  compaction_job.io_status().PermitUncheckedError();
  c->ReleaseCompactionFiles(s);
+#ifndef ROCKSDB_LITE
  // Need to make sure SstFileManager does its bookkeeping
  auto sfm = static_cast<SstFileManagerImpl*>(
      immutable_db_options_.sst_file_manager.get());
  if (sfm && sfm_reserved_compact_space) {
    sfm->OnCompactionCompletion(c.get());
  }
+#endif  // ROCKSDB_LITE

  ReleaseFileNumberFromPendingOutputs(pending_outputs_inserted_elem);

@@ -1525,6 +1576,7 @@ Status DBImpl::CompactFilesImpl(

  return status;
 }
+#endif  // ROCKSDB_LITE

 Status DBImpl::PauseBackgroundWork() {
  InstrumentedMutexLock guard_lock(&mutex_);
@@ -1558,6 +1610,7 @@ void DBImpl::NotifyOnCompactionBegin(ColumnFamilyData* cfd, Compaction* c,
                                     const Status& st,
                                     const CompactionJobStats& job_stats,
                                     int job_id) {
+#ifndef ROCKSDB_LITE
  if (immutable_db_options_.listeners.empty()) {
    return;
  }
@@ -1586,11 +1639,19 @@ void DBImpl::NotifyOnCompactionBegin(ColumnFamilyData* cfd, Compaction* c,
  }
  mutex_.Lock();
  current->Unref();
+#else
+  (void)cfd;
+  (void)c;
+  (void)st;
+  (void)job_stats;
+  (void)job_id;
+#endif  // ROCKSDB_LITE
 }

 void DBImpl::NotifyOnCompactionCompleted(
    ColumnFamilyData* cfd, Compaction* c, const Status& st,
    const CompactionJobStats& compaction_job_stats, const int job_id) {
+#ifndef ROCKSDB_LITE
  if (immutable_db_options_.listeners.size() == 0U) {
    return;
  }
@@ -1620,6 +1681,13 @@ void DBImpl::NotifyOnCompactionCompleted(
  current->Unref();
  // no need to signal bg_cv_ as it will be signaled at the end of the
  // flush process.
+#else
+  (void)cfd;
+  (void)c;
+  (void)st;
+  (void)compaction_job_stats;
+  (void)job_id;
+#endif  // ROCKSDB_LITE
 }

 // REQUIREMENT: block all background work by calling PauseBackgroundWork()
@@ -1794,8 +1862,8 @@ Status DBImpl::Flush(const FlushOptions& flush_options,
                 cfh->GetName().c_str());
  Status s;
  if (immutable_db_options_.atomic_flush) {
-    s = AtomicFlushMemTables(flush_options, FlushReason::kManualFlush,
-                             {cfh->cfd()});
+    s = AtomicFlushMemTables({cfh->cfd()}, flush_options,
+                             FlushReason::kManualFlush);
  } else {
    s = FlushMemTable(cfh->cfd(), flush_options, FlushReason::kManualFlush);
  }
@@ -1833,7 +1901,7 @@ Status DBImpl::Flush(const FlushOptions& flush_options,
                    auto cfh = static_cast<ColumnFamilyHandleImpl*>(elem);
                    cfds.emplace_back(cfh->cfd());
                  });
-    s = AtomicFlushMemTables(flush_options, FlushReason::kManualFlush, cfds);
+    s = AtomicFlushMemTables(cfds, flush_options, FlushReason::kManualFlush);
    ROCKS_LOG_INFO(immutable_db_options_.info_log,
                   "Manual atomic flush finished, status: %s\n"
                   "=====Column families:=====",
@@ -2217,9 +2285,11 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
  return s;
 }

+// Flush all elements in 'column_family_datas'
+// and atomically record the result to the MANIFEST.
 Status DBImpl::AtomicFlushMemTables(
+    const autovector<ColumnFamilyData*>& column_family_datas,
    const FlushOptions& flush_options, FlushReason flush_reason,
-    const autovector<ColumnFamilyData*>& provided_candidate_cfds,
    bool entered_write_thread) {
  assert(immutable_db_options_.atomic_flush);
  if (!flush_options.wait && write_controller_.IsStopped()) {
@@ -2229,48 +2299,18 @@ Status DBImpl::AtomicFlushMemTables(
    return Status::TryAgain(oss.str());
  }
  Status s;
-  autovector<ColumnFamilyData*> candidate_cfds;
-  if (provided_candidate_cfds.empty()) {
-    // Generate candidate cfds if not provided
-    {
-      InstrumentedMutexLock l(&mutex_);
-      for (ColumnFamilyData* cfd : *versions_->GetColumnFamilySet()) {
-        if (!cfd->IsDropped() && cfd->initialized()) {
-          cfd->Ref();
-          candidate_cfds.push_back(cfd);
-        }
-      }
-    }
-  } else {
-    candidate_cfds = provided_candidate_cfds;
-  }
-
  if (!flush_options.allow_write_stall) {
    int num_cfs_to_flush = 0;
-    for (auto cfd : candidate_cfds) {
+    for (auto cfd : column_family_datas) {
      bool flush_needed = true;
      s = WaitUntilFlushWouldNotStallWrites(cfd, &flush_needed);
      if (!s.ok()) {
-        // Unref the newly generated candidate cfds (when not provided) in
-        // `candidate_cfds`
-        if (provided_candidate_cfds.empty()) {
-          for (auto candidate_cfd : candidate_cfds) {
-            candidate_cfd->UnrefAndTryDelete();
-          }
-        }
        return s;
      } else if (flush_needed) {
        ++num_cfs_to_flush;
      }
    }
    if (0 == num_cfs_to_flush) {
-      // Unref the newly generated candidate cfds (when not provided) in
-      // `candidate_cfds`
-      if (provided_candidate_cfds.empty()) {
-        for (auto candidate_cfd : candidate_cfds) {
-          candidate_cfd->UnrefAndTryDelete();
-        }
-      }
      return s;
    }
  }
@@ -2291,16 +2331,15 @@ Status DBImpl::AtomicFlushMemTables(
    }
    WaitForPendingWrites();

-    SelectColumnFamiliesForAtomicFlush(&cfds, candidate_cfds);
-
-    // Unref the newly generated candidate cfds (when not provided) in
-    // `candidate_cfds`
-    if (provided_candidate_cfds.empty()) {
-      for (auto candidate_cfd : candidate_cfds) {
-        candidate_cfd->UnrefAndTryDelete();
+    for (auto cfd : column_family_datas) {
+      if (cfd->IsDropped()) {
+        continue;
+      }
+      if (cfd->imm()->NumNotFlushed() != 0 || !cfd->mem()->IsEmpty() ||
+          !cached_recoverable_state_empty_.load()) {
+        cfds.emplace_back(cfd);
      }
    }
-
    for (auto cfd : cfds) {
      if ((cfd->mem()->IsEmpty() && cached_recoverable_state_empty_.load()) ||
          flush_reason == FlushReason::kErrorRecoveryRetryFlush) {
@@ -2931,7 +2970,7 @@ Status DBImpl::BackgroundFlush(bool* made_progress, JobContext* job_context,
 // All the CFD/bg_flush_arg in the FlushReq must have the same flush reason, so
 // just grab the first one
 #ifndef NDEBUG
-    for (const auto& bg_flush_arg : bg_flush_args) {
+    for (const auto bg_flush_arg : bg_flush_args) {
      assert(bg_flush_arg.flush_reason_ == bg_flush_args[0].flush_reason_);
    }
 #endif /* !NDEBUG */
@@ -3562,12 +3601,14 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
    c->ReleaseCompactionFiles(status);
    *made_progress = true;

+#ifndef ROCKSDB_LITE
    // Need to make sure SstFileManager does its bookkeeping
    auto sfm = static_cast<SstFileManagerImpl*>(
        immutable_db_options_.sst_file_manager.get());
    if (sfm && sfm_reserved_compact_space) {
      sfm->OnCompactionCompletion(c.get());
    }
+#endif  // ROCKSDB_LITE

    NotifyOnCompactionCompleted(c->column_family_data(), c.get(), status,
                                compaction_job_stats, job_context->job_id);
@@ -3744,6 +3785,7 @@ bool DBImpl::MCOverlap(ManualCompactionState* m, ManualCompactionState* m1) {
  return false;
 }

+#ifndef ROCKSDB_LITE
 void DBImpl::BuildCompactionJobInfo(
    const ColumnFamilyData* cfd, Compaction* c, const Status& st,
    const CompactionJobStats& compaction_job_stats, const int job_id,
@@ -3812,6 +3854,7 @@ void DBImpl::BuildCompactionJobInfo(
        std::move(blob_file_garbage_info));
  }
 }
+#endif

 // SuperVersionContext gets created and destructed outside of the lock --
 // we use this conveniently to:
@@ -155,10 +155,8 @@ Status DBImpl::TEST_FlushMemTable(ColumnFamilyData* cfd,
 }

 Status DBImpl::TEST_AtomicFlushMemTables(
-    const autovector<ColumnFamilyData*>& provided_candidate_cfds,
-    const FlushOptions& flush_opts) {
-  return AtomicFlushMemTables(flush_opts, FlushReason::kTest,
-                              provided_candidate_cfds);
+    const autovector<ColumnFamilyData*>& cfds, const FlushOptions& flush_opts) {
+  return AtomicFlushMemTables(cfds, flush_opts, FlushReason::kTest);
 }

 Status DBImpl::TEST_WaitForBackgroundWork() {
@@ -200,8 +198,6 @@ void DBImpl::TEST_LockMutex() { mutex_.Lock(); }

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

-void DBImpl::TEST_SignalAllBgCv() { bg_cv_.SignalAll(); }
-
 void* DBImpl::TEST_BeginWrite() {
  auto w = new WriteThread::Writer();
  write_thread_.EnterUnbatched(w, &mutex_);
@@ -293,6 +289,7 @@ size_t DBImpl::TEST_GetWalPreallocateBlockSize(
  return GetWalPreallocateBlockSize(write_buffer_size);
 }

+#ifndef ROCKSDB_LITE
 void DBImpl::TEST_WaitForPeriodicTaskRun(std::function<void()> callback) const {
  periodic_task_scheduler_.TEST_WaitForRun(callback);
 }
@@ -306,6 +303,7 @@ SeqnoToTimeMapping DBImpl::TEST_GetSeqnoToTimeMapping() const {
  return seqno_time_mapping_;
 }

+#endif  // !ROCKSDB_LITE

 size_t DBImpl::TEST_EstimateInMemoryStatsHistorySize() const {
  return EstimateInMemoryStatsHistorySize();
@@ -20,6 +20,7 @@

 namespace ROCKSDB_NAMESPACE {

+#ifndef ROCKSDB_LITE
 Status DBImpl::SuggestCompactRange(ColumnFamilyHandle* column_family,
                                   const Slice* begin, const Slice* end) {
  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
@@ -153,5 +154,6 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {

  return status;
 }
+#endif  // ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE
@@ -361,6 +361,8 @@ void DBImpl::DeleteObsoleteFileImpl(int job_id, const std::string& fname,
  }
  TEST_SYNC_POINT_CALLBACK("DBImpl::DeleteObsoleteFileImpl:AfterDeletion",
                           &file_deletion_status);
+  TEST_SYNC_POINT_CALLBACK("DBImpl::DeleteObsoleteFileImpl:AfterDeletion2",
+                           const_cast<std::string*>(&fname));
  if (file_deletion_status.ok()) {
    ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
                    "[JOB %d] Delete %s type=%d #%" PRIu64 " -- %s\n", job_id,
@@ -597,11 +599,13 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
              to_delete;
    }

+#ifndef ROCKSDB_LITE
    if (type == kWalFile && (immutable_db_options_.WAL_ttl_seconds > 0 ||
                             immutable_db_options_.WAL_size_limit_MB > 0)) {
      wal_manager_.ArchiveWALFile(fname, number);
      continue;
    }
+#endif  // !ROCKSDB_LITE

    // If I do not own these files, e.g. secondary instance with max_open_files
    // = -1, then no need to delete or schedule delete these files since they
@@ -665,7 +669,9 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
      }
    }
  }
+#ifndef ROCKSDB_LITE
  wal_manager_.PurgeObsoleteWALFiles();
+#endif  // ROCKSDB_LITE
  LogFlush(immutable_db_options_.info_log);
  InstrumentedMutexLock l(&mutex_);
  --pending_purge_obsolete_files_;
@@ -153,6 +153,7 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src,
    result.avoid_flush_during_recovery = false;
  }

+#ifndef ROCKSDB_LITE
  ImmutableDBOptions immutable_db_options(result);
  if (!immutable_db_options.IsWalDirSameAsDBPath()) {
    // Either the WAL dir and db_paths[0]/db_name are not the same, or we
@@ -194,6 +195,7 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src,
        NewSstFileManager(result.env, result.info_log));
    result.sst_file_manager = sst_file_manager;
  }
+#endif  // !ROCKSDB_LITE

  // Supported wal compression types
  if (!StreamingCompressionTypeSupported(result.wal_compression)) {
@@ -216,7 +218,7 @@ Status ValidateOptionsByTable(
    const DBOptions& db_opts,
    const std::vector<ColumnFamilyDescriptor>& column_families) {
  Status s;
-  for (auto& cf : column_families) {
+  for (auto cf : column_families) {
    s = ValidateOptions(db_opts, cf.options);
    if (!s.ok()) {
      return s;
@@ -843,6 +845,7 @@ Status DBImpl::LogAndApplyForRecovery(const RecoveryContext& recovery_ctx) {
 }

 void DBImpl::InvokeWalFilterIfNeededOnColumnFamilyToWalNumberMap() {
+#ifndef ROCKSDB_LITE
  if (immutable_db_options_.wal_filter == nullptr) {
    return;
  }
@@ -860,6 +863,7 @@ void DBImpl::InvokeWalFilterIfNeededOnColumnFamilyToWalNumberMap() {
  }

  wal_filter.ColumnFamilyLogNumberMap(cf_lognumber_map, cf_name_id_map);
+#endif  // !ROCKSDB_LITE
 }

 bool DBImpl::InvokeWalFilterIfNeededOnWalRecord(uint64_t wal_number,
@@ -868,6 +872,7 @@ bool DBImpl::InvokeWalFilterIfNeededOnWalRecord(uint64_t wal_number,
                                                Status& status,
                                                bool& stop_replay,
                                                WriteBatch& batch) {
+#ifndef ROCKSDB_LITE
  if (immutable_db_options_.wal_filter == nullptr) {
    return true;
  }
@@ -953,6 +958,15 @@ bool DBImpl::InvokeWalFilterIfNeededOnWalRecord(uint64_t wal_number,
    batch = new_batch;
  }
  return true;
+#else   // !ROCKSDB_LITE
+  (void)wal_number;
+  (void)wal_fname;
+  (void)reporter;
+  (void)status;
+  (void)stop_replay;
+  (void)batch;
+  return true;
+#endif  // ROCKSDB_LITE
 }

 // REQUIRES: wal_numbers are sorted in ascending order
@@ -1967,6 +1981,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
  }
  impl->mutex_.Unlock();

+#ifndef ROCKSDB_LITE
  auto sfm = static_cast<SstFileManagerImpl*>(
      impl->immutable_db_options_.sst_file_manager.get());
  if (s.ok() && sfm) {
@@ -2050,6 +2065,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
                           impl->immutable_db_options_.db_paths[0].path);
  }

+#endif  // !ROCKSDB_LITE

  if (s.ok()) {
    ROCKS_LOG_HEADER(impl->immutable_db_options_.info_log, "DB pointer %p",
@@ -16,6 +16,7 @@

 namespace ROCKSDB_NAMESPACE {

+#ifndef ROCKSDB_LITE

 DBImplReadOnly::DBImplReadOnly(const DBOptions& db_options,
                               const std::string& dbname)
@@ -320,5 +321,21 @@ Status DBImplReadOnly::OpenForReadOnlyWithoutCheck(
  return s;
 }

+#else   // !ROCKSDB_LITE
+
+Status DB::OpenForReadOnly(const Options& /*options*/,
+                           const std::string& /*dbname*/, DB** /*dbptr*/,
+                           bool /*error_if_wal_file_exists*/) {
+  return Status::NotSupported("Not supported in ROCKSDB_LITE.");
+}
+
+Status DB::OpenForReadOnly(
+    const DBOptions& /*db_options*/, const std::string& /*dbname*/,
+    const std::vector<ColumnFamilyDescriptor>& /*column_families*/,
+    std::vector<ColumnFamilyHandle*>* /*handles*/, DB** /*dbptr*/,
+    bool /*error_if_wal_file_exists*/) {
+  return Status::NotSupported("Not supported in ROCKSDB_LITE.");
+}
+#endif  // !ROCKSDB_LITE

 }  // namespace ROCKSDB_NAMESPACE
@@ -5,6 +5,7 @@

 #pragma once

+#ifndef ROCKSDB_LITE

 #include <string>
 #include <vector>
@@ -145,10 +146,12 @@ class DBImplReadOnly : public DBImpl {
  // FIXME: some missing overrides for more "write" functions

 protected:
+#ifndef ROCKSDB_LITE
  Status FlushForGetLiveFiles() override {
    // No-op for read-only DB
    return Status::OK();
  }
+#endif  // !ROCKSDB_LITE

 private:
  // A "helper" function for DB::OpenForReadOnly without column families
@@ -164,3 +167,4 @@ class DBImplReadOnly : public DBImpl {
 };
 }  // namespace ROCKSDB_NAMESPACE

+#endif  // !ROCKSDB_LITE
@@ -17,6 +17,7 @@

 namespace ROCKSDB_NAMESPACE {

+#ifndef ROCKSDB_LITE
 DBImplSecondary::DBImplSecondary(const DBOptions& db_options,
                                 const std::string& dbname,
                                 std::string secondary_path)
@@ -945,5 +946,22 @@ Status DB::OpenAndCompact(
                        output, override_options);
 }

+#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_NAMESPACE
@@ -5,6 +5,7 @@

 #pragma once

+#ifndef ROCKSDB_LITE

 #include <string>
 #include <vector>
@@ -268,10 +269,12 @@ class DBImplSecondary : public DBImpl {
 #endif  // NDEBUG

 protected:
+#ifndef ROCKSDB_LITE
  Status FlushForGetLiveFiles() override {
    // No-op for read-only DB
    return Status::OK();
  }
+#endif  // !ROCKSDB_LITE

  // ColumnFamilyCollector is a write batch handler which does nothing
  // except recording unique column family IDs
@@ -404,3 +407,4 @@ class DBImplSecondary : public DBImpl {

 }  // namespace ROCKSDB_NAMESPACE

+#endif  // !ROCKSDB_LITE
Author	SHA1	Message	Date
anand76	3258b5c3e2	Update version.h to 7.10.2	2023-02-23 10:00:36 -08:00
anand76	3a04cd558e	Update HISTORY.md for 7.10.2	2023-02-10 14:59:29 -08:00
anand76	0ffa8db9b1	Fix bug in WAL streaming uncompression (#11198 ) Summary: Fix a bug in the calculation of the input buffer address/offset in log_reader.cc. The bug is when consecutive fragments of a compressed record are located at the same offset in the log reader buffer, the second fragment input buffer is treated as a leftover from the previous input buffer. As a result, the offset in the `ZSTD_inBuffer` is not reset. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11198 Test Plan: Add a unit test in log_test.cc that fails without the fix and passes with it. Reviewed By: ajkr, cbi42 Differential Revision: D43102692 Pulled By: anand1976 fbshipit-source-id: aa2648f4802c33991b76a3233c5a58d4cc9e77fd	2023-02-10 14:58:13 -08:00
anand76	1185bb75ca	Return any errors returned by ReadAsync to the MultiGet caller (#11171 ) Summary: Currently, we incorrectly return a Status::Corruption to the MultiGet caller if the file system ReadAsync cannot issue a read and returns an error for some reason, such as IOStatus::NotSupported(). In this PR, we copy the ReadAsync error to the request status so it can be returned to the user. Tests: Update existing unit tests and add a new one for this scenario Pull Request resolved: https://github.com/facebook/rocksdb/pull/11171 Reviewed By: akankshamahajan15 Differential Revision: D42950057 Pulled By: anand1976 fbshipit-source-id: 85ffcb015fa6c064c311f8a28488fec78c487869	2023-02-02 17:28:08 -08:00
Andrew Kryczka	d57ec3f896	update HISTORY.md and version.h for 7.10.1	2023-02-02 08:51:12 -08:00
Andrew Kryczka	8a354a1197	add release note for GetMergeOperands() fix	2023-02-01 17:05:04 -08:00
Andrew Kryczka	fcb0580b08	Fix GetMergeOperands() returning MergeInProgress (#11136 ) Summary: Pull Request resolved: https://github.com/facebook/rocksdb/pull/11136 Test Plan: the provided unit test used to fail due to `GetMergeOperands()` returning `Status::MergeInProgress()`; it passes now because the `GetMergeOperands()` call returns `Status::OK()` Reviewed By: pdillinger Differential Revision: D42759198 Pulled By: ajkr fbshipit-source-id: 878f9f40ccc1d7e2fe7b1352814bae3a49c19939	2023-02-01 17:03:31 -08:00
Andrew Kryczka	06765b5131	Allow canceling manual compaction while waiting for conflicting compaction (#11165 ) Summary: This PR adds logic to the `RunManualCompaction()` loop to check for cancellation before waiting on any conflicting compactions to finish. In case of cancellation, `RunManualCompaction()` no longer waits on conflicting compactions Pull Request resolved: https://github.com/facebook/rocksdb/pull/11165 Test Plan: repro test case Reviewed By: cbi42 Differential Revision: D42864058 Pulled By: ajkr fbshipit-source-id: ea4dd1a8f294abe212905495a8fbe8f07fca3f5a	2023-02-01 17:00:51 -08:00
Hui Xiao	fa13962e0c	Fix data race on `ColumnFamilyData::flush_reason` by letting FlushRequest/Job owns flush_reason instead of CFD (#11111 ) Summary: Context: Concurrent flushes on the same CF can set on `ColumnFamilyData::flush_reason` before each other flush finishes. An symptom is one CF has different flush_reason with others though all of them are in an atomic flush `db_stress: db/db_impl/db_impl_compaction_flush.cc:423: rocksdb::Status rocksdb::DBImpl::AtomicFlushMemTablesToOutputFiles(const rocksdb::autovector<rocksdb::DBImpl::BGFlushArg>&, bool, rocksdb::JobContext, rocksdb::LogBuffer, rocksdb::Env::Priority): Assertion cfd->GetFlushReason() == cfds[0]->GetFlushReason() failed. ` Summary:* Suggested by ltamasi, we now refactor and let FlushRequest/Job to own flush_reason as there is no good way to define `ColumnFamilyData::flush_reason` in face of concurrent flushes on the same CF (which wasn't the case a long time ago when `ColumnFamilyData::flush_reason ` first introduced`) Tets: - new unit test - make check - aggressive crash test rehearsal Pull Request resolved: https://github.com/facebook/rocksdb/pull/11111 Reviewed By: ajkr Differential Revision: D42644600 Pulled By: hx235 fbshipit-source-id: 8589c8184869d3415e5b780c887f877818a5ebaf	2023-02-01 16:59:41 -08:00
Peter Dillinger	e5dcebf756	Fix DelayWrite() calls for two_write_queues (#11130 ) Summary: PR https://github.com/facebook/rocksdb/issues/11020 fixed a case where it was easy to deadlock the DB with LockWAL() but introduced a bug showing up as a rare assertion failure in the stress test. Specifically, `assert(w->state == STATE_INIT)` in `WriteThread::LinkOne()` called from `BeginWriteStall()`, `DelayWrite()`, `WriteImplWALOnly()`. I haven't been about to generate a unit test that reproduces this failure but I believe the root cause is that DelayWrite() was never meant to be re-entrant, only called from the DB's write_thread_ leader. https://github.com/facebook/rocksdb/issues/11020 introduced a call to DelayWrite() from the nonmem_write_thread_ group leader. This fix is to make DelayWrite() apply to the specific write queue that it is being called from (inject a dummy write stall entry to the head of the appropriate write queue). WriteController is re-entrant, based on polling and state changes signalled with bg_cv_, so can manage stalling two queues. The only anticipated complication (called out by Andrew in previous PR) is that we don't want timed write delays being injected in parallel for the two queues, because that dimishes the intended throttling effect. Thus, we only allow timed delays for the primary write queue. HISTORY not updated because this is intended for the same release where the bug was introduced. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11130 Test Plan: Although I was not able to reproduce the assertion failure, I was able to reproduce a distinct flaw with what I believe is the same root cause: a kind of deadlock if both write queues need to wake up from stopped writes. Only one will be waiting on bg_cv_ (the other waiting in `LinkOne()` for the write queue to open up), so a single SignalAll() will only unblock one of the queues, with the other re-instating the stop until another signal on bg_cv_. A simple unit test is added for this case. Will also run crash_test_with_multiops_wc_txn for a while looking for issues. Reviewed By: ajkr Differential Revision: D42749330 Pulled By: pdillinger fbshipit-source-id: 4317dd899a93d57c26fd5af7143038f82d4d4d1b	2023-01-25 14:25:33 -08:00
Hui Xiao	ab389242fb	Update history for 7.10.fb	2023-01-23 11:11:36 -08:00