Summary:
**Context:**
We prefetch the tail part of a SST file (i.e, the blocks after data blocks till the end of the file) during each SST file open in hope to prefetch all the stuff at once ahead of time for later read e.g, footer, meta index, filter/index etc. The existing approach to estimate the tail size to prefetch is through `TailPrefetchStats` heuristics introduced in https://github.com/facebook/rocksdb/pull/4156, which has caused small reads in unlucky case (e.g, small read into the tail buffer during table open in thread 1 under the same BlockBasedTableFactory object can make thread 2's tail prefetching use a small size that it shouldn't) and is hard to debug. Therefore we decide to record the exact tail size and use it directly to prefetch tail of the SST instead of relying heuristics.
**Summary:**
- Obtain and record in manifest the tail size in `BlockBasedTableBuilder::Finish()`
- For backward compatibility, we fall back to TailPrefetchStats and last to simple heuristics that the tail size is a linear portion of the file size - see PR conversation for more.
- Make`tail_start_offset` part of the table properties and deduct tail size to record in manifest for external files (e.g, file ingestion, import CF) and db repair (with no access to manifest).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/11406
Test Plan:
1. New UT
2. db bench
Note: db bench on /tmp/ where direct read is supported is too slow to finish and the default pinning setting in db bench is not helpful to profile # sst read of Get. Therefore I hacked the following to obtain the following comparison.
```
diff --git a/table/block_based/block_based_table_reader.cc b/table/block_based/block_based_table_reader.cc
index bd5669f0f..791484c1f 100644
--- a/table/block_based/block_based_table_reader.cc
+++ b/table/block_based/block_based_table_reader.cc
@@ -838,7 +838,7 @@ Status BlockBasedTable::PrefetchTail(
&tail_prefetch_size);
// Try file system prefetch
- if (!file->use_direct_io() && !force_direct_prefetch) {
+ if (false && !file->use_direct_io() && !force_direct_prefetch) {
if (!file->Prefetch(prefetch_off, prefetch_len, ro.rate_limiter_priority)
.IsNotSupported()) {
prefetch_buffer->reset(new FilePrefetchBuffer(
diff --git a/tools/db_bench_tool.cc b/tools/db_bench_tool.cc
index ea40f5fa0..39a0ac385 100644
--- a/tools/db_bench_tool.cc
+++ b/tools/db_bench_tool.cc
@@ -4191,6 +4191,8 @@ class Benchmark {
std::shared_ptr<TableFactory>(NewCuckooTableFactory(table_options));
} else {
BlockBasedTableOptions block_based_options;
+ block_based_options.metadata_cache_options.partition_pinning =
+ PinningTier::kAll;
block_based_options.checksum =
static_cast<ChecksumType>(FLAGS_checksum_type);
if (FLAGS_use_hash_search) {
```
Create DB
```
./db_bench --bloom_bits=3 --use_existing_db=1 --seed=1682546046158958 --partition_index_and_filters=1 --statistics=1 -db=/dev/shm/testdb/ -benchmarks=readrandom -key_size=3200 -value_size=512 -num=1000000 -write_buffer_size=6550000 -disable_auto_compactions=false -target_file_size_base=6550000 -compression_type=none
```
ReadRandom
```
./db_bench --bloom_bits=3 --use_existing_db=1 --seed=1682546046158958 --partition_index_and_filters=1 --statistics=1 -db=/dev/shm/testdb/ -benchmarks=readrandom -key_size=3200 -value_size=512 -num=1000000 -write_buffer_size=6550000 -disable_auto_compactions=false -target_file_size_base=6550000 -compression_type=none
```
(a) Existing (Use TailPrefetchStats for tail size + use seperate prefetch buffer in PartitionedFilter/IndexReader::CacheDependencies())
```
rocksdb.table.open.prefetch.tail.hit COUNT : 3395
rocksdb.sst.read.micros P50 : 5.655570 P95 : 9.931396 P99 : 14.845454 P100 : 585.000000 COUNT : 999905 SUM : 6590614
```
(b) This PR (Record tail size + use the same tail buffer in PartitionedFilter/IndexReader::CacheDependencies())
```
rocksdb.table.open.prefetch.tail.hit COUNT : 14257
rocksdb.sst.read.micros P50 : 5.173347 P95 : 9.015017 P99 : 12.912610 P100 : 228.000000 COUNT : 998547 SUM : 5976540
```
As we can see, we increase the prefetch tail hit count and decrease SST read count with this PR
3. Test backward compatibility by stepping through reading with post-PR code on a db generated pre-PR.
Reviewed By: pdillinger
Differential Revision: D45413346
Pulled By: hx235
fbshipit-source-id: 7d5e36a60a72477218f79905168d688452a4c064
Summary:
compensate file sizes in compaction picking so files with range tombstones are preferred, such that they get compacted down earlier as they tend to delete a lot of data. This PR adds a `compensated_range_deletion_size` field in FileMeta that is computed during Flush/Compaction and persisted in MANIFEST. This value is added to `compensated_file_size` which will be used for compaction picking. Currently, for a file in level L, `compensated_range_deletion_size` is set to the estimated bytes deleted by range tombstone of this file in all levels > L. This helps to reduce space amp when data in older levels are covered by range tombstones in level L.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/10734
Test Plan:
- Added unit tests.
- benchmark to check if the above definition `compensated_range_deletion_size` is reducing space amp as intended, without affecting write amp too much. The experiment set up favorable for this optimization: large range tombstone issued infrequently. Command used:
```
./db_bench -benchmarks=fillrandom,waitforcompaction,stats,levelstats -use_existing_db=false -avoid_flush_during_recovery=true -write_buffer_size=33554432 -level_compaction_dynamic_level_bytes=true -max_background_jobs=8 -max_bytes_for_level_base=134217728 -target_file_size_base=33554432 -writes_per_range_tombstone=500000 -range_tombstone_width=5000000 -num=50000000 -benchmark_write_rate_limit=8388608 -threads=16 -duration=1800 --max_num_range_tombstones=1000000000
```
In this experiment, each thread wrote 16 range tombstones over the duration of 30 minutes, each range tombstone has width 5M that is the 10% of the key space width. Results shows this PR generates a smaller DB size.
Compaction stats from this PR:
```
Level Files Size Score Read(GB) Rn(GB) Rnp1(GB) Write(GB) Wnew(GB) Moved(GB) W-Amp Rd(MB/s) Wr(MB/s) Comp(sec) CompMergeCPU(sec) Comp(cnt) Avg(sec) KeyIn KeyDrop Rblob(GB) Wblob(GB)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
L0 2/0 31.54 MB 0.5 0.0 0.0 0.0 8.4 8.4 0.0 1.0 0.0 63.4 135.56 110.94 544 0.249 0 0 0.0 0.0
L4 3/0 96.55 MB 0.8 18.5 6.7 11.8 18.4 6.6 0.0 2.7 65.3 64.9 290.08 284.03 108 2.686 284M 1957K 0.0 0.0
L5 15/0 404.41 MB 1.0 19.1 7.7 11.4 18.8 7.4 0.3 2.5 66.6 65.7 292.93 285.34 220 1.332 293M 3808K 0.0 0.0
L6 143/0 4.12 GB 0.0 45.0 7.5 37.5 41.6 4.1 0.0 5.5 71.2 65.9 647.00 632.66 251 2.578 739M 47M 0.0 0.0
Sum 163/0 4.64 GB 0.0 82.6 21.9 60.7 87.2 26.5 0.3 10.4 61.9 65.4 1365.58 1312.97 1123 1.216 1318M 52M 0.0 0.0
```
Compaction stats from main:
```
Level Files Size Score Read(GB) Rn(GB) Rnp1(GB) Write(GB) Wnew(GB) Moved(GB) W-Amp Rd(MB/s) Wr(MB/s) Comp(sec) CompMergeCPU(sec) Comp(cnt) Avg(sec) KeyIn KeyDrop Rblob(GB) Wblob(GB)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
L0 0/0 0.00 KB 0.0 0.0 0.0 0.0 8.4 8.4 0.0 1.0 0.0 60.5 142.12 115.89 569 0.250 0 0 0.0 0.0
L4 3/0 85.68 MB 1.0 17.7 6.8 10.9 17.6 6.7 0.0 2.6 62.7 62.3 289.05 281.79 112 2.581 272M 2309K 0.0 0.0
L5 11/0 293.73 MB 1.0 18.8 7.5 11.2 18.5 7.2 0.5 2.5 64.9 63.9 296.07 288.50 220 1.346 288M 4365K 0.0 0.0
L6 130/0 3.94 GB 0.0 51.5 7.6 43.9 47.9 3.9 0.0 6.3 67.2 62.4 784.95 765.92 258 3.042 848M 51M 0.0 0.0
Sum 144/0 4.31 GB 0.0 88.0 21.9 66.0 92.3 26.3 0.5 11.0 59.6 62.5 1512.19 1452.09 1159 1.305 1409M 58M 0.0 0.0```
Reviewed By: ajkr
Differential Revision: D39834713
Pulled By: cbi42
fbshipit-source-id: fe9341040b8704a8fbb10cad5cf5c43e962c7e6b
Summary:
**Context:**
Sorting L0 files by `largest_seqno` has at least two inconvenience:
- File ingestion and compaction involving ingested files can create files of overlapping seqno range with the existing files. `force_consistency_check=true` will catch such overlap seqno range even those harmless overlap.
- For example, consider the following sequence of events ("key@n" indicates key at seqno "n")
- insert k1@1 to memtable m1
- ingest file s1 with k2@2, ingest file s2 with k3@3
- insert k4@4 to m1
- compact files s1, s2 and result in new file s3 of seqno range [2, 3]
- flush m1 and result in new file s4 of seqno range [1, 4]. And `force_consistency_check=true` will think s4 and s3 has file reordering corruption that might cause retuning an old value of k1
- However such caught corruption is a false positive since s1, s2 will not have overlapped keys with k1 or whatever inserted into m1 before ingest file s1 by the requirement of file ingestion (otherwise the m1 will be flushed first before any of the file ingestion completes). Therefore there in fact isn't any file reordering corruption.
- Single delete can decrease a file's largest seqno and ordering by `largest_seqno` can introduce a wrong ordering hence file reordering corruption
- For example, consider the following sequence of events ("key@n" indicates key at seqno "n", Credit to ajkr for this example)
- an existing SST s1 contains only k1@1
- insert k1@2 to memtable m1
- ingest file s2 with k3@3, ingest file s3 with k4@4
- insert single delete k5@5 in m1
- flush m1 and result in new file s4 of seqno range [2, 5]
- compact s1, s2, s3 and result in new file s5 of seqno range [1, 4]
- compact s4 and result in new file s6 of seqno range [2] due to single delete
- By the last step, we have file ordering by largest seqno (">" means "newer") : s5 > s6 while s6 contains a newer version of the k1's value (i.e, k1@2) than s5, which is a real reordering corruption. While this can be caught by `force_consistency_check=true`, there isn't a good way to prevent this from happening if ordering by `largest_seqno`
Therefore, we are redesigning the sorting criteria of L0 files and avoid above inconvenience. Credit to ajkr , we now introduce `epoch_num` which describes the order of a file being flushed or ingested/imported (compaction output file will has the minimum `epoch_num` among input files'). This will avoid the above inconvenience in the following ways:
- In the first case above, there will no longer be overlap seqno range check in `force_consistency_check=true` but `epoch_number` ordering check. This will result in file ordering s1 < s2 < s4 (pre-compaction) and s3 < s4 (post-compaction) which won't trigger false positive corruption. See test class `DBCompactionTestL0FilesMisorderCorruption*` for more.
- In the second case above, this will result in file ordering s1 < s2 < s3 < s4 (pre-compacting s1, s2, s3), s5 < s4 (post-compacting s1, s2, s3), s5 < s6 (post-compacting s4), which are correct file ordering without causing any corruption.
**Summary:**
- Introduce `epoch_number` stored per `ColumnFamilyData` and sort CF's L0 files by their assigned `epoch_number` instead of `largest_seqno`.
- `epoch_number` is increased and assigned upon `VersionEdit::AddFile()` for flush (or similarly for WriteLevel0TableForRecovery) and file ingestion (except for allow_behind_true, which will always get assigned as the `kReservedEpochNumberForFileIngestedBehind`)
- Compaction output file is assigned with the minimum `epoch_number` among input files'
- Refit level: reuse refitted file's epoch_number
- Other paths needing `epoch_number` treatment:
- Import column families: reuse file's epoch_number if exists. If not, assign one based on `NewestFirstBySeqNo`
- Repair: reuse file's epoch_number if exists. If not, assign one based on `NewestFirstBySeqNo`.
- Assigning new epoch_number to a file and adding this file to LSM tree should be atomic. This is guaranteed by us assigning epoch_number right upon `VersionEdit::AddFile()` where this version edit will be apply to LSM tree shape right after by holding the db mutex (e.g, flush, file ingestion, import column family) or by there is only 1 ongoing edit per CF (e.g, WriteLevel0TableForRecovery, Repair).
- Assigning the minimum input epoch number to compaction output file won't misorder L0 files (even through later `Refit(target_level=0)`). It's due to for every key "k" in the input range, a legit compaction will cover a continuous epoch number range of that key. As long as we assign the key "k" the minimum input epoch number, it won't become newer or older than the versions of this key that aren't included in this compaction hence no misorder.
- Persist `epoch_number` of each file in manifest and recover `epoch_number` on db recovery
- Backward compatibility with old db without `epoch_number` support is guaranteed by assigning `epoch_number` to recovered files by `NewestFirstBySeqno` order. See `VersionStorageInfo::RecoverEpochNumbers()` for more
- Forward compatibility with manifest is guaranteed by flexibility of `NewFileCustomTag`
- Replace `force_consistent_check` on L0 with `epoch_number` and remove false positive check like case 1 with `largest_seqno` above
- Due to backward compatibility issue, we might encounter files with missing epoch number at the beginning of db recovery. We will still use old L0 sorting mechanism (`NewestFirstBySeqno`) to check/sort them till we infer their epoch number. See usages of `EpochNumberRequirement`.
- Remove fix https://github.com/facebook/rocksdb/pull/5958#issue-511150930 and their outdated tests to file reordering corruption because such fix can be replaced by this PR.
- Misc:
- update existing tests with `epoch_number` so make check will pass
- update https://github.com/facebook/rocksdb/pull/5958#issue-511150930 tests to verify corruption is fixed using `epoch_number` and cover universal/fifo compaction/CompactRange/CompactFile cases
- assert db_mutex is held for a few places before calling ColumnFamilyData::NewEpochNumber()
Pull Request resolved: https://github.com/facebook/rocksdb/pull/10922
Test Plan:
- `make check`
- New unit tests under `db/db_compaction_test.cc`, `db/db_test2.cc`, `db/version_builder_test.cc`, `db/repair_test.cc`
- Updated tests (i.e, `DBCompactionTestL0FilesMisorderCorruption*`) under https://github.com/facebook/rocksdb/pull/5958#issue-511150930
- [Ongoing] Compatibility test: manually run 36a5686ec0 (with file ingestion off for running the `.orig` binary to prevent this bug affecting upgrade/downgrade formality checking) for 1 hour on `simple black/white box`, `cf_consistency/txn/enable_ts with whitebox + test_best_efforts_recovery with blackbox`
- [Ongoing] normal db stress test
- [Ongoing] db stress test with aggressive value https://github.com/facebook/rocksdb/pull/10761
Reviewed By: ajkr
Differential Revision: D41063187
Pulled By: hx235
fbshipit-source-id: 826cb23455de7beaabe2d16c57682a82733a32a9
Summary:
Ran `find ./db/ -type f | xargs clang-format -i`. Excluded minor changes it tried to make on db/db_impl/. Everything else it changed was directly under db/ directory. Included minor manual touchups mentioned in PR commit history.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/10910
Reviewed By: riversand963
Differential Revision: D40880683
Pulled By: ajkr
fbshipit-source-id: cfe26cda05b3fb9a72e3cb82c286e21d8c5c4174
Summary:
Instead of existing calls to ps from gnu_parallel, call a new wrapper that does ps, looks for unit test like processes, and uses pstack or gdb to print thread stack traces. Also, using `ps -wwf` instead of `ps -wf` ensures output is not cut off.
For security, CircleCI runs with security restrictions on ptrace (/proc/sys/kernel/yama/ptrace_scope = 1), and this change adds a work-around to `InstallStackTraceHandler()` (only used by testing tools) to allow any process from the same user to debug it. (I've also touched >100 files to ensure all the unit tests call this function.)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/10828
Test Plan: local manual + temporary infinite loop in a unit test to observe in CircleCI
Reviewed By: hx235
Differential Revision: D40447634
Pulled By: pdillinger
fbshipit-source-id: 718a4c4a5b54fa0f9af2d01a446162b45e5e84e1
Summary:
FileMetaData::[min|max]_timestamp are not currently being used or
tracked by RocksDB, even when user-defined timestamp is enabled. Each of
them is a std::string which can occupy 32 bytes. Remove them for now.
They may be added back when we have a pressing need for them. When we do
add them back, consider store them in a more compact way, e.g. one
boolean flag and a byte array of size 16.
Per file min/max timestamp bounds are available as table properties.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/10443
Test Plan: make check
Reviewed By: pdillinger
Differential Revision: D38292275
Pulled By: riversand963
fbshipit-source-id: 841dc4e855ad8f8481c80cb020603de9607c9c94
Summary:
Start tracking SST unique id in MANIFEST, which is used to verify with
SST properties to make sure the SST file is not overwritten or
misplaced. A DB option `try_verify_sst_unique_id` is introduced to
enable/disable the verification, if enabled, it opens all SST files
during DB-open to read the unique_id from table properties (default is
false), so it's recommended to use it with `max_open_files = -1` to
pre-open the files.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9990
Test Plan: unittests, format-compatible test, mini-crash
Reviewed By: anand1976
Differential Revision: D36381863
Pulled By: jay-zhuang
fbshipit-source-id: 89ea2eb6b35ed3e80ead9c724eb096083eaba63f
Summary:
ToString() is created as some platform doesn't support std::to_string(). However, we've already used std::to_string() by mistake for 16 months (in db/db_info_dumper.cc). This commit just remove ToString().
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9955
Test Plan: Watch CI tests
Reviewed By: riversand963
Differential Revision: D36176799
fbshipit-source-id: bdb6dcd0e3a3ab96a1ac810f5d0188f684064471
Summary:
The patch replaces `std::map` with a sorted `std::vector` for
`VersionStorageInfo::blob_files_` and preallocates the space
for the `vector` before saving the `BlobFileMetaData` into the
new `VersionStorageInfo` in `VersionBuilder::Rep::SaveBlobFilesTo`.
These changes reduce the time the DB mutex is held while
saving new `Version`s, and using a sorted `vector` also makes
lookups faster thanks to better memory locality.
In addition, the patch introduces helper methods
`VersionStorageInfo::GetBlobFileMetaData` and
`VersionStorageInfo::GetBlobFileMetaDataLB` that can be used by
clients to perform lookups in the `vector`, and does some general
cleanup in the parts of code where blob file metadata are used.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9526
Test Plan:
Ran `make check` and the crash test script for a while.
Performance was tested using a load-optimized benchmark (`fillseq` with vector memtable, no WAL) and small file sizes so that a significant number of files are produced:
```
numactl --interleave=all ./db_bench --benchmarks=fillseq --allow_concurrent_memtable_write=false --level0_file_num_compaction_trigger=4 --level0_slowdown_writes_trigger=20 --level0_stop_writes_trigger=30 --max_background_jobs=8 --max_write_buffer_number=8 --db=/data/ltamasi-dbbench --wal_dir=/data/ltamasi-dbbench --num=800000000 --num_levels=8 --key_size=20 --value_size=400 --block_size=8192 --cache_size=51539607552 --cache_numshardbits=6 --compression_max_dict_bytes=0 --compression_ratio=0.5 --compression_type=lz4 --bytes_per_sync=8388608 --cache_index_and_filter_blocks=1 --cache_high_pri_pool_ratio=0.5 --benchmark_write_rate_limit=0 --write_buffer_size=16777216 --target_file_size_base=16777216 --max_bytes_for_level_base=67108864 --verify_checksum=1 --delete_obsolete_files_period_micros=62914560 --max_bytes_for_level_multiplier=8 --statistics=0 --stats_per_interval=1 --stats_interval_seconds=20 --histogram=1 --memtablerep=skip_list --bloom_bits=10 --open_files=-1 --subcompactions=1 --compaction_style=0 --min_level_to_compress=3 --level_compaction_dynamic_level_bytes=true --pin_l0_filter_and_index_blocks_in_cache=1 --soft_pending_compaction_bytes_limit=167503724544 --hard_pending_compaction_bytes_limit=335007449088 --min_level_to_compress=0 --use_existing_db=0 --sync=0 --threads=1 --memtablerep=vector --allow_concurrent_memtable_write=false --disable_wal=1 --enable_blob_files=1 --blob_file_size=16777216 --min_blob_size=0 --blob_compression_type=lz4 --enable_blob_garbage_collection=1 --seed=<some value>
```
Final statistics before the patch:
```
Cumulative writes: 0 writes, 700M keys, 0 commit groups, 0.0 writes per commit group, ingest: 284.62 GB, 121.27 MB/s
Interval writes: 0 writes, 334K keys, 0 commit groups, 0.0 writes per commit group, ingest: 139.28 MB, 72.46 MB/s
```
With the patch:
```
Cumulative writes: 0 writes, 760M keys, 0 commit groups, 0.0 writes per commit group, ingest: 308.66 GB, 131.52 MB/s
Interval writes: 0 writes, 445K keys, 0 commit groups, 0.0 writes per commit group, ingest: 185.35 MB, 93.15 MB/s
```
Total time to complete the benchmark is 2611 seconds with the patch, down from 2986 secs.
Reviewed By: riversand963
Differential Revision: D34082728
Pulled By: ltamasi
fbshipit-source-id: fc598abf676dce436734d06bb9d2d99a26a004fc
Summary:
The patch does some cleanup in and around `VersionStorageInfo`:
* Renames the method `PrepareApply` to `PrepareAppend` in `Version`
to make it clear that it is to be called before appending the `Version` to
`VersionSet` (via `AppendVersion`), not before applying any `VersionEdit`s.
* Introduces a helper method `VersionStorageInfo::PrepareForVersionAppend`
(called by `Version::PrepareAppend`) that encapsulates the population of the
various derived data structures in `VersionStorageInfo`, and turns the
methods computing the derived structures (`UpdateNumNonEmptyLevels`,
`CalculateBaseBytes` etc.) into private helpers.
* Changes `Version::PrepareAppend` so it only calls `UpdateAccumulatedStats`
if the `update_stats` flag is set. (Earlier, this was checked by the callee.)
Related to this, it also moves the call to `ComputeCompensatedSizes` to
`VersionStorageInfo::PrepareForVersionAppend`.
* Updates and cleans up `version_builder_test`, `version_set_test`, and
`compaction_picker_test` so `PrepareForVersionAppend` is called anytime
a new `VersionStorageInfo` is set up or saved. This cleanup also involves
splitting `VersionStorageInfoTest.MaxBytesForLevelDynamic`
into multiple smaller test cases.
* Fixes up a bunch of comments that were outdated or just plain incorrect.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9494
Test Plan: Ran `make check` and the crash test script for a while.
Reviewed By: riversand963
Differential Revision: D33971666
Pulled By: ltamasi
fbshipit-source-id: fda52faac7783041126e4f8dec0fe01bdcadf65a
Summary:
Fix a bug that causes file temperature not preserved after DB is restarted, or options.max_manifest_file_size is hit.
Also, pass temperature information to NewRandomAccessFile() to allow users to hack a solution where they don't preserve tiering information.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9242
Test Plan: Add a unit test that would fail without the fix.
Reviewed By: jay-zhuang
Differential Revision: D32818150
fbshipit-source-id: 36aa3f148c60107f7b8e9d65b63b039f9e1a1eec
Summary:
Track per-SST user-defined timestamp information in MANIFEST https://github.com/facebook/rocksdb/issues/8957
Rockdb has supported user-defined timestamp feature. Application can specify a timestamp
when writing each k-v pair. When data flush from memory to disk file called SST files, file
creation activity will commit to MANIFEST. This commit is for tracking timestamp info in the
MANIFEST for each file. The changes involved are as follows:
1) Track max/min timestamp in FileMetaData, and fix invoved codes.
2) Add NewFileCustomTag::kMinTimestamp and NewFileCustomTag::kMinTimestamp in
NewFileCustomTag ( in the kNewFile4 part ), and support invoved codes such as
VersionEdit Encode and Decode etc.
3) Add unit test code for VersionEdit EncodeDecodeNewFile4, and fix invoved test codes.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9092
Reviewed By: ajkr, akankshamahajan15
Differential Revision: D32252323
Pulled By: riversand963
fbshipit-source-id: d2642898d6e3ad1fef0eb866b98045408bd4e162
Summary:
Right now, when options.ttl is set, compactions are triggered around the time when TTL is reached. This might cause extra compactions which are often bursty. This commit tries to mitigate it by picking those files earlier in normal compaction picking process. This is only implemented using kMinOverlappingRatio with Leveled compaction as it is the default value and it is more complicated to change other styles.
When a file is aged more than ttl/2, RocksDB starts to boost the compaction priority of files in normal compaction picking process, and hope by the time TTL is reached, very few extra compaction is needed.
In order for this to work, another change is made: during a compaction, if an output level file is older than ttl/2, cut output files based on original boundary (if it is not in the last level). This is to make sure that after an old file is moved to the next level, and new data is merged from the upper level, the new data falling into this range isn't reset with old timestamp. Without this change, in many cases, most files from one level will keep having old timestamp, even if they have newer data and we stuck in it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8749
Test Plan: Add a unit test to test the boosting logic. Will add a unit test to test it end-to-end.
Reviewed By: jay-zhuang
Differential Revision: D30735261
fbshipit-source-id: 503c2d89250b22911eb99e72b379be154de3428e
Summary:
The number or total size of garbage blobs in any given blob file can
never exceed the number or total size of all blobs in the file. (This
would be a similar error to e.g. attempting to remove from the LSM tree
an SST file that has already been removed.) The patch builds on
https://github.com/facebook/rocksdb/issues/9085 and adds a
consistency check to `VersionBuilder` that prevents the above from
happening.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9100
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D32048982
Pulled By: ltamasi
fbshipit-source-id: 6f7e0793bf534ad04c3359cc0f696b8e4e5ef81c
Summary:
Currently, blob file checksums are incorrectly dumped as raw bytes
in the `ldb manifest_dump` output (i.e. they are not printed as hex).
The patch fixes this and also updates some test cases to reflect that
the checksum value field in `BlobFileAddition` and `SharedBlobFileMetaData`
contains the raw checksum and not a hex string.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8437
Test Plan:
`make check`
Tested using `ldb manifest_dump`
Reviewed By: akankshamahajan15
Differential Revision: D29284170
Pulled By: ltamasi
fbshipit-source-id: d11cfb3435b14cd73c8a3d3eb14fa0f9fa1d2228
Summary:
The ImmutableCFOptions contained a bunch of fields that belonged to the ImmutableDBOptions. This change cleans that up by introducing an ImmutableOptions struct. Following the pattern of Options struct, this class inherits from the DB and CFOption structs (of the Immutable form).
Only one structural change (the ImmutableCFOptions::fs was changed to a shared_ptr from a raw one) is in this PR. All of the other changes involve moving the member variables from the ImmutableCFOptions into the ImmutableOptions and changing member variables or function parameters as required for compilation purposes.
Follow-on PRs may do a further clean-up of the code, such as renaming variables (such as "ImmutableOptions cf_options") and potentially eliminating un-needed function parameters (there is no longer a need to pass both an ImmutableDBOptions and an ImmutableOptions to a function).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8262
Reviewed By: pdillinger
Differential Revision: D28226540
Pulled By: mrambacher
fbshipit-source-id: 18ae71eadc879dedbe38b1eb8e6f9ff5c7147dbf
Summary:
This is a PR generated **semi-automatically** by an internal tool to remove unused includes and `using` statements.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7604
Test Plan: make check
Reviewed By: ajkr
Differential Revision: D24579392
Pulled By: riversand963
fbshipit-source-id: c4bfa6c6b08da1de186690d37eb73d8fff45aecd
Summary:
When multiple background jobs are generating blob files in parallel, it is actually
possible for a blob file to be added with a file number that is lower than the
highest one in the base version. (This is a harmless race condition.) The patch
fixes the handling of this case and adds a unit test.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7349
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D23542453
Pulled By: ltamasi
fbshipit-source-id: 4ff6f3654bc58c391d10b9870e1cc40b5e3fa8e4
Summary:
The earlier `VersionBuilder` code only cleaned up blob files that were
marked as entirely consisting of garbage using `VersionEdits` with
`BlobFileGarbage`. This covers the cases when table files go through
regular compaction, where we iterate through the KVs and thus have an
opportunity to calculate the amount of garbage (that is, most cases).
However, it does not help when table files are simply dropped (e.g. deletion
compactions or the `DeleteFile` API). To deal with such cases, the patch
adds logic that cleans up all blob files at the head of the list until the first
one with linked SSTs is found. (As an example, let's assume we have blob files
with numbers 1..10, and the first one with any linked SSTs is number 8.
This means that SSTs in the `Version` only rely on blob files with numbers >= 8,
and thus 1..7 are no longer needed.)
The code change itself is pretty small; however, changing the logic like this
necessitated changes to some tests that have been added recently (namely
to the ones that use blob files in isolation, i.e. without any table files referring
to them). Some of these cases were fixed by bypassing `VersionBuilder` altogether
in order to keep the tests simple (which actually makes them more proper unit tests
as well), while the `VersionBuilder` unit tests were fixed by adding dummy table
files to the test cases as needed.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7001
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D22119474
Pulled By: ltamasi
fbshipit-source-id: c6547141355667d4291d9661d6518eb741e7b54a
Summary:
The `FileMetaData` objects associated with table files already contain the
number of the oldest blob file referenced by the SST in question. This patch
adds the inverse mapping to `BlobFileMetaData`, namely the set of table file
numbers for which the oldest blob file link points to the given blob file (these
are referred to as *linked SSTs*). This mapping will be used by the GC logic.
Implementation-wise, the patch builds on the `BlobFileMetaDataDelta`
functionality introduced in https://github.com/facebook/rocksdb/pull/6835: newly linked/unlinked SSTs are
accumulated in `BlobFileMetaDataDelta`, and the changes to the linked SST set
are applied in one shot when the new `Version` is saved. The patch also reworks
the blob file related consistency checks in `VersionBuilder` so they validate the
consistency of the forward table file -> blob file links and the backward blob file ->
table file links for blob files that are part of the `Version`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6945
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D21912228
Pulled By: ltamasi
fbshipit-source-id: c5bc7acf6e729a8fccbb12672dd5cd00f6f000f8
Summary:
We currently do not have any validation that would ensure that the `FileMetaData`
objects are equivalent when a file gets deleted from the LSM tree and then re-added
(think trivial moves); however, if we did, this test case would be in violation. The patch
changes the values used in the test case so they are consistent.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6942
Test Plan: `make check`
Reviewed By: zhichao-cao
Differential Revision: D21911366
Pulled By: ltamasi
fbshipit-source-id: 2f0486f8337373a6a111b6f28433d70507857104
Summary:
The patch cleans up the code and improves the consistency checks around
adding/deleting table files in `VersionBuilder`. Namely, it makes the checks
stricter and improves them in the following ways:
1) A table file can now only be deleted from the LSM tree using the level it
resides on. Earlier, there was some unnecessary wiggle room for
trivially moved files (they could be deleted using a lower level number than
the actual one).
2) A table file cannot be added to the tree if it is already present in the tree
on any level (not just the target level). The earlier code only had an assertion
(which is a no-op in release builds) that the newly added file is not already
present on the target level.
3) The above consistency checks around state transitions are now mandatory,
as opposed to the earlier `CheckConsistencyForDeletes`, which was a no-op
in release mode unless `force_consistency_checks` was set to `true`. The rationale
here is that assuming that the initial state is consistent, a valid transition leads to a
next state that is also consistent; however, an *invalid* transition offers no such
guarantee. Hence it makes sense to validate the transitions unconditionally,
and save `force_consistency_checks` for the paranoid checks that re-validate
the entire state.
4) The new checks build on the mechanism introduced in https://github.com/facebook/rocksdb/pull/6862,
which enables us to efficiently look up the location (level and position within level)
of files in a `Version` by file number. This makes the consistency checks much more
efficient than the earlier `CheckConsistencyForDeletes`, which essentially
performed a linear search.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6901
Test Plan:
Extended the unit tests and ran:
`make check`
`make whitebox_crash_test`
Reviewed By: ajkr
Differential Revision: D21822714
Pulled By: ltamasi
fbshipit-source-id: e2b29c8b6da1bf0f59004acc889e4870b2d18215
Summary:
The patch adds logic to keep track of obsolete blob files. A blob file becomes
obsolete when the last `shared_ptr` that points to the corresponding
`SharedBlobFileMetaData` object goes away, which, in turn, happens when the
last `Version` that contains the blob file is destroyed. No longer needed blob
files are added to the obsolete list in `VersionSet` using a custom deleter to
avoid unnecessary coupling between `SharedBlobFileMetaData` and `VersionSet`.
Obsolete blob files are returned by `VersionSet::GetObsoleteFiles` and stored
in `JobContext`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6755
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D21233155
Pulled By: ltamasi
fbshipit-source-id: 47757e06fdc0127f27ed57f51abd27893d9a7b7a
Summary:
The patch makes a couple of small cleanups to `SharedBlobFileMetaData` and `BlobFileMetaData`:
* It makes the constructors private and introduces factory methods to ensure these objects are always owned by `shared_ptr`s. Note that `SharedBlobFileMetaData` has an additional factory that takes a deleter object; we can utilize this to e.g. notify `VersionSet` when a blob file becomes obsolete (which is exactly when `SharedBlobFileMetaData` is destroyed).
* It disables move operations explicitly instead of relying on them being suppressed because of a user-declared destructor.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6749
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D21206947
Pulled By: ltamasi
fbshipit-source-id: 9094c14cc335b3e226f883e5a0df4f87a5cdeb95
Summary:
The patch adds a couple of classes to represent metadata about
blob files: `SharedBlobFileMetaData` contains the information elements
that are immutable (once the blob file is closed), e.g. blob file number,
total number and size of blob files, checksum method/value, while
`BlobFileMetaData` contains attributes that can vary across versions like
the amount of garbage in the file. There is a single `SharedBlobFileMetaData`
for each blob file, which is jointly owned by the `BlobFileMetaData` objects
that point to it; `BlobFileMetaData` objects, in turn, are owned by `Version`s
and can also be shared if the (immutable _and_ mutable) state of the blob file
is the same in two versions.
In addition, the patch adds the blob file metadata to `VersionStorageInfo`, and extends
`VersionBuilder` so that it can apply blob file related `VersionEdit`s (i.e. those
containing `BlobFileAddition`s and/or `BlobFileGarbage`), and save blob file metadata
to a new `VersionStorageInfo`. Consistency checks are also extended to ensure
that table files point to blob files that are part of the `Version`, and that all blob files
that are part of any given `Version` have at least some _non_-garbage data in them.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6597
Test Plan: `make check`
Reviewed By: riversand963
Differential Revision: D20656803
Pulled By: ltamasi
fbshipit-source-id: f1f74d135045b3b42d0146f03ee576ef0a4bfd80
Summary:
When dynamically linking two binaries together, different builds of RocksDB from two sources might cause errors. To provide a tool for user to solve the problem, the RocksDB namespace is changed to a flag which can be overridden in build time.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6433
Test Plan: Build release, all and jtest. Try to build with ROCKSDB_NAMESPACE with another flag.
Differential Revision: D19977691
fbshipit-source-id: aa7f2d0972e1c31d75339ac48478f34f6cfcfb3e
Summary:
In the current code base, RocksDB generate the checksum for each block and verify the checksum at usage. Current PR enable SST file checksum. After a SST file is generated by Flush or Compaction, RocksDB generate the SST file checksum and store the checksum value and checksum method name in the vs_info and MANIFEST as part for the FileMetadata.
Added the enable_sst_file_checksum to Options to enable or disable file checksum. Added sst_file_checksum to Options such that user can plugin their own SST file checksum calculate method via overriding the SstFileChecksum class. The checksum information inlcuding uint32_t checksum value and a checksum name (string). A new tool is added to LDB such that user can dump out a list of file checksum information from MANIFEST. If user enables the file checksum but does not provide the sst_file_checksum instance, RocksDB will use the default crc32checksum implemented in table/sst_file_checksum_crc32c.h
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6216
Test Plan: Added the testing case in table_test and ldb_cmd_test to verify checksum is correct in different level. Pass make asan_check.
Differential Revision: D19171461
Pulled By: zhichao-cao
fbshipit-source-id: b2e53479eefc5bb0437189eaa1941670e5ba8b87
Summary:
options.periodic_compaction_seconds isn't supported when options.max_open_files != -1. It's because that the information of file creation time is stored in table properties and are not guaranteed to be loaded unless options.max_open_files = -1. Relax this constraint by storing the information in manifest.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6090
Test Plan: Pass all existing tests; Modify an existing test to force the manifest value to take 0 to simulate backward compatibility case; manually open the DB generated with the change by release 4.2.
Differential Revision: D18702268
fbshipit-source-id: 13e0bd94f546498a04f3dc5fc0d9dff5125ec9eb
Summary:
Previously, options.ttl cannot be set with options.max_open_files = -1, because it makes use of creation_time field in table properties, which is not available unless max_open_files = -1. With this commit, the information will be stored in manifest and when it is available, will be used instead.
Note that, this change will break forward compatibility for release 5.1 and older.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6060
Test Plan: Extend existing test case to options.max_open_files != -1, and simulate backward compatility in one test case by forcing the value to be 0.
Differential Revision: D18631623
fbshipit-source-id: 30c232a8672de5432ce9608bb2488ecc19138830
Summary:
This is groundwork for adding garbage collection support to BlobDB. The
patch adds logic that keeps track of the oldest blob file referred to by
each SST file. The oldest blob file is identified during flush/
compaction (similarly to how the range of keys covered by the SST is
identified), and persisted in the manifest as a custom field of the new
file edit record. Blob indexes with TTL are ignored for the purposes of
identifying the oldest blob file (since such blob files are cleaned up by the
TTL logic in BlobDB).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5903
Test Plan:
Added new unit tests; also ran db_bench in BlobDB mode, inspected the
manifest using ldb, and confirmed (by scanning the SST files using
sst_dump) that the value of the oldest blob file number field matches
the contents of the file for each SST.
Differential Revision: D17859997
Pulled By: ltamasi
fbshipit-source-id: 21662c137c6259a6af70446faaf3a9912c550e90
Summary:
Many logging related source files are under util/. It will be more structured if they are together.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5387
Differential Revision: D15579036
Pulled By: siying
fbshipit-source-id: 3850134ed50b8c0bb40a0c8ae1f184fa4081303f
Summary:
There are too many types of files under util/. Some test related files don't belong to there or just are just loosely related. Mo
ve them to a new directory test_util/, so that util/ is cleaner.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5377
Differential Revision: D15551366
Pulled By: siying
fbshipit-source-id: 0f5c8653832354ef8caa31749c0143815d719e2c
Summary:
RocksDB used to store global_seqno in external SST files written by
SstFileWriter. During file ingestion, RocksDB uses `pwrite` to update the
`global_seqno`. Since random write is not supported in some non-POSIX compliant
file systems, external SST file ingestion is not supported on these file
systems. To address this limitation, we no longer update `global_seqno` during
file ingestion. Later RocksDB uses the MANIFEST and other information in table
properties to deduce global seqno for externally-ingested SST files.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4172
Differential Revision: D8961465
Pulled By: riversand963
fbshipit-source-id: 4382ec85270a96be5bc0cf33758ca2b167b05071
Summary:
We always run consistency checks when compiling in debug mode
allow users to set Options::force_consistency_checks to true to be able to run such checks even when compiling in release mode
Test Plan:
make check -j64
make release
Reviewers: lightmark, sdong, yiwu
Reviewed By: yiwu
Subscribers: hermanlee4, andrewkr, yoshinorim, jkedgar, dhruba
Differential Revision: https://reviews.facebook.net/D64701
Summary:
* Change constructor of MutableCFOptions to depends only on ColumnFamilyOptions.
* Move `max_subcompactions`, `compaction_options_fifo` and `compaction_pri` to ImmutableCFOptions to make it clear that they are immutable.
Test Plan: existing unit tests.
Reviewers: yhchiang, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D63945
Summary:
Add options.compaction_pri, which specifies the policy about which file to compact first.
kCompactionPriByLargestSeq will compact oldest files first.
Verified the behavior in db_bench but did not write unit tests yet. Also need to make it settable through option string and dynamically changeable.
Test Plan: Will write unit tests
Reviewers: igor, rven, anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, MarkCallaghan
Reviewed By: yhchiang
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D45951
Summary:
It is experimental. Allow users to return from a call back function TablePropertiesCollector::NeedCompact(), based on the data in the file.
It can be used to allow users to suggest DB to clear up delete tombstones faster.
Test Plan: Add a unit test.
Reviewers: igor, yhchiang, kradhakrishnan, rven
Reviewed By: rven
Subscribers: yoshinorim, MarkCallaghan, maykov, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D39585
Summary:
This diff updates the logic of how we do trivial move, now trivial move can run on any number of files in input level as long as they are not overlapping
The conditions for trivial move have been updated
Introduced conditions:
- Trivial move cannot happen if we have a compaction filter (except if the compaction is not manual)
- Input level files cannot be overlapping
Removed conditions:
- Trivial move only run when the compaction is not manual
- Input level should can contain only 1 file
More context on what tests failed because of Trivial move
```
DBTest.CompactionsGenerateMultipleFiles
This test is expecting compaction on a file in L0 to generate multiple files in L1, this test will fail with trivial move because we end up with one file in L1
```
```
DBTest.NoSpaceCompactRange
This test expect compaction to fail when we force environment to report running out of space, of course this is not valid in trivial move situation
because trivial move does not need any extra space, and did not check for that
```
```
DBTest.DropWrites
Similar to DBTest.NoSpaceCompactRange
```
```
DBTest.DeleteObsoleteFilesPendingOutputs
This test expect that a file in L2 is deleted after it's moved to L3, this is not valid with trivial move because although the file was moved it is now used by L3
```
```
CuckooTableDBTest.CompactionIntoMultipleFiles
Same as DBTest.CompactionsGenerateMultipleFiles
```
This diff is based on a work by @sdonghttps://reviews.facebook.net/D34149
Test Plan: make -j64 check
Reviewers: rven, sdong, igor
Reviewed By: igor
Subscribers: yhchiang, ott, march, dhruba, sdong
Differential Revision: https://reviews.facebook.net/D34797
Summary:
Our existing test notation is very similar to what is used in gtest. It makes it easy to adopt what is different.
In this diff I modify existing [[ https://code.google.com/p/googletest/wiki/Primer#Test_Fixtures:_Using_the_Same_Data_Configuration_for_Multiple_Te | test fixture ]] classes to inherit from `testing::Test`. Also for unit tests that use fixture class, `TEST` is replaced with `TEST_F` as required in gtest.
There are several custom `main` functions in our existing tests. To make this transition easier, I modify all `main` functions to fallow gtest notation. But eventually we can remove them and use implementation of `main` that gtest provides.
```lang=bash
% cat ~/transform
#!/bin/sh
files=$(git ls-files '*test\.cc')
for file in $files
do
if grep -q "rocksdb::test::RunAllTests()" $file
then
if grep -Eq '^class \w+Test {' $file
then
perl -pi -e 's/^(class \w+Test) {/${1}: public testing::Test {/g' $file
perl -pi -e 's/^(TEST)/${1}_F/g' $file
fi
perl -pi -e 's/(int main.*\{)/${1}::testing::InitGoogleTest(&argc, argv);/g' $file
perl -pi -e 's/rocksdb::test::RunAllTests/RUN_ALL_TESTS/g' $file
fi
done
% sh ~/transform
% make format
```
Second iteration of this diff contains only scripted changes.
Third iteration contains manual changes to fix last errors and make it compilable.
Test Plan:
Build and notice no errors.
```lang=bash
% USE_CLANG=1 make check -j55
```
Tests are still testing.
Reviewers: meyering, sdong, rven, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D35157
Summary:
When having fixed max_bytes_for_level_base, the ratio of size of largest level and the second one can range from 0 to the multiplier. This makes LSM tree frequently irregular and unpredictable. It can also cause poor space amplification in some cases.
In this improvement (proposed by Igor Kabiljo), we introduce a parameter option.level_compaction_use_dynamic_max_bytes. When turning it on, RocksDB is free to pick a level base in the range of (options.max_bytes_for_level_base/options.max_bytes_for_level_multiplier, options.max_bytes_for_level_base] so that real level ratios are close to options.max_bytes_for_level_multiplier.
Test Plan: New unit tests and pass tests suites including valgrind.
Reviewers: MarkCallaghan, rven, yhchiang, igor, ikabiljo
Reviewed By: ikabiljo
Subscribers: yoshinorim, ikabiljo, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D31437
Summary: Added requirement that ComputeCompactionScore() be executed in mutex, since it's accessing being_compacted bool, which can be mutated by other threads. Also added more comments about thread safety of FileMetaData, since it was a bit confusing. However, it seems that FileMetaData doesn't have data races (except being_compacted)
Test Plan: Ran 100 ConvertCompactionStyle tests with thread sanitizer. On master -- some failures. With this patch -- none.
Reviewers: yhchiang, rven, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D32283
Summary: During recovery, VersionBuilder::Apply() was called multiple times. If the DB is open for long enough, most of files added earlier will be deleted by later deletes. In current solution, sorting added file happens first and then deletes are applied. In this patch, deletes are applied when possible inside Apply(), which can significantly reduce the sorting time in some cases.
Test Plan:
Add unit tests in version_builder
valgrind_check
Open a manifest of 50MB, with 9K live files. The manifest read time reduced from 1.6 seconds to 0.7 seconds.
Reviewers: rven, yhchiang, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D30765
Hui Xiao