Summary:
The new DB::GetApproximateSizes with SizeApproximationOptions argument, which allows to add more options/knobs to the DB::GetApproximateSizes call (beyond only the include_flags)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5626
Differential Revision: D16496913
Pulled By: elipoz
fbshipit-source-id: ee8c6c182330a285fa056ecfc3905a592b451720
Summary:
The ObjectRegistry class replaces the Registrar and NewCustomObjects. Objects are registered with the registry by Type (the class must implement the static const char *Type() method).
This change is necessary for a few reasons:
- By having a class (rather than static template instances), the class can be passed between compilation units, meaning that objects could be registered and shared from a dynamic library with an executable.
- By having a class with instances, different units could have different objects registered. This could be useful if, for example, one Option allowed for a dynamic library and one did not.
When combined with some other PRs (being able to load shared libraries, a Configurable interface to configure objects to/from string), this code will allow objects in external shared libraries to be added to a RocksDB image at run-time, rather than requiring every new extension to be built into the main library and called explicitly by every program.
Test plan (on riversand963's devserver)
```
$COMPILE_WITH_ASAN=1 make -j32 all && sleep 1 && make check
```
All tests pass.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5293
Differential Revision: D16363396
Pulled By: riversand963
fbshipit-source-id: fbe4acb615bfc11103eef40a0b288845791c0180
Summary:
RocksDB has historically stored uncompression dictionary objects in the block
cache as opposed to storing just the block contents. This neccesitated
evicting the object upon table close. With the new code, only the raw blocks
are stored in the cache, eliminating the need for eviction.
In addition, the patch makes the following improvements:
1) Compression dictionary blocks are now prefetched/pinned similarly to
index/filter blocks.
2) A copy operation got eliminated when the uncompression dictionary is
retrieved.
3) Errors related to retrieving the uncompression dictionary are propagated as
opposed to silently ignored.
Note: the patch temporarily breaks the compression dictionary evicition stats.
They will be fixed in a separate phase.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5584
Test Plan: make asan_check
Differential Revision: D16344151
Pulled By: ltamasi
fbshipit-source-id: 2962b295f5b19628f9da88a3fcebbce5a5017a7b
Summary:
Right now, ldb cannot scan a DB with merge operands with default ldb. There is no hard to give a general merge operator so that it can at least print out something
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5607
Test Plan: Run ldb against a DB with merge operands and see the outputs.
Differential Revision: D16442634
fbshipit-source-id: c66c414ec07f219cfc6e6ec2cc14c783ee95df54
Summary:
There are concerns about the correctness of this patch. Disabling by default until the concerns are resolved.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5606
Differential Revision: D16428064
Pulled By: maysamyabandeh
fbshipit-source-id: a89280f0ea85796c9c9dfbfd9a8e91dad9b000b3
Summary:
Right now, users cannot take advantage of row cache, unless no snapshot is used, or Get() is repeated for the same snapshots. This limits the usage of row cache.
This change eliminate this restriction in some cases. If the snapshot used is newer than the largest sequence number in the file, and write callback function is not registered, the same row cache key is used as no snapshot is given. We still need the callback function restriction for now because the callback function may filter out different keys for different snapshots even if the snapshots are new.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5600
Test Plan: Add a unit test.
Differential Revision: D16386616
fbshipit-source-id: 6b7d214bd215d191b03ccf55926ad4b703ec2e53
Summary:
Currently, when the block cache is used for the filter block, it is not
really the block itself that is stored in the cache but a FilterBlockReader
object. Since this object is not pure data (it has, for instance, pointers that
might dangle, including in one case a back pointer to the TableReader), it's not
really sharable. To avoid the issues around this, the current code erases the
cache entries when the TableReader is closed (which, BTW, is not sufficient
since a concurrent TableReader might have picked up the object in the meantime).
Instead of doing this, the patch moves the FilterBlockReader out of the cache
altogether, and decouples the filter reader object from the filter block.
In particular, instead of the TableReader owning, or caching/pinning the
FilterBlockReader (based on the customer's settings), with the change the
TableReader unconditionally owns the FilterBlockReader, which in turn
owns/caches/pins the filter block. This change also enables us to reuse the code
paths historically used for data blocks for filters as well.
Note:
Eviction statistics for filter blocks are temporarily broken. We plan to fix this in a
separate phase.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5504
Test Plan: make asan_check
Differential Revision: D16036974
Pulled By: ltamasi
fbshipit-source-id: 770f543c5fb4ed126fd1e04bfd3809cf4ff9c091
Summary:
Right now ldb can open running DB through read-only DB. However, it might leave info logs files to the read-only DB directory. Add an option to open the DB as secondary to avoid it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5537
Test Plan:
Run
./ldb scan --max_keys=10 --db=/tmp/rocksdbtest-2491/dbbench --secondary_path=/tmp --no_value --hex
and
./ldb get 0x00000000000000103030303030303030 --hex --db=/tmp/rocksdbtest-2491/dbbench --secondary_path=/tmp
against a normal db_bench run and observe the output changes. Also observe that no new info logs files are created under /tmp/rocksdbtest-2491/dbbench.
Run without --secondary_path and observe that new info logs created under /tmp/rocksdbtest-2491/dbbench.
Differential Revision: D16113886
fbshipit-source-id: 4e09dec47c2528f6ca08a9e7a7894ba2d9daebbb
Summary:
Previously `GetAllKeyVersions()` supports default column family only. This PR add support for other column families.
Test plan (devserver):
```
$make clean && COMPILE_WITH_ASAN=1 make -j32 db_basic_test
$./db_basic_test --gtest_filter=DBBasicTest.GetAllKeyVersions
```
All other unit tests must pass.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5544
Differential Revision: D16147551
Pulled By: riversand963
fbshipit-source-id: 5a61aece2a32d789e150226a9b8d53f4a5760168
Summary:
1. Cleanup WAL trash files on open
2. Don't apply deletion rate limit if WAL dir is different from db dir
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5520
Test Plan: Add new unit tests and make check
Differential Revision: D16096750
Pulled By: anand1976
fbshipit-source-id: 6f07858ad864b754b711db416f0389c45ede599b
Summary:
Sometimes it is helpful to fetch the whole history of stats after benchmark runs. Add such an option
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5532
Test Plan: Run the benchmark manually and observe the output is as expected.
Differential Revision: D16097764
fbshipit-source-id: 10b5b735a22a18be198b8f348be11f11f8806904
Summary:
Enhancement to MultiGet batching to read data blocks required for keys in a batch in parallel from disk. It uses Env::MultiRead() API to read multiple blocks and reduce latency.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5464
Test Plan:
1. make check
2. make asan_check
3. make asan_crash
Differential Revision: D15911771
Pulled By: anand1976
fbshipit-source-id: 605036b9af0f90ca0020dc87c3a86b4da6e83394
Summary:
Mid-point insertion is a useful feature and is mature now. Make it default. Also changed cache_index_and_filter_blocks_with_high_priority=true as default accordingly, so that we won't evict index and filter blocks easier after the change, to avoid too many surprises to users.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5508
Test Plan: Run all existing tests.
Differential Revision: D16021179
fbshipit-source-id: ce8456e8d43b3bfb48df6c304b5290a9d19817eb
Summary:
Add C binding for secondary instance as well as unit test.
Test plan (on devserver)
```
$make clean && COMPILE_WITH_ASAN=1 make -j20 all
$./c_test
$make check
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5505
Differential Revision: D16000043
Pulled By: riversand963
fbshipit-source-id: 3361ef6bfdf4ce12438cee7290a0ac203b5250bd
Summary:
The first key is used to defer reading the data block until this file gets to the top of merging iterator's heap. For short range scans, most files never make it to the top of the heap, so this change can reduce read amplification by a lot sometimes.
Consider the following workload. There are a few data streams (we'll be calling them "logs"), each stream consisting of a sequence of blobs (we'll be calling them "records"). Each record is identified by log ID and a sequence number within the log. RocksDB key is concatenation of log ID and sequence number (big endian). Reads are mostly relatively short range scans, each within a single log. Writes are mostly sequential for each log, but writes to different logs are randomly interleaved. Compactions are disabled; instead, when we accumulate a few tens of sst files, we create a new column family and start writing to it.
So, a typical sst file consists of a few ranges of blocks, each range corresponding to one log ID (we use FlushBlockPolicy to cut blocks at log boundaries). A typical read would go like this. First, iterator Seek() reads one block from each sst file. Then a series of Next()s move through one sst file (since writes to each log are mostly sequential) until the subiterator reaches the end of this log in this sst file; then Next() switches to the next sst file and reads sequentially from that, and so on. Often a range scan will only return records from a small number of blocks in small number of sst files; in this case, the cost of initial Seek() reading one block from each file may be bigger than the cost of reading the actually useful blocks.
Neither iterate_upper_bound nor bloom filters can prevent reading one block from each file in Seek(). But this PR can: if the index contains first key from each block, we don't have to read the block until this block actually makes it to the top of merging iterator's heap, so for short range scans we won't read any blocks from most of the sst files.
This PR does the deferred block loading inside value() call. This is not ideal: there's no good way to report an IO error from inside value(). As discussed with siying offline, it would probably be better to change InternalIterator's interface to explicitly fetch deferred value and get status. I'll do it in a separate PR.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5289
Differential Revision: D15256423
Pulled By: al13n321
fbshipit-source-id: 750e4c39ce88e8d41662f701cf6275d9388ba46a
Summary:
It it not safe to assume application had sync the SST file before ingest it into DB. Also the directory to put the ingested file needs to be fsync, otherwise the file can be lost. For integrity of RocksDB we need to sync the ingested file and directory before apply the change to manifest.
Also syncing after writing global sequence when write_global_seqno=true was removed in https://github.com/facebook/rocksdb/issues/4172. Adding it back.
Fixes https://github.com/facebook/rocksdb/issues/5287.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5435
Test Plan:
Test ingest file with ldb command and observe fsync/fdatasync in strace output. Tried both move_files=true and move_files=false.
https://gist.github.com/yiwu-arbug/650a4023f57979056d83485fa863bef9
More test suggestions are welcome.
Differential Revision: D15941675
Pulled By: riversand963
fbshipit-source-id: 389533f3923065a96df2cdde23ff4724a1810d78
Summary:
When tailing the WAL with TransactionLogIterator, it used to return Corruption status to indicate that the WAL has new tail that is not visible to the iterator, which is a misleading status. The patch replaces it with TryAgain which is more descriptive of a status, indicating that the user needs to create a new iterator to fetch the recent tail.
Fixes https://github.com/facebook/rocksdb/issues/5455
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5474
Differential Revision: D15898953
Pulled By: maysamyabandeh
fbshipit-source-id: 40966f6457cb539e1aeb104daeada6b0e46059fc
Summary:
The patch brings the semantics of per-block-type read performance
context counters in sync with the generic block_read_count by only
incrementing the counter if the block was actually read from the file.
It also fixes index_block_read_count, which fell victim to the
refactoring in PR https://github.com/facebook/rocksdb/issues/5298.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5484
Test Plan: Extended the unit tests.
Differential Revision: D15887431
Pulled By: ltamasi
fbshipit-source-id: a3889759d0ac5759d56625d692cd828d1b9207a6
Summary:
While the secondary is replaying after the primary, the primary may switch to a new MANIFEST. The secondary is already able to detect and follow the primary to the new MANIFEST. However, the current implementation has a bug, described as follows.
The new MANIFEST's first records have been generated by VersionSet::WriteSnapshot to describe the current state of the column families and the db as of the MANIFEST creation. Since the secondary instance has already finished recovering upon start, there is no need for the secondary to process these records. Actually, if the secondary were to replay these records, the secondary may end up adding the same SST files **again** to each column family, causing consistency checks done by VersionBuilder to fail. Therefore, we record the number of records to skip at the beginning of the new MANIFEST and ignore them.
Test plan (on dev server)
```
$make clean && make -j32 all
$./db_secondary_test
```
All existing unit tests must pass as well.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5472
Differential Revision: D15866771
Pulled By: riversand963
fbshipit-source-id: a1eec4837fb2ad13059398efb0f437e74fd53bed
Summary:
It seems like CF Options are not properly validated when creating a new column family with `CreateColumnFamily` API; only a selected few checks are done. Calling `ColumnFamilyData::ValidateOptions`, which is the single source for all CFOptions validations, will help fix this. (`ColumnFamilyData::ValidateOptions` is already called at the time of `DB::Open`).
**Test Plan:**
Added a new test: `DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions`
```
TEST_TMPDIR=/dev/shm ./db_test --gtest_filter=DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions
```
Also ran gtest-parallel to make sure the new test is not flaky.
```
TEST_TMPDIR=/dev/shm ~/gtest-parallel/gtest-parallel ./db_test --gtest_filter=DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions --repeat=10000
[10000/10000] DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions (15 ms)
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5453
Differential Revision: D15816851
Pulled By: sagar0
fbshipit-source-id: 9e702b9850f5c4a7e0ef8d39e1e6f9b81e7fe1e5
Summary:
In regular RocksDB instance, `MemTable::earliest_seqno_` is "db sequence number at the time of creation". However, we cannot use the db sequence number to set the value of `MemTable::earliest_seqno_` for secondary instance, i.e. `DBImplSecondary` due to the logic of MANIFEST and WAL replay.
When replaying the log files of the primary, the secondary instance first replays MANIFEST and updates the db sequence number if necessary. Next, the secondary replays WAL files, creates new memtables if necessary and inserts key-value pairs into memtables. The following can occur when the db has two or more column families.
Assume the db has column family "default" and "cf1". At a certain in time, both "default" and "cf1" have data in memtables.
1. Primary triggers a flush and flushes "cf1". "default" is **not** flushed.
2. Secondary replays the MANIFEST updates its db sequence number to the latest value learned from the MANIFEST.
3. Secondary starts to replay WAL that contains the writes to "default". It is possible that the write batches' sequence numbers are smaller than the db sequence number. In this case, these write batches will be skipped, and these updates will not be visible to reader until "default" is later flushed.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5413
Differential Revision: D15637407
Pulled By: riversand963
fbshipit-source-id: 3de3fe35cfc6f1b9f844f3f926f0df29717b6580
Summary:
The patch cleans up the handling of cache hit/miss/insertion related
performance counters, get context counters, and statistics by
eliminating some code duplication and factoring out the affected logic
into separate methods. In addition, it makes the semantics of cache hit
metrics more consistent by changing the code so that accessing a
partition of partitioned indexes/filters through a pinned reference no
longer counts as a cache hit.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5408
Differential Revision: D15610883
Pulled By: ltamasi
fbshipit-source-id: ee749c18965077aca971d8f8bee8b24ed8fa76f1
Summary:
It's useful to be able to (optionally) associate key-value pairs with user-provided timestamps. This PR is an early effort towards this goal and continues the work of facebook#4942. A suite of new unit tests exist in DBBasicTestWithTimestampWithParam. Support for timestamp requires the user to provide timestamp as a slice in `ReadOptions` and `WriteOptions`. All timestamps of the same database must share the same length, format, etc. The format of the timestamp is the same throughout the same database, and the user is responsible for providing a comparator function (Comparator) to order the <key, timestamp> tuples. Once created, the format and length of the timestamp cannot change (at least for now).
Test plan (on devserver):
```
$COMPILE_WITH_ASAN=1 make -j32 all
$./db_basic_test --gtest_filter=Timestamp/DBBasicTestWithTimestampWithParam.PutAndGet/*
$make check
```
All tests must pass.
We also run the following db_bench tests to verify whether there is regression on Get/Put while timestamp is not enabled.
```
$TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=fillseq,readrandom -num=1000000
$TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=fillrandom -num=1000000
```
Repeat for 6 times for both versions.
Results are as follows:
```
| | readrandom | fillrandom |
| master | 16.77 MB/s | 47.05 MB/s |
| PR5079 | 16.44 MB/s | 47.03 MB/s |
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5079
Differential Revision: D15132946
Pulled By: riversand963
fbshipit-source-id: 833a0d657eac21182f0f206c910a6438154c742c
Summary:
Flush/compaction use `MergeUntil` which has a special code path to
handle a merge ending with a non-`Merge` point key. In particular if
that key is a `Put` we forgot to check whether it is covered by a range
tombstone. If it is covered then we must not include it in the following call
to `TimedFullMerge`.
Fixes#5392.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5406
Differential Revision: D15611144
Pulled By: sagar0
fbshipit-source-id: ba6a7863ca2d043f591de78fd0c4f4561f0c500e
Summary:
Right now, with auto roll logger, options.keep_log_file_num enforcement is triggered by events like DB reopen or full obsolete scan happens. In the mean time, the size and number of log files can grow without a limit. We put a stronger enforcement to the option, so that the number of log files can always under control.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5370
Differential Revision: D15570413
Pulled By: siying
fbshipit-source-id: 0916c3c4d42ab8fdd29389ee7fd7e1557b03176e
Summary:
1. Fix a bug in WAL replay in which write batches with old sequence numbers are mistakenly inserted into memtables.
2. Add support for benchmarking secondary instance to db_bench_tool.
With changes made in this PR, we can start benchmarking secondary instance
using two processes. It is also possible to vary the frequency at which the
secondary instance tries to catch up with the primary. The info log of the
secondary can be found in a directory whose path can be specified with
'-secondary_path'.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5170
Differential Revision: D15564608
Pulled By: riversand963
fbshipit-source-id: ce97688ed3d33f69d3a0b9266ebbbbf887aa0ec8
Summary:
Currently, when the block cache is used for index blocks as well, it is
not really the index block that is stored in the cache but an
IndexReader object. Since this object is not pure data (it has, for
instance, pointers that might dangle), it's not really sharable. To
avoid the issues around this, the current code uses a dummy unique cache
key for each TableReader to store the IndexReader, and erases the
IndexReader entry when the TableReader is closed. Instead of doing this,
the new code moves the IndexReader out of the cache altogether. In
particular, instead of the TableReader owning, or caching/pinning the
IndexReader based on the customer's settings, the TableReader
unconditionally owns the IndexReader, which in turn owns/caches/pins
the index block (which is itself sharable and thus can be safely put in
the cache without any hacks).
Note: the change has two side effects:
1) Partitions of partitioned indexes no longer affect the read
amplification statistics.
2) Eviction statistics for index blocks are temporarily broken. We plan to fix
this in a separate phase.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5298
Differential Revision: D15303203
Pulled By: ltamasi
fbshipit-source-id: 935a69ba59d87d5e44f42e2310619b790c366e47
Summary:
RocksDB always tries to perform a hard link operation on the external SST file to ingest. This operation can fail if the external SST resides on a different device/FS, or the underlying FS does not support hard link. Currently RocksDB assumes that if the link fails, the user is willing to perform file copy, which is not true according to the post. This commit provides an option named 'failed_move_fall_back_to_copy' for users to choose which behavior they want.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5333
Differential Revision: D15457597
Pulled By: HaoyuHuang
fbshipit-source-id: f3626e13f845db4f7ed970a53ec8a2b1f0d62214
Summary:
Right now, in log writer, we call flush after writing each physical record. I don't see the necessarity of it. Right now, the underlying writer has a buffer, so there isn't a concern that the write request is too large either. On the other hand, in an Env where every flush is expensive, the current approach is significantly slower than only flushing after a whole record finishes, when the record is very large.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5328
Differential Revision: D15425032
Pulled By: siying
fbshipit-source-id: 440ebef002dfbb60c59d8388c9ddfc83d79700aa
Summary:
WritePrepared transactions when configured with two_write_queues=true offers higher throughput with unordered_write feature without however compromising the rocksdb guarantees. This is because it performs ordering among writes in a 2nd step that is not tied to memtable write speed. The 2nd step is naturally provided by 2PC when the commit phase does the ordering as well. Without 2PC, the 2nd step would only be provided when we use two_write_queues=true, where WritePrepared after performing the writes, in a 2nd step uses the 2nd queue to assign order to the writes.
The patch clarifies the need for two_write_queues=true in the HISTORY and inline comments of unordered_writes. Moreover it extends the stress tests of WritePrepared to unordred_write.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5313
Differential Revision: D15379977
Pulled By: maysamyabandeh
fbshipit-source-id: 5b6f05b9b59285dcbf3b0532215ba9fe7d926e00
Summary:
RocksDB secondary can replay both MANIFEST and WAL now.
On the one hand, the memory usage by memtables will grow after replaying WAL for sometime. On the other hand, replaying the MANIFEST can bring the database persistent data to a more recent point in time, giving us the opportunity to discard some memtables containing out-dated data.
This PR coordinates the MANIFEST and WAL replay, using the updates from MANIFEST replay to update the active memtable and immutable memtable list of each column family.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5305
Differential Revision: D15386512
Pulled By: riversand963
fbshipit-source-id: a3ea6fc415f8382d8cf624f52a71ebdcffa3e355
Summary:
Previously if iterator upper/lower bound presents, `DBIter` will check the bound for every key. This patch turns the check into per-file or per-data block check when applicable, by checking against either file largest/smallest key or block index key.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5111
Differential Revision: D15330061
Pulled By: siying
fbshipit-source-id: 8a653fe3cd50d94d81eb2d13b087326c58ee2024
Summary:
After cherry-pick a bug fix to 6.2.fb branch, update the HISTORY.md file to reflect this change.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5309
Differential Revision: D15358002
Pulled By: riversand963
fbshipit-source-id: 5a60510ec6dd444ce5ffaefc69b2e4c38914a921
Summary:
Performing unordered writes in rocksdb when unordered_write option is set to true. When enabled the writes to memtable are done without joining any write thread. This offers much higher write throughput since the upcoming writes would not have to wait for the slowest memtable write to finish. The tradeoff is that the writes visible to a snapshot might change over time. If the application cannot tolerate that, it should implement its own mechanisms to work around that. Using TransactionDB with WRITE_PREPARED write policy is one way to achieve that. Doing so increases the max throughput by 2.2x without however compromising the snapshot guarantees.
The patch is prepared based on an original by siying
Existing unit tests are extended to include unordered_write option.
Benchmark Results:
```
TEST_TMPDIR=/dev/shm/ ./db_bench_unordered --benchmarks=fillrandom --threads=32 --num=10000000 -max_write_buffer_number=16 --max_background_jobs=64 --batch_size=8 --writes=3000000 -level0_file_num_compaction_trigger=99999 --level0_slowdown_writes_trigger=99999 --level0_stop_writes_trigger=99999 -enable_pipelined_write=false -disable_auto_compactions --unordered_write=1
```
With WAL
- Vanilla RocksDB: 78.6 MB/s
- WRITER_PREPARED with unordered_write: 177.8 MB/s (2.2x)
- unordered_write: 368.9 MB/s (4.7x with relaxed snapshot guarantees)
Without WAL
- Vanilla RocksDB: 111.3 MB/s
- WRITER_PREPARED with unordered_write: 259.3 MB/s MB/s (2.3x)
- unordered_write: 645.6 MB/s (5.8x with relaxed snapshot guarantees)
- WRITER_PREPARED with unordered_write disable concurrency control: 185.3 MB/s MB/s (2.35x)
Limitations:
- The feature is not yet extended to `max_successive_merges` > 0. The feature is also incompatible with `enable_pipelined_write` = true as well as with `allow_concurrent_memtable_write` = false.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5218
Differential Revision: D15219029
Pulled By: maysamyabandeh
fbshipit-source-id: 38f2abc4af8780148c6128acdba2b3227bc81759
Summary:
Previous code may call `~ColumnFamilyData` in `DBImpl::AtomicFlushMemTablesToOutputFiles` if the column family is dropped or `cfd->IsFlushPending() == false`. In `~ColumnFamilyData`, the db mutex is released briefly and re-acquired. This can cause correctness issue. The reason is as follows.
Assume there are more bg flush threads. After bg_flush_thr1 releases the db mutex, bg_flush_thr2 can grab it and pop an element from the flush queue. This will cause bg_flush_thr2 to accidentally pick some memtables which should have been picked by bg_flush_thr1. To make the matter worse, bg_flush_thr2 can clear `flush_requested_` flag for the memtable list, causing a subsequent call to `MemTableList::IsFlushPending()` by bg_flush_thr1 to return false, which is wrong.
The fix is to delay `ColumnFamilyData::Unref` and `~ColumnFamilyData` for column families not selected for flush until `AtomicFlushMemTablesToOutputFiles` returns. Furthermore, a bg flush thread should not clear `MemTableList::flush_requested_` in `MemTableList::PickMemtablesToFlush` unless atomic flush is not used **or** the memtable list does not have unpicked memtables.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5294
Differential Revision: D15295297
Pulled By: riversand963
fbshipit-source-id: 03b101205ca22c242647cbf488bcf0ed80b2ecbd
Summary:
When reseek happens in merging iterator, reseeking a child iterator can be avoided if:
(1) the iterator represents imutable data
(2) reseek() to a larger key than the current key
(3) the current key of the child iterator is larger than the seek key
because it is guaranteed that the result will fall into the same position.
This optimization will be useful for use cases where users keep seeking to keys nearby in ascending order.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5286
Differential Revision: D15283635
Pulled By: siying
fbshipit-source-id: 35f79ffd5ce3609146faa8cd55f2bfd733502f83
Summary:
Right now, DBIter::Next() always checks whether an entry is for the same user key as the previous entry to see whether the key should be hidden to the user. However, if previous entry's sequence number is 0, the check is not needed because 0 is the oldest possible sequence number.
We could extend it from seqnum 0 case to simply prev_seqno >= current_seqno. However, it is less robust with bug or unexpected situations, while the gain is relatively low. We can always extend it later when needed.
In a readseq benchmark with full formed LSM-tree, number of key comparisons called is reduced from 2.981 to 2.165. readseq against a fully compacted DB, no key comparison is called. Performance in this benchmark didn't show obvious improvement, which is expected because key comparisons only takes small percentage of CPU. But it may show up to be more effective if users have an expensive customized comparator.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5244
Differential Revision: D15067257
Pulled By: siying
fbshipit-source-id: b7e1ef3ec4fa928cba509683d2b3246e35d270d9
Summary:
Part of compaction cpu goes to processing snapshot list, the larger the list the bigger the overhead. Although the lifetime of most of the snapshots is much shorter than the lifetime of compactions, the compaction conservatively operates on the list of snapshots that it initially obtained. This patch allows the snapshot list to be updated via a callback if the compaction is taking long. This should let the compaction to continue more efficiently with much smaller snapshot list.
For simplicity, to avoid the feature is disabled in two cases: i) When more than one sub-compaction are sharing the same snapshot list, ii) when Range Delete is used in which the range delete aggregator has its own copy of snapshot list.
This fixes the reverted https://github.com/facebook/rocksdb/pull/5099 issue with range deletes.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5278
Differential Revision: D15203291
Pulled By: maysamyabandeh
fbshipit-source-id: fa645611e606aa222c7ce53176dc5bb6f259c258
Summary:
Right now, when Seek() is called again, RocksDB always does a binary search against the files and index blocks, even if they end up with the same file/block. Improve it as following:
1. in LevelIterator, reseek first try to check the boundary of the current file. If it falls into the same file, skip the binary search to find the file
2. in block based table iterator, reseek skip to reseek the iterator block if the seek key is larger than the current key and lower than the index key (boundary of the current block and the next block).
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5256
Differential Revision: D15105072
Pulled By: siying
fbshipit-source-id: 39634bdb4a881082451fa39cecd7ecf12160bf80
Summary:
Sometimes, users might make mistake of not releasing snapshots before closing the DB. This is undocumented use of RocksDB and the behavior is unknown. We return DB::Close() to provide a way to check it for the users. Aborted() will be returned to users when they call DB::Close().
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5272
Differential Revision: D15159713
Pulled By: siying
fbshipit-source-id: 39369def612398d9f239d83d396b5a28e5af65cd
Summary:
Our daily stress tests are failing after this feature. Reverting temporarily until we figure the reason for test failures.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5269
Differential Revision: D15151285
Pulled By: maysamyabandeh
fbshipit-source-id: e4002b99690a97df30d4b4b58bf0f61e9591bc6e
Summary:
Improve the iterators performance when the user explicitly sets the readahead size via `ReadOptions.readahead_size`.
1. Stop creating new table readers when the user explicitly sets readahead size.
2. Make use of an internal buffer based on `FilePrefetchBuffer` instead of using `ReadaheadRandomAccessFileReader`, to handle the user readahead requests (for both buffered and direct io cases).
3. Add `readahead_size` to db_bench.
**Benchmarks:**
https://gist.github.com/sagar0/53693edc320a18abeaeca94ca32f5737
For 1 MB readahead, Buffered IO performance improves by 28% and Direct IO performance improves by 50%.
For 512KB readahead, Buffered IO performance improves by 30% and Direct IO performance improves by 67%.
**Test Plan:**
Updated `DBIteratorTest.ReadAhead` test to make sure that:
- no new table readers are created for iterators on setting ReadOptions.readahead_size
- At least "readahead" number of bytes are actually getting read on each iterator read.
TODO later:
- Use similar logic for compactions as well.
- This ties in nicely with #4052 and paves the way for removing ReadaheadRandomAcessFile later.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5246
Differential Revision: D15107946
Pulled By: sagar0
fbshipit-source-id: 2c1149729ca7d779e4e8b7710ba6f4e8cbfd3bea
Summary:
Part of compaction cpu goes to processing snapshot list, the larger the list the bigger the overhead. Although the lifetime of most of the snapshots is much shorter than the lifetime of compactions, the compaction conservatively operates on the list of snapshots that it initially obtained. This patch allows the snapshot list to be updated via a callback if the compaction is taking long. This should let the compaction to continue more efficiently with much smaller snapshot list.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5099
Differential Revision: D15086710
Pulled By: maysamyabandeh
fbshipit-source-id: 7649f56c3b6b2fb334962048150142a3bf9c1a12
Summary:
- By providing the "env" field in any text-based options (i.e., string, map, or file), we can use `NewCustomObject` to deserialize the text value into an actual `Env` object.
- Currently factory functions for `Env` registered with object registry should only return pointer to static `Env` objects. That's because `DBOptions::env` is a raw pointer so we cannot easily delegate cleanup.
- Note I did not add `env` to `db_option_type_info`. It wasn't needed for (de)serialization, and I believe we don't want to do verification on `env`, even by checking name. That's because the user should be able to copy their DB from Linux to Windows, change envs, and not see an option verification error.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5237
Differential Revision: D15056360
Pulled By: siying
fbshipit-source-id: 4b5f0b83297a5058f8949ec955dbf27d98d73d7e
Summary:
Currently one thread in RocksDB keeps a WAL file open while another thread
deletes it. Although the first thread never writes to the WAL again, it still
tries to close it in the end. This is fine on POSIX, but can be problematic on
other platforms, e.g. HDFS, etc.. It will either cause a lot of warning messages or
throw exceptions. The solution is to let the second thread close the WAL before deleting it.
RocksDB keeps the writers of the logs to delete in `logs_to_free_`, which is passed to `job_context` during `FindObsoleteFiles` (holding mutex). Then in `PurgeObsoleteFiles` (without mutex), these writers should close the logs.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5233
Differential Revision: D15032670
Pulled By: riversand963
fbshipit-source-id: c55e8a612db8cc2306644001a5e6d53842a8f754
Eli Pozniansky