Summary:
This diff adds support for concurrent adds to the skiplist memtable
implementations. Memory allocation is made thread-safe by the addition of
a spinlock, with small per-core buffers to avoid contention. Concurrent
memtable writes are made via an additional method and don't impose a
performance overhead on the non-concurrent case, so parallelism can be
selected on a per-batch basis.
Write thread synchronization is an increasing bottleneck for higher levels
of concurrency, so this diff adds --enable_write_thread_adaptive_yield
(default off). This feature causes threads joining a write batch
group to spin for a short time (default 100 usec) using sched_yield,
rather than going to sleep on a mutex. If the timing of the yield calls
indicates that another thread has actually run during the yield then
spinning is avoided. This option improves performance for concurrent
situations even without parallel adds, although it has the potential to
increase CPU usage (and the heuristic adaptation is not yet mature).
Parallel writes are not currently compatible with
inplace updates, update callbacks, or delete filtering.
Enable it with --allow_concurrent_memtable_write (and
--enable_write_thread_adaptive_yield). Parallel memtable writes
are performance neutral when there is no actual parallelism, and in
my experiments (SSD server-class Linux and varying contention and key
sizes for fillrandom) they are always a performance win when there is
more than one thread.
Statistics are updated earlier in the write path, dropping the number
of DB mutex acquisitions from 2 to 1 for almost all cases.
This diff was motivated and inspired by Yahoo's cLSM work. It is more
conservative than cLSM: RocksDB's write batch group leader role is
preserved (along with all of the existing flush and write throttling
logic) and concurrent writers are blocked until all memtable insertions
have completed and the sequence number has been advanced, to preserve
linearizability.
My test config is "db_bench -benchmarks=fillrandom -threads=$T
-batch_size=1 -memtablerep=skip_list -value_size=100 --num=1000000/$T
-level0_slowdown_writes_trigger=9999 -level0_stop_writes_trigger=9999
-disable_auto_compactions --max_write_buffer_number=8
-max_background_flushes=8 --disable_wal --write_buffer_size=160000000
--block_size=16384 --allow_concurrent_memtable_write" on a two-socket
Xeon E5-2660 @ 2.2Ghz with lots of memory and an SSD hard drive. With 1
thread I get ~440Kops/sec. Peak performance for 1 socket (numactl
-N1) is slightly more than 1Mops/sec, at 16 threads. Peak performance
across both sockets happens at 30 threads, and is ~900Kops/sec, although
with fewer threads there is less performance loss when the system has
background work.
Test Plan:
1. concurrent stress tests for InlineSkipList and DynamicBloom
2. make clean; make check
3. make clean; DISABLE_JEMALLOC=1 make valgrind_check; valgrind db_bench
4. make clean; COMPILE_WITH_TSAN=1 make all check; db_bench
5. make clean; COMPILE_WITH_ASAN=1 make all check; db_bench
6. make clean; OPT=-DROCKSDB_LITE make check
7. verify no perf regressions when disabled
Reviewers: igor, sdong
Reviewed By: sdong
Subscribers: MarkCallaghan, IslamAbdelRahman, anthony, yhchiang, rven, sdong, guyg8, kradhakrishnan, dhruba
Differential Revision: https://reviews.facebook.net/D50589
Summary:
When Get() or NewIterator() trigger file loads, skip caching the filter block if
(1) optimize_filters_for_hits is set and (2) the file is on the bottommost
level. Also skip checking filters under the same conditions, which means that
for a preloaded file or a file that was trivially-moved to the bottom level, its
filter block will eventually expire from the cache.
- added parameters/instance variables in various places in order to propagate the config ("skip_filters") from version_set to block_based_table_reader
- in BlockBasedTable::Rep, this optimization prevents filter from being loaded when the file is opened simply by setting filter_policy = nullptr
- in BlockBasedTable::Get/BlockBasedTable::NewIterator, this optimization prevents filter from being used (even if it was loaded already) by setting filter = nullptr
Test Plan:
updated unit test:
$ ./db_test --gtest_filter=DBTest.OptimizeFiltersForHits
will also run 'make check'
Reviewers: sdong, igor, paultuckfield, anthony, rven, kradhakrishnan, IslamAbdelRahman, yhchiang
Reviewed By: yhchiang
Subscribers: leveldb
Differential Revision: https://reviews.facebook.net/D51633
Summary:
I added this line in my previous patch D48999 (which is incorrect)
We should not release the iterator since releasing it will evict the blocks from cache
Test Plan:
Run the test under valgrind
make check
Reviewers: rven, yhchiang, sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D52161
Summary:
This patch update the Iterator API to introduce new functions that allow users to keep the Slices returned by key() valid as long as the Iterator is not deleted
ReadOptions::pin_data : If true keep loaded blocks in memory as long as the iterator is not deleted
Iterator::IsKeyPinned() : If true, this mean that the Slice returned by key() is valid as long as the iterator is not deleted
Also add a new option BlockBasedTableOptions::use_delta_encoding to allow users to disable delta_encoding if needed.
Benchmark results (using https://phabricator.fb.com/P20083553)
```
// $ du -h /home/tec/local/normal.4K.Snappy/db10077
// 6.1G /home/tec/local/normal.4K.Snappy/db10077
// $ du -h /home/tec/local/zero.8K.LZ4/db10077
// 6.4G /home/tec/local/zero.8K.LZ4/db10077
// Benchmarks for shard db10077
// _build/opt/rocks/benchmark/rocks_copy_benchmark \
// --normal_db_path="/home/tec/local/normal.4K.Snappy/db10077" \
// --zero_db_path="/home/tec/local/zero.8K.LZ4/db10077"
// First run
// ============================================================================
// rocks/benchmark/RocksCopyBenchmark.cpp relative time/iter iters/s
// ============================================================================
// BM_StringCopy 1.73s 576.97m
// BM_StringPiece 103.74% 1.67s 598.55m
// ============================================================================
// Match rate : 1000000 / 1000000
// Second run
// ============================================================================
// rocks/benchmark/RocksCopyBenchmark.cpp relative time/iter iters/s
// ============================================================================
// BM_StringCopy 611.99ms 1.63
// BM_StringPiece 203.76% 300.35ms 3.33
// ============================================================================
// Match rate : 1000000 / 1000000
```
Test Plan: Unit tests
Reviewers: sdong, igor, anthony, yhchiang, rven
Reviewed By: rven
Subscribers: dhruba, lovro, adsharma
Differential Revision: https://reviews.facebook.net/D48999
Summary:
Currently, transactions can fail even if there is no actual write conflict. This is due to relying on only the memtables to check for write-conflicts. Users have to tune memtable settings to try to avoid this, but it's hard to figure out exactly how to tune these settings.
With this diff, TransactionDB will use both memtables and SST files to determine if there are any write conflicts. This relies on the fact that BlockBasedTable stores sequence numbers for all writes that happen after any open snapshot. Also, D50295 is needed to prevent SingleDelete from disappearing writes (the TODOs in this test code will be fixed once the other diff is approved and merged).
Note that Optimistic transactions will still rely on tuning memtable settings as we do not want to read from SST while on the write thread. Also, memtable settings can still be used to reduce how often TransactionDB needs to read SST files.
Test Plan: unit tests, db bench
Reviewers: rven, yhchiang, kradhakrishnan, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb, yoshinorim
Differential Revision: https://reviews.facebook.net/D50475
Summary: In non-mmap mode, plain table reader can issue two pread() for index checking and reading the actual data, although it's for the same location. By reusing the key decoder, we reuse the buffer used for the two to avoid it.
Test Plan: Run unit tests. Run table_reader_bench and see from strace the repeat read cases to disappear.
Reviewers: anthony, yhchiang, rven, kradhakrishnan, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D50949
Summary:
The commit of option helper refactor broken the build:
(1) a git merge problem
(2) some uncaught compiler warning
Fix it.
Test Plan: Make sure "make all" passes
Reviewers: anthony, IslamAbdelRahman, rven, kradhakrishnan, yhchiang
Reviewed By: yhchiang
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D50943
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
Summary:
Reverting https://reviews.facebook.net/D34269 for now
after I landed it a flaky test started continuously failing, I am almost sure this patch is not related to the test but I will revert it until I figure out why it's failing
Test Plan: make check
Reviewers: kradhakrishnan
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D50385
Enable C4307 'operator' : integral constant overflow
Longs and ints on Windows are 32-bit hence the overflow
Enable C4309 'conversion' : truncation of constant value
Enable C4512 'class' : assignment operator could not be generated
Enable C4701 Potentially uninitialized local variable 'name' used
Summary:
This diff introduce new table properties that will be written for block based tables
These properties are
- comparator name
- merge operator name
- property collectors names
Test Plan:
- Added a new unit test to verify that these tests are written/read correctly
- Running all other tests right now (wont land until all tests finish)
Reviewers: rven, kradhakrishnan, igor, sdong, anthony, yhchiang
Reviewed By: yhchiang
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D34269
On Windows two tests fail that use MockTable:
flush_job_test and compaction_job_test with the following message:
compaction_job_test_je.exe : Assertion failed: result.size() == 4,
file c:\dev\rocksdb\rocksdb\table\mock_table.cc, line 110
Investigation reveals that this failure occurs when a 4 byte
ID written to a beginning of the physically open file (main
contents remains in a in-memory map) can not be read back.
The reason for the failure is that the ID is written directly
to a WritableFile bypassing WritableFileWriter. The side effect of that
is that pending_sync_ never becomes true so the file is never flushed,
however, the direct cause of the failure is that the filesize_ member
of the WritableFileWriter remains zero. At Close() the file is truncated
to that size and the file becomes empty so the ID can not be read back.
Summary:
We are cleaning up dependencies.
This diff takes a first step at moving memtable files to their own
directory called memtable. In future diffs, we will move other memtable
files from db to memtable.
Test Plan: make check
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D48915
Summary:
Separate a new class InternalIterator from class Iterator, when the look-up is done internally, which also means they operate on key with sequence ID and type.
This change will enable potential future optimizations but for now InternalIterator's functions are still the same as Iterator's.
At the same time, separate the cleanup function to a separate class and let both of InternalIterator and Iterator inherit from it.
Test Plan: Run all existing tests.
Reviewers: igor, yhchiang, anthony, kradhakrishnan, IslamAbdelRahman, rven
Reviewed By: rven
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D48549
Summary:
Fix tests that compile under ROCKSDB_LITE but currently failing.
table_test:
RandomizedLongDB test is using internal stats which is not supported in ROCKSDB_LITE
compaction_job_test:
Using CompactionJobStats which is not supported
perf_context_test:
KeyComparisonCount test try to open DB in ReadOnly mode which is not supported
Test Plan: run the tests under ROCKSDB_LITE
Reviewers: yhchiang, sdong, igor
Reviewed By: igor
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48585
Summary: It would be nice to have and access to table properties in compaction callbacks. In MyRocks project, it will make possible to update optimizer statistics online.
Test Plan: ran the unit test. Ran myrocks with the new way of collecting stats.
Reviewers: igor, rven, yhchiang
Reviewed By: yhchiang
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48267
Summary: Pass column family ID through TablePropertiesCollectorFactory::CreateTablePropertiesCollector() so that users can identify which column family this file is for and handle it differently.
Test Plan: Add unit test scenarios in tests related to table properties collectors to verify the information passed in is correct.
Reviewers: rven, yhchiang, anthony, kradhakrishnan, igor, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: yoshinorim, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D48411
Summary: Add 2 new counters BLOCK_CACHE_BYTES_WRITE, BLOCK_CACHE_BYTES_READ to keep track of how many bytes were written to the cache and how many bytes that we read from cache
Test Plan: make check
Reviewers: sdong, yhchiang, igor
Reviewed By: igor
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48195
Summary:
hit and miss bloom filter stats for memtable and SST
stats added to perf_context struct
key matches and prefix matches combined into one stat
Test Plan: unit test veryfing the functionality added, see BloomStatsTest in db_test.cc for details
Reviewers: yhchiang, igor, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D47859
Summary:
Since Andres' internship is over, I took over https://reviews.facebook.net/D42555 and rebased and simplified it a bit.
The behavior in this diff is a bit simpler than in D42555:
* only merge operators are passed through FilterMergeValue(). If fitler function returns true, the merge operator is ignored
* compaction filter is *not* called on: 1) results of merge operations and 2) base values that are getting merged with merge operands (the second case was also true in previous diff)
Do we also need a compaction filter to get called on merge results?
Test Plan: make && make check
Reviewers: lovro, tnovak, rven, yhchiang, sdong
Reviewed By: sdong
Subscribers: noetzli, kolmike, leveldb, dhruba, sdong
Differential Revision: https://reviews.facebook.net/D47847
Summary: Recent patch introduced a bug that if non-mmap mode is used, in prefix encoding case, there is a resizing of cur_key_ within the same prefix, we still read prefix from the released buffer. It fails ASAN tests and this commit fixes it.
Test Plan: Run the ASAN tests for the failing test case.
Reviewers: IslamAbdelRahman, yhchiang, anthony, igor, kradhakrishnan, rven
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D47457
Summary:
This is an initial version of bulk load feature
This diff allow us to create sst files, and then bulk load them later, right now the restrictions for loading an sst file are
(1) Memtables are empty
(2) Added sst files have sequence number = 0, and existing values in database have sequence number = 0
(3) Added sst files values are not overlapping
Test Plan: unit testing
Reviewers: igor, ott, sdong
Reviewed By: sdong
Subscribers: leveldb, ott, dhruba
Differential Revision: https://reviews.facebook.net/D39081
Summary:
PlainTableReader now only allows mmap-mode. Add the support to non-mmap mode for more flexibility.
Refactor the codes to move all logic of reading data to PlainTableKeyDecoder, and consolidate the calls to Read() call and ReadVarint32() call. Implement the calls for both of mmap and non-mmap case seperately. For non-mmap mode, make copy of keys in several places when we need to move the buffer after reading the keys.
Test Plan: Add the mode of non-mmap case in plain_table_db_test. Run it in valgrind mode too.
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D47187
Summary: RandomAccessFileReader unnecessarily inherited RandomAccessFile, which can introduce unnecessarily extra costs. Remove it.
Test Plan: Run all existing tests
Reviewers: yhchiang, anthony, igor, kradhakrishnan, rven, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D47409
Summary:
This patch fixes#7460559. It introduces SingleDelete as a new database
operation. This operation can be used to delete keys that were never
overwritten (no put following another put of the same key). If an overwritten
key is single deleted the behavior is undefined. Single deletion of a
non-existent key has no effect but multiple consecutive single deletions are
not allowed (see limitations).
In contrast to the conventional Delete() operation, the deletion entry is
removed along with the value when the two are lined up in a compaction. Note:
The semantics are similar to @igor's prototype that allowed to have this
behavior on the granularity of a column family (
https://reviews.facebook.net/D42093 ). This new patch, however, is more
aggressive when it comes to removing tombstones: It removes the SingleDelete
together with the value whenever there is no snapshot between them while the
older patch only did this when the sequence number of the deletion was older
than the earliest snapshot.
Most of the complex additions are in the Compaction Iterator, all other changes
should be relatively straightforward. The patch also includes basic support for
single deletions in db_stress and db_bench.
Limitations:
- Not compatible with cuckoo hash tables
- Single deletions cannot be used in combination with merges and normal
deletions on the same key (other keys are not affected by this)
- Consecutive single deletions are currently not allowed (and older version of
this patch supported this so it could be resurrected if needed)
Test Plan: make all check
Reviewers: yhchiang, sdong, rven, anthony, yoshinorim, igor
Reviewed By: igor
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D43179
Summary: Missed one file in the previous commit
Test Plan: compiles
Reviewers: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D47055
Summary:
Refactoring NewTableReader to accept TableReaderOptions
This will make it easier to add new options in the future, for example in this diff https://reviews.facebook.net/D46071
Test Plan: run existing tests
Reviewers: igor, yhchiang, anthony, rven, sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D46179
Summary:
In some cases, equality comparisons can be done more efficiently than three-way
comparisons. There are quite a few places in the code where we only care about
equality. This patch adds an Equal() method that defaults to using the
Compare() method.
Test Plan: make clean all check
Reviewers: rven, anthony, yhchiang, igor, sdong
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46233
Summary:
This optimizes the case when (cache_index_and_filter_blocks=1) and bloom filter is not present in the cache.
Previously we did:
1. Read meta block from file
2. Read the filter position from the meta block
3. Read the filter
Now, we pre-load the filter position on Table::Open(), so we can skip steps (1) and (2) on bloom filter cache miss. Instead of 2 IOs, we do only 1.
Test Plan: make check
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46047
Summary:
This diff is a collection of cleanups that were initially part of D43179.
Additionally it adds a unified way of defining key-value maps that use a
Comparator for sorting (this was previously implemented in four different
places).
Test Plan: make clean check all
Reviewers: rven, anthony, yhchiang, sdong, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D45993
Summary: Add ZSTD compression type. The same way as adding LZ4.
Test Plan: run all tests. Generate files in db_bench. Make sure reads succeed. But the SST files cannot be opened in older versions. Also some other adhoc tests.
Reviewers: rven, anthony, IslamAbdelRahman, kradhakrishnan, igor
Reviewed By: igor
Subscribers: MarkCallaghan, maykov, yoshinorim, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D45747
Summary:
This patch adds "rocksdb.aggregated-table-properties"
and "rocksdb.aggregated-table-properties-at-levelN", the former
returns the aggreated table properties of a column family,
while the later returns the aggregated table properties
of the specified level N.
Test Plan: Added tests in db_test
Reviewers: igor, sdong, IslamAbdelRahman, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D45087
Summary: Add more test cases of data race causing wrong iterating results. Tag tests not passing as DISABLED_
Test Plan: Run the tests
Reviewers: igor, rven, IslamAbdelRahman, anthony, kradhakrishnan, yhchiang
Reviewed By: yhchiang
Subscribers: tnovak, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D44907
Summary: Iterator has a bug: if a child iterator reaches its end, and user issues a Prev(), and just before SeekToLast() of the child iterator is called, some extra rows is added in the end, the position of iterator can be misplaced.
Test Plan: Run the tests with or without valgrind
Reviewers: rven, yhchiang, IslamAbdelRahman, anthony
Reviewed By: anthony
Subscribers: tnovak, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D43671
Summary:
Changed compaction_job_test to support better/more thorough
tests and added two tests. Also changed MockFileContents
to order using InternalKeyComparator.
Test Plan: make compaction_job_test && ./compaction_job_test; make all && make check
Reviewers: sdong, rven, igor, yhchiang, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D42837
Summary:
While working on https://reviews.facebook.net/D43179 , I found
duplicate code in the tests. This patch removes it.
Test Plan: make clean all check
Reviewers: igor, sdong, rven, anthony, yhchiang
Reviewed By: yhchiang
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D43263
