rocksdb

fork of https://github.com/oxigraph/rocksdb and https://github.com/facebook/rocksdb for nextgraph and oxigraph

Peter Dillinger 9d0cae7104 Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 ) Summary: When MultiGet() determines that multiple query keys can be served by examining the same data block in block cache (one Lookup()), each PinnableSlice referring to data in that data block needs to hold on to the block in cache so that they can be released at arbitrary times by the API user. Historically this is accomplished with extra calls to Ref() on the Handle from Lookup(), with each PinnableSlice cleanup calling Release() on the Handle, but this creates extra contention on the block cache for the extra Ref()s and Release()es, especially because they hit the same cache shard repeatedly. In the case of merge operands (possibly more cases?), the problem was compounded by doing an extra Ref()+eventual Release() for each merge operand for a key reusing a block (which could be the same key!), rather than one Ref() per key. (Note: the non-shared case with `biter` was already one per key.) This change optimizes MultiGet not to rely on these extra, contentious Ref()+Release() calls by instead, in the shared block case, wrapping the cache Release() cleanup in a refcounted object referenced by the PinnableSlices, such that after the last wrapped reference is released, the cache entry is Release()ed. Relaxed atomic refcounts should be much faster than mutex-guarded Ref() and Release(), and much less prone to a performance cliff when MultiGet() does a lot of block sharing. Note that I did not use std::shared_ptr, because that would require an extra indirection object (shared_ptr itself new/delete) in order to associate a ref increment/decrement with a Cleanable cleanup entry. (If I assumed it was the size of two pointers, I could do some hackery to make it work without the extra indirection, but that's too fragile.) Some details: * Fixed (removed) extra block cache tracing entries in cases of cache entry reuse in MultiGet, but it's likely that in some other cases traces are missing (XXX comment inserted) * Moved existing implementations for cleanable.h from iterator.cc to new cleanable.cc * Improved API comments on Cleanable * Added a public SharedCleanablePtr class to cleanable.h in case others could benefit from the same pattern (potentially many Cleanables and/or smart pointers referencing a shared Cleanable) * Add a typedef for MultiGetContext::Mask * Some variable renaming for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/9899 Test Plan: Added unit tests for SharedCleanablePtr. Greatly enhanced ability of existing tests to detect cache use-after-free. * Release PinnableSlices from MultiGet as they are read rather than in bulk (in db_test_util wrapper). * In ASAN build, default to using a trivially small LRUCache for block_cache so that entries are immediately erased when unreferenced. (Updated two tests that depend on caching.) New ASAN testsuite running time seems OK to me. If I introduce a bug into my implementation where we skip the shared cleanups on block reuse, ASAN detects the bug in `db_basic_test MultiGet`. If I remove either of the above testing enhancements, the bug is not detected. Consider for follow-up work: manipulate or randomize ordering of PinnableSlice use and release from MultiGet db_test_util wrapper. But in typical cases, natural ordering gives pretty good functional coverage. Performance test: In the extreme (but possible) case of MultiGetting the same or adjacent keys in a batch, throughput can improve by an order of magnitude. `./db_bench -benchmarks=multireadrandom -db=/dev/shm/testdb -readonly -num=5 -duration=10 -threads=20 -multiread_batched -batch_size=200` Before ops/sec, num=5: 1,384,394 Before ops/sec, num=500: 6,423,720 After ops/sec, num=500: 10,658,794 After ops/sec, num=5: 16,027,257 Also note that previously, with high parallelism, having query keys concentrated in a single block was worse than spreading them out a bit. Now concentrated in a single block is faster than spread out, which is hopefully consistent with natural expectation. Random query performance: with num=1000000, over 999 x 10s runs running before & after simultaneously (each -threads=12): Before: multireadrandom [AVG 999 runs] : 1088699 (± 7344) ops/sec; 120.4 (± 0.8 ) MB/sec After: multireadrandom [AVG 999 runs] : 1090402 (± 7230) ops/sec; 120.6 (± 0.8 ) MB/sec Possibly better, possibly in the noise. Reviewed By: anand1976 Differential Revision: D35907003 Pulled By: pdillinger fbshipit-source-id: bbd244d703649a8ca12d476f2d03853ed9d1a17e		3 years ago
.circleci	Misc CI improvements / additions (#9859 )	3 years ago
.github/workflows	Use released clang-format instead of the one from dev branch (#9646 )	3 years ago
buckifier	Remove flaky servicelab metrics DBPut P95/P99 (#9844 )	3 years ago
build_tools	Add "no compression" job to CircleCI (#9850 )	3 years ago
cache	Expose `CacheEntryRole` and map keys for block cache stat collections (#9838 )	3 years ago
cmake	gcc-11 and cmake related cleanup (#9286 )	3 years ago
coverage	Fix commit_prereq and other targets (#9797 )	3 years ago
db	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
db_stress_tool	Add WAL compression to stress tests (#9811 )	3 years ago
docs	Bump nokogiri from 1.13.3 to 1.13.4 in /docs (#9831 )	3 years ago
env	Add experimental new FS API AbortIO to cancel read request (#9901 )	3 years ago
examples	Remove BlockBasedTableOptions.hash_index_allow_collision (#9454 )	3 years ago
file	Add stats related to async prefetching (#9845 )	3 years ago
fuzz	Fix compilation errors and add fuzzers to CircleCI (#9420 )	3 years ago
include/rocksdb	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
java	Revert open logic changes in #9634 (#9906 )	3 years ago
logging	Use system-wide thread ID in info log lines (#9164 )	3 years ago
memory	Meta-internal folly integration with F14FastMap (#9546 )	3 years ago
memtable	Account memory of big memory users in BlockBasedTable in global memory limit (#9748 )	3 years ago
microbench	Add mmap DBGet microbench parameters (#9903 )	3 years ago
monitoring	Revert open logic changes in #9634 (#9906 )	3 years ago
options	Misc CI improvements / additions (#9859 )	3 years ago
plugin	Add initial CMake support to plugin (#9214 )	3 years ago
port	Meta-internal folly integration with F14FastMap (#9546 )	3 years ago
table	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
test_util	Remove BlockBasedTableOptions.hash_index_allow_collision (#9454 )	3 years ago
third-party	Meta-internal folly integration with F14FastMap (#9546 )	3 years ago
tools	Add 95% confidence intervals to db_bench output (#9882 )	3 years ago
trace_replay	Added `TraceOptions::preserve_write_order` (#9334 )	3 years ago
util	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
utilities	Add rollback_deletion_type_callback to TxnDBOptions (#9873 )	3 years ago
.clang-format	A script that automatically reformat affected lines	11 years ago
.gitignore	Meta-internal folly integration with F14FastMap (#9546 )	3 years ago
.lgtm.yml	Create lgtm.yml for LGTM.com C/C++ analysis (#4058 )	6 years ago
.travis.yml	Fix remaining uses of "backupable" (#9792 )	3 years ago
.watchmanconfig	Added .watchmanconfig file to rocksdb repo (#5593 )	5 years ago
AUTHORS	Update RocksDB Authors File	7 years ago
CMakeLists.txt	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
CODE_OF_CONDUCT.md	Adopt Contributor Covenant	5 years ago
CONTRIBUTING.md	Add Code of Conduct	7 years ago
COPYING	Add GPLv2 as an alternative license.	8 years ago
DEFAULT_OPTIONS_HISTORY.md	Add Options::DisableExtraChecks, clarify force_consistency_checks (#9363 )	3 years ago
DUMP_FORMAT.md	First version of rocksdb_dump and rocksdb_undump.	10 years ago
HISTORY.md	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
INSTALL.md	Update supported VS versions in INSTALL.md (#9823 )	3 years ago
LANGUAGE-BINDINGS.md	Update branch name to "main" in README/LANGUAGE_BINDINGS (#8727 )	3 years ago
LICENSE.Apache	Change RocksDB License	7 years ago
LICENSE.leveldb	Add back the LevelDB license file	7 years ago
Makefile	Misc CI improvements / additions (#9859 )	3 years ago
PLUGINS.md	Move RADOS support to separate repo (#9206 )	3 years ago
README.md	README: De-list slack channel, list Google group (#9387 )	3 years ago
ROCKSDB_LITE.md	Fix remaining uses of "backupable" (#9792 )	3 years ago
TARGETS	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
USERS.md	Add Solana's RocksDB use case in USERS.md (#9558 )	3 years ago
Vagrantfile	Adding CentOS 7 Vagrantfile & build script	7 years ago
WINDOWS_PORT.md	Update branch name in WINDOWS_PORT.md (#8745 )	3 years ago
crash_test.mk	Add new determinators for multiops transactions stress test (#9708 )	3 years ago
issue_template.md	Add Google Group to Issue Template	5 years ago
python.mk	crash_test Makefile refactoring, add to CircleCI (#9702 )	3 years ago
src.mk	Eliminate unnecessary (slow) block cache Ref()ing in MultiGet (#9899 )	3 years ago
thirdparty.inc	Fix build jemalloc api (#5470 )	5 years ago

README.md

RocksDB: A Persistent Key-Value Store for Flash and RAM Storage

RocksDB is developed and maintained by Facebook Database Engineering Team. It is built on earlier work on LevelDB by Sanjay Ghemawat (sanjay@google.com) and Jeff Dean (jeff@google.com)

This code is a library that forms the core building block for a fast key-value server, especially suited for storing data on flash drives. It has a Log-Structured-Merge-Database (LSM) design with flexible tradeoffs between Write-Amplification-Factor (WAF), Read-Amplification-Factor (RAF) and Space-Amplification-Factor (SAF). It has multi-threaded compactions, making it especially suitable for storing multiple terabytes of data in a single database.

Start with example usage here: https://github.com/facebook/rocksdb/tree/main/examples

See the github wiki for more explanation.

The public interface is in include/. Callers should not include or rely on the details of any other header files in this package. Those internal APIs may be changed without warning.

Questions and discussions are welcome on the RocksDB Developers Public Facebook group and email list on Google Groups.

License

RocksDB is dual-licensed under both the GPLv2 (found in the COPYING file in the root directory) and Apache 2.0 License (found in the LICENSE.Apache file in the root directory). You may select, at your option, one of the above-listed licenses.