rocksdb/db/column_family.cc

1760 lines
68 KiB
C++
Raw Normal View History

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/column_family.h"
#include <algorithm>
#include <cinttypes>
#include <limits>
#include <sstream>
#include <string>
#include <vector>
#include "db/blob/blob_file_cache.h"
#include "db/blob/blob_source.h"
#include "db/compaction/compaction_picker.h"
#include "db/compaction/compaction_picker_fifo.h"
#include "db/compaction/compaction_picker_level.h"
#include "db/compaction/compaction_picker_universal.h"
#include "db/db_impl/db_impl.h"
#include "db/internal_stats.h"
#include "db/job_context.h"
#include "db/range_del_aggregator.h"
#include "db/table_properties_collector.h"
#include "db/version_set.h"
#include "db/write_controller.h"
#include "file/sst_file_manager_impl.h"
#include "logging/logging.h"
#include "monitoring/thread_status_util.h"
#include "options/options_helper.h"
Auto enable Periodic Compactions if a Compaction Filter is used (#5865) Summary: - Periodic compactions are auto-enabled if a compaction filter or a compaction filter factory is set, in Level Compaction. - The default value of `periodic_compaction_seconds` is changed to UINT64_MAX, which lets RocksDB auto-tune periodic compactions as needed. An explicit value of 0 will still work as before ie. to disable periodic compactions completely. For now, on seeing a compaction filter along with a UINT64_MAX value for `periodic_compaction_seconds`, RocksDB will make SST files older than 30 days to go through periodic copmactions. Some RocksDB users make use of compaction filters to control when their data can be deleted, usually with a custom TTL logic. But it is occasionally possible that the compactions get delayed by considerable time due to factors like low writes to a key range, data reaching bottom level, etc before the TTL expiry. Periodic Compactions feature was originally built to help such cases. Now periodic compactions are auto enabled by default when compaction filters or compaction filter factories are used, as it is generally helpful to all cases to collect garbage. `periodic_compaction_seconds` is set to a large value, 30 days, in `SanitizeOptions` when RocksDB sees that a `compaction_filter` or `compaction_filter_factory` is used. This is done only for Level Compaction style. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5865 Test Plan: - Added a new test `DBCompactionTest.LevelPeriodicCompactionWithCompactionFilters` to make sure that `periodic_compaction_seconds` is set if either `compaction_filter` or `compaction_filter_factory` options are set. - `COMPILE_WITH_ASAN=1 make check` Differential Revision: D17659180 Pulled By: sagar0 fbshipit-source-id: 4887b9cf2e53cf2dc93a7b658c6b15e1181217ee
2019-10-29 22:04:26 +00:00
#include "port/port.h"
#include "rocksdb/convenience.h"
#include "rocksdb/table.h"
#include "table/merging_iterator.h"
#include "util/autovector.h"
#include "util/cast_util.h"
#include "util/compression.h"
namespace ROCKSDB_NAMESPACE {
ColumnFamilyHandleImpl::ColumnFamilyHandleImpl(
ColumnFamilyData* column_family_data, DBImpl* db, InstrumentedMutex* mutex)
: cfd_(column_family_data), db_(db), mutex_(mutex) {
if (cfd_ != nullptr) {
cfd_->Ref();
}
}
ColumnFamilyHandleImpl::~ColumnFamilyHandleImpl() {
if (cfd_ != nullptr) {
for (auto& listener : cfd_->ioptions()->listeners) {
listener->OnColumnFamilyHandleDeletionStarted(this);
}
// Job id == 0 means that this is not our background process, but rather
// user thread
// Need to hold some shared pointers owned by the initial_cf_options
// before final cleaning up finishes.
ColumnFamilyOptions initial_cf_options_copy = cfd_->initial_cf_options();
JobContext job_context(0);
mutex_->Lock();
bool dropped = cfd_->IsDropped();
if (cfd_->UnrefAndTryDelete()) {
if (dropped) {
db_->FindObsoleteFiles(&job_context, false, true);
}
}
mutex_->Unlock();
if (job_context.HaveSomethingToDelete()) {
bool defer_purge =
db_->immutable_db_options().avoid_unnecessary_blocking_io;
db_->PurgeObsoleteFiles(job_context, defer_purge);
}
job_context.Clean();
}
}
uint32_t ColumnFamilyHandleImpl::GetID() const { return cfd()->GetID(); }
const std::string& ColumnFamilyHandleImpl::GetName() const {
return cfd()->GetName();
}
Status ColumnFamilyHandleImpl::GetDescriptor(ColumnFamilyDescriptor* desc) {
// accessing mutable cf-options requires db mutex.
InstrumentedMutexLock l(mutex_);
*desc = ColumnFamilyDescriptor(cfd()->GetName(), cfd()->GetLatestCFOptions());
return Status::OK();
}
const Comparator* ColumnFamilyHandleImpl::GetComparator() const {
return cfd()->user_comparator();
}
void GetIntTblPropCollectorFactory(
const ImmutableCFOptions& ioptions,
IntTblPropCollectorFactories* int_tbl_prop_collector_factories) {
assert(int_tbl_prop_collector_factories);
auto& collector_factories = ioptions.table_properties_collector_factories;
for (size_t i = 0; i < ioptions.table_properties_collector_factories.size();
++i) {
assert(collector_factories[i]);
int_tbl_prop_collector_factories->emplace_back(
new UserKeyTablePropertiesCollectorFactory(collector_factories[i]));
}
}
Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
if (!cf_options.compression_per_level.empty()) {
for (size_t level = 0; level < cf_options.compression_per_level.size();
++level) {
if (!CompressionTypeSupported(cf_options.compression_per_level[level])) {
return Status::InvalidArgument(
"Compression type " +
CompressionTypeToString(cf_options.compression_per_level[level]) +
" is not linked with the binary.");
}
}
} else {
if (!CompressionTypeSupported(cf_options.compression)) {
return Status::InvalidArgument(
"Compression type " +
CompressionTypeToString(cf_options.compression) +
" is not linked with the binary.");
}
}
if (cf_options.compression_opts.zstd_max_train_bytes > 0) {
Support using ZDICT_finalizeDictionary to generate zstd dictionary (#9857) Summary: An untrained dictionary is currently simply the concatenation of several samples. The ZSTD API, ZDICT_finalizeDictionary(), can improve such a dictionary's effectiveness at low cost. This PR changes how dictionary is created by calling the ZSTD ZDICT_finalizeDictionary() API instead of creating raw content dictionary (when max_dict_buffer_bytes > 0), and pass in all buffered uncompressed data blocks as samples. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9857 Test Plan: #### db_bench test for cpu/memory of compression+decompression and space saving on synthetic data: Set up: change the parameter [here](https://github.com/facebook/rocksdb/blob/fb9a167a55e0970b1ef6f67c1600c8d9c4c6114f/tools/db_bench_tool.cc#L1766) to 16384 to make synthetic data more compressible. ``` # linked local ZSTD with version 1.5.2 # DEBUG_LEVEL=0 ROCKSDB_NO_FBCODE=1 ROCKSDB_DISABLE_ZSTD=1 EXTRA_CXXFLAGS="-DZSTD_STATIC_LINKING_ONLY -DZSTD -I/data/users/changyubi/install/include/" EXTRA_LDFLAGS="-L/data/users/changyubi/install/lib/ -l:libzstd.a" make -j32 db_bench dict_bytes=16384 train_bytes=1048576 echo "========== No Dictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== Raw Content Dictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench_main -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench_main -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== FinalizeDictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== TrainDictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total # Result: TrainDictionary is much better on space saving, but FinalizeDictionary seems to use less memory. # before compression data size: 1.2GB dict_bytes=16384 max_dict_buffer_bytes = 1048576 space cpu/memory No Dictionary 468M 14.93user 1.00system 0:15.92elapsed 100%CPU (0avgtext+0avgdata 23904maxresident)k Raw Dictionary 251M 15.81user 0.80system 0:16.56elapsed 100%CPU (0avgtext+0avgdata 156808maxresident)k FinalizeDictionary 236M 11.93user 0.64system 0:12.56elapsed 100%CPU (0avgtext+0avgdata 89548maxresident)k TrainDictionary 84M 7.29user 0.45system 0:07.75elapsed 100%CPU (0avgtext+0avgdata 97288maxresident)k ``` #### Benchmark on 10 sample SST files for spacing saving and CPU time on compression: FinalizeDictionary is comparable to TrainDictionary in terms of space saving, and takes less time in compression. ``` dict_bytes=16384 train_bytes=1048576 for sst_file in `ls ../temp/myrock-sst/` do echo "********** $sst_file **********" echo "========== No Dictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD echo "========== Raw Content Dictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes echo "========== FinalizeDictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes --compression_use_zstd_finalize_dict echo "========== TrainDictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes done 010240.sst (Size/Time) 011029.sst 013184.sst 021552.sst 185054.sst 185137.sst 191666.sst 7560381.sst 7604174.sst 7635312.sst No Dictionary 28165569 / 2614419 32899411 / 2976832 32977848 / 3055542 31966329 / 2004590 33614351 / 1755877 33429029 / 1717042 33611933 / 1776936 33634045 / 2771417 33789721 / 2205414 33592194 / 388254 Raw Content Dictionary 28019950 / 2697961 33748665 / 3572422 33896373 / 3534701 26418431 / 2259658 28560825 / 1839168 28455030 / 1846039 28494319 / 1861349 32391599 / 3095649 33772142 / 2407843 33592230 / 474523 FinalizeDictionary 27896012 / 2650029 33763886 / 3719427 33904283 / 3552793 26008225 / 2198033 28111872 / 1869530 28014374 / 1789771 28047706 / 1848300 32296254 / 3204027 33698698 / 2381468 33592344 / 517433 TrainDictionary 28046089 / 2740037 33706480 / 3679019 33885741 / 3629351 25087123 / 2204558 27194353 / 1970207 27234229 / 1896811 27166710 / 1903119 32011041 / 3322315 32730692 / 2406146 33608631 / 570593 ``` #### Decompression/Read test: With FinalizeDictionary/TrainDictionary, some data structure used for decompression are in stored in dictionary, so they are expected to be faster in terms of decompression/reads. ``` dict_bytes=16384 train_bytes=1048576 echo "No Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=0 > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=0 2>&1 | grep MB/s echo "Raw Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes 2>&1 | grep MB/s echo "FinalizeDict" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false 2>&1 | grep MB/s echo "Train Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes 2>&1 | grep MB/s No Dictionary readrandom : 12.183 micros/op 82082 ops/sec 12.183 seconds 1000000 operations; 9.1 MB/s (1000000 of 1000000 found) Raw Dictionary readrandom : 12.314 micros/op 81205 ops/sec 12.314 seconds 1000000 operations; 9.0 MB/s (1000000 of 1000000 found) FinalizeDict readrandom : 9.787 micros/op 102180 ops/sec 9.787 seconds 1000000 operations; 11.3 MB/s (1000000 of 1000000 found) Train Dictionary readrandom : 9.698 micros/op 103108 ops/sec 9.699 seconds 1000000 operations; 11.4 MB/s (1000000 of 1000000 found) ``` Reviewed By: ajkr Differential Revision: D35720026 Pulled By: cbi42 fbshipit-source-id: 24d230fdff0fd28a1bb650658798f00dfcfb2a1f
2022-05-20 19:09:09 +00:00
if (cf_options.compression_opts.use_zstd_dict_trainer) {
if (!ZSTD_TrainDictionarySupported()) {
return Status::InvalidArgument(
"zstd dictionary trainer cannot be used because ZSTD 1.1.3+ "
"is not linked with the binary.");
}
} else if (!ZSTD_FinalizeDictionarySupported()) {
return Status::InvalidArgument(
Support using ZDICT_finalizeDictionary to generate zstd dictionary (#9857) Summary: An untrained dictionary is currently simply the concatenation of several samples. The ZSTD API, ZDICT_finalizeDictionary(), can improve such a dictionary's effectiveness at low cost. This PR changes how dictionary is created by calling the ZSTD ZDICT_finalizeDictionary() API instead of creating raw content dictionary (when max_dict_buffer_bytes > 0), and pass in all buffered uncompressed data blocks as samples. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9857 Test Plan: #### db_bench test for cpu/memory of compression+decompression and space saving on synthetic data: Set up: change the parameter [here](https://github.com/facebook/rocksdb/blob/fb9a167a55e0970b1ef6f67c1600c8d9c4c6114f/tools/db_bench_tool.cc#L1766) to 16384 to make synthetic data more compressible. ``` # linked local ZSTD with version 1.5.2 # DEBUG_LEVEL=0 ROCKSDB_NO_FBCODE=1 ROCKSDB_DISABLE_ZSTD=1 EXTRA_CXXFLAGS="-DZSTD_STATIC_LINKING_ONLY -DZSTD -I/data/users/changyubi/install/include/" EXTRA_LDFLAGS="-L/data/users/changyubi/install/lib/ -l:libzstd.a" make -j32 db_bench dict_bytes=16384 train_bytes=1048576 echo "========== No Dictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== Raw Content Dictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench_main -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench_main -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== FinalizeDictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total echo "========== TrainDictionary ==========" TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1 TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 2>&1 | grep elapsed du -hc /dev/shm/dbbench/*sst | grep total # Result: TrainDictionary is much better on space saving, but FinalizeDictionary seems to use less memory. # before compression data size: 1.2GB dict_bytes=16384 max_dict_buffer_bytes = 1048576 space cpu/memory No Dictionary 468M 14.93user 1.00system 0:15.92elapsed 100%CPU (0avgtext+0avgdata 23904maxresident)k Raw Dictionary 251M 15.81user 0.80system 0:16.56elapsed 100%CPU (0avgtext+0avgdata 156808maxresident)k FinalizeDictionary 236M 11.93user 0.64system 0:12.56elapsed 100%CPU (0avgtext+0avgdata 89548maxresident)k TrainDictionary 84M 7.29user 0.45system 0:07.75elapsed 100%CPU (0avgtext+0avgdata 97288maxresident)k ``` #### Benchmark on 10 sample SST files for spacing saving and CPU time on compression: FinalizeDictionary is comparable to TrainDictionary in terms of space saving, and takes less time in compression. ``` dict_bytes=16384 train_bytes=1048576 for sst_file in `ls ../temp/myrock-sst/` do echo "********** $sst_file **********" echo "========== No Dictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD echo "========== Raw Content Dictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes echo "========== FinalizeDictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes --compression_use_zstd_finalize_dict echo "========== TrainDictionary ==========" ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes done 010240.sst (Size/Time) 011029.sst 013184.sst 021552.sst 185054.sst 185137.sst 191666.sst 7560381.sst 7604174.sst 7635312.sst No Dictionary 28165569 / 2614419 32899411 / 2976832 32977848 / 3055542 31966329 / 2004590 33614351 / 1755877 33429029 / 1717042 33611933 / 1776936 33634045 / 2771417 33789721 / 2205414 33592194 / 388254 Raw Content Dictionary 28019950 / 2697961 33748665 / 3572422 33896373 / 3534701 26418431 / 2259658 28560825 / 1839168 28455030 / 1846039 28494319 / 1861349 32391599 / 3095649 33772142 / 2407843 33592230 / 474523 FinalizeDictionary 27896012 / 2650029 33763886 / 3719427 33904283 / 3552793 26008225 / 2198033 28111872 / 1869530 28014374 / 1789771 28047706 / 1848300 32296254 / 3204027 33698698 / 2381468 33592344 / 517433 TrainDictionary 28046089 / 2740037 33706480 / 3679019 33885741 / 3629351 25087123 / 2204558 27194353 / 1970207 27234229 / 1896811 27166710 / 1903119 32011041 / 3322315 32730692 / 2406146 33608631 / 570593 ``` #### Decompression/Read test: With FinalizeDictionary/TrainDictionary, some data structure used for decompression are in stored in dictionary, so they are expected to be faster in terms of decompression/reads. ``` dict_bytes=16384 train_bytes=1048576 echo "No Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=0 > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=0 2>&1 | grep MB/s echo "Raw Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes 2>&1 | grep MB/s echo "FinalizeDict" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false 2>&1 | grep MB/s echo "Train Dictionary" TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes > /dev/null 2>&1 TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes 2>&1 | grep MB/s No Dictionary readrandom : 12.183 micros/op 82082 ops/sec 12.183 seconds 1000000 operations; 9.1 MB/s (1000000 of 1000000 found) Raw Dictionary readrandom : 12.314 micros/op 81205 ops/sec 12.314 seconds 1000000 operations; 9.0 MB/s (1000000 of 1000000 found) FinalizeDict readrandom : 9.787 micros/op 102180 ops/sec 9.787 seconds 1000000 operations; 11.3 MB/s (1000000 of 1000000 found) Train Dictionary readrandom : 9.698 micros/op 103108 ops/sec 9.699 seconds 1000000 operations; 11.4 MB/s (1000000 of 1000000 found) ``` Reviewed By: ajkr Differential Revision: D35720026 Pulled By: cbi42 fbshipit-source-id: 24d230fdff0fd28a1bb650658798f00dfcfb2a1f
2022-05-20 19:09:09 +00:00
"zstd finalizeDictionary cannot be used because ZSTD 1.4.5+ "
"is not linked with the binary.");
}
if (cf_options.compression_opts.max_dict_bytes == 0) {
return Status::InvalidArgument(
"The dictionary size limit (`CompressionOptions::max_dict_bytes`) "
"should be nonzero if we're using zstd's dictionary generator.");
}
}
if (!CompressionTypeSupported(cf_options.blob_compression_type)) {
std::ostringstream oss;
oss << "The specified blob compression type "
<< CompressionTypeToString(cf_options.blob_compression_type)
<< " is not available.";
return Status::InvalidArgument(oss.str());
}
return Status::OK();
}
support for concurrent adds to memtable 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
2015-08-14 23:59:07 +00:00
Status CheckConcurrentWritesSupported(const ColumnFamilyOptions& cf_options) {
if (cf_options.inplace_update_support) {
return Status::InvalidArgument(
"In-place memtable updates (inplace_update_support) is not compatible "
"with concurrent writes (allow_concurrent_memtable_write)");
}
if (!cf_options.memtable_factory->IsInsertConcurrentlySupported()) {
return Status::InvalidArgument(
"Memtable doesn't concurrent writes (allow_concurrent_memtable_write)");
}
support for concurrent adds to memtable 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
2015-08-14 23:59:07 +00:00
return Status::OK();
}
Status CheckCFPathsSupported(const DBOptions& db_options,
const ColumnFamilyOptions& cf_options) {
// More than one cf_paths are supported only in universal
// and level compaction styles. This function also checks the case
// in which cf_paths is not specified, which results in db_paths
// being used.
if ((cf_options.compaction_style != kCompactionStyleUniversal) &&
(cf_options.compaction_style != kCompactionStyleLevel)) {
if (cf_options.cf_paths.size() > 1) {
return Status::NotSupported(
"More than one CF paths are only supported in "
"universal and level compaction styles. ");
} else if (cf_options.cf_paths.empty() && db_options.db_paths.size() > 1) {
return Status::NotSupported(
"More than one DB paths are only supported in "
"universal and level compaction styles. ");
}
}
return Status::OK();
}
namespace {
const uint64_t kDefaultTtl = 0xfffffffffffffffe;
const uint64_t kDefaultPeriodicCompSecs = 0xfffffffffffffffe;
} // anonymous namespace
ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
const ColumnFamilyOptions& src) {
ColumnFamilyOptions result = src;
size_t clamp_max = std::conditional<
sizeof(size_t) == 4, std::integral_constant<size_t, 0xffffffff>,
std::integral_constant<uint64_t, 64ull << 30>>::type::value;
ClipToRange(&result.write_buffer_size, (static_cast<size_t>(64)) << 10,
clamp_max);
// if user sets arena_block_size, we trust user to use this value. Otherwise,
// calculate a proper value from writer_buffer_size;
if (result.arena_block_size <= 0) {
result.arena_block_size =
std::min(size_t{1024 * 1024}, result.write_buffer_size / 8);
// Align up to 4k
const size_t align = 4 * 1024;
result.arena_block_size =
((result.arena_block_size + align - 1) / align) * align;
}
result.min_write_buffer_number_to_merge =
std::min(result.min_write_buffer_number_to_merge,
result.max_write_buffer_number - 1);
if (result.min_write_buffer_number_to_merge < 1) {
result.min_write_buffer_number_to_merge = 1;
}
if (db_options.atomic_flush && result.min_write_buffer_number_to_merge > 1) {
ROCKS_LOG_WARN(
db_options.logger,
"Currently, if atomic_flush is true, then triggering flush for any "
"column family internally (non-manual flush) will trigger flushing "
"all column families even if the number of memtables is smaller "
"min_write_buffer_number_to_merge. Therefore, configuring "
"min_write_buffer_number_to_merge > 1 is not compatible and should "
"be satinized to 1. Not doing so will lead to data loss and "
"inconsistent state across multiple column families when WAL is "
"disabled, which is a common setting for atomic flush");
result.min_write_buffer_number_to_merge = 1;
}
if (result.num_levels < 1) {
result.num_levels = 1;
}
if (result.compaction_style == kCompactionStyleLevel &&
result.num_levels < 2) {
result.num_levels = 2;
}
if (result.compaction_style == kCompactionStyleUniversal &&
db_options.allow_ingest_behind && result.num_levels < 3) {
result.num_levels = 3;
}
if (result.max_write_buffer_number < 2) {
result.max_write_buffer_number = 2;
}
Refactor trimming logic for immutable memtables (#5022) Summary: MyRocks currently sets `max_write_buffer_number_to_maintain` in order to maintain enough history for transaction conflict checking. The effectiveness of this approach depends on the size of memtables. When memtables are small, it may not keep enough history; when memtables are large, this may consume too much memory. We are proposing a new way to configure memtable list history: by limiting the memory usage of immutable memtables. The new option is `max_write_buffer_size_to_maintain` and it will take precedence over the old `max_write_buffer_number_to_maintain` if they are both set to non-zero values. The new option accounts for the total memory usage of flushed immutable memtables and mutable memtable. When the total usage exceeds the limit, RocksDB may start dropping immutable memtables (which is also called trimming history), starting from the oldest one. The semantics of the old option actually works both as an upper bound and lower bound. History trimming will start if number of immutable memtables exceeds the limit, but it will never go below (limit-1) due to history trimming. In order the mimic the behavior with the new option, history trimming will stop if dropping the next immutable memtable causes the total memory usage go below the size limit. For example, assuming the size limit is set to 64MB, and there are 3 immutable memtables with sizes of 20, 30, 30. Although the total memory usage is 80MB > 64MB, dropping the oldest memtable will reduce the memory usage to 60MB < 64MB, so in this case no memtable will be dropped. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5022 Differential Revision: D14394062 Pulled By: miasantreble fbshipit-source-id: 60457a509c6af89d0993f988c9b5c2aa9e45f5c5
2019-08-23 20:54:09 +00:00
// fall back max_write_buffer_number_to_maintain if
// max_write_buffer_size_to_maintain is not set
if (result.max_write_buffer_size_to_maintain < 0) {
result.max_write_buffer_size_to_maintain =
result.max_write_buffer_number *
static_cast<int64_t>(result.write_buffer_size);
} else if (result.max_write_buffer_size_to_maintain == 0 &&
result.max_write_buffer_number_to_maintain < 0) {
Support saving history in memtable_list Summary: For transactions, we are using the memtables to validate that there are no write conflicts. But after flushing, we don't have any memtables, and transactions could fail to commit. So we want to someone keep around some extra history to use for conflict checking. In addition, we want to provide a way to increase the size of this history if too many transactions fail to commit. After chatting with people, it seems like everyone prefers just using Memtables to store this history (instead of a separate history structure). It seems like the best place for this is abstracted inside the memtable_list. I decide to create a separate list in MemtableListVersion as using the same list complicated the flush/installalflushresults logic too much. This diff adds a new parameter to control how much memtable history to keep around after flushing. However, it sounds like people aren't too fond of adding new parameters. So I am making the default size of flushed+not-flushed memtables be set to max_write_buffers. This should not change the maximum amount of memory used, but make it more likely we're using closer the the limit. (We are now postponing deleting flushed memtables until the max_write_buffer limit is reached). So while we might use more memory on average, we are still obeying the limit set (and you could argue it's better to go ahead and use up memory now instead of waiting for a write stall to happen to test this limit). However, if people are opposed to this default behavior, we can easily set it to 0 and require this parameter be set in order to use transactions. Test Plan: Added a xfunc test to play around with setting different values of this parameter in all tests. Added testing in memtablelist_test and planning on adding more testing here. Reviewers: sdong, rven, igor Reviewed By: igor Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D37443
2015-05-28 23:34:24 +00:00
result.max_write_buffer_number_to_maintain = result.max_write_buffer_number;
}
// bloom filter size shouldn't exceed 1/4 of memtable size.
if (result.memtable_prefix_bloom_size_ratio > 0.25) {
result.memtable_prefix_bloom_size_ratio = 0.25;
} else if (result.memtable_prefix_bloom_size_ratio < 0) {
result.memtable_prefix_bloom_size_ratio = 0;
}
Support saving history in memtable_list Summary: For transactions, we are using the memtables to validate that there are no write conflicts. But after flushing, we don't have any memtables, and transactions could fail to commit. So we want to someone keep around some extra history to use for conflict checking. In addition, we want to provide a way to increase the size of this history if too many transactions fail to commit. After chatting with people, it seems like everyone prefers just using Memtables to store this history (instead of a separate history structure). It seems like the best place for this is abstracted inside the memtable_list. I decide to create a separate list in MemtableListVersion as using the same list complicated the flush/installalflushresults logic too much. This diff adds a new parameter to control how much memtable history to keep around after flushing. However, it sounds like people aren't too fond of adding new parameters. So I am making the default size of flushed+not-flushed memtables be set to max_write_buffers. This should not change the maximum amount of memory used, but make it more likely we're using closer the the limit. (We are now postponing deleting flushed memtables until the max_write_buffer limit is reached). So while we might use more memory on average, we are still obeying the limit set (and you could argue it's better to go ahead and use up memory now instead of waiting for a write stall to happen to test this limit). However, if people are opposed to this default behavior, we can easily set it to 0 and require this parameter be set in order to use transactions. Test Plan: Added a xfunc test to play around with setting different values of this parameter in all tests. Added testing in memtablelist_test and planning on adding more testing here. Reviewers: sdong, rven, igor Reviewed By: igor Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D37443
2015-05-28 23:34:24 +00:00
if (!result.prefix_extractor) {
assert(result.memtable_factory);
Slice name = result.memtable_factory->Name();
if (name.compare("HashSkipListRepFactory") == 0 ||
name.compare("HashLinkListRepFactory") == 0) {
result.memtable_factory = std::make_shared<SkipListFactory>();
}
}
if (result.compaction_style == kCompactionStyleFIFO) {
// since we delete level0 files in FIFO compaction when there are too many
// of them, these options don't really mean anything
result.level0_slowdown_writes_trigger = std::numeric_limits<int>::max();
result.level0_stop_writes_trigger = std::numeric_limits<int>::max();
}
if (result.max_bytes_for_level_multiplier <= 0) {
result.max_bytes_for_level_multiplier = 1;
}
if (result.level0_file_num_compaction_trigger == 0) {
ROCKS_LOG_WARN(db_options.logger,
"level0_file_num_compaction_trigger cannot be 0");
result.level0_file_num_compaction_trigger = 1;
}
if (result.level0_stop_writes_trigger <
result.level0_slowdown_writes_trigger ||
result.level0_slowdown_writes_trigger <
result.level0_file_num_compaction_trigger) {
ROCKS_LOG_WARN(db_options.logger,
"This condition must be satisfied: "
"level0_stop_writes_trigger(%d) >= "
"level0_slowdown_writes_trigger(%d) >= "
"level0_file_num_compaction_trigger(%d)",
result.level0_stop_writes_trigger,
result.level0_slowdown_writes_trigger,
result.level0_file_num_compaction_trigger);
if (result.level0_slowdown_writes_trigger <
result.level0_file_num_compaction_trigger) {
result.level0_slowdown_writes_trigger =
result.level0_file_num_compaction_trigger;
}
if (result.level0_stop_writes_trigger <
result.level0_slowdown_writes_trigger) {
result.level0_stop_writes_trigger = result.level0_slowdown_writes_trigger;
}
ROCKS_LOG_WARN(db_options.logger,
"Adjust the value to "
"level0_stop_writes_trigger(%d)"
"level0_slowdown_writes_trigger(%d)"
"level0_file_num_compaction_trigger(%d)",
result.level0_stop_writes_trigger,
result.level0_slowdown_writes_trigger,
result.level0_file_num_compaction_trigger);
}
if (result.soft_pending_compaction_bytes_limit == 0) {
result.soft_pending_compaction_bytes_limit =
result.hard_pending_compaction_bytes_limit;
} else if (result.hard_pending_compaction_bytes_limit > 0 &&
result.soft_pending_compaction_bytes_limit >
result.hard_pending_compaction_bytes_limit) {
result.soft_pending_compaction_bytes_limit =
result.hard_pending_compaction_bytes_limit;
}
// When the DB is stopped, it's possible that there are some .trash files that
// were not deleted yet, when we open the DB we will find these .trash files
// and schedule them to be deleted (or delete immediately if SstFileManager
// was not used)
auto sfm =
static_cast<SstFileManagerImpl*>(db_options.sst_file_manager.get());
for (size_t i = 0; i < result.cf_paths.size(); i++) {
DeleteScheduler::CleanupDirectory(db_options.env, sfm,
result.cf_paths[i].path)
.PermitUncheckedError();
}
if (result.cf_paths.empty()) {
result.cf_paths = db_options.db_paths;
}
if (result.level_compaction_dynamic_level_bytes) {
if (result.compaction_style != kCompactionStyleLevel) {
ROCKS_LOG_WARN(db_options.info_log.get(),
"level_compaction_dynamic_level_bytes only makes sense"
"for level-based compaction");
result.level_compaction_dynamic_level_bytes = false;
} else if (result.cf_paths.size() > 1U) {
// we don't yet know how to make both of this feature and multiple
// DB path work.
ROCKS_LOG_WARN(db_options.info_log.get(),
"multiple cf_paths/db_paths and"
"level_compaction_dynamic_level_bytes"
"can't be used together");
result.level_compaction_dynamic_level_bytes = false;
}
}
if (result.max_compaction_bytes == 0) {
result.max_compaction_bytes = result.target_file_size_base * 25;
}
bool is_block_based_table = (result.table_factory->IsInstanceOf(
TableFactory::kBlockBasedTableName()));
const uint64_t kAdjustedTtl = 30 * 24 * 60 * 60;
if (result.ttl == kDefaultTtl) {
if (is_block_based_table) {
// FIFO also requires max_open_files=-1, which is checked in
// ValidateOptions().
result.ttl = kAdjustedTtl;
} else {
result.ttl = 0;
}
}
const uint64_t kAdjustedPeriodicCompSecs = 30 * 24 * 60 * 60;
if (result.compaction_style == kCompactionStyleLevel) {
if ((result.compaction_filter != nullptr ||
result.compaction_filter_factory != nullptr) &&
result.periodic_compaction_seconds == kDefaultPeriodicCompSecs &&
is_block_based_table) {
result.periodic_compaction_seconds = kAdjustedPeriodicCompSecs;
}
} else if (result.compaction_style == kCompactionStyleUniversal) {
if (result.periodic_compaction_seconds == kDefaultPeriodicCompSecs &&
is_block_based_table) {
result.periodic_compaction_seconds = kAdjustedPeriodicCompSecs;
}
} else if (result.compaction_style == kCompactionStyleFIFO) {
if (result.periodic_compaction_seconds != kDefaultPeriodicCompSecs) {
ROCKS_LOG_WARN(
db_options.info_log.get(),
"periodic_compaction_seconds does not support FIFO compaction. You"
"may want to set option TTL instead.");
}
Auto enable Periodic Compactions if a Compaction Filter is used (#5865) Summary: - Periodic compactions are auto-enabled if a compaction filter or a compaction filter factory is set, in Level Compaction. - The default value of `periodic_compaction_seconds` is changed to UINT64_MAX, which lets RocksDB auto-tune periodic compactions as needed. An explicit value of 0 will still work as before ie. to disable periodic compactions completely. For now, on seeing a compaction filter along with a UINT64_MAX value for `periodic_compaction_seconds`, RocksDB will make SST files older than 30 days to go through periodic copmactions. Some RocksDB users make use of compaction filters to control when their data can be deleted, usually with a custom TTL logic. But it is occasionally possible that the compactions get delayed by considerable time due to factors like low writes to a key range, data reaching bottom level, etc before the TTL expiry. Periodic Compactions feature was originally built to help such cases. Now periodic compactions are auto enabled by default when compaction filters or compaction filter factories are used, as it is generally helpful to all cases to collect garbage. `periodic_compaction_seconds` is set to a large value, 30 days, in `SanitizeOptions` when RocksDB sees that a `compaction_filter` or `compaction_filter_factory` is used. This is done only for Level Compaction style. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5865 Test Plan: - Added a new test `DBCompactionTest.LevelPeriodicCompactionWithCompactionFilters` to make sure that `periodic_compaction_seconds` is set if either `compaction_filter` or `compaction_filter_factory` options are set. - `COMPILE_WITH_ASAN=1 make check` Differential Revision: D17659180 Pulled By: sagar0 fbshipit-source-id: 4887b9cf2e53cf2dc93a7b658c6b15e1181217ee
2019-10-29 22:04:26 +00:00
}
// For universal compaction, `ttl` and `periodic_compaction_seconds` mean the
// same thing, take the stricter value.
if (result.compaction_style == kCompactionStyleUniversal) {
if (result.periodic_compaction_seconds == 0) {
result.periodic_compaction_seconds = result.ttl;
} else if (result.ttl != 0) {
result.periodic_compaction_seconds =
std::min(result.ttl, result.periodic_compaction_seconds);
}
}
if (result.periodic_compaction_seconds == kDefaultPeriodicCompSecs) {
result.periodic_compaction_seconds = 0;
}
return result;
}
int SuperVersion::dummy = 0;
void* const SuperVersion::kSVInUse = &SuperVersion::dummy;
void* const SuperVersion::kSVObsolete = nullptr;
SuperVersion::~SuperVersion() {
for (auto td : to_delete) {
delete td;
}
}
SuperVersion* SuperVersion::Ref() {
refs.fetch_add(1, std::memory_order_relaxed);
return this;
}
bool SuperVersion::Unref() {
// fetch_sub returns the previous value of ref
uint32_t previous_refs = refs.fetch_sub(1);
assert(previous_refs > 0);
return previous_refs == 1;
}
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
2021-07-10 00:16:00 +00:00
void SuperVersion::Cleanup() {
assert(refs.load(std::memory_order_relaxed) == 0);
Memtable "MemPurge" prototype (#8454) Summary: Implement an experimental feature called "MemPurge", which consists in purging "garbage" bytes out of a memtable and reuse the memtable struct instead of making it immutable and eventually flushing its content to storage. The prototype is by default deactivated and is not intended for use. It is intended for correctness and validation testing. At the moment, the "MemPurge" feature can be switched on by using the `options.experimental_allow_mempurge` flag. For this early stage, when the allow_mempurge flag is set to `true`, all the flush operations will be rerouted to perform a MemPurge. This is a temporary design decision that will give us the time to explore meaningful heuristics to use MemPurge at the right time for relevant workloads . Moreover, the current MemPurge operation only supports `Puts`, `Deletes`, `DeleteRange` operations, and handles `Iterators` as well as `CompactionFilter`s that are invoked at flush time . Three unit tests are added to `db_flush_test.cc` to test if MemPurge works correctly (and checks that the previously mentioned operations are fully supported thoroughly tested). One noticeable design decision is the timing of the MemPurge operation in the memtable workflow: for this prototype, the mempurge happens when the memtable is switched (and usually made immutable). This is an inefficient process because it implies that the entirety of the MemPurge operation happens while holding the db_mutex. Future commits will make the MemPurge operation a background task (akin to the regular flush operation) and aim at drastically enhancing the performance of this operation. The MemPurge is also not fully "WAL-compatible" yet, but when the WAL is full, or when the regular MemPurge operation fails (or when the purged memtable still needs to be flushed), a regular flush operation takes place. Later commits will also correct these behaviors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8454 Reviewed By: anand1976 Differential Revision: D29433971 Pulled By: bjlemaire fbshipit-source-id: 6af48213554e35048a7e03816955100a80a26dc5
2021-07-02 12:22:03 +00:00
// Since this SuperVersion object is being deleted,
// decrement reference to the immutable MemtableList
// this SV object was pointing to.
imm->Unref(&to_delete);
MemTable* m = mem->Unref();
if (m != nullptr) {
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
2021-07-10 00:16:00 +00:00
auto* memory_usage = current->cfd()->imm()->current_memory_usage();
assert(*memory_usage >= m->ApproximateMemoryUsage());
*memory_usage -= m->ApproximateMemoryUsage();
to_delete.push_back(m);
}
current->Unref();
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
cfd->UnrefAndTryDelete();
}
void SuperVersion::Init(ColumnFamilyData* new_cfd, MemTable* new_mem,
MemTableListVersion* new_imm, Version* new_current) {
cfd = new_cfd;
mem = new_mem;
imm = new_imm;
current = new_current;
cfd->Ref();
mem->Ref();
imm->Ref();
current->Ref();
refs.store(1, std::memory_order_relaxed);
}
namespace {
void SuperVersionUnrefHandle(void* ptr) {
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
// UnrefHandle is called when a thread exits or a ThreadLocalPtr gets
// destroyed. When the former happens, the thread shouldn't see kSVInUse.
// When the latter happens, only super_version_ holds a reference
// to ColumnFamilyData, so no further queries are possible.
SuperVersion* sv = static_cast<SuperVersion*>(ptr);
bool was_last_ref __attribute__((__unused__));
was_last_ref = sv->Unref();
// Thread-local SuperVersions can't outlive ColumnFamilyData::super_version_.
// This is important because we can't do SuperVersion cleanup here.
// That would require locking DB mutex, which would deadlock because
// SuperVersionUnrefHandle is called with locked ThreadLocalPtr mutex.
assert(!was_last_ref);
}
} // anonymous namespace
std::vector<std::string> ColumnFamilyData::GetDbPaths() const {
std::vector<std::string> paths;
paths.reserve(ioptions_.cf_paths.size());
for (const DbPath& db_path : ioptions_.cf_paths) {
paths.emplace_back(db_path.path);
}
return paths;
}
const uint32_t ColumnFamilyData::kDummyColumnFamilyDataId =
std::numeric_limits<uint32_t>::max();
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
ColumnFamilyData::ColumnFamilyData(
uint32_t id, const std::string& name, Version* _dummy_versions,
Cache* _table_cache, WriteBufferManager* write_buffer_manager,
const ColumnFamilyOptions& cf_options, const ImmutableDBOptions& db_options,
const FileOptions* file_options, ColumnFamilySet* column_family_set,
BlockCacheTracer* const block_cache_tracer,
const std::shared_ptr<IOTracer>& io_tracer, const std::string& db_id,
const std::string& db_session_id)
: id_(id),
name_(name),
dummy_versions_(_dummy_versions),
current_(nullptr),
refs_(0),
initialized_(false),
dropped_(false),
internal_comparator_(cf_options.comparator),
initial_cf_options_(SanitizeOptions(db_options, cf_options)),
ioptions_(db_options, initial_cf_options_),
mutable_cf_options_(initial_cf_options_),
is_delete_range_supported_(
cf_options.table_factory->IsDeleteRangeSupported()),
write_buffer_manager_(write_buffer_manager),
mem_(nullptr),
imm_(ioptions_.min_write_buffer_number_to_merge,
Refactor trimming logic for immutable memtables (#5022) Summary: MyRocks currently sets `max_write_buffer_number_to_maintain` in order to maintain enough history for transaction conflict checking. The effectiveness of this approach depends on the size of memtables. When memtables are small, it may not keep enough history; when memtables are large, this may consume too much memory. We are proposing a new way to configure memtable list history: by limiting the memory usage of immutable memtables. The new option is `max_write_buffer_size_to_maintain` and it will take precedence over the old `max_write_buffer_number_to_maintain` if they are both set to non-zero values. The new option accounts for the total memory usage of flushed immutable memtables and mutable memtable. When the total usage exceeds the limit, RocksDB may start dropping immutable memtables (which is also called trimming history), starting from the oldest one. The semantics of the old option actually works both as an upper bound and lower bound. History trimming will start if number of immutable memtables exceeds the limit, but it will never go below (limit-1) due to history trimming. In order the mimic the behavior with the new option, history trimming will stop if dropping the next immutable memtable causes the total memory usage go below the size limit. For example, assuming the size limit is set to 64MB, and there are 3 immutable memtables with sizes of 20, 30, 30. Although the total memory usage is 80MB > 64MB, dropping the oldest memtable will reduce the memory usage to 60MB < 64MB, so in this case no memtable will be dropped. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5022 Differential Revision: D14394062 Pulled By: miasantreble fbshipit-source-id: 60457a509c6af89d0993f988c9b5c2aa9e45f5c5
2019-08-23 20:54:09 +00:00
ioptions_.max_write_buffer_number_to_maintain,
ioptions_.max_write_buffer_size_to_maintain),
super_version_(nullptr),
super_version_number_(0),
local_sv_(new ThreadLocalPtr(&SuperVersionUnrefHandle)),
next_(nullptr),
prev_(nullptr),
log_number_(0),
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
column_family_set_(column_family_set),
queued_for_flush_(false),
queued_for_compaction_(false),
prev_compaction_needed_bytes_(0),
allow_2pc_(db_options.allow_2pc),
last_memtable_id_(0),
Dynamically changeable `MemPurge` option (#10011) Summary: **Summary** Make the mempurge option flag a Mutable Column Family option flag. Therefore, the mempurge feature can be dynamically toggled. **Motivation** RocksDB users prefer having the ability to switch features on and off without having to close and reopen the DB. This is particularly important if the feature causes issues and needs to be turned off. Dynamically changing a DB option flag does not seem currently possible. Moreover, with this new change, the MemPurge feature can be toggled on or off independently between column families, which we see as a major improvement. **Content of this PR** This PR includes removal of the `experimental_mempurge_threshold` flag as a DB option flag, and its re-introduction as a `MutableCFOption` flag. I updated the code to handle dynamic changes of the flag (in particular inside the `FlushJob` file). Additionally, this PR includes a new test to demonstrate the capacity of the code to toggle the MemPurge feature on and off, as well as the addition in the `db_stress` module of 2 different mempurge threshold values (0.0 and 1.0) that can be randomly changed with the `set_option_one_in` flag. This is useful to stress test the dynamic changes. **Benchmarking** I will add numbers to prove that there is no performance impact within the next 12 hours. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10011 Reviewed By: pdillinger Differential Revision: D36462357 Pulled By: bjlemaire fbshipit-source-id: 5e3d63bdadf085c0572ecc2349e7dd9729ce1802
2022-06-23 16:42:18 +00:00
db_paths_registered_(false),
Sort L0 files by newly introduced epoch_num (#10922) 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 https://github.com/ajkr/rocksdb/commit/36a5686ec012f35a4371e409aa85c404ca1c210d (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
2022-12-13 21:29:37 +00:00
mempurge_used_(false),
next_epoch_number_(1) {
if (id_ != kDummyColumnFamilyDataId) {
// TODO(cc): RegisterDbPaths can be expensive, considering moving it
// outside of this constructor which might be called with db mutex held.
// TODO(cc): considering using ioptions_.fs, currently some tests rely on
// EnvWrapper, that's the main reason why we use env here.
Status s = ioptions_.env->RegisterDbPaths(GetDbPaths());
if (s.ok()) {
db_paths_registered_ = true;
} else {
ROCKS_LOG_ERROR(
ioptions_.logger,
"Failed to register data paths of column family (id: %d, name: %s)",
id_, name_.c_str());
}
}
Ref();
// Convert user defined table properties collector factories to internal ones.
GetIntTblPropCollectorFactory(ioptions_, &int_tbl_prop_collector_factories_);
// if _dummy_versions is nullptr, then this is a dummy column family.
if (_dummy_versions != nullptr) {
internal_stats_.reset(
new InternalStats(ioptions_.num_levels, ioptions_.clock, this));
Introduce a new storage specific Env API (#5761) Summary: The current Env API encompasses both storage/file operations, as well as OS related operations. Most of the APIs return a Status, which does not have enough metadata about an error, such as whether its retry-able or not, scope (i.e fault domain) of the error etc., that may be required in order to properly handle a storage error. The file APIs also do not provide enough control over the IO SLA, such as timeout, prioritization, hinting about placement and redundancy etc. This PR separates out the file/storage APIs from Env into a new FileSystem class. The APIs are updated to return an IOStatus with metadata about the error, as well as to take an IOOptions structure as input in order to allow more control over the IO. The user can set both ```options.env``` and ```options.file_system``` to specify that RocksDB should use the former for OS related operations and the latter for storage operations. Internally, a ```CompositeEnvWrapper``` has been introduced that inherits from ```Env``` and redirects individual methods to either an ```Env``` implementation or the ```FileSystem``` as appropriate. When options are sanitized during ```DB::Open```, ```options.env``` is replaced with a newly allocated ```CompositeEnvWrapper``` instance if both env and file_system have been specified. This way, the rest of the RocksDB code can continue to function as before. This PR also ports PosixEnv to the new API by splitting it into two - PosixEnv and PosixFileSystem. PosixEnv is defined as a sub-class of CompositeEnvWrapper, and threading/time functions are overridden with Posix specific implementations in order to avoid an extra level of indirection. The ```CompositeEnvWrapper``` translates ```IOStatus``` return code to ```Status```, and sets the severity to ```kSoftError``` if the io_status is retryable. The error handling code in RocksDB can then recover the DB automatically. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5761 Differential Revision: D18868376 Pulled By: anand1976 fbshipit-source-id: 39efe18a162ea746fabac6360ff529baba48486f
2019-12-13 22:47:08 +00:00
table_cache_.reset(new TableCache(ioptions_, file_options, _table_cache,
block_cache_tracer, io_tracer,
db_session_id));
blob_file_cache_.reset(
new BlobFileCache(_table_cache, ioptions(), soptions(), id_,
internal_stats_->GetBlobFileReadHist(), io_tracer));
blob_source_.reset(new BlobSource(ioptions(), db_id, db_session_id,
blob_file_cache_.get()));
if (ioptions_.compaction_style == kCompactionStyleLevel) {
compaction_picker_.reset(
new LevelCompactionPicker(ioptions_, &internal_comparator_));
} else if (ioptions_.compaction_style == kCompactionStyleUniversal) {
compaction_picker_.reset(
new UniversalCompactionPicker(ioptions_, &internal_comparator_));
} else if (ioptions_.compaction_style == kCompactionStyleFIFO) {
compaction_picker_.reset(
new FIFOCompactionPicker(ioptions_, &internal_comparator_));
} else if (ioptions_.compaction_style == kCompactionStyleNone) {
compaction_picker_.reset(
new NullCompactionPicker(ioptions_, &internal_comparator_));
ROCKS_LOG_WARN(ioptions_.logger,
"Column family %s does not use any background compaction. "
"Compactions can only be done via CompactFiles\n",
GetName().c_str());
} else {
ROCKS_LOG_ERROR(ioptions_.logger,
"Unable to recognize the specified compaction style %d. "
"Column family %s will use kCompactionStyleLevel.\n",
ioptions_.compaction_style, GetName().c_str());
compaction_picker_.reset(
new LevelCompactionPicker(ioptions_, &internal_comparator_));
}
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
if (column_family_set_->NumberOfColumnFamilies() < 10) {
ROCKS_LOG_INFO(ioptions_.logger,
"--------------- Options for column family [%s]:\n",
name.c_str());
initial_cf_options_.Dump(ioptions_.logger);
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
} else {
ROCKS_LOG_INFO(ioptions_.logger, "\t(skipping printing options)\n");
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
}
}
RecalculateWriteStallConditions(mutable_cf_options_);
Account memory of FileMetaData in global memory limit (#9924) Summary: **Context/Summary:** As revealed by heap profiling, allocation of `FileMetaData` for [newly created file added to a Version](https://github.com/facebook/rocksdb/pull/9924/files#diff-a6aa385940793f95a2c5b39cc670bd440c4547fa54fd44622f756382d5e47e43R774) can consume significant heap memory. This PR is to account that toward our global memory limit based on block cache capacity. Pull Request resolved: https://github.com/facebook/rocksdb/pull/9924 Test Plan: - Previous `make check` verified there are only 2 places where the memory of the allocated `FileMetaData` can be released - New unit test `TEST_P(ChargeFileMetadataTestWithParam, Basic)` - db bench (CPU cost of `charge_file_metadata` in write and compact) - **write micros/op: -0.24%** : `TEST_TMPDIR=/dev/shm/testdb ./db_bench -benchmarks=fillseq -db=$TEST_TMPDIR -charge_file_metadata=1 (remove this option for pre-PR) -disable_auto_compactions=1 -write_buffer_size=100000 -num=4000000 | egrep 'fillseq'` - **compact micros/op -0.87%** : `TEST_TMPDIR=/dev/shm/testdb ./db_bench -benchmarks=fillseq -db=$TEST_TMPDIR -charge_file_metadata=1 -disable_auto_compactions=1 -write_buffer_size=100000 -num=4000000 -numdistinct=1000 && ./db_bench -benchmarks=compact -db=$TEST_TMPDIR -use_existing_db=1 -charge_file_metadata=1 -disable_auto_compactions=1 | egrep 'compact'` table 1 - write #-run | (pre-PR) avg micros/op | std micros/op | (post-PR) micros/op | std micros/op | change (%) -- | -- | -- | -- | -- | -- 10 | 3.9711 | 0.264408 | 3.9914 | 0.254563 | 0.5111933721 20 | 3.83905 | 0.0664488 | 3.8251 | 0.0695456 | -0.3633711465 40 | 3.86625 | 0.136669 | 3.8867 | 0.143765 | 0.5289363078 80 | 3.87828 | 0.119007 | 3.86791 | 0.115674 | **-0.2673865734** 160 | 3.87677 | 0.162231 | 3.86739 | 0.16663 | **-0.2419539978** table 2 - compact #-run | (pre-PR) avg micros/op | std micros/op | (post-PR) micros/op | std micros/op | change (%) -- | -- | -- | -- | -- | -- 10 | 2,399,650.00 | 96,375.80 | 2,359,537.00 | 53,243.60 | -1.67 20 | 2,410,480.00 | 89,988.00 | 2,433,580.00 | 91,121.20 | 0.96 40 | 2.41E+06 | 121811 | 2.39E+06 | 131525 | **-0.96** 80 | 2.40E+06 | 134503 | 2.39E+06 | 108799 | **-0.78** - stress test: `python3 tools/db_crashtest.py blackbox --charge_file_metadata=1 --cache_size=1` killed as normal Reviewed By: ajkr Differential Revision: D36055583 Pulled By: hx235 fbshipit-source-id: b60eab94707103cb1322cf815f05810ef0232625
2022-06-14 20:06:40 +00:00
if (cf_options.table_factory->IsInstanceOf(
TableFactory::kBlockBasedTableName()) &&
cf_options.table_factory->GetOptions<BlockBasedTableOptions>()) {
const BlockBasedTableOptions* bbto =
cf_options.table_factory->GetOptions<BlockBasedTableOptions>();
const auto& options_overrides = bbto->cache_usage_options.options_overrides;
const auto file_metadata_charged =
options_overrides.at(CacheEntryRole::kFileMetadata).charged;
if (bbto->block_cache &&
file_metadata_charged == CacheEntryRoleOptions::Decision::kEnabled) {
// TODO(hx235): Add a `ConcurrentCacheReservationManager` at DB scope
// responsible for reservation of `ObsoleteFileInfo` so that we can keep
// this `file_metadata_cache_res_mgr_` nonconcurrent
file_metadata_cache_res_mgr_.reset(new ConcurrentCacheReservationManager(
std::make_shared<
CacheReservationManagerImpl<CacheEntryRole::kFileMetadata>>(
bbto->block_cache)));
}
}
}
// DB mutex held
ColumnFamilyData::~ColumnFamilyData() {
assert(refs_.load(std::memory_order_relaxed) == 0);
// remove from linked list
auto prev = prev_;
auto next = next_;
prev->next_ = next;
next->prev_ = prev;
2015-01-06 20:44:21 +00:00
if (!dropped_ && column_family_set_ != nullptr) {
// If it's dropped, it's already removed from column family set
// If column_family_set_ == nullptr, this is dummy CFD and not in
// ColumnFamilySet
column_family_set_->RemoveColumnFamily(this);
}
if (current_ != nullptr) {
current_->Unref();
}
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
// It would be wrong if this ColumnFamilyData is in flush_queue_ or
// compaction_queue_ and we destroyed it
assert(!queued_for_flush_);
assert(!queued_for_compaction_);
assert(super_version_ == nullptr);
if (dummy_versions_ != nullptr) {
// List must be empty
assert(dummy_versions_->Next() == dummy_versions_);
bool deleted __attribute__((__unused__));
deleted = dummy_versions_->Unref();
assert(deleted);
}
if (mem_ != nullptr) {
delete mem_->Unref();
}
autovector<MemTable*> to_delete;
imm_.current()->Unref(&to_delete);
for (MemTable* m : to_delete) {
delete m;
}
if (db_paths_registered_) {
// TODO(cc): considering using ioptions_.fs, currently some tests rely on
// EnvWrapper, that's the main reason why we use env here.
Status s = ioptions_.env->UnregisterDbPaths(GetDbPaths());
if (!s.ok()) {
ROCKS_LOG_ERROR(
ioptions_.logger,
"Failed to unregister data paths of column family (id: %d, name: %s)",
id_, name_.c_str());
}
}
}
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
bool ColumnFamilyData::UnrefAndTryDelete() {
int old_refs = refs_.fetch_sub(1);
assert(old_refs > 0);
if (old_refs == 1) {
assert(super_version_ == nullptr);
delete this;
return true;
}
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
if (old_refs == 2 && super_version_ != nullptr) {
// Only the super_version_ holds me
SuperVersion* sv = super_version_;
super_version_ = nullptr;
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
// Release SuperVersion references kept in ThreadLocalPtr.
local_sv_.reset();
if (sv->Unref()) {
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
// Note: sv will delete this ColumnFamilyData during Cleanup()
assert(sv->cfd == this);
sv->Cleanup();
delete sv;
return true;
}
}
return false;
}
2015-01-06 20:44:21 +00:00
void ColumnFamilyData::SetDropped() {
// can't drop default CF
assert(id_ != 0);
dropped_ = true;
write_controller_token_.reset();
// remove from column_family_set
column_family_set_->RemoveColumnFamily(this);
}
ColumnFamilyOptions ColumnFamilyData::GetLatestCFOptions() const {
return BuildColumnFamilyOptions(initial_cf_options_, mutable_cf_options_);
}
uint64_t ColumnFamilyData::OldestLogToKeep() {
auto current_log = GetLogNumber();
if (allow_2pc_) {
auto imm_prep_log = imm()->PrecomputeMinLogContainingPrepSection();
auto mem_prep_log = mem()->GetMinLogContainingPrepSection();
if (imm_prep_log > 0 && imm_prep_log < current_log) {
current_log = imm_prep_log;
}
if (mem_prep_log > 0 && mem_prep_log < current_log) {
current_log = mem_prep_log;
}
}
return current_log;
}
const double kIncSlowdownRatio = 0.8;
const double kDecSlowdownRatio = 1 / kIncSlowdownRatio;
const double kNearStopSlowdownRatio = 0.6;
const double kDelayRecoverSlowdownRatio = 1.4;
namespace {
// If penalize_stop is true, we further reduce slowdown rate.
std::unique_ptr<WriteControllerToken> SetupDelay(
WriteController* write_controller, uint64_t compaction_needed_bytes,
uint64_t prev_compaction_need_bytes, bool penalize_stop,
bool auto_compactions_disabled) {
const uint64_t kMinWriteRate = 16 * 1024u; // Minimum write rate 16KB/s.
uint64_t max_write_rate = write_controller->max_delayed_write_rate();
uint64_t write_rate = write_controller->delayed_write_rate();
if (auto_compactions_disabled) {
// When auto compaction is disabled, always use the value user gave.
write_rate = max_write_rate;
} else if (write_controller->NeedsDelay() && max_write_rate > kMinWriteRate) {
// If user gives rate less than kMinWriteRate, don't adjust it.
//
// If already delayed, need to adjust based on previous compaction debt.
// When there are two or more column families require delay, we always
// increase or reduce write rate based on information for one single
// column family. It is likely to be OK but we can improve if there is a
// problem.
// Ignore compaction_needed_bytes = 0 case because compaction_needed_bytes
// is only available in level-based compaction
//
// If the compaction debt stays the same as previously, we also further slow
// down. It usually means a mem table is full. It's mainly for the case
// where both of flush and compaction are much slower than the speed we
// insert to mem tables, so we need to actively slow down before we get
// feedback signal from compaction and flushes to avoid the full stop
// because of hitting the max write buffer number.
//
// If DB just falled into the stop condition, we need to further reduce
// the write rate to avoid the stop condition.
if (penalize_stop) {
// Penalize the near stop or stop condition by more aggressive slowdown.
// This is to provide the long term slowdown increase signal.
// The penalty is more than the reward of recovering to the normal
// condition.
write_rate = static_cast<uint64_t>(static_cast<double>(write_rate) *
kNearStopSlowdownRatio);
if (write_rate < kMinWriteRate) {
write_rate = kMinWriteRate;
}
} else if (prev_compaction_need_bytes > 0 &&
prev_compaction_need_bytes <= compaction_needed_bytes) {
write_rate = static_cast<uint64_t>(static_cast<double>(write_rate) *
kIncSlowdownRatio);
if (write_rate < kMinWriteRate) {
write_rate = kMinWriteRate;
}
} else if (prev_compaction_need_bytes > compaction_needed_bytes) {
// We are speeding up by ratio of kSlowdownRatio when we have paid
// compaction debt. But we'll never speed up to faster than the write rate
// given by users.
write_rate = static_cast<uint64_t>(static_cast<double>(write_rate) *
kDecSlowdownRatio);
if (write_rate > max_write_rate) {
write_rate = max_write_rate;
}
}
}
return write_controller->GetDelayToken(write_rate);
}
int GetL0ThresholdSpeedupCompaction(int level0_file_num_compaction_trigger,
int level0_slowdown_writes_trigger) {
// SanitizeOptions() ensures it.
assert(level0_file_num_compaction_trigger <= level0_slowdown_writes_trigger);
if (level0_file_num_compaction_trigger < 0) {
return std::numeric_limits<int>::max();
}
const int64_t twice_level0_trigger =
static_cast<int64_t>(level0_file_num_compaction_trigger) * 2;
const int64_t one_fourth_trigger_slowdown =
static_cast<int64_t>(level0_file_num_compaction_trigger) +
((level0_slowdown_writes_trigger - level0_file_num_compaction_trigger) /
4);
assert(twice_level0_trigger >= 0);
assert(one_fourth_trigger_slowdown >= 0);
// 1/4 of the way between L0 compaction trigger threshold and slowdown
// condition.
// Or twice as compaction trigger, if it is smaller.
int64_t res = std::min(twice_level0_trigger, one_fourth_trigger_slowdown);
if (res >= std::numeric_limits<int32_t>::max()) {
return std::numeric_limits<int32_t>::max();
} else {
// res fits in int
return static_cast<int>(res);
}
}
} // anonymous namespace
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
std::pair<WriteStallCondition, WriteStallCause>
ColumnFamilyData::GetWriteStallConditionAndCause(
int num_unflushed_memtables, int num_l0_files,
uint64_t num_compaction_needed_bytes,
Fix checkpoint stuck (#7921) Summary: ## 1. Bug description: When RocksDB Checkpoint, it may be stuck in `WaitUntilFlushWouldNotStallWrites` method. ## 2. Simple analysis of the reasons: ### 2.1 Configuration parameters: ```yaml Compaction Style : Universal max_write_buffer_number : 4 min_write_buffer_number_to_merge : 3 ``` Checkpoint is usually very fast. When the Checkpoint is executed, `WaitUntilFlushWouldNotStallWrites` is called. If there are 2 Immutable MemTables, which are less than `min_write_buffer_number_to_merge`, they will not be flushed. But will enter this code. ```c++ // method: GetWriteStallConditionAndCause if (mutable_cf_options.max_write_buffer_number> 3 && num_unflushed_memtables >= mutable_cf_options.max_write_buffer_number-1) { return {WriteStallCondition::kDelayed, WriteStallCause::kMemtableLimit}; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/column_family.cc#L847 Checkpoint thought there was a FlushJob, but it didn't. So will always wait. ### 2.2 solution: Increase the restriction: the `number of Immutable MemTable` >= `min_write_buffer_number_to_merge will wait`. If there are other better solutions, you can correct me. ### 2.3 Code that can reproduce the problem: https://github.com/1996fanrui/fanrui-learning/blob/flink-1.12/module-java/src/main/java/com/dream/rocksdb/RocksDBCheckpointStuck.java ## 3. Interesting point This bug will be triggered only when `the number of sorted runs >= level0_file_num_compaction_trigger`. Because there is a break in WaitUntilFlushWouldNotStallWrites. ```c++ if (cfd->imm()->NumNotFlushed() < cfd->ioptions()->min_write_buffer_number_to_merge && vstorage->l0_delay_trigger_count() < mutable_cf_options.level0_file_num_compaction_trigger) { break; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/db_impl/db_impl_compaction_flush.cc#L1974 Universal may have `l0_delay_trigger_count() >= level0_file_num_compaction_trigger`, so this bug is triggered. Pull Request resolved: https://github.com/facebook/rocksdb/pull/7921 Reviewed By: jay-zhuang Differential Revision: D26900559 Pulled By: ajkr fbshipit-source-id: 133c1252dad7393753f04a47590b68c7d8e670df
2021-03-09 10:19:28 +00:00
const MutableCFOptions& mutable_cf_options,
const ImmutableCFOptions& immutable_cf_options) {
if (num_unflushed_memtables >= mutable_cf_options.max_write_buffer_number) {
return {WriteStallCondition::kStopped, WriteStallCause::kMemtableLimit};
} else if (!mutable_cf_options.disable_auto_compactions &&
num_l0_files >= mutable_cf_options.level0_stop_writes_trigger) {
return {WriteStallCondition::kStopped, WriteStallCause::kL0FileCountLimit};
} else if (!mutable_cf_options.disable_auto_compactions &&
mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
num_compaction_needed_bytes >=
mutable_cf_options.hard_pending_compaction_bytes_limit) {
return {WriteStallCondition::kStopped,
WriteStallCause::kPendingCompactionBytes};
} else if (mutable_cf_options.max_write_buffer_number > 3 &&
num_unflushed_memtables >=
Fix checkpoint stuck (#7921) Summary: ## 1. Bug description: When RocksDB Checkpoint, it may be stuck in `WaitUntilFlushWouldNotStallWrites` method. ## 2. Simple analysis of the reasons: ### 2.1 Configuration parameters: ```yaml Compaction Style : Universal max_write_buffer_number : 4 min_write_buffer_number_to_merge : 3 ``` Checkpoint is usually very fast. When the Checkpoint is executed, `WaitUntilFlushWouldNotStallWrites` is called. If there are 2 Immutable MemTables, which are less than `min_write_buffer_number_to_merge`, they will not be flushed. But will enter this code. ```c++ // method: GetWriteStallConditionAndCause if (mutable_cf_options.max_write_buffer_number> 3 && num_unflushed_memtables >= mutable_cf_options.max_write_buffer_number-1) { return {WriteStallCondition::kDelayed, WriteStallCause::kMemtableLimit}; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/column_family.cc#L847 Checkpoint thought there was a FlushJob, but it didn't. So will always wait. ### 2.2 solution: Increase the restriction: the `number of Immutable MemTable` >= `min_write_buffer_number_to_merge will wait`. If there are other better solutions, you can correct me. ### 2.3 Code that can reproduce the problem: https://github.com/1996fanrui/fanrui-learning/blob/flink-1.12/module-java/src/main/java/com/dream/rocksdb/RocksDBCheckpointStuck.java ## 3. Interesting point This bug will be triggered only when `the number of sorted runs >= level0_file_num_compaction_trigger`. Because there is a break in WaitUntilFlushWouldNotStallWrites. ```c++ if (cfd->imm()->NumNotFlushed() < cfd->ioptions()->min_write_buffer_number_to_merge && vstorage->l0_delay_trigger_count() < mutable_cf_options.level0_file_num_compaction_trigger) { break; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/db_impl/db_impl_compaction_flush.cc#L1974 Universal may have `l0_delay_trigger_count() >= level0_file_num_compaction_trigger`, so this bug is triggered. Pull Request resolved: https://github.com/facebook/rocksdb/pull/7921 Reviewed By: jay-zhuang Differential Revision: D26900559 Pulled By: ajkr fbshipit-source-id: 133c1252dad7393753f04a47590b68c7d8e670df
2021-03-09 10:19:28 +00:00
mutable_cf_options.max_write_buffer_number - 1 &&
num_unflushed_memtables - 1 >=
immutable_cf_options.min_write_buffer_number_to_merge) {
return {WriteStallCondition::kDelayed, WriteStallCause::kMemtableLimit};
} else if (!mutable_cf_options.disable_auto_compactions &&
mutable_cf_options.level0_slowdown_writes_trigger >= 0 &&
num_l0_files >=
mutable_cf_options.level0_slowdown_writes_trigger) {
return {WriteStallCondition::kDelayed, WriteStallCause::kL0FileCountLimit};
} else if (!mutable_cf_options.disable_auto_compactions &&
mutable_cf_options.soft_pending_compaction_bytes_limit > 0 &&
num_compaction_needed_bytes >=
mutable_cf_options.soft_pending_compaction_bytes_limit) {
return {WriteStallCondition::kDelayed,
WriteStallCause::kPendingCompactionBytes};
}
return {WriteStallCondition::kNormal, WriteStallCause::kNone};
}
WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
const MutableCFOptions& mutable_cf_options) {
auto write_stall_condition = WriteStallCondition::kNormal;
if (current_ != nullptr) {
auto* vstorage = current_->storage_info();
auto write_controller = column_family_set_->write_controller_;
uint64_t compaction_needed_bytes =
vstorage->estimated_compaction_needed_bytes();
auto write_stall_condition_and_cause = GetWriteStallConditionAndCause(
imm()->NumNotFlushed(), vstorage->l0_delay_trigger_count(),
Fix checkpoint stuck (#7921) Summary: ## 1. Bug description: When RocksDB Checkpoint, it may be stuck in `WaitUntilFlushWouldNotStallWrites` method. ## 2. Simple analysis of the reasons: ### 2.1 Configuration parameters: ```yaml Compaction Style : Universal max_write_buffer_number : 4 min_write_buffer_number_to_merge : 3 ``` Checkpoint is usually very fast. When the Checkpoint is executed, `WaitUntilFlushWouldNotStallWrites` is called. If there are 2 Immutable MemTables, which are less than `min_write_buffer_number_to_merge`, they will not be flushed. But will enter this code. ```c++ // method: GetWriteStallConditionAndCause if (mutable_cf_options.max_write_buffer_number> 3 && num_unflushed_memtables >= mutable_cf_options.max_write_buffer_number-1) { return {WriteStallCondition::kDelayed, WriteStallCause::kMemtableLimit}; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/column_family.cc#L847 Checkpoint thought there was a FlushJob, but it didn't. So will always wait. ### 2.2 solution: Increase the restriction: the `number of Immutable MemTable` >= `min_write_buffer_number_to_merge will wait`. If there are other better solutions, you can correct me. ### 2.3 Code that can reproduce the problem: https://github.com/1996fanrui/fanrui-learning/blob/flink-1.12/module-java/src/main/java/com/dream/rocksdb/RocksDBCheckpointStuck.java ## 3. Interesting point This bug will be triggered only when `the number of sorted runs >= level0_file_num_compaction_trigger`. Because there is a break in WaitUntilFlushWouldNotStallWrites. ```c++ if (cfd->imm()->NumNotFlushed() < cfd->ioptions()->min_write_buffer_number_to_merge && vstorage->l0_delay_trigger_count() < mutable_cf_options.level0_file_num_compaction_trigger) { break; } ``` code link: https://github.com/facebook/rocksdb/blob/fbed72f03c3d9e4fdca3e5993587ef2559ba6ab9/db/db_impl/db_impl_compaction_flush.cc#L1974 Universal may have `l0_delay_trigger_count() >= level0_file_num_compaction_trigger`, so this bug is triggered. Pull Request resolved: https://github.com/facebook/rocksdb/pull/7921 Reviewed By: jay-zhuang Differential Revision: D26900559 Pulled By: ajkr fbshipit-source-id: 133c1252dad7393753f04a47590b68c7d8e670df
2021-03-09 10:19:28 +00:00
vstorage->estimated_compaction_needed_bytes(), mutable_cf_options,
*ioptions());
write_stall_condition = write_stall_condition_and_cause.first;
auto write_stall_cause = write_stall_condition_and_cause.second;
bool was_stopped = write_controller->IsStopped();
bool needed_delay = write_controller->NeedsDelay();
if (write_stall_condition == WriteStallCondition::kStopped &&
write_stall_cause == WriteStallCause::kMemtableLimit) {
write_controller_token_ = write_controller->GetStopToken();
internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_STOPS, 1);
ROCKS_LOG_WARN(
ioptions_.logger,
"[%s] Stopping writes because we have %d immutable memtables "
"(waiting for flush), max_write_buffer_number is set to %d",
Support saving history in memtable_list Summary: For transactions, we are using the memtables to validate that there are no write conflicts. But after flushing, we don't have any memtables, and transactions could fail to commit. So we want to someone keep around some extra history to use for conflict checking. In addition, we want to provide a way to increase the size of this history if too many transactions fail to commit. After chatting with people, it seems like everyone prefers just using Memtables to store this history (instead of a separate history structure). It seems like the best place for this is abstracted inside the memtable_list. I decide to create a separate list in MemtableListVersion as using the same list complicated the flush/installalflushresults logic too much. This diff adds a new parameter to control how much memtable history to keep around after flushing. However, it sounds like people aren't too fond of adding new parameters. So I am making the default size of flushed+not-flushed memtables be set to max_write_buffers. This should not change the maximum amount of memory used, but make it more likely we're using closer the the limit. (We are now postponing deleting flushed memtables until the max_write_buffer limit is reached). So while we might use more memory on average, we are still obeying the limit set (and you could argue it's better to go ahead and use up memory now instead of waiting for a write stall to happen to test this limit). However, if people are opposed to this default behavior, we can easily set it to 0 and require this parameter be set in order to use transactions. Test Plan: Added a xfunc test to play around with setting different values of this parameter in all tests. Added testing in memtablelist_test and planning on adding more testing here. Reviewers: sdong, rven, igor Reviewed By: igor Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D37443
2015-05-28 23:34:24 +00:00
name_.c_str(), imm()->NumNotFlushed(),
mutable_cf_options.max_write_buffer_number);
} else if (write_stall_condition == WriteStallCondition::kStopped &&
write_stall_cause == WriteStallCause::kL0FileCountLimit) {
write_controller_token_ = write_controller->GetStopToken();
internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_STOPS, 1);
if (compaction_picker_->IsLevel0CompactionInProgress()) {
internal_stats_->AddCFStats(
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
InternalStats::L0_FILE_COUNT_LIMIT_STOPS_WITH_ONGOING_COMPACTION,
1);
}
ROCKS_LOG_WARN(ioptions_.logger,
"[%s] Stopping writes because we have %d level-0 files",
name_.c_str(), vstorage->l0_delay_trigger_count());
} else if (write_stall_condition == WriteStallCondition::kStopped &&
write_stall_cause == WriteStallCause::kPendingCompactionBytes) {
write_controller_token_ = write_controller->GetStopToken();
internal_stats_->AddCFStats(
InternalStats::PENDING_COMPACTION_BYTES_LIMIT_STOPS, 1);
ROCKS_LOG_WARN(
ioptions_.logger,
"[%s] Stopping writes because of estimated pending compaction "
"bytes %" PRIu64,
name_.c_str(), compaction_needed_bytes);
} else if (write_stall_condition == WriteStallCondition::kDelayed &&
write_stall_cause == WriteStallCause::kMemtableLimit) {
write_controller_token_ =
SetupDelay(write_controller, compaction_needed_bytes,
prev_compaction_needed_bytes_, was_stopped,
mutable_cf_options.disable_auto_compactions);
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_DELAYS, 1);
ROCKS_LOG_WARN(
ioptions_.logger,
"[%s] Stalling writes because we have %d immutable memtables "
"(waiting for flush), max_write_buffer_number is set to %d "
"rate %" PRIu64,
name_.c_str(), imm()->NumNotFlushed(),
mutable_cf_options.max_write_buffer_number,
write_controller->delayed_write_rate());
} else if (write_stall_condition == WriteStallCondition::kDelayed &&
write_stall_cause == WriteStallCause::kL0FileCountLimit) {
// L0 is the last two files from stopping.
bool near_stop = vstorage->l0_delay_trigger_count() >=
mutable_cf_options.level0_stop_writes_trigger - 2;
write_controller_token_ =
SetupDelay(write_controller, compaction_needed_bytes,
prev_compaction_needed_bytes_, was_stopped || near_stop,
mutable_cf_options.disable_auto_compactions);
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_DELAYS, 1);
if (compaction_picker_->IsLevel0CompactionInProgress()) {
internal_stats_->AddCFStats(
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
InternalStats::L0_FILE_COUNT_LIMIT_DELAYS_WITH_ONGOING_COMPACTION,
1);
}
ROCKS_LOG_WARN(ioptions_.logger,
"[%s] Stalling writes because we have %d level-0 files "
"rate %" PRIu64,
name_.c_str(), vstorage->l0_delay_trigger_count(),
write_controller->delayed_write_rate());
} else if (write_stall_condition == WriteStallCondition::kDelayed &&
write_stall_cause == WriteStallCause::kPendingCompactionBytes) {
// If the distance to hard limit is less than 1/4 of the gap between soft
// and
// hard bytes limit, we think it is near stop and speed up the slowdown.
bool near_stop =
mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
(compaction_needed_bytes -
mutable_cf_options.soft_pending_compaction_bytes_limit) >
3 *
(mutable_cf_options.hard_pending_compaction_bytes_limit -
mutable_cf_options.soft_pending_compaction_bytes_limit) /
4;
write_controller_token_ =
SetupDelay(write_controller, compaction_needed_bytes,
prev_compaction_needed_bytes_, was_stopped || near_stop,
mutable_cf_options.disable_auto_compactions);
internal_stats_->AddCFStats(
New stat rocksdb.{cf|db}-write-stall-stats exposed in a structural way (#11300) Summary: **Context/Summary:** Users are interested in figuring out what has caused write stall. - Refactor write stall related stats from property `kCFStats` into its own db property `rocksdb.cf-write-stall-stats` as a map or string. For now, this only contains count of different combination of (CF-scope `WriteStallCause`) + (`WriteStallCondition`) - Add new `WriteStallCause::kWriteBufferManagerLimit` to reflect write stall caused by write buffer manager - Add new `rocksdb.db-write-stall-stats`. For now, this only contains `WriteStallCause::kWriteBufferManagerLimit` + `WriteStallCondition::kStopped` - Expose functions in new class `WriteStallStatsMapKeys` for examining the above two properties returned as map - Misc: rename/comment some write stall InternalStats for clarity Pull Request resolved: https://github.com/facebook/rocksdb/pull/11300 Test Plan: - New UT - Stress test `python3 tools/db_crashtest.py blackbox --simple --get_property_one_in=1` - Perf test: Both converge very slowly at similar rates but post-change has higher average ops/sec than pre-change even though they are run at the same time. ``` ./db_bench -seed=1679014417652004 -db=/dev/shm/testdb/ -statistics=false -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=100000 -db_write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 ``` pre-change: ``` fillseq [AVG 15 runs] : 1176 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1052.671 micros/op 949 ops/sec 105.267 seconds 100000 operations; 0.5 MB/s fillseq [AVG 16 runs] : 1162 (± 685) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1387.330 micros/op 720 ops/sec 138.733 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1136 (± 646) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1232.011 micros/op 811 ops/sec 123.201 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1118 (± 610) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1282.567 micros/op 779 ops/sec 128.257 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1100 (± 578) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1914.336 micros/op 522 ops/sec 191.434 seconds 100000 operations; 0.3 MB/s fillseq [AVG 20 runs] : 1071 (± 551) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1227.510 micros/op 814 ops/sec 122.751 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1059 (± 525) ops/sec; 0.5 (± 0.3) MB/sec ``` post-change: ``` fillseq [AVG 15 runs] : 1226 (± 732) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1323.825 micros/op 755 ops/sec 132.383 seconds 100000 operations; 0.4 MB/s fillseq [AVG 16 runs] : 1196 (± 687) ops/sec; 0.6 (± 0.4) MB/sec fillseq : 1223.905 micros/op 817 ops/sec 122.391 seconds 100000 operations; 0.4 MB/s fillseq [AVG 17 runs] : 1174 (± 647) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1168.996 micros/op 855 ops/sec 116.900 seconds 100000 operations; 0.4 MB/s fillseq [AVG 18 runs] : 1156 (± 611) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1348.729 micros/op 741 ops/sec 134.873 seconds 100000 operations; 0.4 MB/s fillseq [AVG 19 runs] : 1134 (± 579) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1196.887 micros/op 835 ops/sec 119.689 seconds 100000 operations; 0.4 MB/s fillseq [AVG 20 runs] : 1119 (± 550) ops/sec; 0.6 (± 0.3) MB/sec fillseq : 1193.697 micros/op 837 ops/sec 119.370 seconds 100000 operations; 0.4 MB/s fillseq [AVG 21 runs] : 1106 (± 524) ops/sec; 0.6 (± 0.3) MB/sec ``` Reviewed By: ajkr Differential Revision: D44159541 Pulled By: hx235 fbshipit-source-id: 8d29efb70001fdc52d34535eeb3364fc3e71e40b
2023-03-18 16:51:58 +00:00
InternalStats::PENDING_COMPACTION_BYTES_LIMIT_DELAYS, 1);
ROCKS_LOG_WARN(
ioptions_.logger,
"[%s] Stalling writes because of estimated pending compaction "
"bytes %" PRIu64 " rate %" PRIu64,
name_.c_str(), vstorage->estimated_compaction_needed_bytes(),
write_controller->delayed_write_rate());
} else {
assert(write_stall_condition == WriteStallCondition::kNormal);
if (vstorage->l0_delay_trigger_count() >=
GetL0ThresholdSpeedupCompaction(
mutable_cf_options.level0_file_num_compaction_trigger,
mutable_cf_options.level0_slowdown_writes_trigger)) {
write_controller_token_ =
write_controller->GetCompactionPressureToken();
ROCKS_LOG_INFO(
ioptions_.logger,
"[%s] Increasing compaction threads because we have %d level-0 "
"files ",
name_.c_str(), vstorage->l0_delay_trigger_count());
} else if (vstorage->estimated_compaction_needed_bytes() >=
mutable_cf_options.soft_pending_compaction_bytes_limit / 4) {
// Increase compaction threads if bytes needed for compaction exceeds
// 1/4 of threshold for slowing down.
// If soft pending compaction byte limit is not set, always speed up
// compaction.
write_controller_token_ =
write_controller->GetCompactionPressureToken();
if (mutable_cf_options.soft_pending_compaction_bytes_limit > 0) {
ROCKS_LOG_INFO(
ioptions_.logger,
"[%s] Increasing compaction threads because of estimated pending "
"compaction "
"bytes %" PRIu64,
name_.c_str(), vstorage->estimated_compaction_needed_bytes());
}
} else {
write_controller_token_.reset();
}
// If the DB recovers from delay conditions, we reward with reducing
// double the slowdown ratio. This is to balance the long term slowdown
// increase signal.
if (needed_delay) {
uint64_t write_rate = write_controller->delayed_write_rate();
write_controller->set_delayed_write_rate(static_cast<uint64_t>(
static_cast<double>(write_rate) * kDelayRecoverSlowdownRatio));
// Set the low pri limit to be 1/4 the delayed write rate.
// Note we don't reset this value even after delay condition is relased.
// Low-pri rate will continue to apply if there is a compaction
// pressure.
write_controller->low_pri_rate_limiter()->SetBytesPerSecond(write_rate /
4);
}
}
prev_compaction_needed_bytes_ = compaction_needed_bytes;
}
return write_stall_condition;
}
Introduce a new storage specific Env API (#5761) Summary: The current Env API encompasses both storage/file operations, as well as OS related operations. Most of the APIs return a Status, which does not have enough metadata about an error, such as whether its retry-able or not, scope (i.e fault domain) of the error etc., that may be required in order to properly handle a storage error. The file APIs also do not provide enough control over the IO SLA, such as timeout, prioritization, hinting about placement and redundancy etc. This PR separates out the file/storage APIs from Env into a new FileSystem class. The APIs are updated to return an IOStatus with metadata about the error, as well as to take an IOOptions structure as input in order to allow more control over the IO. The user can set both ```options.env``` and ```options.file_system``` to specify that RocksDB should use the former for OS related operations and the latter for storage operations. Internally, a ```CompositeEnvWrapper``` has been introduced that inherits from ```Env``` and redirects individual methods to either an ```Env``` implementation or the ```FileSystem``` as appropriate. When options are sanitized during ```DB::Open```, ```options.env``` is replaced with a newly allocated ```CompositeEnvWrapper``` instance if both env and file_system have been specified. This way, the rest of the RocksDB code can continue to function as before. This PR also ports PosixEnv to the new API by splitting it into two - PosixEnv and PosixFileSystem. PosixEnv is defined as a sub-class of CompositeEnvWrapper, and threading/time functions are overridden with Posix specific implementations in order to avoid an extra level of indirection. The ```CompositeEnvWrapper``` translates ```IOStatus``` return code to ```Status```, and sets the severity to ```kSoftError``` if the io_status is retryable. The error handling code in RocksDB can then recover the DB automatically. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5761 Differential Revision: D18868376 Pulled By: anand1976 fbshipit-source-id: 39efe18a162ea746fabac6360ff529baba48486f
2019-12-13 22:47:08 +00:00
const FileOptions* ColumnFamilyData::soptions() const {
return &(column_family_set_->file_options_);
}
void ColumnFamilyData::SetCurrent(Version* current_version) {
current_ = current_version;
}
uint64_t ColumnFamilyData::GetNumLiveVersions() const {
return VersionSet::GetNumLiveVersions(dummy_versions_);
}
uint64_t ColumnFamilyData::GetTotalSstFilesSize() const {
return VersionSet::GetTotalSstFilesSize(dummy_versions_);
}
Add DB properties for BlobDB (#8734) Summary: RocksDB exposes certain internal statistics via the DB property interface. However, there are currently no properties related to BlobDB. For starters, we would like to add the following BlobDB properties: `rocksdb.num-blob-files`: number of blob files in the current Version (kind of like `num-files-at-level` but note this is not per level, since blob files are not part of the LSM tree). `rocksdb.blob-stats`: this could return the total number and size of all blob files, and potentially also the total amount of garbage (in bytes) in the blob files in the current Version. `rocksdb.total-blob-file-size`: the total size of all blob files (as a blob counterpart for `total-sst-file-size`) of all Versions. `rocksdb.live-blob-file-size`: the total size of all blob files in the current Version. `rocksdb.estimate-live-data-size`: this is actually an existing property that we can extend so it considers blob files as well. When it comes to blobs, we actually have an exact value for live bytes. Namely, live bytes can be computed simply as total bytes minus garbage bytes, summed over the entire set of blob files in the Version. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8734 Test Plan: ``` ➜ rocksdb git:(new_feature_blobDB_properties) ./db_blob_basic_test [==========] Running 16 tests from 2 test cases. [----------] Global test environment set-up. [----------] 10 tests from DBBlobBasicTest [ RUN ] DBBlobBasicTest.GetBlob [ OK ] DBBlobBasicTest.GetBlob (12 ms) [ RUN ] DBBlobBasicTest.MultiGetBlobs [ OK ] DBBlobBasicTest.MultiGetBlobs (11 ms) [ RUN ] DBBlobBasicTest.GetBlob_CorruptIndex [ OK ] DBBlobBasicTest.GetBlob_CorruptIndex (10 ms) [ RUN ] DBBlobBasicTest.GetBlob_InlinedTTLIndex [ OK ] DBBlobBasicTest.GetBlob_InlinedTTLIndex (12 ms) [ RUN ] DBBlobBasicTest.GetBlob_IndexWithInvalidFileNumber [ OK ] DBBlobBasicTest.GetBlob_IndexWithInvalidFileNumber (9 ms) [ RUN ] DBBlobBasicTest.GenerateIOTracing [ OK ] DBBlobBasicTest.GenerateIOTracing (11 ms) [ RUN ] DBBlobBasicTest.BestEffortsRecovery_MissingNewestBlobFile [ OK ] DBBlobBasicTest.BestEffortsRecovery_MissingNewestBlobFile (13 ms) [ RUN ] DBBlobBasicTest.GetMergeBlobWithPut [ OK ] DBBlobBasicTest.GetMergeBlobWithPut (11 ms) [ RUN ] DBBlobBasicTest.MultiGetMergeBlobWithPut [ OK ] DBBlobBasicTest.MultiGetMergeBlobWithPut (14 ms) [ RUN ] DBBlobBasicTest.BlobDBProperties [ OK ] DBBlobBasicTest.BlobDBProperties (21 ms) [----------] 10 tests from DBBlobBasicTest (124 ms total) [----------] 6 tests from DBBlobBasicTest/DBBlobBasicIOErrorTest [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.GetBlob_IOError/0 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.GetBlob_IOError/0 (12 ms) [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.GetBlob_IOError/1 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.GetBlob_IOError/1 (10 ms) [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.MultiGetBlobs_IOError/0 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.MultiGetBlobs_IOError/0 (10 ms) [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.MultiGetBlobs_IOError/1 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.MultiGetBlobs_IOError/1 (10 ms) [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.CompactionFilterReadBlob_IOError/0 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.CompactionFilterReadBlob_IOError/0 (1011 ms) [ RUN ] DBBlobBasicTest/DBBlobBasicIOErrorTest.CompactionFilterReadBlob_IOError/1 [ OK ] DBBlobBasicTest/DBBlobBasicIOErrorTest.CompactionFilterReadBlob_IOError/1 (1013 ms) [----------] 6 tests from DBBlobBasicTest/DBBlobBasicIOErrorTest (2066 ms total) [----------] Global test environment tear-down [==========] 16 tests from 2 test cases ran. (2190 ms total) [ PASSED ] 16 tests. ``` Reviewed By: ltamasi Differential Revision: D30690849 Pulled By: Zhiyi-Zhang fbshipit-source-id: a7567319487ad76bd1a2e24bf143afdbbd9e4346
2021-09-08 19:19:01 +00:00
uint64_t ColumnFamilyData::GetTotalBlobFileSize() const {
return VersionSet::GetTotalBlobFileSize(dummy_versions_);
}
uint64_t ColumnFamilyData::GetLiveSstFilesSize() const {
return current_->GetSstFilesSize();
}
MemTable* ColumnFamilyData::ConstructNewMemtable(
const MutableCFOptions& mutable_cf_options, SequenceNumber earliest_seq) {
return new MemTable(internal_comparator_, ioptions_, mutable_cf_options,
write_buffer_manager_, earliest_seq, id_);
}
void ColumnFamilyData::CreateNewMemtable(
const MutableCFOptions& mutable_cf_options, SequenceNumber earliest_seq) {
if (mem_ != nullptr) {
delete mem_->Unref();
}
SetMemtable(ConstructNewMemtable(mutable_cf_options, earliest_seq));
mem_->Ref();
}
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
bool ColumnFamilyData::NeedsCompaction() const {
return !mutable_cf_options_.disable_auto_compactions &&
compaction_picker_->NeedsCompaction(current_->storage_info());
Rewritten system for scheduling background work Summary: When scaling to higher number of column families, the worst bottleneck was MaybeScheduleFlushOrCompaction(), which did a for loop over all column families while holding a mutex. This patch addresses the issue. The approach is similar to our earlier efforts: instead of a pull-model, where we do something for every column family, we can do a push-based model -- when we detect that column family is ready to be flushed/compacted, we add it to the flush_queue_/compaction_queue_. That way we don't need to loop over every column family in MaybeScheduleFlushOrCompaction. Here are the performance results: Command: ./db_bench --write_buffer_size=268435456 --db_write_buffer_size=268435456 --db=/fast-rocksdb-tmp/rocks_lots_of_cf --use_existing_db=0 --open_files=55000 --statistics=1 --histogram=1 --disable_data_sync=1 --max_write_buffer_number=2 --sync=0 --benchmarks=fillrandom --threads=16 --num_column_families=5000 --disable_wal=1 --max_background_flushes=16 --max_background_compactions=16 --level0_file_num_compaction_trigger=2 --level0_slowdown_writes_trigger=2 --level0_stop_writes_trigger=3 --hard_rate_limit=1 --num=33333333 --writes=33333333 Before the patch: fillrandom : 26.950 micros/op 37105 ops/sec; 4.1 MB/s After the patch: fillrandom : 17.404 micros/op 57456 ops/sec; 6.4 MB/s Next bottleneck is VersionSet::AddLiveFiles, which is painfully slow when we have a lot of files. This is coming in the next patch, but when I removed that code, here's what I got: fillrandom : 7.590 micros/op 131758 ops/sec; 14.6 MB/s Test Plan: make check two stress tests: Big number of compactions and flushes: ./db_stress --threads=30 --ops_per_thread=20000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=15 --max_background_compactions=10 --max_background_flushes=10 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 max_background_flushes=0, to verify that this case also works correctly ./db_stress --threads=30 --ops_per_thread=2000000 --max_key=10000 --column_families=20 --clear_column_family_one_in=10000000 --verify_before_write=0 --reopen=3 --max_background_compactions=3 --max_background_flushes=0 --db=/fast-rocksdb-tmp/db_stress --prefixpercent=0 --iterpercent=0 --writepercent=75 --db_write_buffer_size=2000000 Reviewers: ljin, rven, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D30123
2014-12-19 19:38:12 +00:00
}
Compaction* ColumnFamilyData::PickCompaction(
const MutableCFOptions& mutable_options,
const MutableDBOptions& mutable_db_options, LogBuffer* log_buffer) {
auto* result = compaction_picker_->PickCompaction(
GetName(), mutable_options, mutable_db_options, current_->storage_info(),
Sort L0 files by newly introduced epoch_num (#10922) 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 https://github.com/ajkr/rocksdb/commit/36a5686ec012f35a4371e409aa85c404ca1c210d (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
2022-12-13 21:29:37 +00:00
log_buffer);
if (result != nullptr) {
result->SetInputVersion(current_);
}
return result;
}
bool ColumnFamilyData::RangeOverlapWithCompaction(
const Slice& smallest_user_key, const Slice& largest_user_key,
int level) const {
return compaction_picker_->RangeOverlapWithCompaction(
smallest_user_key, largest_user_key, level);
}
Status ColumnFamilyData::RangesOverlapWithMemtables(
const autovector<Range>& ranges, SuperVersion* super_version,
bool allow_data_in_errors, bool* overlap) {
assert(overlap != nullptr);
*overlap = false;
// Create an InternalIterator over all unflushed memtables
Arena arena;
Group rocksdb.sst.read.micros stat by IOActivity flush and compaction (#11288) Summary: **Context:** The existing stat rocksdb.sst.read.micros does not reflect each of compaction and flush cases but aggregate them, which is not so helpful for us to understand IO read behavior of each of them. **Summary** - Update `StopWatch` and `RandomAccessFileReader` to record `rocksdb.sst.read.micros` and `rocksdb.file.{flush/compaction}.read.micros` - Fixed the default histogram in `RandomAccessFileReader` - New field `ReadOptions/IOOptions::io_activity`; Pass `ReadOptions` through paths under db open, flush and compaction to where we can prepare `IOOptions` and pass it to `RandomAccessFileReader` - Use `thread_status_util` for assertion in `DbStressFSWrapper` for continuous testing on we are passing correct `io_activity` under db open, flush and compaction Pull Request resolved: https://github.com/facebook/rocksdb/pull/11288 Test Plan: - **Stress test** - **Db bench 1: rocksdb.sst.read.micros COUNT ≈ sum of rocksdb.file.read.flush.micros's and rocksdb.file.read.compaction.micros's.** (without blob) - May not be exactly the same due to `HistogramStat::Add` only guarantees atomic not accuracy across threads. ``` ./db_bench -db=/dev/shm/testdb/ -statistics=true -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -target_file_size_base=655 -disable_auto_compactions=false -compression_type=none -bloom_bits=3 (-use_plain_table=1 -prefix_size=10) ``` ``` // BlockBasedTable rocksdb.sst.read.micros P50 : 2.009374 P95 : 4.968548 P99 : 8.110362 P100 : 43.000000 COUNT : 40456 SUM : 114805 rocksdb.file.read.flush.micros P50 : 1.871841 P95 : 3.872407 P99 : 5.540541 P100 : 43.000000 COUNT : 2250 SUM : 6116 rocksdb.file.read.compaction.micros P50 : 2.023109 P95 : 5.029149 P99 : 8.196910 P100 : 26.000000 COUNT : 38206 SUM : 108689 // PlainTable Does not apply ``` - **Db bench 2: performance** **Read** SETUP: db with 900 files ``` ./db_bench -db=/dev/shm/testdb/ -benchmarks="fillseq" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=true -target_file_size_base=655 -compression_type=none ```run till convergence ``` ./db_bench -seed=1678564177044286 -use_existing_db=true -db=/dev/shm/testdb -benchmarks=readrandom[-X60] -statistics=true -num=1000000 -disable_auto_compactions=true -compression_type=none -bloom_bits=3 ``` Pre-change `readrandom [AVG 60 runs] : 21568 (± 248) ops/sec` Post-change (no regression, -0.3%) `readrandom [AVG 60 runs] : 21486 (± 236) ops/sec` **Compaction/Flush**run till convergence ``` ./db_bench -db=/dev/shm/testdb2/ -seed=1678564177044286 -benchmarks="fillseq[-X60]" -key_size=32 -value_size=512 -num=50000 -write_buffer_size=655 -disable_auto_compactions=false -target_file_size_base=655 -compression_type=none rocksdb.sst.read.micros COUNT : 33820 rocksdb.sst.read.flush.micros COUNT : 1800 rocksdb.sst.read.compaction.micros COUNT : 32020 ``` Pre-change `fillseq [AVG 46 runs] : 1391 (± 214) ops/sec; 0.7 (± 0.1) MB/sec` Post-change (no regression, ~-0.4%) `fillseq [AVG 46 runs] : 1385 (± 216) ops/sec; 0.7 (± 0.1) MB/sec` Reviewed By: ajkr Differential Revision: D44007011 Pulled By: hx235 fbshipit-source-id: a54c89e4846dfc9a135389edf3f3eedfea257132
2023-04-21 16:07:18 +00:00
// TODO: plumb Env::IOActivity
ReadOptions read_opts;
read_opts.total_order_seek = true;
MergeIteratorBuilder merge_iter_builder(&internal_comparator_, &arena);
merge_iter_builder.AddIterator(
super_version->mem->NewIterator(read_opts, &arena));
super_version->imm->AddIterators(read_opts, &merge_iter_builder,
false /* add_range_tombstone_iter */);
ScopedArenaIterator memtable_iter(merge_iter_builder.Finish());
auto read_seq = super_version->current->version_set()->LastSequence();
ReadRangeDelAggregator range_del_agg(&internal_comparator_, read_seq);
Fragment memtable range tombstone in the write path (#10380) Summary: - Right now each read fragments the memtable range tombstones https://github.com/facebook/rocksdb/issues/4808. This PR explores the idea of fragmenting memtable range tombstones in the write path and reads can just read this cached fragmented tombstone without any fragmenting cost. This PR only does the caching for immutable memtable, and does so right before a memtable is added to an immutable memtable list. The fragmentation is done without holding mutex to minimize its performance impact. - db_bench is updated to print out the number of range deletions executed if there is any. Pull Request resolved: https://github.com/facebook/rocksdb/pull/10380 Test Plan: - CI, added asserts in various places to check whether a fragmented range tombstone list should have been constructed. - Benchmark: as this PR only optimizes immutable memtable path, the number of writes in the benchmark is chosen such an immutable memtable is created and range tombstones are in that memtable. ``` single thread: ./db_bench --benchmarks=fillrandom,readrandom --writes_per_range_tombstone=1 --max_write_buffer_number=100 --min_write_buffer_number_to_merge=100 --writes=500000 --reads=100000 --max_num_range_tombstones=100 multi_thread ./db_bench --benchmarks=fillrandom,readrandom --writes_per_range_tombstone=1 --max_write_buffer_number=100 --min_write_buffer_number_to_merge=100 --writes=15000 --reads=20000 --threads=32 --max_num_range_tombstones=100 ``` Commit 99cdf16464a057ca44de2f747541dedf651bae9e is included in benchmark result. It was an earlier attempt where tombstones are fragmented for each write operation. Reader threads share it using a shared_ptr which would slow down multi-thread read performance as seen in benchmark results. Results are averaged over 5 runs. Single thread result: | Max # tombstones | main fillrandom micros/op | 99cdf16464a057ca44de2f747541dedf651bae9e | Post PR | main readrandom micros/op | 99cdf16464a057ca44de2f747541dedf651bae9e | Post PR | | ------------- | ------------- |------------- |------------- |------------- |------------- |------------- | | 0 |6.68 |6.57 |6.72 |4.72 |4.79 |4.54 | | 1 |6.67 |6.58 |6.62 |5.41 |4.74 |4.72 | | 10 |6.59 |6.5 |6.56 |7.83 |4.69 |4.59 | | 100 |6.62 |6.75 |6.58 |29.57 |5.04 |5.09 | | 1000 |6.54 |6.82 |6.61 |320.33 |5.22 |5.21 | 32-thread result: note that "Max # tombstones" is per thread. | Max # tombstones | main fillrandom micros/op | 99cdf16464a057ca44de2f747541dedf651bae9e | Post PR | main readrandom micros/op | 99cdf16464a057ca44de2f747541dedf651bae9e | Post PR | | ------------- | ------------- |------------- |------------- |------------- |------------- |------------- | | 0 |234.52 |260.25 |239.42 |5.06 |5.38 |5.09 | | 1 |236.46 |262.0 |231.1 |19.57 |22.14 |5.45 | | 10 |236.95 |263.84 |251.49 |151.73 |21.61 |5.73 | | 100 |268.16 |296.8 |280.13 |2308.52 |22.27 |6.57 | Reviewed By: ajkr Differential Revision: D37916564 Pulled By: cbi42 fbshipit-source-id: 05d6d2e16df26c374c57ddcca13a5bfe9d5b731e
2022-08-05 19:02:33 +00:00
auto* active_range_del_iter = super_version->mem->NewRangeTombstoneIterator(
read_opts, read_seq, false /* immutable_memtable */);
range_del_agg.AddTombstones(
std::unique_ptr<FragmentedRangeTombstoneIterator>(active_range_del_iter));
Status status;
status = super_version->imm->AddRangeTombstoneIterators(
read_opts, nullptr /* arena */, &range_del_agg);
// AddRangeTombstoneIterators always return Status::OK.
assert(status.ok());
for (size_t i = 0; i < ranges.size() && status.ok() && !*overlap; ++i) {
auto* vstorage = super_version->current->storage_info();
auto* ucmp = vstorage->InternalComparator()->user_comparator();
InternalKey range_start(ranges[i].start, kMaxSequenceNumber,
kValueTypeForSeek);
memtable_iter->Seek(range_start.Encode());
status = memtable_iter->status();
ParsedInternalKey seek_result;
if (status.ok() && memtable_iter->Valid()) {
status = ParseInternalKey(memtable_iter->key(), &seek_result,
allow_data_in_errors);
}
if (status.ok()) {
if (memtable_iter->Valid() &&
ucmp->Compare(seek_result.user_key, ranges[i].limit) <= 0) {
*overlap = true;
} else if (range_del_agg.IsRangeOverlapped(ranges[i].start,
ranges[i].limit)) {
*overlap = true;
}
}
}
return status;
}
const int ColumnFamilyData::kCompactAllLevels = -1;
const int ColumnFamilyData::kCompactToBaseLevel = -2;
Compaction* ColumnFamilyData::CompactRange(
const MutableCFOptions& mutable_cf_options,
const MutableDBOptions& mutable_db_options, int input_level,
int output_level, const CompactRangeOptions& compact_range_options,
const InternalKey* begin, const InternalKey* end,
InternalKey** compaction_end, bool* conflict,
uint64_t max_file_num_to_ignore, const std::string& trim_ts) {
auto* result = compaction_picker_->CompactRange(
GetName(), mutable_cf_options, mutable_db_options,
current_->storage_info(), input_level, output_level,
compact_range_options, begin, end, compaction_end, conflict,
max_file_num_to_ignore, trim_ts);
if (result != nullptr) {
result->SetInputVersion(current_);
}
Add missing range conflict check between file ingestion and RefitLevel() (#10988) Summary: **Context:** File ingestion never checks whether the key range it acts on overlaps with an ongoing RefitLevel() (used in `CompactRange()` with `change_level=true`). That's because RefitLevel() doesn't register and make its key range known to file ingestion. Though it checks overlapping with other compactions by https://github.com/facebook/rocksdb/blob/7.8.fb/db/external_sst_file_ingestion_job.cc#L998. RefitLevel() (used in `CompactRange()` with `change_level=true`) doesn't check whether the key range it acts on overlaps with an ongoing file ingestion. That's because file ingestion does not register and make its key range known to other compactions. - Note that non-refitlevel-compaction (e.g, manual compaction w/o RefitLevel() or general compaction) also does not check key range overlap with ongoing file ingestion for the same reason. - But it's fine. Credited to cbi42's discovery, `WaitForIngestFile` was called by background and foreground compactions. They were introduced in https://github.com/facebook/rocksdb/commit/0f88160f67d36ea30e3aca3a3cef924c3a009be6, https://github.com/facebook/rocksdb/commit/5c64fb67d2fc198f1a73ff3ae543749a6a41f513 and https://github.com/facebook/rocksdb/commit/87dfc1d23e0e16ff73e15f63c6fa0fb3b3fc8c8c. - Regardless, this PR registers file ingestion like a compaction is a general approach that will also add range conflict check between file ingestion and non-refitlevel-compaction, though it has not been the issue motivated this PR. Above are bugs resulting in two bad consequences: - If file ingestion and RefitLevel() creates files in the same level, then range-overlapped files will be created at that level and caught as corruption by `force_consistency_checks=true` - If file ingestion and RefitLevel() creates file in different levels, then with one further compaction on the ingested file, it can result in two same keys both with seqno 0 in two different levels. Then with iterator's [optimization](https://github.com/facebook/rocksdb/blame/c62f3221698fd273b673d4f7e54eabb8329a4369/db/db_iter.cc#L342-L343) that assumes no two same keys both with seqno 0, it will either break this assertion in debug build or, even worst, return value of this same key for the key after it, which is the wrong value to return, in release build. Therefore we decide to introduce range conflict check for file ingestion and RefitLevel() inspired from the existing range conflict check among compactions. **Summary:** - Treat file ingestion job and RefitLevel() as `Compaction` of new compaction reasons: `CompactionReason::kExternalSstIngestion` and `CompactionReason::kRefitLevel` and register/unregister them. File ingestion is treated as compaction from L0 to different levels and RefitLevel() as compaction from source level to target level. - Check for `RangeOverlapWithCompaction` with other ongoing compactions, `RegisterCompaction()` on this "compaction" before changing the LSM state in `VersionStorageInfo`, and `UnregisterCompaction()` after changing. - Replace scattered fixes (https://github.com/facebook/rocksdb/commit/0f88160f67d36ea30e3aca3a3cef924c3a009be6, https://github.com/facebook/rocksdb/commit/5c64fb67d2fc198f1a73ff3ae543749a6a41f513 and https://github.com/facebook/rocksdb/commit/87dfc1d23e0e16ff73e15f63c6fa0fb3b3fc8c8c.) that prevents overlapping between file ingestion and non-refit-level compaction with this fix cuz those practices are easy to overlook. - Misc: logic cleanup, see PR comments Pull Request resolved: https://github.com/facebook/rocksdb/pull/10988 Test Plan: - New unit test `DBCompactionTestWithOngoingFileIngestionParam*` that failed pre-fix and passed afterwards. - Made compatible with existing tests, see PR comments - make check - [Ongoing] Stress test rehearsal with normal value and aggressive CI value https://github.com/facebook/rocksdb/pull/10761 Reviewed By: cbi42 Differential Revision: D41535685 Pulled By: hx235 fbshipit-source-id: 549833a577ba1496d20a870583d4caa737da1258
2022-12-29 23:05:36 +00:00
TEST_SYNC_POINT("ColumnFamilyData::CompactRange:Return");
return result;
}
SuperVersion* ColumnFamilyData::GetReferencedSuperVersion(DBImpl* db) {
SuperVersion* sv = GetThreadLocalSuperVersion(db);
sv->Ref();
if (!ReturnThreadLocalSuperVersion(sv)) {
// This Unref() corresponds to the Ref() in GetThreadLocalSuperVersion()
// when the thread-local pointer was populated. So, the Ref() earlier in
// this function still prevents the returned SuperVersion* from being
// deleted out from under the caller.
sv->Unref();
}
return sv;
}
SuperVersion* ColumnFamilyData::GetThreadLocalSuperVersion(DBImpl* db) {
// The SuperVersion is cached in thread local storage to avoid acquiring
// mutex when SuperVersion does not change since the last use. When a new
// SuperVersion is installed, the compaction or flush thread cleans up
// cached SuperVersion in all existing thread local storage. To avoid
// acquiring mutex for this operation, we use atomic Swap() on the thread
// local pointer to guarantee exclusive access. If the thread local pointer
// is being used while a new SuperVersion is installed, the cached
// SuperVersion can become stale. In that case, the background thread would
// have swapped in kSVObsolete. We re-check the value at when returning
// SuperVersion back to thread local, with an atomic compare and swap.
// The superversion will need to be released if detected to be stale.
void* ptr = local_sv_->Swap(SuperVersion::kSVInUse);
// Invariant:
// (1) Scrape (always) installs kSVObsolete in ThreadLocal storage
// (2) the Swap above (always) installs kSVInUse, ThreadLocal storage
// should only keep kSVInUse before ReturnThreadLocalSuperVersion call
// (if no Scrape happens).
assert(ptr != SuperVersion::kSVInUse);
SuperVersion* sv = static_cast<SuperVersion*>(ptr);
if (sv == SuperVersion::kSVObsolete) {
RecordTick(ioptions_.stats, NUMBER_SUPERVERSION_ACQUIRES);
db->mutex()->Lock();
sv = super_version_->Ref();
db->mutex()->Unlock();
}
assert(sv != nullptr);
return sv;
}
bool ColumnFamilyData::ReturnThreadLocalSuperVersion(SuperVersion* sv) {
assert(sv != nullptr);
// Put the SuperVersion back
void* expected = SuperVersion::kSVInUse;
if (local_sv_->CompareAndSwap(static_cast<void*>(sv), expected)) {
// When we see kSVInUse in the ThreadLocal, we are sure ThreadLocal
2015-04-25 09:14:27 +00:00
// storage has not been altered and no Scrape has happened. The
// SuperVersion is still current.
return true;
} else {
// ThreadLocal scrape happened in the process of this GetImpl call (after
// thread local Swap() at the beginning and before CompareAndSwap()).
// This means the SuperVersion it holds is obsolete.
assert(expected == SuperVersion::kSVObsolete);
}
return false;
}
void ColumnFamilyData::InstallSuperVersion(SuperVersionContext* sv_context,
InstrumentedMutex* db_mutex) {
db_mutex->AssertHeld();
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
return InstallSuperVersion(sv_context, mutable_cf_options_);
}
void ColumnFamilyData::InstallSuperVersion(
Fix a race in ColumnFamilyData::UnrefAndTryDelete (#8605) Summary: The `ColumnFamilyData::UnrefAndTryDelete` code currently on the trunk unlocks the DB mutex before destroying the `ThreadLocalPtr` holding the per-thread `SuperVersion` pointers when the only remaining reference is the back reference from `super_version_`. The idea behind this was to break the circular dependency between `ColumnFamilyData` and `SuperVersion`: when the penultimate reference goes away, `ColumnFamilyData` can clean up the `SuperVersion`, which can in turn clean up `ColumnFamilyData`. (Assuming there is a `SuperVersion` and it is not referenced by anything else.) However, unlocking the mutex throws a wrench in this plan by making it possible for another thread to jump in and take another reference to the `ColumnFamilyData`, keeping the object alive in a zombie `ThreadLocalPtr`-less state. This can cause issues like https://github.com/facebook/rocksdb/issues/8440 , https://github.com/facebook/rocksdb/issues/8382 , and might also explain the `was_last_ref` assertion failures from the `ColumnFamilySet` destructor we sometimes observe during close in our stress tests. Digging through the archives, this unlocking goes way back to 2014 (or earlier). The original rationale was that `SuperVersionUnrefHandle` used to lock the mutex so it can call `SuperVersion::Cleanup`; however, this logic turned out to be deadlock-prone. https://github.com/facebook/rocksdb/pull/3510 fixed the deadlock but left the unlocking in place. https://github.com/facebook/rocksdb/pull/6147 then introduced the circular dependency and associated cleanup logic described above (in order to enable iterators to keep the `ColumnFamilyData` for dropped column families alive), and moved the unlocking-relocking snippet to its present location in `UnrefAndTryDelete`. Finally, https://github.com/facebook/rocksdb/pull/7749 fixed a memory leak but apparently exacerbated the race by (otherwise correctly) switching to `UnrefAndTryDelete` in `SuperVersion::Cleanup`. The patch simply eliminates the unlocking and relocking, which has been unnecessary ever since https://github.com/facebook/rocksdb/issues/3510 made `SuperVersionUnrefHandle` lock-free. This closes the window during which another thread could increase the reference count, and hopefully fixes the issues above. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8605 Test Plan: Ran `make check` and stress tests locally. Reviewed By: pdillinger Differential Revision: D30051035 Pulled By: ltamasi fbshipit-source-id: 8fe559e4b4ad69fc142579f8bc393ef525918528
2021-08-03 01:10:57 +00:00
SuperVersionContext* sv_context,
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
2021-07-10 00:16:00 +00:00
const MutableCFOptions& mutable_cf_options) {
SuperVersion* new_superversion = sv_context->new_superversion.release();
new_superversion->mutable_cf_options = mutable_cf_options;
new_superversion->Init(this, mem_, imm_.current(), current_);
SuperVersion* old_superversion = super_version_;
super_version_ = new_superversion;
++super_version_number_;
super_version_->version_number = super_version_number_;
if (old_superversion == nullptr || old_superversion->current != current() ||
old_superversion->mem != mem_ ||
old_superversion->imm != imm_.current()) {
// Should not recalculate slow down condition if nothing has changed, since
// currently RecalculateWriteStallConditions() treats it as further slowing
// down is needed.
super_version_->write_stall_condition =
RecalculateWriteStallConditions(mutable_cf_options);
} else {
super_version_->write_stall_condition =
old_superversion->write_stall_condition;
}
if (old_superversion != nullptr) {
// Reset SuperVersions cached in thread local storage.
// This should be done before old_superversion->Unref(). That's to ensure
// that local_sv_ never holds the last reference to SuperVersion, since
// it has no means to safely do SuperVersion cleanup.
ResetThreadLocalSuperVersions();
if (old_superversion->mutable_cf_options.write_buffer_size !=
mutable_cf_options.write_buffer_size) {
mem_->UpdateWriteBufferSize(mutable_cf_options.write_buffer_size);
}
if (old_superversion->write_stall_condition !=
new_superversion->write_stall_condition) {
sv_context->PushWriteStallNotification(
old_superversion->write_stall_condition,
new_superversion->write_stall_condition, GetName(), ioptions());
}
if (old_superversion->Unref()) {
Make mempurge a background process (equivalent to in-memory compaction). (#8505) Summary: In https://github.com/facebook/rocksdb/issues/8454, I introduced a new process baptized `MemPurge` (memtable garbage collection). This new PR is built upon this past mempurge prototype. In this PR, I made the `mempurge` process a background task, which provides superior performance since the mempurge process does not cling on the db_mutex anymore, and addresses severe restrictions from the past iteration (including a scenario where the past mempurge was failling, when a memtable was mempurged but was still referred to by an iterator/snapshot/...). Now the mempurge process ressembles an in-memory compaction process: the stack of immutable memtables is filtered out, and the useful payload is used to populate an output memtable. If the output memtable is filled at more than 60% capacity (arbitrary heuristic) the mempurge process is aborted and a regular flush process takes place, else the output memtable is kept in the immutable memtable stack. Note that adding this output memtable to the `imm()` memtable stack does not trigger another flush process, so that the flush thread can go to sleep at the end of a successful mempurge. MemPurge is activated by making the `experimental_allow_mempurge` flag `true`. When activated, the `MemPurge` process will always happen when the flush reason is `kWriteBufferFull`. The 3 unit tests confirm that this process supports `Put`, `Get`, `Delete`, `DeleteRange` operators and is compatible with `Iterators` and `CompactionFilters`. Pull Request resolved: https://github.com/facebook/rocksdb/pull/8505 Reviewed By: pdillinger Differential Revision: D29619283 Pulled By: bjlemaire fbshipit-source-id: 8a99bee76b63a8211bff1a00e0ae32360aaece95
2021-07-10 00:16:00 +00:00
old_superversion->Cleanup();
sv_context->superversions_to_free.push_back(old_superversion);
}
}
}
void ColumnFamilyData::ResetThreadLocalSuperVersions() {
autovector<void*> sv_ptrs;
local_sv_->Scrape(&sv_ptrs, SuperVersion::kSVObsolete);
for (auto ptr : sv_ptrs) {
assert(ptr);
if (ptr == SuperVersion::kSVInUse) {
continue;
}
auto sv = static_cast<SuperVersion*>(ptr);
bool was_last_ref __attribute__((__unused__));
was_last_ref = sv->Unref();
// sv couldn't have been the last reference because
// ResetThreadLocalSuperVersions() is called before
// unref'ing super_version_.
assert(!was_last_ref);
}
}
Status ColumnFamilyData::ValidateOptions(
const DBOptions& db_options, const ColumnFamilyOptions& cf_options) {
Status s;
s = CheckCompressionSupported(cf_options);
if (s.ok() && db_options.allow_concurrent_memtable_write) {
s = CheckConcurrentWritesSupported(cf_options);
}
if (s.ok() && db_options.unordered_write &&
cf_options.max_successive_merges != 0) {
s = Status::InvalidArgument(
"max_successive_merges > 0 is incompatible with unordered_write");
}
if (s.ok()) {
s = CheckCFPathsSupported(db_options, cf_options);
}
if (!s.ok()) {
return s;
}
if (cf_options.ttl > 0 && cf_options.ttl != kDefaultTtl) {
if (!cf_options.table_factory->IsInstanceOf(
TableFactory::kBlockBasedTableName())) {
return Status::NotSupported(
"TTL is only supported in Block-Based Table format. ");
}
}
Auto enable Periodic Compactions if a Compaction Filter is used (#5865) Summary: - Periodic compactions are auto-enabled if a compaction filter or a compaction filter factory is set, in Level Compaction. - The default value of `periodic_compaction_seconds` is changed to UINT64_MAX, which lets RocksDB auto-tune periodic compactions as needed. An explicit value of 0 will still work as before ie. to disable periodic compactions completely. For now, on seeing a compaction filter along with a UINT64_MAX value for `periodic_compaction_seconds`, RocksDB will make SST files older than 30 days to go through periodic copmactions. Some RocksDB users make use of compaction filters to control when their data can be deleted, usually with a custom TTL logic. But it is occasionally possible that the compactions get delayed by considerable time due to factors like low writes to a key range, data reaching bottom level, etc before the TTL expiry. Periodic Compactions feature was originally built to help such cases. Now periodic compactions are auto enabled by default when compaction filters or compaction filter factories are used, as it is generally helpful to all cases to collect garbage. `periodic_compaction_seconds` is set to a large value, 30 days, in `SanitizeOptions` when RocksDB sees that a `compaction_filter` or `compaction_filter_factory` is used. This is done only for Level Compaction style. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5865 Test Plan: - Added a new test `DBCompactionTest.LevelPeriodicCompactionWithCompactionFilters` to make sure that `periodic_compaction_seconds` is set if either `compaction_filter` or `compaction_filter_factory` options are set. - `COMPILE_WITH_ASAN=1 make check` Differential Revision: D17659180 Pulled By: sagar0 fbshipit-source-id: 4887b9cf2e53cf2dc93a7b658c6b15e1181217ee
2019-10-29 22:04:26 +00:00
if (cf_options.periodic_compaction_seconds > 0 &&
cf_options.periodic_compaction_seconds != kDefaultPeriodicCompSecs) {
if (!cf_options.table_factory->IsInstanceOf(
TableFactory::kBlockBasedTableName())) {
return Status::NotSupported(
"Periodic Compaction is only supported in "
"Block-Based Table format. ");
}
}
Integrated blob garbage collection: relocate blobs (#7694) Summary: The patch adds basic garbage collection support to the integrated BlobDB implementation. Valid blobs residing in the oldest blob files are relocated as they are encountered during compaction. The threshold that determines which blob files qualify is computed based on the configuration option `blob_garbage_collection_age_cutoff`, which was introduced in https://github.com/facebook/rocksdb/issues/7661 . Once a blob is retrieved for the purposes of relocation, it passes through the same logic that extracts large values to blob files in general. This means that if, for instance, the size threshold for key-value separation (`min_blob_size`) got changed or writing blob files got disabled altogether, it is possible for the value to be moved back into the LSM tree. In particular, one way to re-inline all blob values if needed would be to perform a full manual compaction with `enable_blob_files` set to `false`, `enable_blob_garbage_collection` set to `true`, and `blob_file_garbage_collection_age_cutoff` set to `1.0`. Some TODOs that I plan to address in separate PRs: 1) We'll have to measure the amount of new garbage in each blob file and log `BlobFileGarbage` entries as part of the compaction job's `VersionEdit`. (For the time being, blob files are cleaned up solely based on the `oldest_blob_file_number` relationships.) 2) When compression is used for blobs, the compression type hasn't changed, and the blob still qualifies for being written to a blob file, we can simply copy the compressed blob to the new file instead of going through decompression and compression. 3) We need to update the formula for computing write amplification to account for the amount of data read from blob files as part of GC. Pull Request resolved: https://github.com/facebook/rocksdb/pull/7694 Test Plan: `make check` Reviewed By: riversand963 Differential Revision: D25069663 Pulled By: ltamasi fbshipit-source-id: bdfa8feb09afcf5bca3b4eba2ba72ce2f15cd06a
2020-11-24 05:07:01 +00:00
Respect cutoff timestamp during flush (#11599) Summary: Make flush respect the cutoff timestamp `full_history_ts_low` as much as possible for the user-defined timestamps in Memtables only feature. We achieve this by not proceeding with the actual flushing but instead reschedule the same `FlushRequest` so a follow up flush job can continue with the check after some interval. This approach doesn't work well for atomic flush, so this feature currently is not supported in combination with atomic flush. Furthermore, this approach also requires a customized method to get the next immediately bigger user-defined timestamp. So currently it's limited to comparator that use uint64_t as the user-defined timestamp format. This support can be extended when we add such a customized method to `AdvancedColumnFamilyOptions`. For non atomic flush request, at any single time, a column family can only have as many as one FlushRequest for it in the `flush_queue_`. There is deduplication done at `FlushRequest` enqueueing(`SchedulePendingFlush`) and dequeueing time (`PopFirstFromFlushQueue`). We hold the db mutex between when a `FlushRequest` is popped from the queue and the same FlushRequest get rescheduled, so no other `FlushRequest` with a higher `max_memtable_id` can be added to the `flush_queue_` blocking us from re-enqueueing the same `FlushRequest`. Flush is continued nevertheless if there is risk of entering write stall mode had the flush being postponed, e.g. due to accumulation of write buffers, exceeding the `max_write_buffer_number` setting. When this happens, the newest user-defined timestamp in the involved Memtables need to be tracked and we use it to increase the `full_history_ts_low`, which is an inclusive cutoff timestamp for which RocksDB promises to keep all user-defined timestamps equal to and newer than it. Tet plan: ``` ./column_family_test --gtest_filter="*RetainUDT*" ./memtable_list_test --gtest_filter="*WithTimestamp*" ./flush_job_test --gtest_filter="*WithTimestamp*" ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11599 Reviewed By: ajkr Differential Revision: D47561586 Pulled By: jowlyzhang fbshipit-source-id: 9400445f983dd6eac489e9dd0fb5d9b99637fe89
2023-07-26 23:25:06 +00:00
const auto* ucmp = cf_options.comparator;
assert(ucmp);
if (ucmp->timestamp_size() > 0 &&
!cf_options.persist_user_defined_timestamps) {
if (db_options.atomic_flush) {
return Status::NotSupported(
"Not persisting user-defined timestamps feature is not supported"
"in combination with atomic flush.");
}
if (db_options.allow_concurrent_memtable_write) {
return Status::NotSupported(
"Not persisting user-defined timestamps feature is not supported"
" in combination with concurrent memtable write.");
}
const char* comparator_name = cf_options.comparator->Name();
size_t name_size = strlen(comparator_name);
const char* suffix = ".u64ts";
size_t suffix_size = strlen(suffix);
if (name_size <= suffix_size ||
strcmp(comparator_name + name_size - suffix_size, suffix) != 0) {
return Status::NotSupported(
"Not persisting user-defined timestamps"
"feature only support user-defined timestamps formatted as "
"uint64_t.");
}
}
Make it possible to force the garbage collection of the oldest blob files (#8994) Summary: The current BlobDB garbage collection logic works by relocating the valid blobs from the oldest blob files as they are encountered during compaction, and cleaning up blob files once they contain nothing but garbage. However, with sufficiently skewed workloads, it is theoretically possible to end up in a situation when few or no compactions get scheduled for the SST files that contain references to the oldest blob files, which can lead to increased space amp due to the lack of GC. In order to efficiently handle such workloads, the patch adds a new BlobDB configuration option called `blob_garbage_collection_force_threshold`, which signals to BlobDB to schedule targeted compactions for the SST files that keep alive the oldest batch of blob files if the overall ratio of garbage in the given blob files meets the threshold *and* all the given blob files are eligible for GC based on `blob_garbage_collection_age_cutoff`. (For example, if the new option is set to 0.9, targeted compactions will get scheduled if the sum of garbage bytes meets or exceeds 90% of the sum of total bytes in the oldest blob files, assuming all affected blob files are below the age-based cutoff.) The net result of these targeted compactions is that the valid blobs in the oldest blob files are relocated and the oldest blob files themselves cleaned up (since *all* SST files that rely on them get compacted away). These targeted compactions are similar to periodic compactions in the sense that they force certain SST files that otherwise would not get picked up to undergo compaction and also in the sense that instead of merging files from multiple levels, they target a single file. (Note: such compactions might still include neighboring files from the same level due to the need of having a "clean cut" boundary but they never include any files from any other level.) This functionality is currently only supported with the leveled compaction style and is inactive by default (since the default value is set to 1.0, i.e. 100%). Pull Request resolved: https://github.com/facebook/rocksdb/pull/8994 Test Plan: Ran `make check` and tested using `db_bench` and the stress/crash tests. Reviewed By: riversand963 Differential Revision: D31489850 Pulled By: ltamasi fbshipit-source-id: 44057d511726a0e2a03c5d9313d7511b3f0c4eab
2021-10-12 01:00:44 +00:00
if (cf_options.enable_blob_garbage_collection) {
if (cf_options.blob_garbage_collection_age_cutoff < 0.0 ||
cf_options.blob_garbage_collection_age_cutoff > 1.0) {
return Status::InvalidArgument(
"The age cutoff for blob garbage collection should be in the range "
"[0.0, 1.0].");
}
if (cf_options.blob_garbage_collection_force_threshold < 0.0 ||
cf_options.blob_garbage_collection_force_threshold > 1.0) {
return Status::InvalidArgument(
"The garbage ratio threshold for forcing blob garbage collection "
"should be in the range [0.0, 1.0].");
}
Integrated blob garbage collection: relocate blobs (#7694) Summary: The patch adds basic garbage collection support to the integrated BlobDB implementation. Valid blobs residing in the oldest blob files are relocated as they are encountered during compaction. The threshold that determines which blob files qualify is computed based on the configuration option `blob_garbage_collection_age_cutoff`, which was introduced in https://github.com/facebook/rocksdb/issues/7661 . Once a blob is retrieved for the purposes of relocation, it passes through the same logic that extracts large values to blob files in general. This means that if, for instance, the size threshold for key-value separation (`min_blob_size`) got changed or writing blob files got disabled altogether, it is possible for the value to be moved back into the LSM tree. In particular, one way to re-inline all blob values if needed would be to perform a full manual compaction with `enable_blob_files` set to `false`, `enable_blob_garbage_collection` set to `true`, and `blob_file_garbage_collection_age_cutoff` set to `1.0`. Some TODOs that I plan to address in separate PRs: 1) We'll have to measure the amount of new garbage in each blob file and log `BlobFileGarbage` entries as part of the compaction job's `VersionEdit`. (For the time being, blob files are cleaned up solely based on the `oldest_blob_file_number` relationships.) 2) When compression is used for blobs, the compression type hasn't changed, and the blob still qualifies for being written to a blob file, we can simply copy the compressed blob to the new file instead of going through decompression and compression. 3) We need to update the formula for computing write amplification to account for the amount of data read from blob files as part of GC. Pull Request resolved: https://github.com/facebook/rocksdb/pull/7694 Test Plan: `make check` Reviewed By: riversand963 Differential Revision: D25069663 Pulled By: ltamasi fbshipit-source-id: bdfa8feb09afcf5bca3b4eba2ba72ce2f15cd06a
2020-11-24 05:07:01 +00:00
}
if (cf_options.compaction_style == kCompactionStyleFIFO &&
db_options.max_open_files != -1 && cf_options.ttl > 0) {
return Status::NotSupported(
"FIFO compaction only supported with max_open_files = -1.");
}
std::vector<uint32_t> supported{0, 1, 2, 4, 8};
if (std::find(supported.begin(), supported.end(),
cf_options.memtable_protection_bytes_per_key) ==
supported.end()) {
return Status::NotSupported(
"Memtable per key-value checksum protection only supports 0, 1, 2, 4 "
"or 8 bytes per key.");
}
Block per key-value checksum (#11287) Summary: add option `block_protection_bytes_per_key` and implementation for block per key-value checksum. The main changes are 1. checksum construction and verification in block.cc/h 2. pass the option `block_protection_bytes_per_key` around (mainly for methods defined in table_cache.h) 3. unit tests/crash test updates Tests: * Added unit tests * Crash test: `python3 tools/db_crashtest.py blackbox --simple --block_protection_bytes_per_key=1 --write_buffer_size=1048576` Follow up (maybe as a separate PR): make sure corruption status returned from BlockIters are correctly handled. Performance: Turning on block per KV protection has a non-trivial negative impact on read performance and costs additional memory. For memory, each block includes additional 24 bytes for checksum-related states beside checksum itself. For CPU, I set up a DB of size ~1.2GB with 5M keys (32 bytes key and 200 bytes value) which compacts to ~5 SST files (target file size 256 MB) in L6 without compression. I tested readrandom performance with various block cache size (to mimic various cache hit rates): ``` SETUP make OPTIMIZE_LEVEL="-O3" USE_LTO=1 DEBUG_LEVEL=0 -j32 db_bench ./db_bench -benchmarks=fillseq,compact0,waitforcompaction,compact,waitforcompaction -write_buffer_size=33554432 -level_compaction_dynamic_level_bytes=true -max_background_jobs=8 -target_file_size_base=268435456 --num=5000000 --key_size=32 --value_size=200 --compression_type=none BENCHMARK ./db_bench --use_existing_db -benchmarks=readtocache,readrandom[-X10] --num=5000000 --key_size=32 --disable_auto_compactions --reads=1000000 --block_protection_bytes_per_key=[0|1] --cache_size=$CACHESIZE The readrandom ops/sec looks like the following: Block cache size: 2GB 1.2GB * 0.9 1.2GB * 0.8 1.2GB * 0.5 8MB Main 240805 223604 198176 161653 139040 PR prot_bytes=0 238691 226693 200127 161082 141153 PR prot_bytes=1 214983 193199 178532 137013 108211 prot_bytes=1 vs -10% -15% -10.8% -15% -23% prot_bytes=0 ``` The benchmark has a lot of variance, but there was a 5% to 25% regression in this benchmark with different cache hit rates. Pull Request resolved: https://github.com/facebook/rocksdb/pull/11287 Reviewed By: ajkr Differential Revision: D43970708 Pulled By: cbi42 fbshipit-source-id: ef98d898b71779846fa74212b9ec9e08b7183940
2023-04-25 19:08:23 +00:00
if (std::find(supported.begin(), supported.end(),
cf_options.block_protection_bytes_per_key) == supported.end()) {
return Status::NotSupported(
"Block per key-value checksum protection only supports 0, 1, 2, 4 "
"or 8 bytes per key.");
}
if (!cf_options.compaction_options_fifo.file_temperature_age_thresholds
.empty()) {
if (cf_options.compaction_style != kCompactionStyleFIFO) {
return Status::NotSupported(
"Option file_temperature_age_thresholds only supports FIFO "
"compaction.");
} else if (cf_options.num_levels > 1) {
return Status::NotSupported(
"Option file_temperature_age_thresholds is only supported when "
"num_levels = 1.");
} else {
const auto& ages =
cf_options.compaction_options_fifo.file_temperature_age_thresholds;
assert(ages.size() >= 1);
// check that age is sorted
for (size_t i = 0; i < ages.size() - 1; ++i) {
if (ages[i].age >= ages[i + 1].age) {
return Status::NotSupported(
"Option file_temperature_age_thresholds requires elements to be "
"sorted in increasing order with respect to `age` field.");
}
}
}
}
return s;
}
Status ColumnFamilyData::SetOptions(
const DBOptions& db_opts,
const std::unordered_map<std::string, std::string>& options_map) {
ColumnFamilyOptions cf_opts =
BuildColumnFamilyOptions(initial_cf_options_, mutable_cf_options_);
ConfigOptions config_opts;
config_opts.mutable_options_only = true;
Status s = GetColumnFamilyOptionsFromMap(config_opts, cf_opts, options_map,
&cf_opts);
if (s.ok()) {
s = ValidateOptions(db_opts, cf_opts);
}
if (s.ok()) {
mutable_cf_options_ = MutableCFOptions(cf_opts);
mutable_cf_options_.RefreshDerivedOptions(ioptions_);
}
return s;
}
// REQUIRES: DB mutex held
Env::WriteLifeTimeHint ColumnFamilyData::CalculateSSTWriteHint(int level) {
if (initial_cf_options_.compaction_style != kCompactionStyleLevel) {
return Env::WLTH_NOT_SET;
}
if (level == 0) {
return Env::WLTH_MEDIUM;
}
int base_level = current_->storage_info()->base_level();
// L1: medium, L2: long, ...
if (level - base_level >= 2) {
return Env::WLTH_EXTREME;
} else if (level < base_level) {
// There is no restriction which prevents level passed in to be smaller
// than base_level.
return Env::WLTH_MEDIUM;
}
return static_cast<Env::WriteLifeTimeHint>(
level - base_level + static_cast<int>(Env::WLTH_MEDIUM));
}
Status ColumnFamilyData::AddDirectories(
std::map<std::string, std::shared_ptr<FSDirectory>>* created_dirs) {
Status s;
assert(created_dirs != nullptr);
assert(data_dirs_.empty());
for (auto& p : ioptions_.cf_paths) {
auto existing_dir = created_dirs->find(p.path);
if (existing_dir == created_dirs->end()) {
std::unique_ptr<FSDirectory> path_directory;
s = DBImpl::CreateAndNewDirectory(ioptions_.fs.get(), p.path,
&path_directory);
if (!s.ok()) {
return s;
}
assert(path_directory != nullptr);
data_dirs_.emplace_back(path_directory.release());
(*created_dirs)[p.path] = data_dirs_.back();
} else {
data_dirs_.emplace_back(existing_dir->second);
}
}
assert(data_dirs_.size() == ioptions_.cf_paths.size());
return s;
}
FSDirectory* ColumnFamilyData::GetDataDir(size_t path_id) const {
if (data_dirs_.empty()) {
return nullptr;
}
assert(path_id < data_dirs_.size());
return data_dirs_[path_id].get();
}
Respect cutoff timestamp during flush (#11599) Summary: Make flush respect the cutoff timestamp `full_history_ts_low` as much as possible for the user-defined timestamps in Memtables only feature. We achieve this by not proceeding with the actual flushing but instead reschedule the same `FlushRequest` so a follow up flush job can continue with the check after some interval. This approach doesn't work well for atomic flush, so this feature currently is not supported in combination with atomic flush. Furthermore, this approach also requires a customized method to get the next immediately bigger user-defined timestamp. So currently it's limited to comparator that use uint64_t as the user-defined timestamp format. This support can be extended when we add such a customized method to `AdvancedColumnFamilyOptions`. For non atomic flush request, at any single time, a column family can only have as many as one FlushRequest for it in the `flush_queue_`. There is deduplication done at `FlushRequest` enqueueing(`SchedulePendingFlush`) and dequeueing time (`PopFirstFromFlushQueue`). We hold the db mutex between when a `FlushRequest` is popped from the queue and the same FlushRequest get rescheduled, so no other `FlushRequest` with a higher `max_memtable_id` can be added to the `flush_queue_` blocking us from re-enqueueing the same `FlushRequest`. Flush is continued nevertheless if there is risk of entering write stall mode had the flush being postponed, e.g. due to accumulation of write buffers, exceeding the `max_write_buffer_number` setting. When this happens, the newest user-defined timestamp in the involved Memtables need to be tracked and we use it to increase the `full_history_ts_low`, which is an inclusive cutoff timestamp for which RocksDB promises to keep all user-defined timestamps equal to and newer than it. Tet plan: ``` ./column_family_test --gtest_filter="*RetainUDT*" ./memtable_list_test --gtest_filter="*WithTimestamp*" ./flush_job_test --gtest_filter="*WithTimestamp*" ``` Pull Request resolved: https://github.com/facebook/rocksdb/pull/11599 Reviewed By: ajkr Differential Revision: D47561586 Pulled By: jowlyzhang fbshipit-source-id: 9400445f983dd6eac489e9dd0fb5d9b99637fe89
2023-07-26 23:25:06 +00:00
bool ColumnFamilyData::ShouldPostponeFlushToRetainUDT(
uint64_t max_memtable_id) {
const Comparator* ucmp = user_comparator();
const size_t ts_sz = ucmp->timestamp_size();
if (ts_sz == 0 || ioptions_.persist_user_defined_timestamps) {
return false;
}
// If users set the `persist_user_defined_timestamps` flag to false, they
// should also set the `full_history_ts_low` flag to indicate the range of
// user-defined timestamps to retain in memory. Otherwise, we do not
// explicitly postpone flush to retain UDTs.
const std::string& full_history_ts_low = GetFullHistoryTsLow();
if (full_history_ts_low.empty()) {
return false;
}
#ifndef NDEBUG
Slice last_table_newest_udt;
#endif /* !NDEBUG */
for (const Slice& table_newest_udt :
imm()->GetTablesNewestUDT(max_memtable_id)) {
assert(table_newest_udt.size() == full_history_ts_low.size());
assert(last_table_newest_udt.empty() ||
ucmp->CompareTimestamp(table_newest_udt, last_table_newest_udt) >=
0);
// Checking the newest UDT contained in MemTable with ascending ID up to
// `max_memtable_id`. MemTable with bigger ID will have newer UDT, return
// immediately on finding the first MemTable that needs postponing.
if (ucmp->CompareTimestamp(table_newest_udt, full_history_ts_low) >= 0) {
return true;
}
#ifndef NDEBUG
last_table_newest_udt = table_newest_udt;
#endif /* !NDEBUG */
}
return false;
}
Sort L0 files by newly introduced epoch_num (#10922) 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 https://github.com/ajkr/rocksdb/commit/36a5686ec012f35a4371e409aa85c404ca1c210d (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
2022-12-13 21:29:37 +00:00
void ColumnFamilyData::RecoverEpochNumbers() {
assert(current_);
auto* vstorage = current_->storage_info();
assert(vstorage);
vstorage->RecoverEpochNumbers(this);
}
ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
const ImmutableDBOptions* db_options,
Introduce a new storage specific Env API (#5761) Summary: The current Env API encompasses both storage/file operations, as well as OS related operations. Most of the APIs return a Status, which does not have enough metadata about an error, such as whether its retry-able or not, scope (i.e fault domain) of the error etc., that may be required in order to properly handle a storage error. The file APIs also do not provide enough control over the IO SLA, such as timeout, prioritization, hinting about placement and redundancy etc. This PR separates out the file/storage APIs from Env into a new FileSystem class. The APIs are updated to return an IOStatus with metadata about the error, as well as to take an IOOptions structure as input in order to allow more control over the IO. The user can set both ```options.env``` and ```options.file_system``` to specify that RocksDB should use the former for OS related operations and the latter for storage operations. Internally, a ```CompositeEnvWrapper``` has been introduced that inherits from ```Env``` and redirects individual methods to either an ```Env``` implementation or the ```FileSystem``` as appropriate. When options are sanitized during ```DB::Open```, ```options.env``` is replaced with a newly allocated ```CompositeEnvWrapper``` instance if both env and file_system have been specified. This way, the rest of the RocksDB code can continue to function as before. This PR also ports PosixEnv to the new API by splitting it into two - PosixEnv and PosixFileSystem. PosixEnv is defined as a sub-class of CompositeEnvWrapper, and threading/time functions are overridden with Posix specific implementations in order to avoid an extra level of indirection. The ```CompositeEnvWrapper``` translates ```IOStatus``` return code to ```Status```, and sets the severity to ```kSoftError``` if the io_status is retryable. The error handling code in RocksDB can then recover the DB automatically. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5761 Differential Revision: D18868376 Pulled By: anand1976 fbshipit-source-id: 39efe18a162ea746fabac6360ff529baba48486f
2019-12-13 22:47:08 +00:00
const FileOptions& file_options,
Cache* table_cache,
WriteBufferManager* _write_buffer_manager,
WriteController* _write_controller,
BlockCacheTracer* const block_cache_tracer,
const std::shared_ptr<IOTracer>& io_tracer,
const std::string& db_id,
const std::string& db_session_id)
: max_column_family_(0),
file_options_(file_options),
dummy_cfd_(new ColumnFamilyData(
ColumnFamilyData::kDummyColumnFamilyDataId, "", nullptr, nullptr,
nullptr, ColumnFamilyOptions(), *db_options, &file_options_, nullptr,
block_cache_tracer, io_tracer, db_id, db_session_id)),
default_cfd_cache_(nullptr),
db_name_(dbname),
db_options_(db_options),
table_cache_(table_cache),
write_buffer_manager_(_write_buffer_manager),
write_controller_(_write_controller),
block_cache_tracer_(block_cache_tracer),
io_tracer_(io_tracer),
db_id_(db_id),
db_session_id_(db_session_id) {
// initialize linked list
dummy_cfd_->prev_ = dummy_cfd_;
dummy_cfd_->next_ = dummy_cfd_;
}
ColumnFamilySet::~ColumnFamilySet() {
while (column_family_data_.size() > 0) {
// cfd destructor will delete itself from column_family_data_
auto cfd = column_family_data_.begin()->second;
bool last_ref __attribute__((__unused__));
last_ref = cfd->UnrefAndTryDelete();
assert(last_ref);
}
bool dummy_last_ref __attribute__((__unused__));
dummy_last_ref = dummy_cfd_->UnrefAndTryDelete();
assert(dummy_last_ref);
}
ColumnFamilyData* ColumnFamilySet::GetDefault() const {
assert(default_cfd_cache_ != nullptr);
return default_cfd_cache_;
}
ColumnFamilyData* ColumnFamilySet::GetColumnFamily(uint32_t id) const {
auto cfd_iter = column_family_data_.find(id);
if (cfd_iter != column_family_data_.end()) {
return cfd_iter->second;
} else {
return nullptr;
}
}
ColumnFamilyData* ColumnFamilySet::GetColumnFamily(
const std::string& name) const {
auto cfd_iter = column_families_.find(name);
if (cfd_iter != column_families_.end()) {
auto cfd = GetColumnFamily(cfd_iter->second);
assert(cfd != nullptr);
return cfd;
} else {
return nullptr;
}
}
uint32_t ColumnFamilySet::GetNextColumnFamilyID() {
return ++max_column_family_;
}
uint32_t ColumnFamilySet::GetMaxColumnFamily() { return max_column_family_; }
void ColumnFamilySet::UpdateMaxColumnFamily(uint32_t new_max_column_family) {
max_column_family_ = std::max(new_max_column_family, max_column_family_);
}
size_t ColumnFamilySet::NumberOfColumnFamilies() const {
return column_families_.size();
}
2015-01-06 20:44:21 +00:00
// under a DB mutex AND write thread
ColumnFamilyData* ColumnFamilySet::CreateColumnFamily(
const std::string& name, uint32_t id, Version* dummy_versions,
const ColumnFamilyOptions& options) {
assert(column_families_.find(name) == column_families_.end());
ColumnFamilyData* new_cfd = new ColumnFamilyData(
id, name, dummy_versions, table_cache_, write_buffer_manager_, options,
*db_options_, &file_options_, this, block_cache_tracer_, io_tracer_,
db_id_, db_session_id_);
column_families_.insert({name, id});
column_family_data_.insert({id, new_cfd});
auto ucmp = new_cfd->user_comparator();
assert(ucmp);
size_t ts_sz = ucmp->timestamp_size();
running_ts_sz_.insert({id, ts_sz});
if (ts_sz > 0) {
ts_sz_for_record_.insert({id, ts_sz});
}
max_column_family_ = std::max(max_column_family_, id);
// add to linked list
new_cfd->next_ = dummy_cfd_;
auto prev = dummy_cfd_->prev_;
new_cfd->prev_ = prev;
prev->next_ = new_cfd;
dummy_cfd_->prev_ = new_cfd;
if (id == 0) {
default_cfd_cache_ = new_cfd;
}
return new_cfd;
}
2015-01-06 20:44:21 +00:00
// under a DB mutex AND from a write thread
void ColumnFamilySet::RemoveColumnFamily(ColumnFamilyData* cfd) {
uint32_t cf_id = cfd->GetID();
auto cfd_iter = column_family_data_.find(cf_id);
assert(cfd_iter != column_family_data_.end());
column_family_data_.erase(cfd_iter);
column_families_.erase(cfd->GetName());
running_ts_sz_.erase(cf_id);
ts_sz_for_record_.erase(cf_id);
}
2015-01-06 20:44:21 +00:00
// under a DB mutex OR from a write thread
bool ColumnFamilyMemTablesImpl::Seek(uint32_t column_family_id) {
if (column_family_id == 0) {
// optimization for common case
current_ = column_family_set_->GetDefault();
} else {
current_ = column_family_set_->GetColumnFamily(column_family_id);
}
handle_.SetCFD(current_);
return current_ != nullptr;
}
uint64_t ColumnFamilyMemTablesImpl::GetLogNumber() const {
assert(current_ != nullptr);
return current_->GetLogNumber();
}
MemTable* ColumnFamilyMemTablesImpl::GetMemTable() const {
assert(current_ != nullptr);
return current_->mem();
}
ColumnFamilyHandle* ColumnFamilyMemTablesImpl::GetColumnFamilyHandle() {
assert(current_ != nullptr);
return &handle_;
}
uint32_t GetColumnFamilyID(ColumnFamilyHandle* column_family) {
uint32_t column_family_id = 0;
if (column_family != nullptr) {
auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
column_family_id = cfh->GetID();
}
return column_family_id;
}
const Comparator* GetColumnFamilyUserComparator(
ColumnFamilyHandle* column_family) {
if (column_family != nullptr) {
return column_family->GetComparator();
}
return nullptr;
}
} // namespace ROCKSDB_NAMESPACE