andreacavalli/rocksdb
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
1b4c104a67
rocksdb/monitoring/statistics.cc

407 lines
18 KiB
C++
Raw Normal View History

Updated all copyright headers to the new format.
2016-02-10 00:12:00 +01:00
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
Change RocksDB License Summary: Closes https://github.com/facebook/rocksdb/pull/2589 Differential Revision: D5431502 Pulled By: siying fbshipit-source-id: 8ebf8c87883daa9daa54b2303d11ce01ab1f6f75
2017-07-16 01:03:42 +02:00
// 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).
Add appropriate LICENSE and Copyright message. Summary: Add appropriate LICENSE and Copyright message. Test Plan: make check Reviewers: CC: Task ID: # Blame Rev:
2013-10-16 23:59:46 +02:00
//
Move some files under util/ to separate dirs Summary: Move some files under util/ to new directories env/, monitoring/ options/ and cache/ Closes https://github.com/facebook/rocksdb/pull/2090 Differential Revision: D4833681 Pulled By: siying fbshipit-source-id: 2fd8bef
2017-04-06 04:02:00 +02:00
#include "monitoring/statistics.h"
Fixing broken Mac build Summary: Made some small changes to fix the broken mac build Test Plan: make check all in both linux and mac. All tests pass. Reviewers: sdong, igor, ljin, yhchiang Reviewed By: ljin, yhchiang Subscribers: leveldb Differential Revision: https://reviews.facebook.net/D20895
2014-08-01 05:52:13 +02:00
simplify include directive involving inttypes (#5402) Summary: When using `PRIu64` type of printf specifier, current code base does the following: ``` #ifndef __STDC_FORMAT_MACROS #define __STDC_FORMAT_MACROS #endif #include <inttypes.h> ``` However, this can be simplified to ``` #include <cinttypes> ``` as long as flag `-std=c++11` is used. This should solve issues like https://github.com/facebook/rocksdb/issues/5159 Pull Request resolved: https://github.com/facebook/rocksdb/pull/5402 Differential Revision: D15701195 Pulled By: miasantreble fbshipit-source-id: 6dac0a05f52aadb55e9728038599d3d2e4b59d03
2019-06-06 22:52:39 +02:00
#include <cinttypes>
Replace include/leveldb with include/rocksdb. Summary: Replace include/leveldb with include/rocksdb. Test Plan: make clean; make check make clean; make release Differential Revision: https://reviews.facebook.net/D12489
2013-08-23 17:38:13 +02:00
#include "rocksdb/statistics.h"
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
#include "port/likely.h"
Fix a Statistics-related unit test faulure Summary: In my MacOS, the member variables are populated with random numbers after initialization. This diff fixes it by fill these arrays with 0. Test Plan: make && ./table_test Reviewers: igor CC: leveldb Differential Revision: https://reviews.facebook.net/D15315
2014-01-22 02:51:36 +01:00
#include <algorithm>
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
#include <cstdio>
Change namespace from leveldb to rocksdb Summary: Change namespace from leveldb to rocksdb. This allows a single application to link in open-source leveldb code as well as rocksdb code into the same process. Test Plan: compile rocksdb Reviewers: emayanke Reviewed By: emayanke CC: leveldb Differential Revision: https://reviews.facebook.net/D13287
2013-10-04 06:49:15 +02:00
namespace rocksdb {
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
// The order of items listed in Tickers should be the same as
// the order listed in TickersNameMap
const std::vector<std::pair<Tickers, std::string>> TickersNameMap = {
{BLOCK_CACHE_MISS, "rocksdb.block.cache.miss"},
{BLOCK_CACHE_HIT, "rocksdb.block.cache.hit"},
{BLOCK_CACHE_ADD, "rocksdb.block.cache.add"},
{BLOCK_CACHE_ADD_FAILURES, "rocksdb.block.cache.add.failures"},
{BLOCK_CACHE_INDEX_MISS, "rocksdb.block.cache.index.miss"},
{BLOCK_CACHE_INDEX_HIT, "rocksdb.block.cache.index.hit"},
{BLOCK_CACHE_INDEX_ADD, "rocksdb.block.cache.index.add"},
{BLOCK_CACHE_INDEX_BYTES_INSERT, "rocksdb.block.cache.index.bytes.insert"},
{BLOCK_CACHE_INDEX_BYTES_EVICT, "rocksdb.block.cache.index.bytes.evict"},
{BLOCK_CACHE_FILTER_MISS, "rocksdb.block.cache.filter.miss"},
{BLOCK_CACHE_FILTER_HIT, "rocksdb.block.cache.filter.hit"},
{BLOCK_CACHE_FILTER_ADD, "rocksdb.block.cache.filter.add"},
{BLOCK_CACHE_FILTER_BYTES_INSERT,
"rocksdb.block.cache.filter.bytes.insert"},
{BLOCK_CACHE_FILTER_BYTES_EVICT, "rocksdb.block.cache.filter.bytes.evict"},
{BLOCK_CACHE_DATA_MISS, "rocksdb.block.cache.data.miss"},
{BLOCK_CACHE_DATA_HIT, "rocksdb.block.cache.data.hit"},
{BLOCK_CACHE_DATA_ADD, "rocksdb.block.cache.data.add"},
{BLOCK_CACHE_DATA_BYTES_INSERT, "rocksdb.block.cache.data.bytes.insert"},
{BLOCK_CACHE_BYTES_READ, "rocksdb.block.cache.bytes.read"},
{BLOCK_CACHE_BYTES_WRITE, "rocksdb.block.cache.bytes.write"},
{BLOOM_FILTER_USEFUL, "rocksdb.bloom.filter.useful"},
{BLOOM_FILTER_FULL_POSITIVE, "rocksdb.bloom.filter.full.positive"},
{BLOOM_FILTER_FULL_TRUE_POSITIVE,
"rocksdb.bloom.filter.full.true.positive"},
Introduce a new MultiGet batching implementation (#5011) Summary: This PR introduces a new MultiGet() API, with the underlying implementation grouping keys based on SST file and batching lookups in a file. The reason for the new API is twofold - the definition allows callers to allocate storage for status and values on stack instead of std::vector, as well as return values as PinnableSlices in order to avoid copying, and it keeps the original MultiGet() implementation intact while we experiment with batching. Batching is useful when there is some spatial locality to the keys being queries, as well as larger batch sizes. The main benefits are due to - 1. Fewer function calls, especially to BlockBasedTableReader::MultiGet() and FullFilterBlockReader::KeysMayMatch() 2. Bloom filter cachelines can be prefetched, hiding the cache miss latency The next step is to optimize the binary searches in the level_storage_info, index blocks and data blocks, since we could reduce the number of key comparisons if the keys are relatively close to each other. The batching optimizations also need to be extended to other formats, such as PlainTable and filter formats. This also needs to be added to db_stress. Benchmark results from db_bench for various batch size/locality of reference combinations are given below. Locality was simulated by offsetting the keys in a batch by a stride length. Each SST file is about 8.6MB uncompressed and key/value size is 16/100 uncompressed. To focus on the cpu benefit of batching, the runs were single threaded and bound to the same cpu to eliminate interference from other system events. The results show a 10-25% improvement in micros/op from smaller to larger batch sizes (4 - 32). Batch Sizes 1 | 2 | 4 | 8 | 16 | 32 Random pattern (Stride length 0) 4.158 | 4.109 | 4.026 | 4.05 | 4.1 | 4.074 - Get 4.438 | 4.302 | 4.165 | 4.122 | 4.096 | 4.075 - MultiGet (no batching) 4.461 | 4.256 | 4.277 | 4.11 | 4.182 | 4.14 - MultiGet (w/ batching) Good locality (Stride length 16) 4.048 | 3.659 | 3.248 | 2.99 | 2.84 | 2.753 4.429 | 3.728 | 3.406 | 3.053 | 2.911 | 2.781 4.452 | 3.45 | 2.833 | 2.451 | 2.233 | 2.135 Good locality (Stride length 256) 4.066 | 3.786 | 3.581 | 3.447 | 3.415 | 3.232 4.406 | 4.005 | 3.644 | 3.49 | 3.381 | 3.268 4.393 | 3.649 | 3.186 | 2.882 | 2.676 | 2.62 Medium locality (Stride length 4096) 4.012 | 3.922 | 3.768 | 3.61 | 3.582 | 3.555 4.364 | 4.057 | 3.791 | 3.65 | 3.57 | 3.465 4.479 | 3.758 | 3.316 | 3.077 | 2.959 | 2.891 dbbench command used (on a DB with 4 levels, 12 million keys)- TEST_TMPDIR=/dev/shm numactl -C 10 ./db_bench.tmp -use_existing_db=true -benchmarks="readseq,multireadrandom" -write_buffer_size=4194304 -target_file_size_base=4194304 -max_bytes_for_level_base=16777216 -num=12000000 -reads=12000000 -duration=90 -threads=1 -compression_type=none -cache_size=4194304000 -batch_size=32 -disable_auto_compactions=true -bloom_bits=10 -cache_index_and_filter_blocks=true -pin_l0_filter_and_index_blocks_in_cache=true -multiread_batched=true -multiread_stride=4 Pull Request resolved: https://github.com/facebook/rocksdb/pull/5011 Differential Revision: D14348703 Pulled By: anand1976 fbshipit-source-id: 774406dab3776d979c809522a67bedac6c17f84b
2019-04-11 23:24:09 +02:00
{BLOOM_FILTER_MICROS, "rocksdb.bloom.filter.micros"},
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
{PERSISTENT_CACHE_HIT, "rocksdb.persistent.cache.hit"},
{PERSISTENT_CACHE_MISS, "rocksdb.persistent.cache.miss"},
{SIM_BLOCK_CACHE_HIT, "rocksdb.sim.block.cache.hit"},
{SIM_BLOCK_CACHE_MISS, "rocksdb.sim.block.cache.miss"},
{MEMTABLE_HIT, "rocksdb.memtable.hit"},
{MEMTABLE_MISS, "rocksdb.memtable.miss"},
{GET_HIT_L0, "rocksdb.l0.hit"},
{GET_HIT_L1, "rocksdb.l1.hit"},
{GET_HIT_L2_AND_UP, "rocksdb.l2andup.hit"},
{COMPACTION_KEY_DROP_NEWER_ENTRY, "rocksdb.compaction.key.drop.new"},
{COMPACTION_KEY_DROP_OBSOLETE, "rocksdb.compaction.key.drop.obsolete"},
{COMPACTION_KEY_DROP_RANGE_DEL, "rocksdb.compaction.key.drop.range_del"},
{COMPACTION_KEY_DROP_USER, "rocksdb.compaction.key.drop.user"},
{COMPACTION_RANGE_DEL_DROP_OBSOLETE,
"rocksdb.compaction.range_del.drop.obsolete"},
{COMPACTION_OPTIMIZED_DEL_DROP_OBSOLETE,
"rocksdb.compaction.optimized.del.drop.obsolete"},
{COMPACTION_CANCELLED, "rocksdb.compaction.cancelled"},
{NUMBER_KEYS_WRITTEN, "rocksdb.number.keys.written"},
{NUMBER_KEYS_READ, "rocksdb.number.keys.read"},
{NUMBER_KEYS_UPDATED, "rocksdb.number.keys.updated"},
{BYTES_WRITTEN, "rocksdb.bytes.written"},
{BYTES_READ, "rocksdb.bytes.read"},
{NUMBER_DB_SEEK, "rocksdb.number.db.seek"},
{NUMBER_DB_NEXT, "rocksdb.number.db.next"},
{NUMBER_DB_PREV, "rocksdb.number.db.prev"},
{NUMBER_DB_SEEK_FOUND, "rocksdb.number.db.seek.found"},
{NUMBER_DB_NEXT_FOUND, "rocksdb.number.db.next.found"},
{NUMBER_DB_PREV_FOUND, "rocksdb.number.db.prev.found"},
{ITER_BYTES_READ, "rocksdb.db.iter.bytes.read"},
{NO_FILE_CLOSES, "rocksdb.no.file.closes"},
{NO_FILE_OPENS, "rocksdb.no.file.opens"},
{NO_FILE_ERRORS, "rocksdb.no.file.errors"},
{STALL_L0_SLOWDOWN_MICROS, "rocksdb.l0.slowdown.micros"},
{STALL_MEMTABLE_COMPACTION_MICROS, "rocksdb.memtable.compaction.micros"},
{STALL_L0_NUM_FILES_MICROS, "rocksdb.l0.num.files.stall.micros"},
{STALL_MICROS, "rocksdb.stall.micros"},
{DB_MUTEX_WAIT_MICROS, "rocksdb.db.mutex.wait.micros"},
{RATE_LIMIT_DELAY_MILLIS, "rocksdb.rate.limit.delay.millis"},
{NO_ITERATORS, "rocksdb.num.iterators"},
{NUMBER_MULTIGET_CALLS, "rocksdb.number.multiget.get"},
{NUMBER_MULTIGET_KEYS_READ, "rocksdb.number.multiget.keys.read"},
{NUMBER_MULTIGET_BYTES_READ, "rocksdb.number.multiget.bytes.read"},
{NUMBER_FILTERED_DELETES, "rocksdb.number.deletes.filtered"},
{NUMBER_MERGE_FAILURES, "rocksdb.number.merge.failures"},
{BLOOM_FILTER_PREFIX_CHECKED, "rocksdb.bloom.filter.prefix.checked"},
{BLOOM_FILTER_PREFIX_USEFUL, "rocksdb.bloom.filter.prefix.useful"},
{NUMBER_OF_RESEEKS_IN_ITERATION, "rocksdb.number.reseeks.iteration"},
{GET_UPDATES_SINCE_CALLS, "rocksdb.getupdatessince.calls"},
{BLOCK_CACHE_COMPRESSED_MISS, "rocksdb.block.cachecompressed.miss"},
{BLOCK_CACHE_COMPRESSED_HIT, "rocksdb.block.cachecompressed.hit"},
{BLOCK_CACHE_COMPRESSED_ADD, "rocksdb.block.cachecompressed.add"},
{BLOCK_CACHE_COMPRESSED_ADD_FAILURES,
"rocksdb.block.cachecompressed.add.failures"},
{WAL_FILE_SYNCED, "rocksdb.wal.synced"},
{WAL_FILE_BYTES, "rocksdb.wal.bytes"},
{WRITE_DONE_BY_SELF, "rocksdb.write.self"},
{WRITE_DONE_BY_OTHER, "rocksdb.write.other"},
{WRITE_TIMEDOUT, "rocksdb.write.timeout"},
{WRITE_WITH_WAL, "rocksdb.write.wal"},
{COMPACT_READ_BYTES, "rocksdb.compact.read.bytes"},
{COMPACT_WRITE_BYTES, "rocksdb.compact.write.bytes"},
{FLUSH_WRITE_BYTES, "rocksdb.flush.write.bytes"},
{NUMBER_DIRECT_LOAD_TABLE_PROPERTIES,
"rocksdb.number.direct.load.table.properties"},
{NUMBER_SUPERVERSION_ACQUIRES, "rocksdb.number.superversion_acquires"},
{NUMBER_SUPERVERSION_RELEASES, "rocksdb.number.superversion_releases"},
{NUMBER_SUPERVERSION_CLEANUPS, "rocksdb.number.superversion_cleanups"},
{NUMBER_BLOCK_COMPRESSED, "rocksdb.number.block.compressed"},
{NUMBER_BLOCK_DECOMPRESSED, "rocksdb.number.block.decompressed"},
{NUMBER_BLOCK_NOT_COMPRESSED, "rocksdb.number.block.not_compressed"},
{MERGE_OPERATION_TOTAL_TIME, "rocksdb.merge.operation.time.nanos"},
{FILTER_OPERATION_TOTAL_TIME, "rocksdb.filter.operation.time.nanos"},
{ROW_CACHE_HIT, "rocksdb.row.cache.hit"},
{ROW_CACHE_MISS, "rocksdb.row.cache.miss"},
{READ_AMP_ESTIMATE_USEFUL_BYTES, "rocksdb.read.amp.estimate.useful.bytes"},
{READ_AMP_TOTAL_READ_BYTES, "rocksdb.read.amp.total.read.bytes"},
{NUMBER_RATE_LIMITER_DRAINS, "rocksdb.number.rate_limiter.drains"},
{NUMBER_ITER_SKIP, "rocksdb.number.iter.skip"},
{BLOB_DB_NUM_PUT, "rocksdb.blobdb.num.put"},
{BLOB_DB_NUM_WRITE, "rocksdb.blobdb.num.write"},
{BLOB_DB_NUM_GET, "rocksdb.blobdb.num.get"},
{BLOB_DB_NUM_MULTIGET, "rocksdb.blobdb.num.multiget"},
{BLOB_DB_NUM_SEEK, "rocksdb.blobdb.num.seek"},
{BLOB_DB_NUM_NEXT, "rocksdb.blobdb.num.next"},
{BLOB_DB_NUM_PREV, "rocksdb.blobdb.num.prev"},
{BLOB_DB_NUM_KEYS_WRITTEN, "rocksdb.blobdb.num.keys.written"},
{BLOB_DB_NUM_KEYS_READ, "rocksdb.blobdb.num.keys.read"},
{BLOB_DB_BYTES_WRITTEN, "rocksdb.blobdb.bytes.written"},
{BLOB_DB_BYTES_READ, "rocksdb.blobdb.bytes.read"},
{BLOB_DB_WRITE_INLINED, "rocksdb.blobdb.write.inlined"},
{BLOB_DB_WRITE_INLINED_TTL, "rocksdb.blobdb.write.inlined.ttl"},
{BLOB_DB_WRITE_BLOB, "rocksdb.blobdb.write.blob"},
{BLOB_DB_WRITE_BLOB_TTL, "rocksdb.blobdb.write.blob.ttl"},
{BLOB_DB_BLOB_FILE_BYTES_WRITTEN, "rocksdb.blobdb.blob.file.bytes.written"},
{BLOB_DB_BLOB_FILE_BYTES_READ, "rocksdb.blobdb.blob.file.bytes.read"},
{BLOB_DB_BLOB_FILE_SYNCED, "rocksdb.blobdb.blob.file.synced"},
{BLOB_DB_BLOB_INDEX_EXPIRED_COUNT,
"rocksdb.blobdb.blob.index.expired.count"},
{BLOB_DB_BLOB_INDEX_EXPIRED_SIZE, "rocksdb.blobdb.blob.index.expired.size"},
{BLOB_DB_BLOB_INDEX_EVICTED_COUNT,
"rocksdb.blobdb.blob.index.evicted.count"},
{BLOB_DB_BLOB_INDEX_EVICTED_SIZE, "rocksdb.blobdb.blob.index.evicted.size"},
{BLOB_DB_GC_NUM_FILES, "rocksdb.blobdb.gc.num.files"},
{BLOB_DB_GC_NUM_NEW_FILES, "rocksdb.blobdb.gc.num.new.files"},
{BLOB_DB_GC_FAILURES, "rocksdb.blobdb.gc.failures"},
{BLOB_DB_GC_NUM_KEYS_OVERWRITTEN, "rocksdb.blobdb.gc.num.keys.overwritten"},
{BLOB_DB_GC_NUM_KEYS_EXPIRED, "rocksdb.blobdb.gc.num.keys.expired"},
{BLOB_DB_GC_NUM_KEYS_RELOCATED, "rocksdb.blobdb.gc.num.keys.relocated"},
{BLOB_DB_GC_BYTES_OVERWRITTEN, "rocksdb.blobdb.gc.bytes.overwritten"},
{BLOB_DB_GC_BYTES_EXPIRED, "rocksdb.blobdb.gc.bytes.expired"},
{BLOB_DB_GC_BYTES_RELOCATED, "rocksdb.blobdb.gc.bytes.relocated"},
{BLOB_DB_FIFO_NUM_FILES_EVICTED, "rocksdb.blobdb.fifo.num.files.evicted"},
{BLOB_DB_FIFO_NUM_KEYS_EVICTED, "rocksdb.blobdb.fifo.num.keys.evicted"},
{BLOB_DB_FIFO_BYTES_EVICTED, "rocksdb.blobdb.fifo.bytes.evicted"},
{TXN_PREPARE_MUTEX_OVERHEAD, "rocksdb.txn.overhead.mutex.prepare"},
{TXN_OLD_COMMIT_MAP_MUTEX_OVERHEAD,
"rocksdb.txn.overhead.mutex.old.commit.map"},
{TXN_DUPLICATE_KEY_OVERHEAD, "rocksdb.txn.overhead.duplicate.key"},
{TXN_SNAPSHOT_MUTEX_OVERHEAD, "rocksdb.txn.overhead.mutex.snapshot"},
WritePrepared: fix Get without snapshot (#5664) Summary: if read_options.snapshot is not set, ::Get will take the last sequence number after taking a super-version and uses that as the sequence number. Theoretically max_eviceted_seq_ could advance this sequence number. This could lead ::IsInSnapshot that will be invoked by the ReadCallback to notice the absence of the snapshot. In this case, the ReadCallback should have passed a non-value to snap_released so that it could be set by the ::IsInSnapshot. The patch does that, and adds a unit test to verify it. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5664 Differential Revision: D16614033 Pulled By: maysamyabandeh fbshipit-source-id: 06fb3fd4aacd75806ed1a1acec7961f5d02486f2
2019-08-05 22:30:56 +02:00
{TXN_GET_TRY_AGAIN, "rocksdb.txn.get.tryagain"},
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
{NUMBER_MULTIGET_KEYS_FOUND, "rocksdb.number.multiget.keys.found"},
{NO_ITERATOR_CREATED, "rocksdb.num.iterator.created"},
{NO_ITERATOR_DELETED, "rocksdb.num.iterator.deleted"},
Cache dictionary used for decompressing data blocks (#4881) Summary: - If block cache disabled or not used for meta-blocks, `BlockBasedTableReader::Rep::uncompression_dict` owns the `UncompressionDict`. It is preloaded during `PrefetchIndexAndFilterBlocks`. - If block cache is enabled and used for meta-blocks, block cache owns the `UncompressionDict`, which holds dictionary and digested dictionary when needed. It is never prefetched though there is a TODO for this in the code. The cache key is simply the compression dictionary block handle. - New stats for compression dictionary accesses in block cache: "BLOCK_CACHE_COMPRESSION_DICT_*" and "compression_dict_block_read_count" Pull Request resolved: https://github.com/facebook/rocksdb/pull/4881 Differential Revision: D13663801 Pulled By: ajkr fbshipit-source-id: bdcc54044e180855cdcc57639b493b0e016c9a3f
2019-01-24 03:11:08 +01:00
{BLOCK_CACHE_COMPRESSION_DICT_MISS,
"rocksdb.block.cache.compression.dict.miss"},
{BLOCK_CACHE_COMPRESSION_DICT_HIT,
"rocksdb.block.cache.compression.dict.hit"},
{BLOCK_CACHE_COMPRESSION_DICT_ADD,
"rocksdb.block.cache.compression.dict.add"},
{BLOCK_CACHE_COMPRESSION_DICT_BYTES_INSERT,
"rocksdb.block.cache.compression.dict.bytes.insert"},
{BLOCK_CACHE_COMPRESSION_DICT_BYTES_EVICT,
"rocksdb.block.cache.compression.dict.bytes.evict"},
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
};
const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = {
{DB_GET, "rocksdb.db.get.micros"},
{DB_WRITE, "rocksdb.db.write.micros"},
{COMPACTION_TIME, "rocksdb.compaction.times.micros"},
Add a new CPU time counter to compaction report (#4889) Summary: Measure CPU time consumed for a compaction and report it in the stats report Enable NowCPUNanos() to work for MacOS Pull Request resolved: https://github.com/facebook/rocksdb/pull/4889 Differential Revision: D13701276 Pulled By: zinoale fbshipit-source-id: 5024e5bbccd4dd10fd90d947870237f436445055
2019-01-30 01:23:21 +01:00
{COMPACTION_CPU_TIME, "rocksdb.compaction.times.cpu_micros"},
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
{SUBCOMPACTION_SETUP_TIME, "rocksdb.subcompaction.setup.times.micros"},
{TABLE_SYNC_MICROS, "rocksdb.table.sync.micros"},
{COMPACTION_OUTFILE_SYNC_MICROS, "rocksdb.compaction.outfile.sync.micros"},
{WAL_FILE_SYNC_MICROS, "rocksdb.wal.file.sync.micros"},
{MANIFEST_FILE_SYNC_MICROS, "rocksdb.manifest.file.sync.micros"},
{TABLE_OPEN_IO_MICROS, "rocksdb.table.open.io.micros"},
{DB_MULTIGET, "rocksdb.db.multiget.micros"},
{READ_BLOCK_COMPACTION_MICROS, "rocksdb.read.block.compaction.micros"},
{READ_BLOCK_GET_MICROS, "rocksdb.read.block.get.micros"},
{WRITE_RAW_BLOCK_MICROS, "rocksdb.write.raw.block.micros"},
{STALL_L0_SLOWDOWN_COUNT, "rocksdb.l0.slowdown.count"},
{STALL_MEMTABLE_COMPACTION_COUNT, "rocksdb.memtable.compaction.count"},
{STALL_L0_NUM_FILES_COUNT, "rocksdb.num.files.stall.count"},
{HARD_RATE_LIMIT_DELAY_COUNT, "rocksdb.hard.rate.limit.delay.count"},
{SOFT_RATE_LIMIT_DELAY_COUNT, "rocksdb.soft.rate.limit.delay.count"},
{NUM_FILES_IN_SINGLE_COMPACTION, "rocksdb.numfiles.in.singlecompaction"},
{DB_SEEK, "rocksdb.db.seek.micros"},
{WRITE_STALL, "rocksdb.db.write.stall"},
{SST_READ_MICROS, "rocksdb.sst.read.micros"},
{NUM_SUBCOMPACTIONS_SCHEDULED, "rocksdb.num.subcompactions.scheduled"},
{BYTES_PER_READ, "rocksdb.bytes.per.read"},
{BYTES_PER_WRITE, "rocksdb.bytes.per.write"},
{BYTES_PER_MULTIGET, "rocksdb.bytes.per.multiget"},
{BYTES_COMPRESSED, "rocksdb.bytes.compressed"},
{BYTES_DECOMPRESSED, "rocksdb.bytes.decompressed"},
{COMPRESSION_TIMES_NANOS, "rocksdb.compression.times.nanos"},
{DECOMPRESSION_TIMES_NANOS, "rocksdb.decompression.times.nanos"},
{READ_NUM_MERGE_OPERANDS, "rocksdb.read.num.merge_operands"},
{BLOB_DB_KEY_SIZE, "rocksdb.blobdb.key.size"},
{BLOB_DB_VALUE_SIZE, "rocksdb.blobdb.value.size"},
{BLOB_DB_WRITE_MICROS, "rocksdb.blobdb.write.micros"},
{BLOB_DB_GET_MICROS, "rocksdb.blobdb.get.micros"},
{BLOB_DB_MULTIGET_MICROS, "rocksdb.blobdb.multiget.micros"},
{BLOB_DB_SEEK_MICROS, "rocksdb.blobdb.seek.micros"},
{BLOB_DB_NEXT_MICROS, "rocksdb.blobdb.next.micros"},
{BLOB_DB_PREV_MICROS, "rocksdb.blobdb.prev.micros"},
{BLOB_DB_BLOB_FILE_WRITE_MICROS, "rocksdb.blobdb.blob.file.write.micros"},
{BLOB_DB_BLOB_FILE_READ_MICROS, "rocksdb.blobdb.blob.file.read.micros"},
{BLOB_DB_BLOB_FILE_SYNC_MICROS, "rocksdb.blobdb.blob.file.sync.micros"},
{BLOB_DB_GC_MICROS, "rocksdb.blobdb.gc.micros"},
{BLOB_DB_COMPRESSION_MICROS, "rocksdb.blobdb.compression.micros"},
{BLOB_DB_DECOMPRESSION_MICROS, "rocksdb.blobdb.decompression.micros"},
{FLUSH_TIME, "rocksdb.db.flush.micros"},
Introduce a new MultiGet batching implementation (#5011) Summary: This PR introduces a new MultiGet() API, with the underlying implementation grouping keys based on SST file and batching lookups in a file. The reason for the new API is twofold - the definition allows callers to allocate storage for status and values on stack instead of std::vector, as well as return values as PinnableSlices in order to avoid copying, and it keeps the original MultiGet() implementation intact while we experiment with batching. Batching is useful when there is some spatial locality to the keys being queries, as well as larger batch sizes. The main benefits are due to - 1. Fewer function calls, especially to BlockBasedTableReader::MultiGet() and FullFilterBlockReader::KeysMayMatch() 2. Bloom filter cachelines can be prefetched, hiding the cache miss latency The next step is to optimize the binary searches in the level_storage_info, index blocks and data blocks, since we could reduce the number of key comparisons if the keys are relatively close to each other. The batching optimizations also need to be extended to other formats, such as PlainTable and filter formats. This also needs to be added to db_stress. Benchmark results from db_bench for various batch size/locality of reference combinations are given below. Locality was simulated by offsetting the keys in a batch by a stride length. Each SST file is about 8.6MB uncompressed and key/value size is 16/100 uncompressed. To focus on the cpu benefit of batching, the runs were single threaded and bound to the same cpu to eliminate interference from other system events. The results show a 10-25% improvement in micros/op from smaller to larger batch sizes (4 - 32). Batch Sizes 1 | 2 | 4 | 8 | 16 | 32 Random pattern (Stride length 0) 4.158 | 4.109 | 4.026 | 4.05 | 4.1 | 4.074 - Get 4.438 | 4.302 | 4.165 | 4.122 | 4.096 | 4.075 - MultiGet (no batching) 4.461 | 4.256 | 4.277 | 4.11 | 4.182 | 4.14 - MultiGet (w/ batching) Good locality (Stride length 16) 4.048 | 3.659 | 3.248 | 2.99 | 2.84 | 2.753 4.429 | 3.728 | 3.406 | 3.053 | 2.911 | 2.781 4.452 | 3.45 | 2.833 | 2.451 | 2.233 | 2.135 Good locality (Stride length 256) 4.066 | 3.786 | 3.581 | 3.447 | 3.415 | 3.232 4.406 | 4.005 | 3.644 | 3.49 | 3.381 | 3.268 4.393 | 3.649 | 3.186 | 2.882 | 2.676 | 2.62 Medium locality (Stride length 4096) 4.012 | 3.922 | 3.768 | 3.61 | 3.582 | 3.555 4.364 | 4.057 | 3.791 | 3.65 | 3.57 | 3.465 4.479 | 3.758 | 3.316 | 3.077 | 2.959 | 2.891 dbbench command used (on a DB with 4 levels, 12 million keys)- TEST_TMPDIR=/dev/shm numactl -C 10 ./db_bench.tmp -use_existing_db=true -benchmarks="readseq,multireadrandom" -write_buffer_size=4194304 -target_file_size_base=4194304 -max_bytes_for_level_base=16777216 -num=12000000 -reads=12000000 -duration=90 -threads=1 -compression_type=none -cache_size=4194304000 -batch_size=32 -disable_auto_compactions=true -bloom_bits=10 -cache_index_and_filter_blocks=true -pin_l0_filter_and_index_blocks_in_cache=true -multiread_batched=true -multiread_stride=4 Pull Request resolved: https://github.com/facebook/rocksdb/pull/5011 Differential Revision: D14348703 Pulled By: anand1976 fbshipit-source-id: 774406dab3776d979c809522a67bedac6c17f84b
2019-04-11 23:24:09 +02:00
{SST_BATCH_SIZE, "rocksdb.sst.batch.size"},
FIX #3278: Move global const object definitions from .h to .cc (#4691) Summary: Summary We should declare constants in headers and define them in source files. But this commit is only aimed at compound types. I don't know if it is necessary to do the same thing to fundamental types. I used this command to find all of the constant definitions in header files. `find . -name "*.h" | xargs grep -e "^const .*=.*"` And here is what I found: ``` ./db/version_edit.h:const uint64_t kFileNumberMask = 0x3FFFFFFFFFFFFFFF; ./include/rocksdb/env.h:const size_t kDefaultPageSize = 4 * 1024; ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./include/rocksdb/table.h:const uint32_t kPlainTableVariableLength = 0; ./include/rocksdb/utilities/transaction_db.h:const uint32_t kInitialMaxDeadlocks = 5; ./port/port_posix.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/port_posix.h:const int kMaxInt32 = std::numeric_limits<int32_t>::max(); ./port/port_posix.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/port_posix.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/port_posix.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const uint32_t kMaxUint32 = UINT32_MAX; ./port/win/port_win.h:const int kMaxInt32 = INT32_MAX; ./port/win/port_win.h:const int64_t kMaxInt64 = INT64_MAX; ./port/win/port_win.h:const uint64_t kMaxUint64 = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT64_MAX; ./port/win/port_win.h:const size_t kMaxSizet = UINT_MAX; ./port/win/port_win.h:const uint32_t kMaxUint32 = std::numeric_limits<uint32_t>::max(); ./port/win/port_win.h:const int kMaxInt32 = std::numeric_limits<int>::max(); ./port/win/port_win.h:const uint64_t kMaxUint64 = std::numeric_limits<uint64_t>::max(); ./port/win/port_win.h:const int64_t kMaxInt64 = std::numeric_limits<int64_t>::max(); ./port/win/port_win.h:const size_t kMaxSizet = std::numeric_limits<size_t>::max(); ./port/win/port_win.h:const bool kLittleEndian = true; ./table/cuckoo_table_factory.h:const uint32_t kCuckooMurmurSeedMultiplier = 816922183; ./table/data_block_hash_index.h:const uint8_t kNoEntry = 255; ./table/data_block_hash_index.h:const uint8_t kCollision = 254; ./table/data_block_hash_index.h:const uint8_t kMaxRestartSupportedByHashIndex = 253; ./table/data_block_hash_index.h:const size_t kMaxBlockSizeSupportedByHashIndex = 1u << 16; ./table/data_block_hash_index.h:const double kDefaultUtilRatio = 0.75; ./table/filter_block.h:const uint64_t kNotValid = ULLONG_MAX; ./table/format.h:const int kMagicNumberLengthByte = 8; ./third-party/fbson/FbsonJsonParser.h:const char* const kJsonDelim = " ,]}\t\r\n"; ./third-party/fbson/FbsonJsonParser.h:const char* const kWhiteSpace = " \t\n\r"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const BiggestInt kMaxBiggestInt = ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestStyleFlag[] = "death_test_style"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kDeathTestUseFork[] = "death_test_use_fork"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char kInternalRunDeathTestFlag[] = "internal_run_death_test"; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const char* pets[] = {"cat", "dog"}; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const size_t kProtobufOneLinerMaxLength = 50; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const int kMaxStackTraceDepth = 100; ./third-party/gtest-1.7.0/fused-src/gtest/gtest.h:const T* WithParamInterface<T>::parameter_ = NULL; ./util/coding.h:const unsigned int kMaxVarint64Length = 10; ./util/filename.h:const size_t kFormatFileNumberBufSize = 38; ./util/testutil.h:const SliceTransform* RandomSliceTransform(Random* rnd, int pre_defined = -1); ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ./util/trace_replay.h:const unsigned int kTraceTimestampSize = 8; ./util/trace_replay.h:const unsigned int kTraceTypeSize = 1; ./util/trace_replay.h:const unsigned int kTracePayloadLengthSize = 4; ./util/trace_replay.h:const unsigned int kTraceMetadataSize = ./utilities/cassandra/serialize.h:const int64_t kCharMask = 0xFFLL; ./utilities/cassandra/serialize.h:const int32_t kBitsPerByte = 8; ``` And these 3 lines are related to this commit: ``` ./include/rocksdb/statistics.h:const std::vector<std::pair<Tickers, std::string>> TickersNameMap = { ./include/rocksdb/statistics.h:const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = { ./util/trace_replay.h:const std::string kTraceMagic = "feedcafedeadbeef"; ``` Any comments would be appreciated. Thanks. Pull Request resolved: https://github.com/facebook/rocksdb/pull/4691 Differential Revision: D13208049 Pulled By: ajkr fbshipit-source-id: e5ee55fdaec5447fc5798c6721e2821e7cdc0d5b
2018-11-27 06:30:12 +01:00
};
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
std::shared_ptr<Statistics> CreateDBStatistics() {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
return std::make_shared<StatisticsImpl>(nullptr);
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
}
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
StatisticsImpl::StatisticsImpl(std::shared_ptr<Statistics> stats)
: stats_(std::move(stats)) {}
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
StatisticsImpl::~StatisticsImpl() {}
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
uint64_t StatisticsImpl::getTickerCount(uint32_t tickerType) const {
Thread-specific ticker statistics Summary: The global atomics we previously used for tickers had poor cache performance since they were typically updated from different threads, causing frequent invalidations. In this diff, - recordTick() updates a local ticker value specific to the thread in which it was called - When a thread exits, its local ticker value is added into merged_sum - getTickerCount() returns the sum of all threads' local ticker values and the merged_sum - setTickerCount() resets all threads' local ticker values and sets merged_sum to the value provided by the caller. In a next diff I will make a similar change for histogram stats. Test Plan: before: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 $ perf report -g --stdio | grep recordTick 7.59% db_bench db_bench [.] rocksdb::StatisticsImpl::recordTick ... after: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 $ perf report -g --stdio | grep recordTick 1.46% db_bench db_bench [.] rocksdb::StatisticsImpl::recordTick ... Reviewers: kradhakrishnan, MarkCallaghan, IslamAbdelRahman, sdong Reviewed By: sdong Subscribers: yiwu, andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D62337
2016-08-25 00:42:31 +02:00
MutexLock lock(&aggregate_lock_);
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
return getTickerCountLocked(tickerType);
}
uint64_t StatisticsImpl::getTickerCountLocked(uint32_t tickerType) const {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(tickerType < TICKER_ENUM_MAX);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
uint64_t res = 0;
for (size_t core_idx = 0; core_idx < per_core_stats_.Size(); ++core_idx) {
res += per_core_stats_.AccessAtCore(core_idx)->tickers_[tickerType];
Fix deadlock when calling getMergedHistogram Summary: When calling StatisticsImpl::HistogramInfo::getMergedHistogram(), if there is a dying thread, which is calling ThreadLocalPtr::StaticMeta::OnThreadExit() to merge its thread values to HistogramInfo, deadlock will occur. Because the former try to hold merge_lock then ThreadMeta::mutex_, but the later try to hold ThreadMeta::mutex_ then merge_lock. In short, the locking order isn't the same. This patch addressed this issue by releasing merge_lock before folding thread values. Closes https://github.com/facebook/rocksdb/pull/1552 Differential Revision: D4211942 Pulled By: ajkr fbshipit-source-id: ef89bcb
2016-11-21 03:14:33 +01:00
}
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
return res;
Thread-specific histogram statistics Summary: To reduce contention for atomics when HistogramStats are shared across threads, this diff makes them thread-specific so updates are faster. This comes at the expense of slower reads (much less frequent), which now require merging all histograms. In this diff, - Thread-specific HistogramImpl is created upon the thread's first measureTime() - Thread-specific HistogramImpl are merged and deleted upon thread termination or ThreadLocalPtr destruction, whichever comes first - getHistogramString() and histogramData() merge all histograms, both thread-specific and previously merged ones Test Plan: unit tests, ran db_bench and verified histograms look similar before: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 ... + 7.63% db_bench db_bench [.] rocksdb::HistogramStat::Add after: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 ... + 0.98% db_bench db_bench [.] rocksdb::HistogramStat::Add Reviewers: sdong, MarkCallaghan, kradhakrishnan, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D62649
2016-08-31 23:02:09 +02:00
}
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
void StatisticsImpl::histogramData(uint32_t histogramType,
HistogramData* const data) const {
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
MutexLock lock(&aggregate_lock_);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
getHistogramImplLocked(histogramType)->Data(data);
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
}
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
std::unique_ptr<HistogramImpl> StatisticsImpl::getHistogramImplLocked(
uint32_t histogramType) const {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(histogramType < HISTOGRAM_ENUM_MAX);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
std::unique_ptr<HistogramImpl> res_hist(new HistogramImpl());
for (size_t core_idx = 0; core_idx < per_core_stats_.Size(); ++core_idx) {
res_hist->Merge(
per_core_stats_.AccessAtCore(core_idx)->histograms_[histogramType]);
}
return res_hist;
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
}
Add Statistics.getHistogramString() to print more detailed outputs of a histogram Summary: Provide a way for users to know more detailed ditribution of a histogram metrics. Example outputs: Manually add statement fprintf(stdout, "%s\n", dbstats->getHistogramString(SST_READ_MICROS).c_str()); Will print out something like: Count: 989151 Average: 1.7659 StdDev: 1.52 Min: 0.0000 Median: 1.2071 Max: 860.0000 Percentiles: P50: 1.21 P75: 1.70 P99: 5.12 P99.9: 13.67 P99.99: 21.70 ------------------------------------------------------ [ 0, 1 ) 390839 39.513% 39.513% ######## [ 1, 2 ) 500918 50.641% 90.154% ########## [ 2, 3 ) 79358 8.023% 98.177% ## [ 3, 4 ) 6297 0.637% 98.813% [ 4, 5 ) 1712 0.173% 98.986% [ 5, 6 ) 1134 0.115% 99.101% [ 6, 7 ) 1222 0.124% 99.224% [ 7, 8 ) 1529 0.155% 99.379% [ 8, 9 ) 1264 0.128% 99.507% [ 9, 10 ) 988 0.100% 99.607% [ 10, 12 ) 1378 0.139% 99.746% [ 12, 14 ) 1828 0.185% 99.931% [ 14, 16 ) 410 0.041% 99.972% [ 16, 18 ) 72 0.007% 99.980% [ 18, 20 ) 67 0.007% 99.986% [ 20, 25 ) 106 0.011% 99.997% [ 25, 30 ) 24 0.002% 99.999% [ 30, 35 ) 1 0.000% 100.000% [ 250, 300 ) 2 0.000% 100.000% [ 300, 350 ) 1 0.000% 100.000% [ 800, 900 ) 1 0.000% 100.000% Test Plan: Manually add a print in db_bench and make sure it prints out as expected. Will add some codes to cover the function Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D43611
2015-08-05 22:14:28 +02:00
std::string StatisticsImpl::getHistogramString(uint32_t histogramType) const {
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
MutexLock lock(&aggregate_lock_);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
return getHistogramImplLocked(histogramType)->ToString();
Thread-specific histogram statistics Summary: To reduce contention for atomics when HistogramStats are shared across threads, this diff makes them thread-specific so updates are faster. This comes at the expense of slower reads (much less frequent), which now require merging all histograms. In this diff, - Thread-specific HistogramImpl is created upon the thread's first measureTime() - Thread-specific HistogramImpl are merged and deleted upon thread termination or ThreadLocalPtr destruction, whichever comes first - getHistogramString() and histogramData() merge all histograms, both thread-specific and previously merged ones Test Plan: unit tests, ran db_bench and verified histograms look similar before: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 ... + 7.63% db_bench db_bench [.] rocksdb::HistogramStat::Add after: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 ... + 0.98% db_bench db_bench [.] rocksdb::HistogramStat::Add Reviewers: sdong, MarkCallaghan, kradhakrishnan, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D62649
2016-08-31 23:02:09 +02:00
}
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
void StatisticsImpl::setTickerCount(uint32_t tickerType, uint64_t count) {
Thread-specific ticker statistics Summary: The global atomics we previously used for tickers had poor cache performance since they were typically updated from different threads, causing frequent invalidations. In this diff, - recordTick() updates a local ticker value specific to the thread in which it was called - When a thread exits, its local ticker value is added into merged_sum - getTickerCount() returns the sum of all threads' local ticker values and the merged_sum - setTickerCount() resets all threads' local ticker values and sets merged_sum to the value provided by the caller. In a next diff I will make a similar change for histogram stats. Test Plan: before: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 $ perf report -g --stdio | grep recordTick 7.59% db_bench db_bench [.] rocksdb::StatisticsImpl::recordTick ... after: $ TEST_TMPDIR=/dev/shm/ perf record -g ./db_bench --benchmarks=readwhilewriting --statistics --num=1000000 --use_existing_db --threads=64 --cache_size=250000000 --compression_type=lz4 $ perf report -g --stdio | grep recordTick 1.46% db_bench db_bench [.] rocksdb::StatisticsImpl::recordTick ... Reviewers: kradhakrishnan, MarkCallaghan, IslamAbdelRahman, sdong Reviewed By: sdong Subscribers: yiwu, andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D62337
2016-08-25 00:42:31 +02:00
{
MutexLock lock(&aggregate_lock_);
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
setTickerCountLocked(tickerType, count);
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
}
if (stats_ && tickerType < TICKER_ENUM_MAX) {
stats_->setTickerCount(tickerType, count);
}
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
}
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
void StatisticsImpl::setTickerCountLocked(uint32_t tickerType, uint64_t count) {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(tickerType < TICKER_ENUM_MAX);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
for (size_t core_idx = 0; core_idx < per_core_stats_.Size(); ++core_idx) {
if (core_idx == 0) {
per_core_stats_.AccessAtCore(core_idx)->tickers_[tickerType] = count;
} else {
per_core_stats_.AccessAtCore(core_idx)->tickers_[tickerType] = 0;
}
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
}
}
Add Statistics::getAndResetTickerCount(). Summary: A convience method to atomically get and reset ticker count. I'm wanting to use it to have a thin wrapper to the statistics object to export ticker counts to ODS for LogDevice (since they don't even use fb303). Test Plan: test in LogDevice shadow cluster. https://fburl.com/461868822 Reviewers: andrewkr, yhchiang, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D64869
2016-10-11 19:54:11 +02:00
uint64_t StatisticsImpl::getAndResetTickerCount(uint32_t tickerType) {
uint64_t sum = 0;
{
MutexLock lock(&aggregate_lock_);
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(tickerType < TICKER_ENUM_MAX);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
for (size_t core_idx = 0; core_idx < per_core_stats_.Size(); ++core_idx) {
sum +=
per_core_stats_.AccessAtCore(core_idx)->tickers_[tickerType].exchange(
0, std::memory_order_relaxed);
Add Statistics::getAndResetTickerCount(). Summary: A convience method to atomically get and reset ticker count. I'm wanting to use it to have a thin wrapper to the statistics object to export ticker counts to ODS for LogDevice (since they don't even use fb303). Test Plan: test in LogDevice shadow cluster. https://fburl.com/461868822 Reviewers: andrewkr, yhchiang, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: andrewkr, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D64869
2016-10-11 19:54:11 +02:00
}
}
if (stats_ && tickerType < TICKER_ENUM_MAX) {
stats_->setTickerCount(tickerType, 0);
}
return sum;
}
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
void StatisticsImpl::recordTick(uint32_t tickerType, uint64_t count) {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(tickerType < TICKER_ENUM_MAX);
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
per_core_stats_.Access()->tickers_[tickerType].fetch_add(
count, std::memory_order_relaxed);
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
if (stats_ && tickerType < TICKER_ENUM_MAX) {
stats_->recordTick(tickerType, count);
}
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
}
Add two more StatsLevel (#5027) Summary: Statistics cost too much CPU for some use cases. Add two stats levels so that people can choose to skip two types of expensive stats, timers and histograms. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5027 Differential Revision: D14252765 Pulled By: siying fbshipit-source-id: 75ecec9eaa44c06118229df4f80c366115346592
2019-02-28 19:14:19 +01:00
void StatisticsImpl::recordInHistogram(uint32_t histogramType, uint64_t value) {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(histogramType < HISTOGRAM_ENUM_MAX);
Make statistics's stats_level change thread-safe (#5030) Summary: Right now, users can change statistics.stats_level while DB is running, but TSAN may report data race. We make stats_level_ to be atomic, and access them using accessors. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5030 Differential Revision: D14267519 Pulled By: siying fbshipit-source-id: 37d7ebeff7a43a406230143422a16af899163f73
2019-03-01 19:39:00 +01:00
if (get_stats_level() <= StatsLevel::kExceptHistogramOrTimers) {
Add two more StatsLevel (#5027) Summary: Statistics cost too much CPU for some use cases. Add two stats levels so that people can choose to skip two types of expensive stats, timers and histograms. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5027 Differential Revision: D14252765 Pulled By: siying fbshipit-source-id: 75ecec9eaa44c06118229df4f80c366115346592
2019-02-28 19:14:19 +01:00
return;
}
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
per_core_stats_.Access()->histograms_[histogramType].Add(value);
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
if (stats_ && histogramType < HISTOGRAM_ENUM_MAX) {
Add two more StatsLevel (#5027) Summary: Statistics cost too much CPU for some use cases. Add two stats levels so that people can choose to skip two types of expensive stats, timers and histograms. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5027 Differential Revision: D14252765 Pulled By: siying fbshipit-source-id: 75ecec9eaa44c06118229df4f80c366115346592
2019-02-28 19:14:19 +01:00
stats_->recordInHistogram(histogramType, value);
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
}
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
}
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
Status StatisticsImpl::Reset() {
MutexLock lock(&aggregate_lock_);
for (uint32_t i = 0; i < TICKER_ENUM_MAX; ++i) {
setTickerCountLocked(i, 0);
}
for (uint32_t i = 0; i < HISTOGRAM_ENUM_MAX; ++i) {
Core-local statistics Summary: This diff changes `StatisticsImpl` from a thread-local approach to a core-local one. The goal is to perform faster aggregations, particularly for applications that have many threads. There should be no behavior change. Closes https://github.com/facebook/rocksdb/pull/2258 Differential Revision: D5016258 Pulled By: ajkr fbshipit-source-id: 7d4d165b4a91d8110f0409d113d1be91f22d31a9
2017-05-23 19:29:14 +02:00
for (size_t core_idx = 0; core_idx < per_core_stats_.Size(); ++core_idx) {
per_core_stats_.AccessAtCore(core_idx)->histograms_[i].Clear();
}
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
}
return Status::OK();
}
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
namespace {
Statistics code cleanup Summary: I'm separating code-cleanup part of https://reviews.facebook.net/D14517. This will make D14517 easier to understand and this diff easier to review. Test Plan: make check Reviewers: haobo, kailiu, sdong, dhruba, tnovak Reviewed By: tnovak CC: leveldb Differential Revision: https://reviews.facebook.net/D15099
2014-01-17 21:46:06 +01:00
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
// a buffer size used for temp string buffers
Move some files under util/ to separate dirs Summary: Move some files under util/ to new directories env/, monitoring/ options/ and cache/ Closes https://github.com/facebook/rocksdb/pull/2090 Differential Revision: D4833681 Pulled By: siying fbshipit-source-id: 2fd8bef
2017-04-06 04:02:00 +02:00
const int kTmpStrBufferSize = 200;
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
} // namespace
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
std::string StatisticsImpl::ToString() const {
Add user stats Reset API Summary: It resets all the ticker and histogram stats to zero. Needed to change the locking a bit since Reset() is the only operation that manipulates multiple tickers/histograms together, and that operation should be seen as atomic by other operations that access tickers/histograms. Closes https://github.com/facebook/rocksdb/pull/2213 Differential Revision: D4952232 Pulled By: ajkr fbshipit-source-id: c0475c3e4c7b940120d53891b69c3091149a0679
2017-04-27 00:19:50 +02:00
MutexLock lock(&aggregate_lock_);
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
std::string res;
res.reserve(20000);
for (const auto& t : TickersNameMap) {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(t.first < TICKER_ENUM_MAX);
char buffer[kTmpStrBufferSize];
snprintf(buffer, kTmpStrBufferSize, "%s COUNT : %" PRIu64 "\n",
t.second.c_str(), getTickerCountLocked(t.first));
res.append(buffer);
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
}
for (const auto& h : HistogramsNameMap) {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
assert(h.first < HISTOGRAM_ENUM_MAX);
char buffer[kTmpStrBufferSize];
HistogramData hData;
getHistogramImplLocked(h.first)->Data(&hData);
// don't handle failures - buffer should always be big enough and arguments
// should be provided correctly
Get `CompactionJobInfo` from CompactFiles Summary: Pull Request resolved: https://github.com/facebook/rocksdb/pull/4716 Differential Revision: D13207677 Pulled By: ajkr fbshipit-source-id: d0ccf5a66df6cbb07288b0c5ebad81fd9df3926b
2018-12-13 23:12:02 +01:00
int ret =
snprintf(buffer, kTmpStrBufferSize,
"%s P50 : %f P95 : %f P99 : %f P100 : %f COUNT : %" PRIu64
" SUM : %" PRIu64 "\n",
h.second.c_str(), hData.median, hData.percentile95,
hData.percentile99, hData.max, hData.count, hData.sum);
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
if (ret < 0 || ret >= kTmpStrBufferSize) {
assert(false);
continue;
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
}
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
res.append(buffer);
[rocksdb][refactor] statistic printing code to one place Summary: $title Test Plan: db_bench --statistics=1 Reviewers: haobo Reviewed By: haobo CC: leveldb Differential Revision: https://reviews.facebook.net/D11373
2013-06-19 05:28:41 +02:00
}
res.shrink_to_fit();
return res;
}
add GetStatsHistory to retrieve stats snapshots (#4748) Summary: This PR adds public `GetStatsHistory` API to retrieve stats history in the form of an std map. The key of the map is the timestamp in microseconds when the stats snapshot is taken, the value is another std map from stats name to stats value (stored in std string). Two DBOptions are introduced: `stats_persist_period_sec` (default 10 minutes) controls the intervals between two snapshots are taken; `max_stats_history_count` (default 10) controls the max number of history snapshots to keep in memory. RocksDB will stop collecting stats snapshots if `stats_persist_period_sec` is set to 0. (This PR is the in-memory part of https://github.com/facebook/rocksdb/pull/4535) Pull Request resolved: https://github.com/facebook/rocksdb/pull/4748 Differential Revision: D13961471 Pulled By: miasantreble fbshipit-source-id: ac836d401ecb84ea92216bf9966f969dedf4ad04
2019-02-21 00:46:59 +01:00
bool StatisticsImpl::getTickerMap(
std::map<std::string, uint64_t>* stats_map) const {
assert(stats_map);
if (!stats_map) return false;
stats_map->clear();
MutexLock lock(&aggregate_lock_);
for (const auto& t : TickersNameMap) {
assert(t.first < TICKER_ENUM_MAX);
(*stats_map)[t.second.c_str()] = getTickerCountLocked(t.first);
}
return true;
}
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
bool StatisticsImpl::HistEnabledForType(uint32_t type) const {
Remove enable_internal_stats (#4714) Summary: Simple patch to address comments in [statistics.h#L65](https://github.com/facebook/rocksdb/blob/master/monitoring/statistics.h#L65|statistics.h#L65) `TODO(ajkr): clean this up since there are no internal stats anymore` Pull Request resolved: https://github.com/facebook/rocksdb/pull/4714 Differential Revision: D13208093 Pulled By: ajkr fbshipit-source-id: 4468badb850592411147539f859082644f5296f6
2018-11-27 21:56:40 +01:00
return type < HISTOGRAM_ENUM_MAX;
make statistics forward-able Summary: Make StatisticsImpl being able to forward stats to provided statistics implementation. The main purpose is to allow us to collect internal stats in the future even when user supplies custom statistics implementation. It avoids intrumenting 2 sets of stats collection code. One immediate use case is tuning advisor, which needs to collect some internal stats, users may not be interested. Test Plan: ran db_bench and see stats show up at the end of run Will run make all check since some tests rely on statistics Reviewers: yhchiang, sdong, igor Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D20145
2014-07-28 21:05:36 +02:00
}
Change namespace from leveldb to rocksdb Summary: Change namespace from leveldb to rocksdb. This allows a single application to link in open-source leveldb code as well as rocksdb code into the same process. Test Plan: compile rocksdb Reviewers: emayanke Reviewed By: emayanke CC: leveldb Differential Revision: https://reviews.facebook.net/D13287
2013-10-04 06:49:15 +02:00
} // namespace rocksdb
Reference in New Issue Copy Permalink