You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
rocksdb/db/compaction_job.h

178 lines
5.9 KiB

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#pragma once
#include <atomic>
#include <deque>
#include <functional>
#include <limits>
#include <set>
#include <string>
#include <utility>
#include <vector>
#include "db/column_family.h"
#include "db/compaction_iterator.h"
#include "db/dbformat.h"
#include "db/flush_scheduler.h"
#include "db/internal_stats.h"
#include "db/job_context.h"
#include "db/log_writer.h"
#include "db/memtable_list.h"
#include "db/range_del_aggregator.h"
#include "db/version_edit.h"
#include "db/write_controller.h"
#include "db/write_thread.h"
#include "options/cf_options.h"
#include "options/db_options.h"
#include "port/port.h"
#include "rocksdb/compaction_filter.h"
#include "rocksdb/compaction_job_stats.h"
#include "rocksdb/db.h"
#include "rocksdb/env.h"
#include "rocksdb/memtablerep.h"
#include "rocksdb/transaction_log.h"
#include "table/scoped_arena_iterator.h"
#include "util/autovector.h"
#include "util/event_logger.h"
#include "util/stop_watch.h"
#include "util/thread_local.h"
namespace rocksdb {
class Arena;
class ErrorHandler;
class MemTable;
class SnapshotChecker;
class TableCache;
class Version;
class VersionEdit;
class VersionSet;
class CompactionJob {
public:
CompactionJob(int job_id, Compaction* compaction,
const ImmutableDBOptions& db_options,
const EnvOptions env_options, VersionSet* versions,
Added support for differential snapshots Summary: The motivation for this PR is to add to RocksDB support for differential (incremental) snapshots, as snapshot of the DB changes between two points in time (one can think of it as diff between to sequence numbers, or the diff D which can be thought of as an SST file or just set of KVs that can be applied to sequence number S1 to get the database to the state at sequence number S2). This feature would be useful for various distributed storages layers built on top of RocksDB, as it should help reduce resources (time and network bandwidth) needed to recover and rebuilt DB instances as replicas in the context of distributed storages. From the API standpoint that would like client app requesting iterator between (start seqnum) and current DB state, and reading the "diff". This is a very draft PR for initial review in the discussion on the approach, i'm going to rework some parts and keep updating the PR. For now, what's done here according to initial discussions: Preserving deletes: - We want to be able to optionally preserve recent deletes for some defined period of time, so that if a delete came in recently and might need to be included in the next incremental snapshot it would't get dropped by a compaction. This is done by adding new param to Options (preserve deletes flag) and new variable to DB Impl where we keep track of the sequence number after which we don't want to drop tombstones, even if they are otherwise eligible for deletion. - I also added a new API call for clients to be able to advance this cutoff seqnum after which we drop deletes; i assume it's more flexible to let clients control this, since otherwise we'd need to keep some kind of timestamp < -- > seqnum mapping inside the DB, which sounds messy and painful to support. Clients could make use of it by periodically calling GetLatestSequenceNumber(), noting the timestamp, doing some calculation and figuring out by how much we need to advance the cutoff seqnum. - Compaction codepath in compaction_iterator.cc has been modified to avoid dropping tombstones with seqnum > cutoff seqnum. Iterator changes: - couple params added to ReadOptions, to optionally allow client to request internal keys instead of user keys (so that client can get the latest value of a key, be it delete marker or a put), as well as min timestamp and min seqnum. TableCache changes: - I modified table_cache code to be able to quickly exclude SST files from iterators heep if creation_time on the file is less then iter_start_ts as passed in ReadOptions. That would help a lot in some DB settings (like reading very recent data only or using FIFO compactions), but not so much for universal compaction with more or less long iterator time span. What's left: - Still looking at how to best plug that inside DBIter codepath. So far it seems that FindNextUserKeyInternal only parses values as UserKeys, and iter->key() call generally returns user key. Can we add new API to DBIter as internal_key(), and modify this internal method to optionally set saved_key_ to point to the full internal key? I don't need to store actual seqnum there, but I do need to store type. Closes https://github.com/facebook/rocksdb/pull/2999 Differential Revision: D6175602 Pulled By: mikhail-antonov fbshipit-source-id: c779a6696ee2d574d86c69cec866a3ae095aa900
7 years ago
const std::atomic<bool>* shutting_down,
const SequenceNumber preserve_deletes_seqnum,
LogBuffer* log_buffer,
Directory* db_directory, Directory* output_directory,
Statistics* stats, InstrumentedMutex* db_mutex,
ErrorHandler* db_error_handler,
std::vector<SequenceNumber> existing_snapshots,
SequenceNumber earliest_write_conflict_snapshot,
const SnapshotChecker* snapshot_checker,
std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
bool paranoid_file_checks, bool measure_io_stats,
const std::string& dbname,
CompactionJobStats* compaction_job_stats);
Allow GetThreadList() to report operation stage. Summary: Allow GetThreadList() to report operation stage. Test Plan: ./thread_list_test ./db_bench --benchmarks=fillrandom --num=100000 --threads=40 \ --max_background_compactions=10 --max_background_flushes=3 \ --thread_status_per_interval=1000 --key_size=16 --value_size=1000 \ --num_column_families=10 export ROCKSDB_TESTS=ThreadStatus ./db_test Sample output ThreadID ThreadType cfName Operation OP_StartTime ElapsedTime Stage State 140116265861184 Low Pri 140116270055488 Low Pri 140116274249792 High Pri column_family_name_000005 Flush 2015/03/10-14:58:11 0 us FlushJob::WriteLevel0Table 140116400078912 Low Pri column_family_name_000004 Compaction 2015/03/10-14:58:11 0 us CompactionJob::FinishCompactionOutputFile 140116358135872 Low Pri column_family_name_000006 Compaction 2015/03/10-14:58:10 1 us CompactionJob::FinishCompactionOutputFile 140116341358656 Low Pri 140116295221312 High Pri default Flush 2015/03/10-14:58:11 0 us FlushJob::WriteLevel0Table 140116324581440 Low Pri column_family_name_000009 Compaction 2015/03/10-14:58:11 0 us CompactionJob::ProcessKeyValueCompaction 140116278444096 Low Pri 140116299415616 Low Pri column_family_name_000008 Compaction 2015/03/10-14:58:11 0 us CompactionJob::FinishCompactionOutputFile 140116291027008 High Pri column_family_name_000001 Flush 2015/03/10-14:58:11 0 us FlushJob::WriteLevel0Table 140116286832704 Low Pri column_family_name_000002 Compaction 2015/03/10-14:58:11 0 us CompactionJob::FinishCompactionOutputFile 140116282638400 Low Pri Reviewers: rven, igor, sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D34683
10 years ago
~CompactionJob();
// no copy/move
CompactionJob(CompactionJob&& job) = delete;
CompactionJob(const CompactionJob& job) = delete;
CompactionJob& operator=(const CompactionJob& job) = delete;
// REQUIRED: mutex held
void Prepare();
// REQUIRED mutex not held
Status Run();
Parallelize L0-L1 Compaction: Restructure Compaction Job Summary: As of now compactions involving files from Level 0 and Level 1 are single threaded because the files in L0, although sorted, are not range partitioned like the other levels. This means that during L0-L1 compaction each file from L1 needs to be merged with potentially all the files from L0. This attempt to parallelize the L0-L1 compaction assigns a thread and a corresponding iterator to each L1 file that then considers only the key range found in that L1 file and only the L0 files that have those keys (and only the specific portion of those L0 files in which those keys are found). In this way the overlap is minimized and potentially eliminated between different iterators focusing on the same files. The first step is to restructure the compaction logic to break L0-L1 compactions into multiple, smaller, sequential compactions. Eventually each of these smaller jobs will be run simultaneously. Areas to pay extra attention to are # Correct aggregation of compaction job statistics across multiple threads # Proper opening/closing of output files (make sure each thread's is unique) # Keys that span multiple L1 files # Skewed distributions of keys within L0 files Test Plan: Make and run db_test (newer version has separate compaction tests) and compaction_job_stats_test Reviewers: igor, noetzli, anthony, sdong, yhchiang Reviewed By: yhchiang Subscribers: MarkCallaghan, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D42699
9 years ago
// REQUIRED: mutex held
Status Install(const MutableCFOptions& mutable_cf_options);
private:
struct SubcompactionState;
void AggregateStatistics();
void GenSubcompactionBoundaries();
Parallelize L0-L1 Compaction: Restructure Compaction Job Summary: As of now compactions involving files from Level 0 and Level 1 are single threaded because the files in L0, although sorted, are not range partitioned like the other levels. This means that during L0-L1 compaction each file from L1 needs to be merged with potentially all the files from L0. This attempt to parallelize the L0-L1 compaction assigns a thread and a corresponding iterator to each L1 file that then considers only the key range found in that L1 file and only the L0 files that have those keys (and only the specific portion of those L0 files in which those keys are found). In this way the overlap is minimized and potentially eliminated between different iterators focusing on the same files. The first step is to restructure the compaction logic to break L0-L1 compactions into multiple, smaller, sequential compactions. Eventually each of these smaller jobs will be run simultaneously. Areas to pay extra attention to are # Correct aggregation of compaction job statistics across multiple threads # Proper opening/closing of output files (make sure each thread's is unique) # Keys that span multiple L1 files # Skewed distributions of keys within L0 files Test Plan: Make and run db_test (newer version has separate compaction tests) and compaction_job_stats_test Reviewers: igor, noetzli, anthony, sdong, yhchiang Reviewed By: yhchiang Subscribers: MarkCallaghan, dhruba, leveldb Differential Revision: https://reviews.facebook.net/D42699
9 years ago
Allow GetThreadList() to report basic compaction operation properties. Summary: Now we're able to show more details about a compaction in GetThreadList() :) This patch allows GetThreadList() to report basic compaction operation properties. Basic compaction properties include: 1. job id 2. compaction input / output level 3. compaction property flags (is_manual, is_deletion, .. etc) 4. total input bytes 5. the number of bytes has been read currently. 6. the number of bytes has been written currently. Flush operation properties will be done in a seperate diff. Test Plan: /db_bench --threads=30 --num=1000000 --benchmarks=fillrandom --thread_status_per_interval=1 Sample output of tracking same job: ThreadID ThreadType cfName Operation ElapsedTime Stage State OperationProperties 140664171987072 Low Pri default Compaction 31.357 ms CompactionJob::FinishCompactionOutputFile BaseInputLevel 1 | BytesRead 2264663 | BytesWritten 1934241 | IsDeletion 0 | IsManual 0 | IsTrivialMove 0 | JobID 277 | OutputLevel 2 | TotalInputBytes 3964158 | ThreadID ThreadType cfName Operation ElapsedTime Stage State OperationProperties 140664171987072 Low Pri default Compaction 59.440 ms CompactionJob::FinishCompactionOutputFile BaseInputLevel 1 | BytesRead 2264663 | BytesWritten 1934241 | IsDeletion 0 | IsManual 0 | IsTrivialMove 0 | JobID 277 | OutputLevel 2 | TotalInputBytes 3964158 | ThreadID ThreadType cfName Operation ElapsedTime Stage State OperationProperties 140664171987072 Low Pri default Compaction 226.375 ms CompactionJob::Install BaseInputLevel 1 | BytesRead 3958013 | BytesWritten 3621940 | IsDeletion 0 | IsManual 0 | IsTrivialMove 0 | JobID 277 | OutputLevel 2 | TotalInputBytes 3964158 | Reviewers: sdong, rven, igor Reviewed By: igor Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D37653
10 years ago
// update the thread status for starting a compaction.
void ReportStartedCompaction(Compaction* compaction);
void AllocateCompactionOutputFileNumbers();
// Call compaction filter. Then iterate through input and compact the
// kv-pairs
void ProcessKeyValueCompaction(SubcompactionState* sub_compact);
Status FinishCompactionOutputFile(
const Status& input_status, SubcompactionState* sub_compact,
CompactionRangeDelAggregator* range_del_agg,
CompactionIterationStats* range_del_out_stats,
const Slice* next_table_min_key = nullptr);
Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options);
void RecordCompactionIOStats();
Status OpenCompactionOutputFile(SubcompactionState* sub_compact);
void CleanupCompaction();
void UpdateCompactionJobStats(
const InternalStats::CompactionStats& stats) const;
void RecordDroppedKeys(const CompactionIterationStats& c_iter_stats,
CompactionJobStats* compaction_job_stats = nullptr);
void UpdateCompactionStats();
void UpdateCompactionInputStatsHelper(
int* num_files, uint64_t* bytes_read, int input_level);
void LogCompaction();
int job_id_;
// CompactionJob state
struct CompactionState;
CompactionState* compact_;
CompactionJobStats* compaction_job_stats_;
InternalStats::CompactionStats compaction_stats_;
// DBImpl state
const std::string& dbname_;
const ImmutableDBOptions& db_options_;
const EnvOptions env_options_;
Env* env_;
// env_option optimized for compaction table reads
EnvOptions env_optiosn_for_read_;
VersionSet* versions_;
const std::atomic<bool>* shutting_down_;
Added support for differential snapshots Summary: The motivation for this PR is to add to RocksDB support for differential (incremental) snapshots, as snapshot of the DB changes between two points in time (one can think of it as diff between to sequence numbers, or the diff D which can be thought of as an SST file or just set of KVs that can be applied to sequence number S1 to get the database to the state at sequence number S2). This feature would be useful for various distributed storages layers built on top of RocksDB, as it should help reduce resources (time and network bandwidth) needed to recover and rebuilt DB instances as replicas in the context of distributed storages. From the API standpoint that would like client app requesting iterator between (start seqnum) and current DB state, and reading the "diff". This is a very draft PR for initial review in the discussion on the approach, i'm going to rework some parts and keep updating the PR. For now, what's done here according to initial discussions: Preserving deletes: - We want to be able to optionally preserve recent deletes for some defined period of time, so that if a delete came in recently and might need to be included in the next incremental snapshot it would't get dropped by a compaction. This is done by adding new param to Options (preserve deletes flag) and new variable to DB Impl where we keep track of the sequence number after which we don't want to drop tombstones, even if they are otherwise eligible for deletion. - I also added a new API call for clients to be able to advance this cutoff seqnum after which we drop deletes; i assume it's more flexible to let clients control this, since otherwise we'd need to keep some kind of timestamp < -- > seqnum mapping inside the DB, which sounds messy and painful to support. Clients could make use of it by periodically calling GetLatestSequenceNumber(), noting the timestamp, doing some calculation and figuring out by how much we need to advance the cutoff seqnum. - Compaction codepath in compaction_iterator.cc has been modified to avoid dropping tombstones with seqnum > cutoff seqnum. Iterator changes: - couple params added to ReadOptions, to optionally allow client to request internal keys instead of user keys (so that client can get the latest value of a key, be it delete marker or a put), as well as min timestamp and min seqnum. TableCache changes: - I modified table_cache code to be able to quickly exclude SST files from iterators heep if creation_time on the file is less then iter_start_ts as passed in ReadOptions. That would help a lot in some DB settings (like reading very recent data only or using FIFO compactions), but not so much for universal compaction with more or less long iterator time span. What's left: - Still looking at how to best plug that inside DBIter codepath. So far it seems that FindNextUserKeyInternal only parses values as UserKeys, and iter->key() call generally returns user key. Can we add new API to DBIter as internal_key(), and modify this internal method to optionally set saved_key_ to point to the full internal key? I don't need to store actual seqnum there, but I do need to store type. Closes https://github.com/facebook/rocksdb/pull/2999 Differential Revision: D6175602 Pulled By: mikhail-antonov fbshipit-source-id: c779a6696ee2d574d86c69cec866a3ae095aa900
7 years ago
const SequenceNumber preserve_deletes_seqnum_;
LogBuffer* log_buffer_;
Directory* db_directory_;
Directory* output_directory_;
Statistics* stats_;
InstrumentedMutex* db_mutex_;
ErrorHandler* db_error_handler_;
// If there were two snapshots with seq numbers s1 and
// s2 and s1 < s2, and if we find two instances of a key k1 then lies
// entirely within s1 and s2, then the earlier version of k1 can be safely
// deleted because that version is not visible in any snapshot.
std::vector<SequenceNumber> existing_snapshots_;
// This is the earliest snapshot that could be used for write-conflict
// checking by a transaction. For any user-key newer than this snapshot, we
// should make sure not to remove evidence that a write occurred.
SequenceNumber earliest_write_conflict_snapshot_;
const SnapshotChecker* const snapshot_checker_;
std::shared_ptr<Cache> table_cache_;
EventLogger* event_logger_;
bool bottommost_level_;
bool paranoid_file_checks_;
bool measure_io_stats_;
// Stores the Slices that designate the boundaries for each subcompaction
std::vector<Slice> boundaries_;
// Stores the approx size of keys covered in the range of each subcompaction
std::vector<uint64_t> sizes_;
Env::WriteLifeTimeHint write_hint_;
};
} // namespace rocksdb