-
Notifications
You must be signed in to change notification settings - Fork 6.4k
/
Copy pathdb_impl.cc
2938 lines (2638 loc) · 99.7 KB
/
db_impl.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/db_impl.h"
#ifndef __STDC_FORMAT_MACROS
#define __STDC_FORMAT_MACROS
#endif
#include <stdint.h>
#ifdef OS_SOLARIS
#include <alloca.h>
#endif
#include <algorithm>
#include <cstdio>
#include <map>
#include <set>
#include <stdexcept>
#include <string>
#include <unordered_map>
#include <unordered_set>
#include <utility>
#include <vector>
#include "db/builder.h"
#include "db/compaction_job.h"
#include "db/db_info_dumper.h"
#include "db/db_iter.h"
#include "db/dbformat.h"
#include "db/event_helpers.h"
#include "db/external_sst_file_ingestion_job.h"
#include "db/flush_job.h"
#include "db/forward_iterator.h"
#include "db/job_context.h"
#include "db/log_reader.h"
#include "db/log_writer.h"
#include "db/malloc_stats.h"
#include "db/managed_iterator.h"
#include "db/memtable.h"
#include "db/memtable_list.h"
#include "db/merge_context.h"
#include "db/merge_helper.h"
#include "db/range_del_aggregator.h"
#include "db/table_cache.h"
#include "db/table_properties_collector.h"
#include "db/transaction_log_impl.h"
#include "db/version_set.h"
#include "db/write_batch_internal.h"
#include "db/write_callback.h"
#include "memtable/hash_linklist_rep.h"
#include "memtable/hash_skiplist_rep.h"
#include "monitoring/iostats_context_imp.h"
#include "monitoring/perf_context_imp.h"
#include "monitoring/thread_status_updater.h"
#include "monitoring/thread_status_util.h"
#include "options/cf_options.h"
#include "options/options_helper.h"
#include "options/options_parser.h"
#include "port/port.h"
#include "rocksdb/cache.h"
#include "rocksdb/compaction_filter.h"
#include "rocksdb/convenience.h"
#include "rocksdb/db.h"
#include "rocksdb/env.h"
#include "rocksdb/merge_operator.h"
#include "rocksdb/statistics.h"
#include "rocksdb/status.h"
#include "rocksdb/table.h"
#include "rocksdb/write_buffer_manager.h"
#include "table/block.h"
#include "table/block_based_table_factory.h"
#include "table/merging_iterator.h"
#include "table/table_builder.h"
#include "table/two_level_iterator.h"
#include "tools/sst_dump_tool_imp.h"
#include "util/auto_roll_logger.h"
#include "util/autovector.h"
#include "util/build_version.h"
#include "util/coding.h"
#include "util/compression.h"
#include "util/crc32c.h"
#include "util/file_reader_writer.h"
#include "util/file_util.h"
#include "util/filename.h"
#include "util/log_buffer.h"
#include "util/logging.h"
#include "util/mutexlock.h"
#include "util/sst_file_manager_impl.h"
#include "util/stop_watch.h"
#include "util/string_util.h"
#include "util/sync_point.h"
namespace rocksdb {
const std::string kDefaultColumnFamilyName("default");
void DumpRocksDBBuildVersion(Logger * log);
CompressionType GetCompressionFlush(
const ImmutableCFOptions& ioptions,
const MutableCFOptions& mutable_cf_options) {
// Compressing memtable flushes might not help unless the sequential load
// optimization is used for leveled compaction. Otherwise the CPU and
// latency overhead is not offset by saving much space.
if (ioptions.compaction_style == kCompactionStyleUniversal) {
if (mutable_cf_options.compaction_options_universal
.compression_size_percent < 0) {
return mutable_cf_options.compression;
} else {
return kNoCompression;
}
} else if (!ioptions.compression_per_level.empty()) {
// For leveled compress when min_level_to_compress != 0.
return ioptions.compression_per_level[0];
} else {
return mutable_cf_options.compression;
}
}
namespace {
void DumpSupportInfo(Logger* logger) {
ROCKS_LOG_HEADER(logger, "Compression algorithms supported:");
for (auto& compression : OptionsHelper::compression_type_string_map) {
if (compression.second != kNoCompression &&
compression.second != kDisableCompressionOption) {
ROCKS_LOG_HEADER(logger, "\t%s supported: %d", compression.first.c_str(),
CompressionTypeSupported(compression.second));
}
}
ROCKS_LOG_HEADER(logger, "Fast CRC32 supported: %s",
crc32c::IsFastCrc32Supported().c_str());
}
int64_t kDefaultLowPriThrottledRate = 2 * 1024 * 1024;
} // namespace
DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
const bool seq_per_batch)
: env_(options.env),
dbname_(dbname),
own_info_log_(options.info_log == nullptr),
initial_db_options_(SanitizeOptions(dbname, options)),
immutable_db_options_(initial_db_options_),
mutable_db_options_(initial_db_options_),
stats_(immutable_db_options_.statistics.get()),
db_lock_(nullptr),
mutex_(stats_, env_, DB_MUTEX_WAIT_MICROS,
immutable_db_options_.use_adaptive_mutex),
shutting_down_(false),
bg_cv_(&mutex_),
logfile_number_(0),
log_dir_synced_(false),
log_empty_(true),
default_cf_handle_(nullptr),
log_sync_cv_(&mutex_),
total_log_size_(0),
max_total_in_memory_state_(0),
is_snapshot_supported_(true),
write_buffer_manager_(immutable_db_options_.write_buffer_manager.get()),
write_thread_(immutable_db_options_),
nonmem_write_thread_(immutable_db_options_),
write_controller_(mutable_db_options_.delayed_write_rate),
// Use delayed_write_rate as a base line to determine the initial
// low pri write rate limit. It may be adjusted later.
low_pri_write_rate_limiter_(NewGenericRateLimiter(std::min(
static_cast<int64_t>(mutable_db_options_.delayed_write_rate / 8),
kDefaultLowPriThrottledRate))),
last_batch_group_size_(0),
unscheduled_flushes_(0),
unscheduled_compactions_(0),
bg_bottom_compaction_scheduled_(0),
bg_compaction_scheduled_(0),
num_running_compactions_(0),
bg_flush_scheduled_(0),
num_running_flushes_(0),
bg_purge_scheduled_(0),
disable_delete_obsolete_files_(0),
pending_purge_obsolete_files_(0),
delete_obsolete_files_last_run_(env_->NowMicros()),
last_stats_dump_time_microsec_(0),
next_job_id_(1),
has_unpersisted_data_(false),
unable_to_flush_oldest_log_(false),
env_options_(BuildDBOptions(immutable_db_options_, mutable_db_options_)),
env_options_for_compaction_(env_->OptimizeForCompactionTableWrite(
env_options_, immutable_db_options_)),
num_running_ingest_file_(0),
#ifndef ROCKSDB_LITE
wal_manager_(immutable_db_options_, env_options_, seq_per_batch),
#endif // ROCKSDB_LITE
event_logger_(immutable_db_options_.info_log.get()),
bg_work_paused_(0),
bg_compaction_paused_(0),
refitting_level_(false),
opened_successfully_(false),
two_write_queues_(options.two_write_queues),
manual_wal_flush_(options.manual_wal_flush),
seq_per_batch_(seq_per_batch),
// last_sequencee_ is always maintained by the main queue that also writes
// to the memtable. When two_write_queues_ is disabled last seq in
// memtable is the same as last seq published to the readers. When it is
// enabled but seq_per_batch_ is disabled, last seq in memtable still
// indicates last published seq since wal-only writes that go to the 2nd
// queue do not consume a sequence number. Otherwise writes performed by
// the 2nd queue could change what is visible to the readers. In this
// cases, last_seq_same_as_publish_seq_==false, the 2nd queue maintains a
// separate variable to indicate the last published sequence.
last_seq_same_as_publish_seq_(
!(seq_per_batch && options.two_write_queues)),
// Since seq_per_batch_ is currently set only by WritePreparedTxn which
// requires a custom gc for compaction, we use that to set use_custom_gc_
// as well.
use_custom_gc_(seq_per_batch),
preserve_deletes_(options.preserve_deletes),
closed_(false) {
env_->GetAbsolutePath(dbname, &db_absolute_path_);
// Reserve ten files or so for other uses and give the rest to TableCache.
// Give a large number for setting of "infinite" open files.
const int table_cache_size = (mutable_db_options_.max_open_files == -1)
? TableCache::kInfiniteCapacity
: mutable_db_options_.max_open_files - 10;
table_cache_ = NewLRUCache(table_cache_size,
immutable_db_options_.table_cache_numshardbits);
versions_.reset(new VersionSet(dbname_, &immutable_db_options_, env_options_,
table_cache_.get(), write_buffer_manager_,
&write_controller_));
column_family_memtables_.reset(
new ColumnFamilyMemTablesImpl(versions_->GetColumnFamilySet()));
DumpRocksDBBuildVersion(immutable_db_options_.info_log.get());
DumpDBFileSummary(immutable_db_options_, dbname_);
immutable_db_options_.Dump(immutable_db_options_.info_log.get());
mutable_db_options_.Dump(immutable_db_options_.info_log.get());
DumpSupportInfo(immutable_db_options_.info_log.get());
// always open the DB with 0 here, which means if preserve_deletes_==true
// we won't drop any deletion markers until SetPreserveDeletesSequenceNumber()
// is called by client and this seqnum is advanced.
preserve_deletes_seqnum_.store(0);
}
// Will lock the mutex_, will wait for completion if wait is true
void DBImpl::CancelAllBackgroundWork(bool wait) {
InstrumentedMutexLock l(&mutex_);
ROCKS_LOG_INFO(immutable_db_options_.info_log,
"Shutdown: canceling all background work");
if (!shutting_down_.load(std::memory_order_acquire) &&
has_unpersisted_data_.load(std::memory_order_relaxed) &&
!mutable_db_options_.avoid_flush_during_shutdown) {
for (auto cfd : *versions_->GetColumnFamilySet()) {
if (!cfd->IsDropped() && cfd->initialized() && !cfd->mem()->IsEmpty()) {
cfd->Ref();
mutex_.Unlock();
FlushMemTable(cfd, FlushOptions());
mutex_.Lock();
cfd->Unref();
}
}
versions_->GetColumnFamilySet()->FreeDeadColumnFamilies();
}
shutting_down_.store(true, std::memory_order_release);
bg_cv_.SignalAll();
if (!wait) {
return;
}
// Wait for background work to finish
while (bg_bottom_compaction_scheduled_ || bg_compaction_scheduled_ ||
bg_flush_scheduled_) {
bg_cv_.Wait();
}
}
Status DBImpl::CloseImpl() {
// CancelAllBackgroundWork called with false means we just set the shutdown
// marker. After this we do a variant of the waiting and unschedule work
// (to consider: moving all the waiting into CancelAllBackgroundWork(true))
CancelAllBackgroundWork(false);
int bottom_compactions_unscheduled =
env_->UnSchedule(this, Env::Priority::BOTTOM);
int compactions_unscheduled = env_->UnSchedule(this, Env::Priority::LOW);
int flushes_unscheduled = env_->UnSchedule(this, Env::Priority::HIGH);
mutex_.Lock();
bg_bottom_compaction_scheduled_ -= bottom_compactions_unscheduled;
bg_compaction_scheduled_ -= compactions_unscheduled;
bg_flush_scheduled_ -= flushes_unscheduled;
// Wait for background work to finish
while (bg_bottom_compaction_scheduled_ || bg_compaction_scheduled_ ||
bg_flush_scheduled_ || bg_purge_scheduled_ ||
pending_purge_obsolete_files_) {
TEST_SYNC_POINT("DBImpl::~DBImpl:WaitJob");
bg_cv_.Wait();
}
EraseThreadStatusDbInfo();
flush_scheduler_.Clear();
while (!flush_queue_.empty()) {
auto cfd = PopFirstFromFlushQueue();
if (cfd->Unref()) {
delete cfd;
}
}
while (!compaction_queue_.empty()) {
auto cfd = PopFirstFromCompactionQueue();
if (cfd->Unref()) {
delete cfd;
}
}
if (default_cf_handle_ != nullptr) {
// we need to delete handle outside of lock because it does its own locking
mutex_.Unlock();
delete default_cf_handle_;
mutex_.Lock();
}
// Clean up obsolete files due to SuperVersion release.
// (1) Need to delete to obsolete files before closing because RepairDB()
// scans all existing files in the file system and builds manifest file.
// Keeping obsolete files confuses the repair process.
// (2) Need to check if we Open()/Recover() the DB successfully before
// deleting because if VersionSet recover fails (may be due to corrupted
// manifest file), it is not able to identify live files correctly. As a
// result, all "live" files can get deleted by accident. However, corrupted
// manifest is recoverable by RepairDB().
if (opened_successfully_) {
JobContext job_context(next_job_id_.fetch_add(1));
FindObsoleteFiles(&job_context, true);
mutex_.Unlock();
// manifest number starting from 2
job_context.manifest_file_number = 1;
if (job_context.HaveSomethingToDelete()) {
PurgeObsoleteFiles(job_context);
}
job_context.Clean();
mutex_.Lock();
}
for (auto l : logs_to_free_) {
delete l;
}
for (auto& log : logs_) {
log.ClearWriter();
}
logs_.clear();
// Table cache may have table handles holding blocks from the block cache.
// We need to release them before the block cache is destroyed. The block
// cache may be destroyed inside versions_.reset(), when column family data
// list is destroyed, so leaving handles in table cache after
// versions_.reset() may cause issues.
// Here we clean all unreferenced handles in table cache.
// Now we assume all user queries have finished, so only version set itself
// can possibly hold the blocks from block cache. After releasing unreferenced
// handles here, only handles held by version set left and inside
// versions_.reset(), we will release them. There, we need to make sure every
// time a handle is released, we erase it from the cache too. By doing that,
// we can guarantee that after versions_.reset(), table cache is empty
// so the cache can be safely destroyed.
table_cache_->EraseUnRefEntries();
for (auto& txn_entry : recovered_transactions_) {
delete txn_entry.second;
}
// versions need to be destroyed before table_cache since it can hold
// references to table_cache.
versions_.reset();
mutex_.Unlock();
if (db_lock_ != nullptr) {
env_->UnlockFile(db_lock_);
}
ROCKS_LOG_INFO(immutable_db_options_.info_log, "Shutdown complete");
LogFlush(immutable_db_options_.info_log);
Status s = Status::OK();
if (immutable_db_options_.info_log && own_info_log_) {
s = immutable_db_options_.info_log->Close();
}
return s;
}
DBImpl::~DBImpl() { Close(); }
void DBImpl::MaybeIgnoreError(Status* s) const {
if (s->ok() || immutable_db_options_.paranoid_checks) {
// No change needed
} else {
ROCKS_LOG_WARN(immutable_db_options_.info_log, "Ignoring error %s",
s->ToString().c_str());
*s = Status::OK();
}
}
const Status DBImpl::CreateArchivalDirectory() {
if (immutable_db_options_.wal_ttl_seconds > 0 ||
immutable_db_options_.wal_size_limit_mb > 0) {
std::string archivalPath = ArchivalDirectory(immutable_db_options_.wal_dir);
return env_->CreateDirIfMissing(archivalPath);
}
return Status::OK();
}
void DBImpl::PrintStatistics() {
auto dbstats = immutable_db_options_.statistics.get();
if (dbstats) {
ROCKS_LOG_WARN(immutable_db_options_.info_log, "STATISTICS:\n %s",
dbstats->ToString().c_str());
}
}
void DBImpl::MaybeDumpStats() {
mutex_.Lock();
unsigned int stats_dump_period_sec =
mutable_db_options_.stats_dump_period_sec;
mutex_.Unlock();
if (stats_dump_period_sec == 0) return;
const uint64_t now_micros = env_->NowMicros();
if (last_stats_dump_time_microsec_ + stats_dump_period_sec * 1000000 <=
now_micros) {
// Multiple threads could race in here simultaneously.
// However, the last one will update last_stats_dump_time_microsec_
// atomically. We could see more than one dump during one dump
// period in rare cases.
last_stats_dump_time_microsec_ = now_micros;
#ifndef ROCKSDB_LITE
const DBPropertyInfo* cf_property_info =
GetPropertyInfo(DB::Properties::kCFStats);
assert(cf_property_info != nullptr);
const DBPropertyInfo* db_property_info =
GetPropertyInfo(DB::Properties::kDBStats);
assert(db_property_info != nullptr);
std::string stats;
{
InstrumentedMutexLock l(&mutex_);
default_cf_internal_stats_->GetStringProperty(
*db_property_info, DB::Properties::kDBStats, &stats);
for (auto cfd : *versions_->GetColumnFamilySet()) {
if (cfd->initialized()) {
cfd->internal_stats()->GetStringProperty(
*cf_property_info, DB::Properties::kCFStatsNoFileHistogram,
&stats);
}
}
for (auto cfd : *versions_->GetColumnFamilySet()) {
if (cfd->initialized()) {
cfd->internal_stats()->GetStringProperty(
*cf_property_info, DB::Properties::kCFFileHistogram, &stats);
}
}
}
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"------- DUMPING STATS -------");
ROCKS_LOG_WARN(immutable_db_options_.info_log, "%s", stats.c_str());
if (immutable_db_options_.dump_malloc_stats) {
stats.clear();
DumpMallocStats(&stats);
if (!stats.empty()) {
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"------- Malloc STATS -------");
ROCKS_LOG_WARN(immutable_db_options_.info_log, "%s", stats.c_str());
}
}
#endif // !ROCKSDB_LITE
PrintStatistics();
}
}
void DBImpl::ScheduleBgLogWriterClose(JobContext* job_context) {
if (!job_context->logs_to_free.empty()) {
for (auto l : job_context->logs_to_free) {
AddToLogsToFreeQueue(l);
}
job_context->logs_to_free.clear();
SchedulePurge();
}
}
Directory* DBImpl::Directories::GetDataDir(size_t path_id) {
assert(path_id < data_dirs_.size());
Directory* ret_dir = data_dirs_[path_id].get();
if (ret_dir == nullptr) {
// Should use db_dir_
return db_dir_.get();
}
return ret_dir;
}
Status DBImpl::SetOptions(ColumnFamilyHandle* column_family,
const std::unordered_map<std::string, std::string>& options_map) {
#ifdef ROCKSDB_LITE
return Status::NotSupported("Not supported in ROCKSDB LITE");
#else
auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family)->cfd();
if (options_map.empty()) {
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"SetOptions() on column family [%s], empty input",
cfd->GetName().c_str());
return Status::InvalidArgument("empty input");
}
MutableCFOptions new_options;
Status s;
Status persist_options_status;
WriteThread::Writer w;
SuperVersionContext sv_context(/* create_superversion */ true);
{
InstrumentedMutexLock l(&mutex_);
s = cfd->SetOptions(options_map);
if (s.ok()) {
new_options = *cfd->GetLatestMutableCFOptions();
// Append new version to recompute compaction score.
VersionEdit dummy_edit;
versions_->LogAndApply(cfd, new_options, &dummy_edit, &mutex_,
directories_.GetDbDir());
// Trigger possible flush/compactions. This has to be before we persist
// options to file, otherwise there will be a deadlock with writer
// thread.
InstallSuperVersionAndScheduleWork(cfd, &sv_context, new_options);
persist_options_status = WriteOptionsFile(
false /*need_mutex_lock*/, true /*need_enter_write_thread*/);
}
}
sv_context.Clean();
ROCKS_LOG_INFO(immutable_db_options_.info_log,
"SetOptions() on column family [%s], inputs:",
cfd->GetName().c_str());
for (const auto& o : options_map) {
ROCKS_LOG_INFO(immutable_db_options_.info_log, "%s: %s\n", o.first.c_str(),
o.second.c_str());
}
if (s.ok()) {
ROCKS_LOG_INFO(immutable_db_options_.info_log,
"[%s] SetOptions() succeeded", cfd->GetName().c_str());
new_options.Dump(immutable_db_options_.info_log.get());
if (!persist_options_status.ok()) {
s = persist_options_status;
}
} else {
ROCKS_LOG_WARN(immutable_db_options_.info_log, "[%s] SetOptions() failed",
cfd->GetName().c_str());
}
LogFlush(immutable_db_options_.info_log);
return s;
#endif // ROCKSDB_LITE
}
Status DBImpl::SetDBOptions(
const std::unordered_map<std::string, std::string>& options_map) {
#ifdef ROCKSDB_LITE
return Status::NotSupported("Not supported in ROCKSDB LITE");
#else
if (options_map.empty()) {
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"SetDBOptions(), empty input.");
return Status::InvalidArgument("empty input");
}
MutableDBOptions new_options;
Status s;
Status persist_options_status;
bool wal_changed = false;
WriteThread::Writer w;
WriteContext write_context;
{
InstrumentedMutexLock l(&mutex_);
s = GetMutableDBOptionsFromStrings(mutable_db_options_, options_map,
&new_options);
if (s.ok()) {
if (new_options.max_background_compactions >
mutable_db_options_.max_background_compactions) {
env_->IncBackgroundThreadsIfNeeded(
new_options.max_background_compactions, Env::Priority::LOW);
MaybeScheduleFlushOrCompaction();
}
write_controller_.set_max_delayed_write_rate(new_options.delayed_write_rate);
table_cache_.get()->SetCapacity(new_options.max_open_files == -1
? TableCache::kInfiniteCapacity
: new_options.max_open_files - 10);
wal_changed = mutable_db_options_.wal_bytes_per_sync !=
new_options.wal_bytes_per_sync;
if (new_options.bytes_per_sync == 0) {
new_options.bytes_per_sync = 1024 * 1024;
}
mutable_db_options_ = new_options;
env_options_for_compaction_ = EnvOptions(
BuildDBOptions(immutable_db_options_, mutable_db_options_));
env_options_for_compaction_ = env_->OptimizeForCompactionTableWrite(
env_options_for_compaction_, immutable_db_options_);
versions_->ChangeEnvOptions(mutable_db_options_);
env_options_for_compaction_ = env_->OptimizeForCompactionTableRead(
env_options_for_compaction_, immutable_db_options_);
env_options_for_compaction_.compaction_readahead_size =
mutable_db_options_.compaction_readahead_size;
write_thread_.EnterUnbatched(&w, &mutex_);
if (total_log_size_ > GetMaxTotalWalSize() || wal_changed) {
Status purge_wal_status = SwitchWAL(&write_context);
if (!purge_wal_status.ok()) {
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"Unable to purge WAL files in SetDBOptions() -- %s",
purge_wal_status.ToString().c_str());
}
}
persist_options_status = WriteOptionsFile(
false /*need_mutex_lock*/, false /*need_enter_write_thread*/);
write_thread_.ExitUnbatched(&w);
}
}
ROCKS_LOG_INFO(immutable_db_options_.info_log, "SetDBOptions(), inputs:");
for (const auto& o : options_map) {
ROCKS_LOG_INFO(immutable_db_options_.info_log, "%s: %s\n", o.first.c_str(),
o.second.c_str());
}
if (s.ok()) {
ROCKS_LOG_INFO(immutable_db_options_.info_log, "SetDBOptions() succeeded");
new_options.Dump(immutable_db_options_.info_log.get());
if (!persist_options_status.ok()) {
if (immutable_db_options_.fail_if_options_file_error) {
s = Status::IOError(
"SetDBOptions() succeeded, but unable to persist options",
persist_options_status.ToString());
}
ROCKS_LOG_WARN(immutable_db_options_.info_log,
"Unable to persist options in SetDBOptions() -- %s",
persist_options_status.ToString().c_str());
}
} else {
ROCKS_LOG_WARN(immutable_db_options_.info_log, "SetDBOptions failed");
}
LogFlush(immutable_db_options_.info_log);
return s;
#endif // ROCKSDB_LITE
}
// return the same level if it cannot be moved
int DBImpl::FindMinimumEmptyLevelFitting(ColumnFamilyData* cfd,
const MutableCFOptions& mutable_cf_options, int level) {
mutex_.AssertHeld();
const auto* vstorage = cfd->current()->storage_info();
int minimum_level = level;
for (int i = level - 1; i > 0; --i) {
// stop if level i is not empty
if (vstorage->NumLevelFiles(i) > 0) break;
// stop if level i is too small (cannot fit the level files)
if (vstorage->MaxBytesForLevel(i) < vstorage->NumLevelBytes(level)) {
break;
}
minimum_level = i;
}
return minimum_level;
}
Status DBImpl::FlushWAL(bool sync) {
{
// We need to lock log_write_mutex_ since logs_ might change concurrently
InstrumentedMutexLock wl(&log_write_mutex_);
log::Writer* cur_log_writer = logs_.back().writer;
auto s = cur_log_writer->WriteBuffer();
if (!s.ok()) {
ROCKS_LOG_ERROR(immutable_db_options_.info_log, "WAL flush error %s",
s.ToString().c_str());
}
if (!sync) {
ROCKS_LOG_DEBUG(immutable_db_options_.info_log, "FlushWAL sync=false");
return s;
}
}
// sync = true
ROCKS_LOG_DEBUG(immutable_db_options_.info_log, "FlushWAL sync=true");
return SyncWAL();
}
Status DBImpl::SyncWAL() {
autovector<log::Writer*, 1> logs_to_sync;
bool need_log_dir_sync;
uint64_t current_log_number;
{
InstrumentedMutexLock l(&mutex_);
assert(!logs_.empty());
// This SyncWAL() call only cares about logs up to this number.
current_log_number = logfile_number_;
while (logs_.front().number <= current_log_number &&
logs_.front().getting_synced) {
log_sync_cv_.Wait();
}
// First check that logs are safe to sync in background.
for (auto it = logs_.begin();
it != logs_.end() && it->number <= current_log_number; ++it) {
if (!it->writer->file()->writable_file()->IsSyncThreadSafe()) {
return Status::NotSupported(
"SyncWAL() is not supported for this implementation of WAL file",
immutable_db_options_.allow_mmap_writes
? "try setting Options::allow_mmap_writes to false"
: Slice());
}
}
for (auto it = logs_.begin();
it != logs_.end() && it->number <= current_log_number; ++it) {
auto& log = *it;
assert(!log.getting_synced);
log.getting_synced = true;
logs_to_sync.push_back(log.writer);
}
need_log_dir_sync = !log_dir_synced_;
}
TEST_SYNC_POINT("DBWALTest::SyncWALNotWaitWrite:1");
RecordTick(stats_, WAL_FILE_SYNCED);
Status status;
for (log::Writer* log : logs_to_sync) {
status = log->file()->SyncWithoutFlush(immutable_db_options_.use_fsync);
if (!status.ok()) {
break;
}
}
if (status.ok() && need_log_dir_sync) {
status = directories_.GetWalDir()->Fsync();
}
TEST_SYNC_POINT("DBWALTest::SyncWALNotWaitWrite:2");
TEST_SYNC_POINT("DBImpl::SyncWAL:BeforeMarkLogsSynced:1");
{
InstrumentedMutexLock l(&mutex_);
MarkLogsSynced(current_log_number, need_log_dir_sync, status);
}
TEST_SYNC_POINT("DBImpl::SyncWAL:BeforeMarkLogsSynced:2");
return status;
}
void DBImpl::MarkLogsSynced(
uint64_t up_to, bool synced_dir, const Status& status) {
mutex_.AssertHeld();
if (synced_dir &&
logfile_number_ == up_to &&
status.ok()) {
log_dir_synced_ = true;
}
for (auto it = logs_.begin(); it != logs_.end() && it->number <= up_to;) {
auto& log = *it;
assert(log.getting_synced);
if (status.ok() && logs_.size() > 1) {
logs_to_free_.push_back(log.ReleaseWriter());
it = logs_.erase(it);
} else {
log.getting_synced = false;
++it;
}
}
assert(!status.ok() || logs_.empty() || logs_[0].number > up_to ||
(logs_.size() == 1 && !logs_[0].getting_synced));
log_sync_cv_.SignalAll();
}
SequenceNumber DBImpl::GetLatestSequenceNumber() const {
return versions_->LastSequence();
}
void DBImpl::SetLastPublishedSequence(SequenceNumber seq) {
versions_->SetLastPublishedSequence(seq);
}
bool DBImpl::SetPreserveDeletesSequenceNumber(SequenceNumber seqnum) {
if (seqnum > preserve_deletes_seqnum_.load()) {
preserve_deletes_seqnum_.store(seqnum);
return true;
} else {
return false;
}
}
InternalIterator* DBImpl::NewInternalIterator(
Arena* arena, RangeDelAggregator* range_del_agg,
ColumnFamilyHandle* column_family) {
ColumnFamilyData* cfd;
if (column_family == nullptr) {
cfd = default_cf_handle_->cfd();
} else {
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
cfd = cfh->cfd();
}
mutex_.Lock();
SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
mutex_.Unlock();
ReadOptions roptions;
return NewInternalIterator(roptions, cfd, super_version, arena,
range_del_agg);
}
void DBImpl::SchedulePurge() {
mutex_.AssertHeld();
assert(opened_successfully_);
// Purge operations are put into High priority queue
bg_purge_scheduled_++;
env_->Schedule(&DBImpl::BGWorkPurge, this, Env::Priority::HIGH, nullptr);
}
void DBImpl::BackgroundCallPurge() {
mutex_.Lock();
// We use one single loop to clear both queues so that after existing the loop
// both queues are empty. This is stricter than what is needed, but can make
// it easier for us to reason the correctness.
while (!purge_queue_.empty() || !logs_to_free_queue_.empty()) {
if (!purge_queue_.empty()) {
auto purge_file = purge_queue_.begin();
auto fname = purge_file->fname;
auto type = purge_file->type;
auto number = purge_file->number;
auto path_id = purge_file->path_id;
auto job_id = purge_file->job_id;
purge_queue_.pop_front();
mutex_.Unlock();
DeleteObsoleteFileImpl(job_id, fname, type, number, path_id);
mutex_.Lock();
} else {
assert(!logs_to_free_queue_.empty());
log::Writer* log_writer = *(logs_to_free_queue_.begin());
logs_to_free_queue_.pop_front();
mutex_.Unlock();
delete log_writer;
mutex_.Lock();
}
}
bg_purge_scheduled_--;
bg_cv_.SignalAll();
// IMPORTANT:there should be no code after calling SignalAll. This call may
// signal the DB destructor that it's OK to proceed with destruction. In
// that case, all DB variables will be dealloacated and referencing them
// will cause trouble.
mutex_.Unlock();
}
namespace {
struct IterState {
IterState(DBImpl* _db, InstrumentedMutex* _mu, SuperVersion* _super_version,
bool _background_purge)
: db(_db),
mu(_mu),
super_version(_super_version),
background_purge(_background_purge) {}
DBImpl* db;
InstrumentedMutex* mu;
SuperVersion* super_version;
bool background_purge;
};
static void CleanupIteratorState(void* arg1, void* arg2) {
IterState* state = reinterpret_cast<IterState*>(arg1);
if (state->super_version->Unref()) {
// Job id == 0 means that this is not our background process, but rather
// user thread
JobContext job_context(0);
state->mu->Lock();
state->super_version->Cleanup();
state->db->FindObsoleteFiles(&job_context, false, true);
if (state->background_purge) {
state->db->ScheduleBgLogWriterClose(&job_context);
}
state->mu->Unlock();
delete state->super_version;
if (job_context.HaveSomethingToDelete()) {
if (state->background_purge) {
// PurgeObsoleteFiles here does not delete files. Instead, it adds the
// files to be deleted to a job queue, and deletes it in a separate
// background thread.
state->db->PurgeObsoleteFiles(job_context, true /* schedule only */);
state->mu->Lock();
state->db->SchedulePurge();
state->mu->Unlock();
} else {
state->db->PurgeObsoleteFiles(job_context);
}
}
job_context.Clean();
}
delete state;
}
} // namespace
InternalIterator* DBImpl::NewInternalIterator(
const ReadOptions& read_options, ColumnFamilyData* cfd,
SuperVersion* super_version, Arena* arena,
RangeDelAggregator* range_del_agg) {
InternalIterator* internal_iter;
assert(arena != nullptr);
assert(range_del_agg != nullptr);
// Need to create internal iterator from the arena.
MergeIteratorBuilder merge_iter_builder(
&cfd->internal_comparator(), arena,
!read_options.total_order_seek &&
cfd->ioptions()->prefix_extractor != nullptr);
// Collect iterator for mutable mem
merge_iter_builder.AddIterator(
super_version->mem->NewIterator(read_options, arena));
std::unique_ptr<InternalIterator> range_del_iter;
Status s;
if (!read_options.ignore_range_deletions) {
range_del_iter.reset(
super_version->mem->NewRangeTombstoneIterator(read_options));
s = range_del_agg->AddTombstones(std::move(range_del_iter));
}
// Collect all needed child iterators for immutable memtables
if (s.ok()) {
super_version->imm->AddIterators(read_options, &merge_iter_builder);
if (!read_options.ignore_range_deletions) {
s = super_version->imm->AddRangeTombstoneIterators(read_options, arena,
range_del_agg);
}
}
TEST_SYNC_POINT_CALLBACK("DBImpl::NewInternalIterator:StatusCallback", &s);
if (s.ok()) {
// Collect iterators for files in L0 - Ln
if (read_options.read_tier != kMemtableTier) {
super_version->current->AddIterators(read_options, env_options_,
&merge_iter_builder, range_del_agg);
}
internal_iter = merge_iter_builder.Finish();
IterState* cleanup =
new IterState(this, &mutex_, super_version,
read_options.background_purge_on_iterator_cleanup);
internal_iter->RegisterCleanup(CleanupIteratorState, cleanup, nullptr);
return internal_iter;
} else {
CleanupSuperVersion(super_version);
}
return NewErrorInternalIterator(s, arena);
}
ColumnFamilyHandle* DBImpl::DefaultColumnFamily() const {
return default_cf_handle_;
}
Status DBImpl::Get(const ReadOptions& read_options,
ColumnFamilyHandle* column_family, const Slice& key,
PinnableSlice* value) {
return GetImpl(read_options, column_family, key, value);
}
Status DBImpl::GetImpl(const ReadOptions& read_options,
ColumnFamilyHandle* column_family, const Slice& key,
PinnableSlice* pinnable_val, bool* value_found,
ReadCallback* callback, bool* is_blob_index) {
assert(pinnable_val != nullptr);
StopWatch sw(env_, stats_, DB_GET);
PERF_TIMER_GUARD(get_snapshot_time);
auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
auto cfd = cfh->cfd();
// Acquire SuperVersion
SuperVersion* sv = GetAndRefSuperVersion(cfd);
TEST_SYNC_POINT("DBImpl::GetImpl:1");
TEST_SYNC_POINT("DBImpl::GetImpl:2");
SequenceNumber snapshot;
if (read_options.snapshot != nullptr) {
// Note: In WritePrepared txns this is not necessary but not harmful either.
// Because prep_seq > snapshot => commit_seq > snapshot so if a snapshot is
// specified we should be fine with skipping seq numbers that are greater
// than that.
snapshot = reinterpret_cast<const SnapshotImpl*>(
read_options.snapshot)->number_;
} else {
// Since we get and reference the super version before getting
// the snapshot number, without a mutex protection, it is possible