forked from facebook/rocksdb
-
Notifications
You must be signed in to change notification settings - Fork 0
/
db_test2.cc
7214 lines (6281 loc) · 256 KB
/
db_test2.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 <atomic>
#include <cstdlib>
#include <functional>
#include <memory>
#include "db/db_test_util.h"
#include "db/read_callback.h"
#include "options/options_helper.h"
#include "port/port.h"
#include "port/stack_trace.h"
#include "rocksdb/experimental.h"
#include "rocksdb/iostats_context.h"
#include "rocksdb/persistent_cache.h"
#include "rocksdb/trace_record.h"
#include "rocksdb/trace_record_result.h"
#include "rocksdb/utilities/replayer.h"
#include "rocksdb/wal_filter.h"
#include "test_util/testutil.h"
#include "util/random.h"
#include "utilities/fault_injection_env.h"
namespace ROCKSDB_NAMESPACE {
class DBTest2 : public DBTestBase {
public:
DBTest2() : DBTestBase("db_test2", /*env_do_fsync=*/true) {}
protected:
#ifndef ROCKSDB_LITE
uint64_t GetSstSizeHelper(Temperature temperature) {
std::string prop;
EXPECT_TRUE(
dbfull()->GetProperty(DB::Properties::kLiveSstFilesSizeAtTemperature +
ToString(static_cast<uint8_t>(temperature)),
&prop));
return static_cast<uint64_t>(std::atoi(prop.c_str()));
}
#endif // ROCKSDB_LITE
};
#ifndef ROCKSDB_LITE
TEST_F(DBTest2, OpenForReadOnly) {
DB* db_ptr = nullptr;
std::string dbname = test::PerThreadDBPath("db_readonly");
Options options = CurrentOptions();
options.create_if_missing = true;
// OpenForReadOnly should fail but will create <dbname> in the file system
ASSERT_NOK(DB::OpenForReadOnly(options, dbname, &db_ptr));
// Since <dbname> is created, we should be able to delete the dir
// We first get the list files under <dbname>
// There should not be any subdirectories -- this is not checked here
std::vector<std::string> files;
ASSERT_OK(env_->GetChildren(dbname, &files));
for (auto& f : files) {
ASSERT_OK(env_->DeleteFile(dbname + "/" + f));
}
// <dbname> should be empty now and we should be able to delete it
ASSERT_OK(env_->DeleteDir(dbname));
options.create_if_missing = false;
// OpenForReadOnly should fail since <dbname> was successfully deleted
ASSERT_NOK(DB::OpenForReadOnly(options, dbname, &db_ptr));
// With create_if_missing false, there should not be a dir in the file system
ASSERT_NOK(env_->FileExists(dbname));
}
TEST_F(DBTest2, OpenForReadOnlyWithColumnFamilies) {
DB* db_ptr = nullptr;
std::string dbname = test::PerThreadDBPath("db_readonly");
Options options = CurrentOptions();
options.create_if_missing = true;
ColumnFamilyOptions cf_options(options);
std::vector<ColumnFamilyDescriptor> column_families;
column_families.push_back(
ColumnFamilyDescriptor(kDefaultColumnFamilyName, cf_options));
column_families.push_back(ColumnFamilyDescriptor("goku", cf_options));
std::vector<ColumnFamilyHandle*> handles;
// OpenForReadOnly should fail but will create <dbname> in the file system
ASSERT_NOK(
DB::OpenForReadOnly(options, dbname, column_families, &handles, &db_ptr));
// Since <dbname> is created, we should be able to delete the dir
// We first get the list files under <dbname>
// There should not be any subdirectories -- this is not checked here
std::vector<std::string> files;
ASSERT_OK(env_->GetChildren(dbname, &files));
for (auto& f : files) {
ASSERT_OK(env_->DeleteFile(dbname + "/" + f));
}
// <dbname> should be empty now and we should be able to delete it
ASSERT_OK(env_->DeleteDir(dbname));
options.create_if_missing = false;
// OpenForReadOnly should fail since <dbname> was successfully deleted
ASSERT_NOK(
DB::OpenForReadOnly(options, dbname, column_families, &handles, &db_ptr));
// With create_if_missing false, there should not be a dir in the file system
ASSERT_NOK(env_->FileExists(dbname));
}
class TestReadOnlyWithCompressedCache
: public DBTestBase,
public testing::WithParamInterface<std::tuple<int, bool>> {
public:
TestReadOnlyWithCompressedCache()
: DBTestBase("test_readonly_with_compressed_cache",
/*env_do_fsync=*/true) {
max_open_files_ = std::get<0>(GetParam());
use_mmap_ = std::get<1>(GetParam());
}
int max_open_files_;
bool use_mmap_;
};
TEST_P(TestReadOnlyWithCompressedCache, ReadOnlyWithCompressedCache) {
if (use_mmap_ && !IsMemoryMappedAccessSupported()) {
ROCKSDB_GTEST_SKIP("Test requires MMAP support");
return;
}
ASSERT_OK(Put("foo", "bar"));
ASSERT_OK(Put("foo2", "barbarbarbarbarbarbarbar"));
ASSERT_OK(Flush());
DB* db_ptr = nullptr;
Options options = CurrentOptions();
options.allow_mmap_reads = use_mmap_;
options.max_open_files = max_open_files_;
options.compression = kSnappyCompression;
BlockBasedTableOptions table_options;
table_options.block_cache_compressed = NewLRUCache(8 * 1024 * 1024);
table_options.no_block_cache = true;
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
options.statistics = CreateDBStatistics();
ASSERT_OK(DB::OpenForReadOnly(options, dbname_, &db_ptr));
std::string v;
ASSERT_OK(db_ptr->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("bar", v);
ASSERT_EQ(0, options.statistics->getTickerCount(BLOCK_CACHE_COMPRESSED_HIT));
ASSERT_OK(db_ptr->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("bar", v);
if (Snappy_Supported()) {
if (use_mmap_) {
ASSERT_EQ(0,
options.statistics->getTickerCount(BLOCK_CACHE_COMPRESSED_HIT));
} else {
ASSERT_EQ(1,
options.statistics->getTickerCount(BLOCK_CACHE_COMPRESSED_HIT));
}
}
delete db_ptr;
}
INSTANTIATE_TEST_CASE_P(TestReadOnlyWithCompressedCache,
TestReadOnlyWithCompressedCache,
::testing::Combine(::testing::Values(-1, 100),
::testing::Bool()));
class PartitionedIndexTestListener : public EventListener {
public:
void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
ASSERT_GT(info.table_properties.index_partitions, 1);
ASSERT_EQ(info.table_properties.index_key_is_user_key, 0);
}
};
TEST_F(DBTest2, PartitionedIndexUserToInternalKey) {
const int kValueSize = 10500;
const int kNumEntriesPerFile = 1000;
const int kNumFiles = 3;
const int kNumDistinctKeys = 30;
BlockBasedTableOptions table_options;
Options options = CurrentOptions();
options.disable_auto_compactions = true;
table_options.index_type = BlockBasedTableOptions::kTwoLevelIndexSearch;
PartitionedIndexTestListener* listener = new PartitionedIndexTestListener();
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
options.listeners.emplace_back(listener);
std::vector<const Snapshot*> snapshots;
Reopen(options);
Random rnd(301);
for (int i = 0; i < kNumFiles; i++) {
for (int j = 0; j < kNumEntriesPerFile; j++) {
int key_id = (i * kNumEntriesPerFile + j) % kNumDistinctKeys;
std::string value = rnd.RandomString(kValueSize);
ASSERT_OK(Put("keykey_" + std::to_string(key_id), value));
snapshots.push_back(db_->GetSnapshot());
}
ASSERT_OK(Flush());
}
for (auto s : snapshots) {
db_->ReleaseSnapshot(s);
}
}
#endif // ROCKSDB_LITE
class PrefixFullBloomWithReverseComparator
: public DBTestBase,
public ::testing::WithParamInterface<bool> {
public:
PrefixFullBloomWithReverseComparator()
: DBTestBase("prefix_bloom_reverse", /*env_do_fsync=*/true) {}
void SetUp() override { if_cache_filter_ = GetParam(); }
bool if_cache_filter_;
};
TEST_P(PrefixFullBloomWithReverseComparator,
PrefixFullBloomWithReverseComparator) {
Options options = last_options_;
options.comparator = ReverseBytewiseComparator();
options.prefix_extractor.reset(NewCappedPrefixTransform(3));
options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
BlockBasedTableOptions bbto;
if (if_cache_filter_) {
bbto.no_block_cache = false;
bbto.cache_index_and_filter_blocks = true;
bbto.block_cache = NewLRUCache(1);
}
bbto.filter_policy.reset(NewBloomFilterPolicy(10, false));
bbto.whole_key_filtering = false;
options.table_factory.reset(NewBlockBasedTableFactory(bbto));
DestroyAndReopen(options);
ASSERT_OK(dbfull()->Put(WriteOptions(), "bar123", "foo"));
ASSERT_OK(dbfull()->Put(WriteOptions(), "bar234", "foo2"));
ASSERT_OK(dbfull()->Put(WriteOptions(), "foo123", "foo3"));
ASSERT_OK(dbfull()->Flush(FlushOptions()));
if (bbto.block_cache) {
bbto.block_cache->EraseUnRefEntries();
}
std::unique_ptr<Iterator> iter(db_->NewIterator(ReadOptions()));
iter->Seek("bar345");
ASSERT_OK(iter->status());
ASSERT_TRUE(iter->Valid());
ASSERT_EQ("bar234", iter->key().ToString());
ASSERT_EQ("foo2", iter->value().ToString());
iter->Next();
ASSERT_TRUE(iter->Valid());
ASSERT_EQ("bar123", iter->key().ToString());
ASSERT_EQ("foo", iter->value().ToString());
iter->Seek("foo234");
ASSERT_OK(iter->status());
ASSERT_TRUE(iter->Valid());
ASSERT_EQ("foo123", iter->key().ToString());
ASSERT_EQ("foo3", iter->value().ToString());
iter->Seek("bar");
ASSERT_OK(iter->status());
ASSERT_TRUE(!iter->Valid());
}
INSTANTIATE_TEST_CASE_P(PrefixFullBloomWithReverseComparator,
PrefixFullBloomWithReverseComparator, testing::Bool());
TEST_F(DBTest2, IteratorPropertyVersionNumber) {
ASSERT_OK(Put("", ""));
Iterator* iter1 = db_->NewIterator(ReadOptions());
ASSERT_OK(iter1->status());
std::string prop_value;
ASSERT_OK(
iter1->GetProperty("rocksdb.iterator.super-version-number", &prop_value));
uint64_t version_number1 =
static_cast<uint64_t>(std::atoi(prop_value.c_str()));
ASSERT_OK(Put("", ""));
ASSERT_OK(Flush());
Iterator* iter2 = db_->NewIterator(ReadOptions());
ASSERT_OK(iter2->status());
ASSERT_OK(
iter2->GetProperty("rocksdb.iterator.super-version-number", &prop_value));
uint64_t version_number2 =
static_cast<uint64_t>(std::atoi(prop_value.c_str()));
ASSERT_GT(version_number2, version_number1);
ASSERT_OK(Put("", ""));
Iterator* iter3 = db_->NewIterator(ReadOptions());
ASSERT_OK(iter3->status());
ASSERT_OK(
iter3->GetProperty("rocksdb.iterator.super-version-number", &prop_value));
uint64_t version_number3 =
static_cast<uint64_t>(std::atoi(prop_value.c_str()));
ASSERT_EQ(version_number2, version_number3);
iter1->SeekToFirst();
ASSERT_OK(
iter1->GetProperty("rocksdb.iterator.super-version-number", &prop_value));
uint64_t version_number1_new =
static_cast<uint64_t>(std::atoi(prop_value.c_str()));
ASSERT_EQ(version_number1, version_number1_new);
delete iter1;
delete iter2;
delete iter3;
}
TEST_F(DBTest2, CacheIndexAndFilterWithDBRestart) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
BlockBasedTableOptions table_options;
table_options.cache_index_and_filter_blocks = true;
table_options.filter_policy.reset(NewBloomFilterPolicy(20));
options.table_factory.reset(NewBlockBasedTableFactory(table_options));
CreateAndReopenWithCF({"pikachu"}, options);
ASSERT_OK(Put(1, "a", "begin"));
ASSERT_OK(Put(1, "z", "end"));
ASSERT_OK(Flush(1));
TryReopenWithColumnFamilies({"default", "pikachu"}, options);
std::string value;
value = Get(1, "a");
}
TEST_F(DBTest2, MaxSuccessiveMergesChangeWithDBRecovery) {
Options options = CurrentOptions();
options.create_if_missing = true;
options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
options.max_successive_merges = 3;
options.merge_operator = MergeOperators::CreatePutOperator();
options.disable_auto_compactions = true;
DestroyAndReopen(options);
ASSERT_OK(Put("poi", "Finch"));
ASSERT_OK(db_->Merge(WriteOptions(), "poi", "Reese"));
ASSERT_OK(db_->Merge(WriteOptions(), "poi", "Shaw"));
ASSERT_OK(db_->Merge(WriteOptions(), "poi", "Root"));
options.max_successive_merges = 2;
Reopen(options);
}
#ifndef ROCKSDB_LITE
class DBTestSharedWriteBufferAcrossCFs
: public DBTestBase,
public testing::WithParamInterface<std::tuple<bool, bool>> {
public:
DBTestSharedWriteBufferAcrossCFs()
: DBTestBase("db_test_shared_write_buffer", /*env_do_fsync=*/true) {}
void SetUp() override {
use_old_interface_ = std::get<0>(GetParam());
cost_cache_ = std::get<1>(GetParam());
}
bool use_old_interface_;
bool cost_cache_;
};
TEST_P(DBTestSharedWriteBufferAcrossCFs, SharedWriteBufferAcrossCFs) {
Options options = CurrentOptions();
options.arena_block_size = 4096;
auto flush_listener = std::make_shared<FlushCounterListener>();
options.listeners.push_back(flush_listener);
// Don't trip the listener at shutdown.
options.avoid_flush_during_shutdown = true;
// Avoid undeterministic value by malloc_usable_size();
// Force arena block size to 1
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"Arena::Arena:0", [&](void* arg) {
size_t* block_size = static_cast<size_t*>(arg);
*block_size = 1;
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"Arena::AllocateNewBlock:0", [&](void* arg) {
std::pair<size_t*, size_t*>* pair =
static_cast<std::pair<size_t*, size_t*>*>(arg);
*std::get<0>(*pair) = *std::get<1>(*pair);
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
// The total soft write buffer size is about 105000
std::shared_ptr<Cache> cache = NewLRUCache(4 * 1024 * 1024, 2);
ASSERT_LT(cache->GetUsage(), 256 * 1024);
if (use_old_interface_) {
options.db_write_buffer_size = 120000; // this is the real limit
} else if (!cost_cache_) {
options.write_buffer_manager.reset(new WriteBufferManager(114285));
} else {
options.write_buffer_manager.reset(new WriteBufferManager(114285, cache));
}
options.write_buffer_size = 500000; // this is never hit
CreateAndReopenWithCF({"pikachu", "dobrynia", "nikitich"}, options);
WriteOptions wo;
wo.disableWAL = true;
std::function<void()> wait_flush = [&]() {
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[0]));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[1]));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[2]));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[3]));
// Ensure background work is fully finished including listener callbacks
// before accessing listener state.
ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
};
// Create some data and flush "default" and "nikitich" so that they
// are newer CFs created.
flush_listener->expected_flush_reason = FlushReason::kManualFlush;
ASSERT_OK(Put(3, Key(1), DummyString(1), wo));
Flush(3);
ASSERT_OK(Put(3, Key(1), DummyString(1), wo));
ASSERT_OK(Put(0, Key(1), DummyString(1), wo));
Flush(0);
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(1));
flush_listener->expected_flush_reason = FlushReason::kWriteBufferManager;
ASSERT_OK(Put(3, Key(1), DummyString(30000), wo));
if (cost_cache_) {
ASSERT_GE(cache->GetUsage(), 256 * 1024);
ASSERT_LE(cache->GetUsage(), 2 * 256 * 1024);
}
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(60000), wo));
if (cost_cache_) {
ASSERT_GE(cache->GetUsage(), 256 * 1024);
ASSERT_LE(cache->GetUsage(), 2 * 256 * 1024);
}
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
// No flush should trigger
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(1));
}
// Trigger a flush. Flushing "nikitich".
ASSERT_OK(Put(3, Key(2), DummyString(30000), wo));
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(1), wo));
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(2));
}
// Without hitting the threshold, no flush should trigger.
ASSERT_OK(Put(2, Key(1), DummyString(30000), wo));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(2));
}
// Hit the write buffer limit again. "default"
// will have been flushed.
ASSERT_OK(Put(2, Key(2), DummyString(10000), wo));
wait_flush();
ASSERT_OK(Put(3, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(1), wo));
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(2));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(2));
}
// Trigger another flush. This time "dobrynia". "pikachu" should not
// be flushed, althrough it was never flushed.
ASSERT_OK(Put(1, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(80000), wo));
wait_flush();
ASSERT_OK(Put(1, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(2));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "pikachu"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "dobrynia"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "nikitich"),
static_cast<uint64_t>(2));
}
if (cost_cache_) {
ASSERT_GE(cache->GetUsage(), 256 * 1024);
Close();
options.write_buffer_manager.reset();
last_options_.write_buffer_manager.reset();
ASSERT_LT(cache->GetUsage(), 256 * 1024);
}
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
}
INSTANTIATE_TEST_CASE_P(DBTestSharedWriteBufferAcrossCFs,
DBTestSharedWriteBufferAcrossCFs,
::testing::Values(std::make_tuple(true, false),
std::make_tuple(false, false),
std::make_tuple(false, true)));
TEST_F(DBTest2, SharedWriteBufferLimitAcrossDB) {
std::string dbname2 = test::PerThreadDBPath("db_shared_wb_db2");
Options options = CurrentOptions();
options.arena_block_size = 4096;
auto flush_listener = std::make_shared<FlushCounterListener>();
options.listeners.push_back(flush_listener);
// Don't trip the listener at shutdown.
options.avoid_flush_during_shutdown = true;
// Avoid undeterministic value by malloc_usable_size();
// Force arena block size to 1
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"Arena::Arena:0", [&](void* arg) {
size_t* block_size = static_cast<size_t*>(arg);
*block_size = 1;
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
"Arena::AllocateNewBlock:0", [&](void* arg) {
std::pair<size_t*, size_t*>* pair =
static_cast<std::pair<size_t*, size_t*>*>(arg);
*std::get<0>(*pair) = *std::get<1>(*pair);
});
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
options.write_buffer_size = 500000; // this is never hit
// Use a write buffer total size so that the soft limit is about
// 105000.
options.write_buffer_manager.reset(new WriteBufferManager(120000));
CreateAndReopenWithCF({"cf1", "cf2"}, options);
ASSERT_OK(DestroyDB(dbname2, options));
DB* db2 = nullptr;
ASSERT_OK(DB::Open(options, dbname2, &db2));
WriteOptions wo;
wo.disableWAL = true;
std::function<void()> wait_flush = [&]() {
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[0]));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[1]));
ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[2]));
ASSERT_OK(static_cast<DBImpl*>(db2)->TEST_WaitForFlushMemTable());
// Ensure background work is fully finished including listener callbacks
// before accessing listener state.
ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
ASSERT_OK(
static_cast_with_check<DBImpl>(db2)->TEST_WaitForBackgroundWork());
};
// Trigger a flush on cf2
flush_listener->expected_flush_reason = FlushReason::kWriteBufferManager;
ASSERT_OK(Put(2, Key(1), DummyString(70000), wo));
wait_flush();
ASSERT_OK(Put(0, Key(1), DummyString(20000), wo));
wait_flush();
// Insert to DB2
ASSERT_OK(db2->Put(wo, Key(2), DummyString(20000)));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
wait_flush();
ASSERT_OK(static_cast<DBImpl*>(db2)->TEST_WaitForFlushMemTable());
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default") +
GetNumberOfSstFilesForColumnFamily(db_, "cf1") +
GetNumberOfSstFilesForColumnFamily(db_, "cf2"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db2, "default"),
static_cast<uint64_t>(0));
}
// Triggering to flush another CF in DB1
ASSERT_OK(db2->Put(wo, Key(2), DummyString(70000)));
wait_flush();
ASSERT_OK(Put(2, Key(1), DummyString(1), wo));
wait_flush();
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "cf1"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "cf2"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db2, "default"),
static_cast<uint64_t>(0));
}
// Triggering flush in DB2.
ASSERT_OK(db2->Put(wo, Key(3), DummyString(40000)));
wait_flush();
ASSERT_OK(db2->Put(wo, Key(1), DummyString(1)));
wait_flush();
ASSERT_OK(static_cast<DBImpl*>(db2)->TEST_WaitForFlushMemTable());
{
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "default"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "cf1"),
static_cast<uint64_t>(0));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db_, "cf2"),
static_cast<uint64_t>(1));
ASSERT_EQ(GetNumberOfSstFilesForColumnFamily(db2, "default"),
static_cast<uint64_t>(1));
}
delete db2;
ASSERT_OK(DestroyDB(dbname2, options));
ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
}
TEST_F(DBTest2, TestWriteBufferNoLimitWithCache) {
Options options = CurrentOptions();
options.arena_block_size = 4096;
std::shared_ptr<Cache> cache =
NewLRUCache(LRUCacheOptions(10000000, 1, false, 0.0));
options.write_buffer_size = 50000; // this is never hit
// Use a write buffer total size so that the soft limit is about
// 105000.
options.write_buffer_manager.reset(new WriteBufferManager(0, cache));
Reopen(options);
ASSERT_OK(Put("foo", "bar"));
// One dummy entry is 256KB.
ASSERT_GT(cache->GetUsage(), 128000);
}
namespace {
void ValidateKeyExistence(DB* db, const std::vector<Slice>& keys_must_exist,
const std::vector<Slice>& keys_must_not_exist) {
// Ensure that expected keys exist
std::vector<std::string> values;
if (keys_must_exist.size() > 0) {
std::vector<Status> status_list =
db->MultiGet(ReadOptions(), keys_must_exist, &values);
for (size_t i = 0; i < keys_must_exist.size(); i++) {
ASSERT_OK(status_list[i]);
}
}
// Ensure that given keys don't exist
if (keys_must_not_exist.size() > 0) {
std::vector<Status> status_list =
db->MultiGet(ReadOptions(), keys_must_not_exist, &values);
for (size_t i = 0; i < keys_must_not_exist.size(); i++) {
ASSERT_TRUE(status_list[i].IsNotFound());
}
}
}
} // namespace
TEST_F(DBTest2, WalFilterTest) {
class TestWalFilter : public WalFilter {
private:
// Processing option that is requested to be applied at the given index
WalFilter::WalProcessingOption wal_processing_option_;
// Index at which to apply wal_processing_option_
// At other indexes default wal_processing_option::kContinueProcessing is
// returned.
size_t apply_option_at_record_index_;
// Current record index, incremented with each record encountered.
size_t current_record_index_;
public:
TestWalFilter(WalFilter::WalProcessingOption wal_processing_option,
size_t apply_option_for_record_index)
: wal_processing_option_(wal_processing_option),
apply_option_at_record_index_(apply_option_for_record_index),
current_record_index_(0) {}
WalProcessingOption LogRecord(const WriteBatch& /*batch*/,
WriteBatch* /*new_batch*/,
bool* /*batch_changed*/) const override {
WalFilter::WalProcessingOption option_to_return;
if (current_record_index_ == apply_option_at_record_index_) {
option_to_return = wal_processing_option_;
}
else {
option_to_return = WalProcessingOption::kContinueProcessing;
}
// Filter is passed as a const object for RocksDB to not modify the
// object, however we modify it for our own purpose here and hence
// cast the constness away.
(const_cast<TestWalFilter*>(this)->current_record_index_)++;
return option_to_return;
}
const char* Name() const override { return "TestWalFilter"; }
};
// Create 3 batches with two keys each
std::vector<std::vector<std::string>> batch_keys(3);
batch_keys[0].push_back("key1");
batch_keys[0].push_back("key2");
batch_keys[1].push_back("key3");
batch_keys[1].push_back("key4");
batch_keys[2].push_back("key5");
batch_keys[2].push_back("key6");
// Test with all WAL processing options
for (int option = 0;
option < static_cast<int>(
WalFilter::WalProcessingOption::kWalProcessingOptionMax);
option++) {
Options options = OptionsForLogIterTest();
DestroyAndReopen(options);
CreateAndReopenWithCF({ "pikachu" }, options);
// Write given keys in given batches
for (size_t i = 0; i < batch_keys.size(); i++) {
WriteBatch batch;
for (size_t j = 0; j < batch_keys[i].size(); j++) {
ASSERT_OK(batch.Put(handles_[0], batch_keys[i][j], DummyString(1024)));
}
ASSERT_OK(dbfull()->Write(WriteOptions(), &batch));
}
WalFilter::WalProcessingOption wal_processing_option =
static_cast<WalFilter::WalProcessingOption>(option);
// Create a test filter that would apply wal_processing_option at the first
// record
size_t apply_option_for_record_index = 1;
TestWalFilter test_wal_filter(wal_processing_option,
apply_option_for_record_index);
// Reopen database with option to use WAL filter
options = OptionsForLogIterTest();
options.wal_filter = &test_wal_filter;
Status status =
TryReopenWithColumnFamilies({ "default", "pikachu" }, options);
if (wal_processing_option ==
WalFilter::WalProcessingOption::kCorruptedRecord) {
ASSERT_NOK(status);
// In case of corruption we can turn off paranoid_checks to reopen
// databse
options.paranoid_checks = false;
ReopenWithColumnFamilies({ "default", "pikachu" }, options);
}
else {
ASSERT_OK(status);
}
// Compute which keys we expect to be found
// and which we expect not to be found after recovery.
std::vector<Slice> keys_must_exist;
std::vector<Slice> keys_must_not_exist;
switch (wal_processing_option) {
case WalFilter::WalProcessingOption::kCorruptedRecord:
case WalFilter::WalProcessingOption::kContinueProcessing: {
fprintf(stderr, "Testing with complete WAL processing\n");
// we expect all records to be processed
for (size_t i = 0; i < batch_keys.size(); i++) {
for (size_t j = 0; j < batch_keys[i].size(); j++) {
keys_must_exist.push_back(Slice(batch_keys[i][j]));
}
}
break;
}
case WalFilter::WalProcessingOption::kIgnoreCurrentRecord: {
fprintf(stderr,
"Testing with ignoring record %" ROCKSDB_PRIszt " only\n",
apply_option_for_record_index);
// We expect the record with apply_option_for_record_index to be not
// found.
for (size_t i = 0; i < batch_keys.size(); i++) {
for (size_t j = 0; j < batch_keys[i].size(); j++) {
if (i == apply_option_for_record_index) {
keys_must_not_exist.push_back(Slice(batch_keys[i][j]));
}
else {
keys_must_exist.push_back(Slice(batch_keys[i][j]));
}
}
}
break;
}
case WalFilter::WalProcessingOption::kStopReplay: {
fprintf(stderr,
"Testing with stopping replay from record %" ROCKSDB_PRIszt
"\n",
apply_option_for_record_index);
// We expect records beyond apply_option_for_record_index to be not
// found.
for (size_t i = 0; i < batch_keys.size(); i++) {
for (size_t j = 0; j < batch_keys[i].size(); j++) {
if (i >= apply_option_for_record_index) {
keys_must_not_exist.push_back(Slice(batch_keys[i][j]));
}
else {
keys_must_exist.push_back(Slice(batch_keys[i][j]));
}
}
}
break;
}
default:
FAIL(); // unhandled case
}
bool checked_after_reopen = false;
while (true) {
// Ensure that expected keys exists
// and not expected keys don't exist after recovery
ValidateKeyExistence(db_, keys_must_exist, keys_must_not_exist);
if (checked_after_reopen) {
break;
}
// reopen database again to make sure previous log(s) are not used
//(even if they were skipped)
// reopn database with option to use WAL filter
options = OptionsForLogIterTest();
ReopenWithColumnFamilies({ "default", "pikachu" }, options);
checked_after_reopen = true;
}
}
}
TEST_F(DBTest2, WalFilterTestWithChangeBatch) {
class ChangeBatchHandler : public WriteBatch::Handler {
private:
// Batch to insert keys in
WriteBatch* new_write_batch_;
// Number of keys to add in the new batch
size_t num_keys_to_add_in_new_batch_;
// Number of keys added to new batch
size_t num_keys_added_;
public:
ChangeBatchHandler(WriteBatch* new_write_batch,
size_t num_keys_to_add_in_new_batch)
: new_write_batch_(new_write_batch),
num_keys_to_add_in_new_batch_(num_keys_to_add_in_new_batch),
num_keys_added_(0) {}
void Put(const Slice& key, const Slice& value) override {
if (num_keys_added_ < num_keys_to_add_in_new_batch_) {
ASSERT_OK(new_write_batch_->Put(key, value));
++num_keys_added_;
}
}
};
class TestWalFilterWithChangeBatch : public WalFilter {
private:
// Index at which to start changing records
size_t change_records_from_index_;
// Number of keys to add in the new batch
size_t num_keys_to_add_in_new_batch_;
// Current record index, incremented with each record encountered.
size_t current_record_index_;
public:
TestWalFilterWithChangeBatch(size_t change_records_from_index,
size_t num_keys_to_add_in_new_batch)
: change_records_from_index_(change_records_from_index),
num_keys_to_add_in_new_batch_(num_keys_to_add_in_new_batch),
current_record_index_(0) {}
WalProcessingOption LogRecord(const WriteBatch& batch,
WriteBatch* new_batch,
bool* batch_changed) const override {
if (current_record_index_ >= change_records_from_index_) {
ChangeBatchHandler handler(new_batch, num_keys_to_add_in_new_batch_);
Status s = batch.Iterate(&handler);
if (s.ok()) {
*batch_changed = true;
} else {
assert(false);
}
}
// Filter is passed as a const object for RocksDB to not modify the
// object, however we modify it for our own purpose here and hence
// cast the constness away.
(const_cast<TestWalFilterWithChangeBatch*>(this)
->current_record_index_)++;
return WalProcessingOption::kContinueProcessing;
}
const char* Name() const override { return "TestWalFilterWithChangeBatch"; }
};
std::vector<std::vector<std::string>> batch_keys(3);
batch_keys[0].push_back("key1");
batch_keys[0].push_back("key2");
batch_keys[1].push_back("key3");
batch_keys[1].push_back("key4");
batch_keys[2].push_back("key5");
batch_keys[2].push_back("key6");
Options options = OptionsForLogIterTest();
DestroyAndReopen(options);
CreateAndReopenWithCF({ "pikachu" }, options);
// Write given keys in given batches
for (size_t i = 0; i < batch_keys.size(); i++) {
WriteBatch batch;
for (size_t j = 0; j < batch_keys[i].size(); j++) {
ASSERT_OK(batch.Put(handles_[0], batch_keys[i][j], DummyString(1024)));
}
ASSERT_OK(dbfull()->Write(WriteOptions(), &batch));
}
// Create a test filter that would apply wal_processing_option at the first
// record
size_t change_records_from_index = 1;
size_t num_keys_to_add_in_new_batch = 1;
TestWalFilterWithChangeBatch test_wal_filter_with_change_batch(
change_records_from_index, num_keys_to_add_in_new_batch);
// Reopen database with option to use WAL filter
options = OptionsForLogIterTest();
options.wal_filter = &test_wal_filter_with_change_batch;
ReopenWithColumnFamilies({ "default", "pikachu" }, options);
// Ensure that all keys exist before change_records_from_index_
// And after that index only single key exists
// as our filter adds only single key for each batch
std::vector<Slice> keys_must_exist;
std::vector<Slice> keys_must_not_exist;
for (size_t i = 0; i < batch_keys.size(); i++) {
for (size_t j = 0; j < batch_keys[i].size(); j++) {
if (i >= change_records_from_index && j >= num_keys_to_add_in_new_batch) {
keys_must_not_exist.push_back(Slice(batch_keys[i][j]));
}
else {
keys_must_exist.push_back(Slice(batch_keys[i][j]));
}
}
}
bool checked_after_reopen = false;
while (true) {
// Ensure that expected keys exists
// and not expected keys don't exist after recovery
ValidateKeyExistence(db_, keys_must_exist, keys_must_not_exist);