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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
|
/**
* Copyright (C) 2016 MongoDB Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License, version 3,
* as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* As a special exception, the copyright holders give permission to link the
* code of portions of this program with the OpenSSL library under certain
* conditions as described in each individual source file and distribute
* linked combinations including the program with the OpenSSL library. You
* must comply with the GNU Affero General Public License in all respects for
* all of the code used other than as permitted herein. If you modify file(s)
* with this exception, you may extend this exception to your version of the
* file(s), but you are not obligated to do so. If you do not wish to do so,
* delete this exception statement from your version. If you delete this
* exception statement from all source files in the program, then also delete
* it in the license file.
*/
#include "mongo/platform/basic.h"
#include "mongo/db/op_observer_impl.h"
#include "mongo/bson/bsonobjbuilder.h"
#include "mongo/db/auth/authorization_manager.h"
#include "mongo/db/catalog/collection_catalog_entry.h"
#include "mongo/db/catalog/collection_options.h"
#include "mongo/db/catalog/database.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/namespace_uuid_cache.h"
#include "mongo/db/commands/feature_compatibility_version.h"
#include "mongo/db/commands/feature_compatibility_version_parser.h"
#include "mongo/db/concurrency/d_concurrency.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/logical_time_validator.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/repl/oplog.h"
#include "mongo/db/repl/oplog_entry_gen.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/s/shard_server_op_observer.h"
#include "mongo/db/server_options.h"
#include "mongo/db/session_catalog.h"
#include "mongo/db/transaction_participant.h"
#include "mongo/db/views/durable_view_catalog.h"
#include "mongo/scripting/engine.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/fail_point_service.h"
namespace mongo {
using repl::OplogEntry;
namespace {
MONGO_FAIL_POINT_DEFINE(failCollectionUpdates);
const auto getDeleteState = OperationContext::declareDecoration<ShardObserverDeleteState>();
repl::OpTime logOperation(OperationContext* opCtx,
const char* opstr,
const NamespaceString& ns,
OptionalCollectionUUID uuid,
const BSONObj& obj,
const BSONObj* o2,
bool fromMigrate,
Date_t wallClockTime,
const OperationSessionInfo& sessionInfo,
StmtId stmtId,
const repl::OplogLink& oplogLink,
bool prepare,
const OplogSlot& oplogSlot) {
auto& times = OpObserver::Times::get(opCtx).reservedOpTimes;
auto opTime = repl::logOp(opCtx,
opstr,
ns,
uuid,
obj,
o2,
fromMigrate,
wallClockTime,
sessionInfo,
stmtId,
oplogLink,
prepare,
oplogSlot);
times.push_back(opTime);
return opTime;
}
/**
* Updates the session state with the last write timestamp and transaction for that session.
*
* In the case of writes with transaction/statement id, this method will be recursively entered a
* second time for the actual write to the transactions table. Since this write does not generate an
* oplog entry, the recursion will stop at this point.
*/
void onWriteOpCompleted(OperationContext* opCtx,
const NamespaceString& nss,
Session* session,
std::vector<StmtId> stmtIdsWritten,
const repl::OpTime& lastStmtIdWriteOpTime,
Date_t lastStmtIdWriteDate,
boost::optional<DurableTxnStateEnum> txnState) {
if (lastStmtIdWriteOpTime.isNull())
return;
if (session) {
session->onWriteOpCompletedOnPrimary(opCtx,
*opCtx->getTxnNumber(),
std::move(stmtIdsWritten),
lastStmtIdWriteOpTime,
lastStmtIdWriteDate,
txnState);
}
}
/**
* Given a raw collMod command object and associated collection metadata, create and return the
* object for the 'o' field of a collMod oplog entry. For TTL index updates, we make sure the oplog
* entry always stores the index name, instead of a key pattern.
*/
BSONObj makeCollModCmdObj(const BSONObj& collModCmd,
const CollectionOptions& oldCollOptions,
boost::optional<TTLCollModInfo> ttlInfo) {
BSONObjBuilder cmdObjBuilder;
std::string ttlIndexFieldName = "index";
// Add all fields from the original collMod command.
for (auto elem : collModCmd) {
// We normalize all TTL collMod oplog entry objects to use the index name, even if the
// command used an index key pattern.
if (elem.fieldNameStringData() == ttlIndexFieldName && ttlInfo) {
BSONObjBuilder ttlIndexObjBuilder;
ttlIndexObjBuilder.append("name", ttlInfo->indexName);
ttlIndexObjBuilder.append("expireAfterSeconds",
durationCount<Seconds>(ttlInfo->expireAfterSeconds));
cmdObjBuilder.append(ttlIndexFieldName, ttlIndexObjBuilder.obj());
} else {
cmdObjBuilder.append(elem);
}
}
return cmdObjBuilder.obj();
}
Date_t getWallClockTimeForOpLog(OperationContext* opCtx) {
auto const clockSource = opCtx->getServiceContext()->getFastClockSource();
return clockSource->now();
}
struct OpTimeBundle {
repl::OpTime writeOpTime;
repl::OpTime prePostImageOpTime;
Date_t wallClockTime;
};
/**
* Write oplog entry(ies) for the update operation.
*/
OpTimeBundle replLogUpdate(OperationContext* opCtx,
Session* session,
const OplogUpdateEntryArgs& args) {
BSONObj storeObj;
if (args.storeDocOption == OplogUpdateEntryArgs::StoreDocOption::PreImage) {
invariant(args.preImageDoc);
storeObj = *args.preImageDoc;
} else if (args.storeDocOption == OplogUpdateEntryArgs::StoreDocOption::PostImage) {
storeObj = args.updatedDoc;
}
OperationSessionInfo sessionInfo;
repl::OplogLink oplogLink;
if (session) {
sessionInfo.setSessionId(*opCtx->getLogicalSessionId());
sessionInfo.setTxnNumber(*opCtx->getTxnNumber());
oplogLink.prevOpTime = session->getLastWriteOpTime(*opCtx->getTxnNumber());
}
OpTimeBundle opTimes;
opTimes.wallClockTime = getWallClockTimeForOpLog(opCtx);
if (!storeObj.isEmpty() && opCtx->getTxnNumber()) {
auto noteUpdateOpTime = logOperation(opCtx,
"n",
args.nss,
args.uuid,
storeObj,
nullptr,
false,
opTimes.wallClockTime,
sessionInfo,
args.stmtId,
{},
false /* prepare */,
OplogSlot());
opTimes.prePostImageOpTime = noteUpdateOpTime;
if (args.storeDocOption == OplogUpdateEntryArgs::StoreDocOption::PreImage) {
oplogLink.preImageOpTime = noteUpdateOpTime;
} else if (args.storeDocOption == OplogUpdateEntryArgs::StoreDocOption::PostImage) {
oplogLink.postImageOpTime = noteUpdateOpTime;
}
}
opTimes.writeOpTime = logOperation(opCtx,
"u",
args.nss,
args.uuid,
args.update,
&args.criteria,
args.fromMigrate,
opTimes.wallClockTime,
sessionInfo,
args.stmtId,
oplogLink,
false /* prepare */,
OplogSlot());
return opTimes;
}
/**
* Write oplog entry(ies) for the delete operation.
*/
OpTimeBundle replLogDelete(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
Session* session,
StmtId stmtId,
bool fromMigrate,
const boost::optional<BSONObj>& deletedDoc) {
OperationSessionInfo sessionInfo;
repl::OplogLink oplogLink;
if (session) {
sessionInfo.setSessionId(*opCtx->getLogicalSessionId());
sessionInfo.setTxnNumber(*opCtx->getTxnNumber());
oplogLink.prevOpTime = session->getLastWriteOpTime(*opCtx->getTxnNumber());
}
OpTimeBundle opTimes;
opTimes.wallClockTime = getWallClockTimeForOpLog(opCtx);
if (deletedDoc && opCtx->getTxnNumber()) {
auto noteOplog = logOperation(opCtx,
"n",
nss,
uuid,
deletedDoc.get(),
nullptr,
false,
opTimes.wallClockTime,
sessionInfo,
stmtId,
{},
false /* prepare */,
OplogSlot());
opTimes.prePostImageOpTime = noteOplog;
oplogLink.preImageOpTime = noteOplog;
}
auto& deleteState = getDeleteState(opCtx);
opTimes.writeOpTime = logOperation(opCtx,
"d",
nss,
uuid,
deleteState.documentKey,
nullptr,
fromMigrate,
opTimes.wallClockTime,
sessionInfo,
stmtId,
oplogLink,
false /* prepare */,
OplogSlot());
return opTimes;
}
/**
* Write oplog entry for applyOps/atomic transaction operations.
*/
OpTimeBundle replLogApplyOps(OperationContext* opCtx,
const NamespaceString& cmdNss,
const BSONObj& applyOpCmd,
const OperationSessionInfo& sessionInfo,
StmtId stmtId,
const repl::OplogLink& oplogLink,
bool prepare,
const OplogSlot& oplogSlot) {
OpTimeBundle times;
times.wallClockTime = getWallClockTimeForOpLog(opCtx);
times.writeOpTime = logOperation(opCtx,
"c",
cmdNss,
{},
applyOpCmd,
nullptr,
false,
times.wallClockTime,
sessionInfo,
stmtId,
oplogLink,
prepare,
oplogSlot);
return times;
}
} // namespace
void OpObserverImpl::onCreateIndex(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
BSONObj indexDoc,
bool fromMigrate) {
const NamespaceString systemIndexes{nss.getSystemIndexesCollection()};
if (uuid) {
BSONObjBuilder builder;
builder.append("createIndexes", nss.coll());
for (const auto& e : indexDoc) {
if (e.fieldNameStringData() != "ns"_sd)
builder.append(e);
}
logOperation(opCtx,
"c",
nss.getCommandNS(),
uuid,
builder.done(),
nullptr,
fromMigrate,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
} else {
logOperation(opCtx,
"i",
systemIndexes,
{},
indexDoc,
nullptr,
fromMigrate,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "i", systemIndexes, indexDoc, nullptr);
}
void OpObserverImpl::onInserts(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
std::vector<InsertStatement>::const_iterator first,
std::vector<InsertStatement>::const_iterator last,
bool fromMigrate) {
auto txnParticipant = TransactionParticipant::get(opCtx);
const bool inMultiDocumentTransaction = txnParticipant && opCtx->writesAreReplicated() &&
txnParticipant->inMultiDocumentTransaction();
Date_t lastWriteDate;
std::vector<repl::OpTime> opTimeList;
repl::OpTime lastOpTime;
if (inMultiDocumentTransaction) {
// Do not add writes to the profile collection to the list of transaction operations, since
// these are done outside the transaction.
if (!opCtx->getWriteUnitOfWork()) {
invariant(nss.isSystemDotProfile());
return;
}
for (auto iter = first; iter != last; iter++) {
auto operation = OplogEntry::makeInsertOperation(nss, uuid, iter->doc);
txnParticipant->addTransactionOperation(opCtx, operation);
}
} else {
lastWriteDate = getWallClockTimeForOpLog(opCtx);
Session* const session = OperationContextSession::get(opCtx);
opTimeList =
repl::logInsertOps(opCtx, nss, uuid, session, first, last, fromMigrate, lastWriteDate);
if (!opTimeList.empty())
lastOpTime = opTimeList.back();
auto& times = OpObserver::Times::get(opCtx).reservedOpTimes;
using std::begin;
using std::end;
times.insert(end(times), begin(opTimeList), end(opTimeList));
std::vector<StmtId> stmtIdsWritten;
std::transform(first,
last,
std::back_inserter(stmtIdsWritten),
[](const InsertStatement& stmt) { return stmt.stmtId; });
onWriteOpCompleted(
opCtx, nss, session, stmtIdsWritten, lastOpTime, lastWriteDate, boost::none);
}
auto* const css = (nss == NamespaceString::kSessionTransactionsTableNamespace || fromMigrate)
? nullptr
: CollectionShardingRuntime::get(opCtx, nss);
size_t index = 0;
for (auto it = first; it != last; it++, index++) {
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "i", nss, it->doc, nullptr);
if (css) {
auto opTime = opTimeList.empty() ? repl::OpTime() : opTimeList[index];
shardObserveInsertOp(opCtx, css, it->doc, opTime);
}
}
if (nss.coll() == "system.js") {
Scope::storedFuncMod(opCtx);
} else if (nss.coll() == DurableViewCatalog::viewsCollectionName()) {
DurableViewCatalog::onExternalChange(opCtx, nss);
} else if (nss == NamespaceString::kServerConfigurationNamespace) {
// We must check server configuration collection writes for featureCompatibilityVersion
// document changes.
for (auto it = first; it != last; it++) {
FeatureCompatibilityVersion::onInsertOrUpdate(opCtx, it->doc);
}
} else if (nss == NamespaceString::kSessionTransactionsTableNamespace && !lastOpTime.isNull()) {
for (auto it = first; it != last; it++) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, it->doc);
}
}
}
void OpObserverImpl::onUpdate(OperationContext* opCtx, const OplogUpdateEntryArgs& args) {
MONGO_FAIL_POINT_BLOCK(failCollectionUpdates, extraData) {
auto collElem = extraData.getData()["collectionNS"];
// If the failpoint specifies no collection or matches the existing one, fail.
if (!collElem || args.nss.ns() == collElem.String()) {
uasserted(40654,
str::stream() << "failCollectionUpdates failpoint enabled, namespace: "
<< args.nss.ns()
<< ", update: "
<< args.update
<< " on document with "
<< args.criteria);
}
}
// Do not log a no-op operation; see SERVER-21738
if (args.update.isEmpty()) {
return;
}
auto txnParticipant = TransactionParticipant::get(opCtx);
const bool inMultiDocumentTransaction = txnParticipant && opCtx->writesAreReplicated() &&
txnParticipant->inMultiDocumentTransaction();
OpTimeBundle opTime;
if (inMultiDocumentTransaction) {
auto operation =
OplogEntry::makeUpdateOperation(args.nss, args.uuid, args.update, args.criteria);
txnParticipant->addTransactionOperation(opCtx, operation);
} else {
Session* const session = OperationContextSession::get(opCtx);
opTime = replLogUpdate(opCtx, session, args);
onWriteOpCompleted(opCtx,
args.nss,
session,
std::vector<StmtId>{args.stmtId},
opTime.writeOpTime,
opTime.wallClockTime,
boost::none);
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "u", args.nss, args.update, &args.criteria);
if (args.nss != NamespaceString::kSessionTransactionsTableNamespace) {
if (!args.fromMigrate) {
auto* const css = CollectionShardingRuntime::get(opCtx, args.nss);
shardObserveUpdateOp(
opCtx, css, args.updatedDoc, opTime.writeOpTime, opTime.prePostImageOpTime);
}
}
if (args.nss.coll() == "system.js") {
Scope::storedFuncMod(opCtx);
} else if (args.nss.coll() == DurableViewCatalog::viewsCollectionName()) {
DurableViewCatalog::onExternalChange(opCtx, args.nss);
} else if (args.nss == NamespaceString::kServerConfigurationNamespace) {
// We must check server configuration collection writes for featureCompatibilityVersion
// document changes.
FeatureCompatibilityVersion::onInsertOrUpdate(opCtx, args.updatedDoc);
} else if (args.nss == NamespaceString::kSessionTransactionsTableNamespace &&
!opTime.writeOpTime.isNull()) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, args.updatedDoc);
}
}
void OpObserverImpl::aboutToDelete(OperationContext* opCtx,
NamespaceString const& nss,
BSONObj const& doc) {
getDeleteState(opCtx) =
ShardObserverDeleteState::make(opCtx, CollectionShardingRuntime::get(opCtx, nss), doc);
}
void OpObserverImpl::onDelete(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
StmtId stmtId,
bool fromMigrate,
const boost::optional<BSONObj>& deletedDoc) {
auto& deleteState = getDeleteState(opCtx);
invariant(!deleteState.documentKey.isEmpty());
auto txnParticipant = TransactionParticipant::get(opCtx);
const bool inMultiDocumentTransaction = txnParticipant && opCtx->writesAreReplicated() &&
txnParticipant->inMultiDocumentTransaction();
OpTimeBundle opTime;
if (inMultiDocumentTransaction) {
auto operation = OplogEntry::makeDeleteOperation(
nss, uuid, deletedDoc ? deletedDoc.get() : deleteState.documentKey);
txnParticipant->addTransactionOperation(opCtx, operation);
} else {
Session* const session = OperationContextSession::get(opCtx);
opTime = replLogDelete(opCtx, nss, uuid, session, stmtId, fromMigrate, deletedDoc);
onWriteOpCompleted(opCtx,
nss,
session,
std::vector<StmtId>{stmtId},
opTime.writeOpTime,
opTime.wallClockTime,
boost::none);
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "d", nss, deleteState.documentKey, nullptr);
if (nss != NamespaceString::kSessionTransactionsTableNamespace) {
if (!fromMigrate) {
auto* const css = CollectionShardingRuntime::get(opCtx, nss);
shardObserveDeleteOp(
opCtx, css, deleteState, opTime.writeOpTime, opTime.prePostImageOpTime);
}
}
if (nss.coll() == "system.js") {
Scope::storedFuncMod(opCtx);
} else if (nss.coll() == DurableViewCatalog::viewsCollectionName()) {
DurableViewCatalog::onExternalChange(opCtx, nss);
} else if (nss.isServerConfigurationCollection()) {
auto _id = deleteState.documentKey["_id"];
if (_id.type() == BSONType::String &&
_id.String() == FeatureCompatibilityVersionParser::kParameterName)
uasserted(40670, "removing FeatureCompatibilityVersion document is not allowed");
} else if (nss == NamespaceString::kSessionTransactionsTableNamespace &&
!opTime.writeOpTime.isNull()) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, deleteState.documentKey);
}
}
void OpObserverImpl::onInternalOpMessage(OperationContext* opCtx,
const NamespaceString& nss,
const boost::optional<UUID> uuid,
const BSONObj& msgObj,
const boost::optional<BSONObj> o2MsgObj) {
const BSONObj* o2MsgPtr = o2MsgObj ? o2MsgObj.get_ptr() : nullptr;
logOperation(opCtx,
"n",
nss,
uuid,
msgObj,
o2MsgPtr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
}
void OpObserverImpl::onCreateCollection(OperationContext* opCtx,
Collection* coll,
const NamespaceString& collectionName,
const CollectionOptions& options,
const BSONObj& idIndex,
const OplogSlot& createOpTime) {
const auto cmdNss = collectionName.getCommandNS();
BSONObjBuilder b;
b.append("create", collectionName.coll().toString());
{
// Don't store the UUID as part of the options, but instead only at the top level
CollectionOptions optionsToStore = options;
optionsToStore.uuid.reset();
b.appendElements(optionsToStore.toBSON());
}
// Include the full _id index spec in the oplog for index versions >= 2.
if (!idIndex.isEmpty()) {
auto versionElem = idIndex[IndexDescriptor::kIndexVersionFieldName];
invariant(versionElem.isNumber());
if (IndexDescriptor::IndexVersion::kV2 <=
static_cast<IndexDescriptor::IndexVersion>(versionElem.numberInt())) {
b.append("idIndex", idIndex);
}
}
const auto cmdObj = b.done();
if (!collectionName.isSystemDotProfile()) {
// do not replicate system.profile modifications
logOperation(opCtx,
"c",
cmdNss,
options.uuid,
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
createOpTime);
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
if (options.uuid) {
opCtx->recoveryUnit()->onRollback([opCtx, collectionName]() {
NamespaceUUIDCache::get(opCtx).evictNamespace(collectionName);
});
}
}
void OpObserverImpl::onCollMod(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
const BSONObj& collModCmd,
const CollectionOptions& oldCollOptions,
boost::optional<TTLCollModInfo> ttlInfo) {
const auto cmdNss = nss.getCommandNS();
// Create the 'o' field object.
const auto cmdObj = makeCollModCmdObj(collModCmd, oldCollOptions, ttlInfo);
// Create the 'o2' field object. We save the old collection metadata and TTL expiration.
BSONObjBuilder o2Builder;
o2Builder.append("collectionOptions_old", oldCollOptions.toBSON());
if (ttlInfo) {
auto oldExpireAfterSeconds = durationCount<Seconds>(ttlInfo->oldExpireAfterSeconds);
o2Builder.append("expireAfterSeconds_old", oldExpireAfterSeconds);
}
const auto o2Obj = o2Builder.done();
if (!nss.isSystemDotProfile()) {
// do not replicate system.profile modifications
logOperation(opCtx,
"c",
cmdNss,
uuid,
cmdObj,
&o2Obj,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
// Make sure the UUID values in the Collection metadata, the Collection object, and the UUID
// catalog are all present and equal.
invariant(opCtx->lockState()->isDbLockedForMode(nss.db(), MODE_X));
Database* db = DatabaseHolder::getDatabaseHolder().get(opCtx, nss.db());
// Some unit tests call the op observer on an unregistered Database.
if (!db) {
return;
}
Collection* coll = db->getCollection(opCtx, nss.ns());
invariant(coll->uuid());
invariant(coll->uuid() == uuid);
CollectionCatalogEntry* entry = coll->getCatalogEntry();
invariant(entry->isEqualToMetadataUUID(opCtx, uuid));
}
void OpObserverImpl::onDropDatabase(OperationContext* opCtx, const std::string& dbName) {
const NamespaceString cmdNss{dbName, "$cmd"};
const auto cmdObj = BSON("dropDatabase" << 1);
logOperation(opCtx,
"c",
cmdNss,
{},
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
uassert(
50714, "dropping the admin database is not allowed.", dbName != NamespaceString::kAdminDb);
if (dbName == NamespaceString::kSessionTransactionsTableNamespace.db()) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, boost::none);
}
NamespaceUUIDCache::get(opCtx).evictNamespacesInDatabase(dbName);
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
}
repl::OpTime OpObserverImpl::onDropCollection(OperationContext* opCtx,
const NamespaceString& collectionName,
OptionalCollectionUUID uuid) {
const auto cmdNss = collectionName.getCommandNS();
const auto cmdObj = BSON("drop" << collectionName.coll());
if (!collectionName.isSystemDotProfile()) {
// Do not replicate system.profile modifications.
logOperation(opCtx,
"c",
cmdNss,
uuid,
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
}
uassert(50715,
"dropping the server configuration collection (admin.system.version) is not allowed.",
collectionName != NamespaceString::kServerConfigurationNamespace);
if (collectionName.coll() == DurableViewCatalog::viewsCollectionName()) {
DurableViewCatalog::onExternalChange(opCtx, collectionName);
} else if (collectionName == NamespaceString::kSessionTransactionsTableNamespace) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, boost::none);
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
// Evict namespace entry from the namespace/uuid cache if it exists.
NamespaceUUIDCache::get(opCtx).evictNamespace(collectionName);
return {};
}
void OpObserverImpl::onDropIndex(OperationContext* opCtx,
const NamespaceString& nss,
OptionalCollectionUUID uuid,
const std::string& indexName,
const BSONObj& indexInfo) {
const auto cmdNss = nss.getCommandNS();
const auto cmdObj = BSON("dropIndexes" << nss.coll() << "index" << indexName);
logOperation(opCtx,
"c",
cmdNss,
uuid,
cmdObj,
&indexInfo,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, &indexInfo);
}
repl::OpTime OpObserverImpl::preRenameCollection(OperationContext* const opCtx,
const NamespaceString& fromCollection,
const NamespaceString& toCollection,
OptionalCollectionUUID uuid,
OptionalCollectionUUID dropTargetUUID,
bool stayTemp) {
const auto cmdNss = fromCollection.getCommandNS();
BSONObjBuilder builder;
builder.append("renameCollection", fromCollection.ns());
builder.append("to", toCollection.ns());
builder.append("stayTemp", stayTemp);
if (dropTargetUUID) {
dropTargetUUID->appendToBuilder(&builder, "dropTarget");
}
const auto cmdObj = builder.done();
logOperation(opCtx,
"c",
cmdNss,
uuid,
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
return {};
}
void OpObserverImpl::postRenameCollection(OperationContext* const opCtx,
const NamespaceString& fromCollection,
const NamespaceString& toCollection,
OptionalCollectionUUID uuid,
OptionalCollectionUUID dropTargetUUID,
bool stayTemp) {
const auto cmdNss = fromCollection.getCommandNS();
BSONObjBuilder builder;
builder.append("renameCollection", fromCollection.ns());
builder.append("to", toCollection.ns());
builder.append("stayTemp", stayTemp);
if (dropTargetUUID) {
dropTargetUUID->appendToBuilder(&builder, "dropTarget");
}
const auto cmdObj = builder.done();
if (fromCollection.isSystemDotViews())
DurableViewCatalog::onExternalChange(opCtx, fromCollection);
if (toCollection.isSystemDotViews())
DurableViewCatalog::onExternalChange(opCtx, toCollection);
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
// Evict namespace entry from the namespace/uuid cache if it exists.
NamespaceUUIDCache& cache = NamespaceUUIDCache::get(opCtx);
cache.evictNamespace(fromCollection);
cache.evictNamespace(toCollection);
opCtx->recoveryUnit()->onRollback(
[&cache, toCollection]() { cache.evictNamespace(toCollection); });
}
void OpObserverImpl::onRenameCollection(OperationContext* const opCtx,
const NamespaceString& fromCollection,
const NamespaceString& toCollection,
OptionalCollectionUUID uuid,
OptionalCollectionUUID dropTargetUUID,
bool stayTemp) {
preRenameCollection(opCtx, fromCollection, toCollection, uuid, dropTargetUUID, stayTemp);
postRenameCollection(opCtx, fromCollection, toCollection, uuid, dropTargetUUID, stayTemp);
}
void OpObserverImpl::onApplyOps(OperationContext* opCtx,
const std::string& dbName,
const BSONObj& applyOpCmd) {
const NamespaceString cmdNss{dbName, "$cmd"};
// Only transactional 'applyOps' commands can be prepared.
constexpr bool prepare = false;
replLogApplyOps(opCtx, cmdNss, applyOpCmd, {}, kUninitializedStmtId, {}, prepare, OplogSlot());
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, applyOpCmd, nullptr);
}
void OpObserverImpl::onEmptyCapped(OperationContext* opCtx,
const NamespaceString& collectionName,
OptionalCollectionUUID uuid) {
const auto cmdNss = collectionName.getCommandNS();
const auto cmdObj = BSON("emptycapped" << collectionName.coll());
if (!collectionName.isSystemDotProfile()) {
// Do not replicate system.profile modifications
logOperation(opCtx,
"c",
cmdNss,
uuid,
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
}
AuthorizationManager::get(opCtx->getServiceContext())
->logOp(opCtx, "c", cmdNss, cmdObj, nullptr);
}
namespace {
OpTimeBundle logApplyOpsForTransaction(OperationContext* opCtx,
Session* const session,
std::vector<repl::ReplOperation> stmts,
const OplogSlot& prepareOplogSlot) {
BSONObjBuilder applyOpsBuilder;
BSONArrayBuilder opsArray(applyOpsBuilder.subarrayStart("applyOps"_sd));
for (auto& stmt : stmts) {
opsArray.append(stmt.toBSON());
}
opsArray.done();
const NamespaceString cmdNss{"admin", "$cmd"};
OperationSessionInfo sessionInfo;
repl::OplogLink oplogLink;
sessionInfo.setSessionId(*opCtx->getLogicalSessionId());
sessionInfo.setTxnNumber(*opCtx->getTxnNumber());
StmtId stmtId(0);
oplogLink.prevOpTime = session->getLastWriteOpTime(*opCtx->getTxnNumber());
// Until we support multiple oplog entries per transaction, prevOpTime should always be null.
invariant(oplogLink.prevOpTime.isNull());
try {
// We are only given an oplog slot for prepared transactions.
auto prepare = !prepareOplogSlot.opTime.isNull();
auto applyOpCmd = applyOpsBuilder.done();
auto times = replLogApplyOps(
opCtx, cmdNss, applyOpCmd, sessionInfo, stmtId, oplogLink, prepare, prepareOplogSlot);
auto txnState = prepare ? DurableTxnStateEnum::kPrepared : DurableTxnStateEnum::kCommitted;
onWriteOpCompleted(
opCtx, cmdNss, session, {stmtId}, times.writeOpTime, times.wallClockTime, txnState);
return times;
} catch (const AssertionException& e) {
// Change the error code to TransactionTooLarge if it is BSONObjectTooLarge.
uassert(ErrorCodes::TransactionTooLarge,
e.reason(),
e.code() != ErrorCodes::BSONObjectTooLarge);
throw;
}
MONGO_UNREACHABLE;
}
} // namespace
void OpObserverImpl::onTransactionCommit(OperationContext* opCtx, bool wasPrepared) {
invariant(opCtx->getTxnNumber());
if (wasPrepared) {
// TODO (SERVER-35865): log commitTransaction oplog entry correctly. We log a dummy entry to
// test the timestamping behavior.
const NamespaceString cmdNss{"admin", "$cmd"};
const auto cmdObj = BSON("commitTransaction" << 1);
TransactionParticipant::SideTransactionBlock sideTxn(opCtx);
WriteUnitOfWork wuow(opCtx);
logOperation(opCtx,
"c",
cmdNss,
boost::none,
cmdObj,
nullptr,
false,
getWallClockTimeForOpLog(opCtx),
{},
kUninitializedStmtId,
{},
false /* prepare */,
OplogSlot());
wuow.commit();
} else {
auto txnParticipant = TransactionParticipant::get(opCtx);
invariant(txnParticipant);
const auto stmts = txnParticipant->endTransactionAndRetrieveOperations(opCtx);
// It is possible that the transaction resulted in no changes. In that case, we should
// not write an empty applyOps entry.
if (stmts.empty())
return;
Session* const session = OperationContextSession::get(opCtx);
invariant(session);
const auto commitOpTime =
logApplyOpsForTransaction(opCtx, session, stmts, OplogSlot()).writeOpTime;
invariant(!commitOpTime.isNull());
}
}
void OpObserverImpl::onTransactionPrepare(OperationContext* opCtx, const OplogSlot& prepareOpTime) {
invariant(opCtx->getTxnNumber());
auto txnParticipant = TransactionParticipant::get(opCtx);
invariant(txnParticipant);
invariant(txnParticipant->inMultiDocumentTransaction());
invariant(!prepareOpTime.opTime.isNull());
auto stmts = txnParticipant->endTransactionAndRetrieveOperations(opCtx);
// We write the oplog entry in a side transaction so that we do not commit the now-prepared
// transaction.
// We write an empty 'applyOps' entry if there were no writes to choose a prepare timestamp
// and allow this transaction to be continued on failover.
{
TransactionParticipant::SideTransactionBlock sideTxn(opCtx);
WriteUnitOfWork wuow(opCtx);
Session* const session = OperationContextSession::get(opCtx);
invariant(session);
logApplyOpsForTransaction(opCtx, session, stmts, prepareOpTime);
wuow.commit();
}
}
void OpObserverImpl::onTransactionAbort(OperationContext* opCtx) {
invariant(opCtx->getTxnNumber());
}
void OpObserverImpl::onReplicationRollback(OperationContext* opCtx,
const RollbackObserverInfo& rbInfo) {
// Invalidate any in-memory auth data if necessary.
const auto& rollbackNamespaces = rbInfo.rollbackNamespaces;
if (rollbackNamespaces.count(AuthorizationManager::versionCollectionNamespace) == 1 ||
rollbackNamespaces.count(AuthorizationManager::usersCollectionNamespace) == 1 ||
rollbackNamespaces.count(AuthorizationManager::rolesCollectionNamespace) == 1) {
AuthorizationManager::get(opCtx->getServiceContext())->invalidateUserCache();
}
// If there were ops rolled back that were part of operations on a session, then invalidate the
// session cache.
if (rbInfo.rollbackSessionIds.size() > 0) {
SessionCatalog::get(opCtx)->invalidateSessions(opCtx, boost::none);
}
// Reset the key manager cache.
auto validator = LogicalTimeValidator::get(opCtx);
if (validator) {
validator->resetKeyManagerCache();
}
// Check if the shard identity document rolled back.
if (rbInfo.shardIdentityRolledBack) {
fassertFailedNoTrace(50712);
}
}
} // namespace mongo
|