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
|
/**
* Copyright (C) 2018-present MongoDB, Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the Server Side Public License, version 1,
* as published by MongoDB, Inc.
*
* 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
* Server Side Public License for more details.
*
* You should have received a copy of the Server Side Public License
* along with this program. If not, see
* <http://www.mongodb.com/licensing/server-side-public-license>.
*
* 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 Server Side 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/db/repl/oplog_applier_impl_test_fixture.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/document_validation.h"
#include "mongo/db/concurrency/exception_util.h"
#include "mongo/db/curop.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/dbdirectclient.h"
#include "mongo/db/index_builds_coordinator.h"
#include "mongo/db/op_observer/op_observer_registry.h"
#include "mongo/db/query/internal_plans.h"
#include "mongo/db/repl/drop_pending_collection_reaper.h"
#include "mongo/db/repl/oplog_applier.h"
#include "mongo/db/repl/replication_consistency_markers_mock.h"
#include "mongo/db/repl/replication_coordinator_mock.h"
#include "mongo/db/repl/replication_process.h"
#include "mongo/db/repl/replication_recovery_mock.h"
#include "mongo/db/repl/storage_interface.h"
#include "mongo/db/repl/storage_interface_impl.h"
#include "mongo/db/session/session_catalog_mongod.h"
#include "mongo/db/transaction/session_catalog_mongod_transaction_interface_impl.h"
#include "mongo/db/vector_clock_mutable.h"
namespace mongo {
namespace repl {
void OplogApplierImplOpObserver::onInserts(OperationContext* opCtx,
const CollectionPtr& coll,
std::vector<InsertStatement>::const_iterator begin,
std::vector<InsertStatement>::const_iterator end,
bool fromMigrate) {
if (!onInsertsFn) {
return;
}
std::vector<BSONObj> docs;
for (auto it = begin; it != end; ++it) {
const InsertStatement& insertStatement = *it;
docs.push_back(insertStatement.doc.getOwned());
}
onInsertsFn(opCtx, coll->ns(), docs);
}
void OplogApplierImplOpObserver::onDelete(OperationContext* opCtx,
const CollectionPtr& coll,
StmtId stmtId,
const OplogDeleteEntryArgs& args) {
if (!onDeleteFn) {
return;
}
onDeleteFn(opCtx, coll, stmtId, args);
}
void OplogApplierImplOpObserver::onUpdate(OperationContext* opCtx,
const OplogUpdateEntryArgs& args) {
if (!onUpdateFn) {
return;
}
onUpdateFn(opCtx, args);
}
void OplogApplierImplOpObserver::onCreateCollection(OperationContext* opCtx,
const CollectionPtr& coll,
const NamespaceString& collectionName,
const CollectionOptions& options,
const BSONObj& idIndex,
const OplogSlot& createOpTime,
bool fromMigrate) {
if (!onCreateCollectionFn) {
return;
}
onCreateCollectionFn(opCtx, coll, collectionName, options, idIndex);
}
void OplogApplierImplOpObserver::onRenameCollection(OperationContext* opCtx,
const NamespaceString& fromCollection,
const NamespaceString& toCollection,
const UUID& uuid,
const boost::optional<UUID>& dropTargetUUID,
std::uint64_t numRecords,
bool stayTemp) {
if (!onRenameCollectionFn) {
return;
}
onRenameCollectionFn(
opCtx, fromCollection, toCollection, uuid, dropTargetUUID, numRecords, stayTemp);
}
void OplogApplierImplOpObserver::onCreateIndex(OperationContext* opCtx,
const NamespaceString& nss,
const UUID& uuid,
BSONObj indexDoc,
bool fromMigrate) {
if (!onCreateIndexFn) {
return;
}
onCreateIndexFn(opCtx, nss, uuid, indexDoc, fromMigrate);
}
void OplogApplierImplOpObserver::onDropIndex(OperationContext* opCtx,
const NamespaceString& nss,
const UUID& uuid,
const std::string& indexName,
const BSONObj& idxDescriptor) {
if (!onDropIndexFn) {
return;
}
onDropIndexFn(opCtx, nss, uuid, indexName, idxDescriptor);
}
void OplogApplierImplOpObserver::onCollMod(OperationContext* opCtx,
const NamespaceString& nss,
const UUID& uuid,
const BSONObj& collModCmd,
const CollectionOptions& oldCollOptions,
boost::optional<IndexCollModInfo> indexInfo) {
if (!onCollModFn) {
return;
}
onCollModFn(opCtx, nss, uuid, collModCmd, oldCollOptions, indexInfo);
}
std::unique_ptr<OpObserver::ApplyOpsOplogSlotAndOperationAssignment>
OplogApplierImplOpObserver::preTransactionPrepare(
OperationContext* opCtx,
const std::vector<OplogSlot>& reservedSlots,
const TransactionOperations& transactionOperations,
Date_t wallClockTime) {
return std::make_unique<ApplyOpsOplogSlotAndOperationAssignment>(/*prepare=*/true);
}
void OplogApplierImplTest::setUp() {
ServiceContextMongoDTest::setUp();
serviceContext = getServiceContext();
_opCtx = cc().makeOperationContext();
ReplicationCoordinator::set(serviceContext,
std::make_unique<ReplicationCoordinatorMock>(serviceContext));
ASSERT_OK(ReplicationCoordinator::get(_opCtx.get())->setFollowerMode(MemberState::RS_PRIMARY));
StorageInterface::set(serviceContext, std::make_unique<StorageInterfaceImpl>());
MongoDSessionCatalog::set(
serviceContext,
std::make_unique<MongoDSessionCatalog>(
std::make_unique<MongoDSessionCatalogTransactionInterfaceImpl>()));
DropPendingCollectionReaper::set(
serviceContext, std::make_unique<DropPendingCollectionReaper>(getStorageInterface()));
repl::createOplog(_opCtx.get());
_consistencyMarkers = std::make_unique<ReplicationConsistencyMarkersMock>();
// Set up an OpObserver to track the documents OplogApplierImpl inserts.
auto opObserver = std::make_unique<OplogApplierImplOpObserver>();
_opObserver = opObserver.get();
auto opObserverRegistry = dynamic_cast<OpObserverRegistry*>(serviceContext->getOpObserver());
opObserverRegistry->addObserver(std::move(opObserver));
// Initialize the featureCompatibilityVersion server parameter. This is necessary because this
// test fixture does not create a featureCompatibilityVersion document from which to initialize
// the server parameter.
// (Generic FCV reference): This FCV reference should exist across LTS binary versions.
serverGlobalParams.mutableFeatureCompatibility.setVersion(multiversion::GenericFCV::kLatest);
// This is necessary to generate ghost timestamps for index builds that are not 0, since 0 is an
// invalid timestamp.
VectorClockMutable::get(_opCtx.get())->tickClusterTimeTo(LogicalTime(Timestamp(1, 0)));
}
void OplogApplierImplTest::tearDown() {
_opCtx.reset();
_consistencyMarkers = {};
DropPendingCollectionReaper::set(serviceContext, {});
StorageInterface::set(serviceContext, {});
ServiceContextMongoDTest::tearDown();
}
ReplicationConsistencyMarkers* OplogApplierImplTest::getConsistencyMarkers() const {
return _consistencyMarkers.get();
}
StorageInterface* OplogApplierImplTest::getStorageInterface() const {
return StorageInterface::get(serviceContext);
}
// Since applyOplogEntryOrGroupedInserts is being tested outside of its calling function
// (applyOplogBatchPerWorker), we recreate the necessary calling context.
Status OplogApplierImplTest::_applyOplogEntryOrGroupedInsertsWrapper(
OperationContext* opCtx,
const OplogEntryOrGroupedInserts& batch,
OplogApplication::Mode oplogApplicationMode) {
UnreplicatedWritesBlock uwb(opCtx);
DisableDocumentValidation validationDisabler(opCtx);
const bool dataIsConsistent = true;
return applyOplogEntryOrGroupedInserts(opCtx, batch, oplogApplicationMode, dataIsConsistent);
}
void OplogApplierImplTest::_testApplyOplogEntryOrGroupedInsertsCrudOperation(
ErrorCodes::Error expectedError,
const OplogEntry& op,
const NamespaceString& targetNss,
bool expectedApplyOpCalled) {
bool applyOpCalled = false;
auto checkOpCtx = [&targetNss](OperationContext* opCtx) {
ASSERT_TRUE(opCtx);
ASSERT_TRUE(opCtx->lockState()->isDbLockedForMode(targetNss.dbName(), MODE_IX));
ASSERT_FALSE(opCtx->lockState()->isDbLockedForMode(targetNss.dbName(), MODE_X));
ASSERT_TRUE(opCtx->lockState()->isCollectionLockedForMode(targetNss, MODE_IX));
ASSERT_FALSE(opCtx->writesAreReplicated());
ASSERT_TRUE(DocumentValidationSettings::get(opCtx).isSchemaValidationDisabled());
};
_opObserver->onInsertsFn =
[&](OperationContext* opCtx, const NamespaceString& nss, const std::vector<BSONObj>& docs) {
applyOpCalled = true;
checkOpCtx(opCtx);
ASSERT_EQUALS(targetNss, nss);
ASSERT_EQUALS(1U, docs.size());
// For upserts we don't know the intended value of the document.
if (op.getOpType() == repl::OpTypeEnum::kInsert) {
ASSERT_BSONOBJ_EQ(op.getObject(), docs[0]);
}
return Status::OK();
};
_opObserver->onDeleteFn = [&](OperationContext* opCtx,
const CollectionPtr& coll,
StmtId stmtId,
const OplogDeleteEntryArgs& args) {
applyOpCalled = true;
checkOpCtx(opCtx);
ASSERT_EQUALS(targetNss, coll->ns());
ASSERT(args.deletedDoc);
ASSERT_BSONOBJ_EQ(op.getObject(), *(args.deletedDoc));
return Status::OK();
};
_opObserver->onUpdateFn = [&](OperationContext* opCtx, const OplogUpdateEntryArgs& args) {
applyOpCalled = true;
checkOpCtx(opCtx);
ASSERT_EQUALS(targetNss, args.coll->ns());
return Status::OK();
};
ASSERT_EQ(_applyOplogEntryOrGroupedInsertsWrapper(
_opCtx.get(), ApplierOperation{&op}, OplogApplication::Mode::kSecondary),
expectedError);
ASSERT_EQ(applyOpCalled, expectedApplyOpCalled);
}
Status failedApplyCommand(OperationContext* opCtx,
const BSONObj& theOperation,
OplogApplication::Mode) {
FAIL("applyCommand unexpectedly invoked.");
return Status::OK();
}
Status OplogApplierImplTest::runOpSteadyState(const OplogEntry& op) {
return runOpsSteadyState({op});
}
Status OplogApplierImplTest::runOpsSteadyState(std::vector<OplogEntry> ops) {
TestApplyOplogGroupApplier oplogApplier(
getConsistencyMarkers(),
getStorageInterface(),
repl::OplogApplier::Options(repl::OplogApplication::Mode::kSecondary));
std::vector<ApplierOperation> opsPtrs;
for (auto& op : ops) {
opsPtrs.emplace_back(&op);
}
WorkerMultikeyPathInfo pathInfo;
const bool dataIsConsistent = true;
return oplogApplier.applyOplogBatchPerWorker(
_opCtx.get(), &opsPtrs, &pathInfo, dataIsConsistent);
}
Status OplogApplierImplTest::runOpInitialSync(const OplogEntry& op) {
return runOpsInitialSync({op});
}
Status OplogApplierImplTest::runOpsInitialSync(std::vector<OplogEntry> ops) {
NoopOplogApplierObserver observer;
auto storageInterface = getStorageInterface();
auto writerPool = makeReplWriterPool();
OplogApplierImpl oplogApplier(
nullptr, // executor
nullptr, // oplogBuffer
&observer,
ReplicationCoordinator::get(_opCtx.get()),
getConsistencyMarkers(),
storageInterface,
repl::OplogApplier::Options(repl::OplogApplication::Mode::kInitialSync),
writerPool.get());
// Idempotency tests apply the same batch of oplog entries multiple times in a loop, which would
// result in out-of-order oplog inserts. So we truncate the oplog collection first before
// calling applyOplogBatch.
ASSERT_OK(
storageInterface->truncateCollection(_opCtx.get(), NamespaceString::kRsOplogNamespace));
// Apply each operation in a batch of one because 'ops' may contain a mix of commands and CRUD
// operations provided by idempotency tests. Applying operations in a batch of one is also
// necessary to work around oplog visibility issues. For example, idempotency tests may contain
// a prepare and a commit that we don't apply both in the same batch in production oplog
// application because the commit needs to read the prepare entry. So we apply each operation in
// its own batch and update oplog visibility after each batch to make sure all previously
// applied entries are visible to subsequent batches.
for (auto& op : ops) {
auto applyResult = oplogApplier.applyOplogBatch(_opCtx.get(), {op});
if (!applyResult.isOK()) {
std::vector<BSONObj> docsFromOps;
for (const auto& opForContext : ops) {
docsFromOps.push_back(opForContext.getEntry().toBSON());
}
auto status = applyResult.getStatus();
return status.withContext(str::stream()
<< "failed to apply operation: " << op.toBSONForLogging()
<< ". " << BSON("ops" << docsFromOps));
}
auto lastApplied = applyResult.getValue();
const bool orderedCommit = true;
// Update oplog visibility by notifying the storage engine of the new oplog entries.
storageInterface->oplogDiskLocRegister(
_opCtx.get(), lastApplied.getTimestamp(), orderedCommit);
}
return Status::OK();
}
void checkTxnTable(OperationContext* opCtx,
const LogicalSessionId& lsid,
const TxnNumber& txnNum,
const repl::OpTime& expectedOpTime,
Date_t expectedWallClock,
boost::optional<repl::OpTime> expectedStartOpTime,
boost::optional<DurableTxnStateEnum> expectedState) {
DBDirectClient client(opCtx);
auto result = client.findOne(NamespaceString::kSessionTransactionsTableNamespace,
BSON(SessionTxnRecord::kSessionIdFieldName << lsid.toBSON()));
ASSERT_FALSE(result.isEmpty());
auto txnRecord = SessionTxnRecord::parse(IDLParserContext("parse txn record for test"), result);
ASSERT_EQ(txnNum, txnRecord.getTxnNum());
ASSERT_EQ(expectedOpTime, txnRecord.getLastWriteOpTime());
ASSERT_EQ(expectedWallClock, txnRecord.getLastWriteDate());
if (expectedStartOpTime) {
ASSERT(txnRecord.getStartOpTime());
ASSERT_EQ(*expectedStartOpTime, *txnRecord.getStartOpTime());
} else {
ASSERT(!txnRecord.getStartOpTime());
}
if (expectedState) {
ASSERT(*expectedState == txnRecord.getState());
}
}
CollectionReader::CollectionReader(OperationContext* opCtx, const NamespaceString& nss)
: _collToScan(opCtx, nss),
_exec(InternalPlanner::collectionScan(opCtx,
&_collToScan.getCollection(),
PlanYieldPolicy::YieldPolicy::NO_YIELD,
InternalPlanner::FORWARD)) {}
StatusWith<BSONObj> CollectionReader::next() {
BSONObj obj;
auto state = _exec->getNext(&obj, nullptr);
if (state == PlanExecutor::IS_EOF) {
return {ErrorCodes::CollectionIsEmpty,
str::stream() << "no more documents in " << _collToScan.getNss()};
}
// PlanExecutors that do not yield should only return ADVANCED or EOF.
invariant(state == PlanExecutor::ADVANCED);
return obj;
}
bool docExists(OperationContext* opCtx, const NamespaceString& nss, const BSONObj& doc) {
DBDirectClient client(opCtx);
auto result = client.findOne(nss, doc);
return !result.isEmpty();
}
/**
* Creates an OplogEntry with given parameters and preset defaults for this test suite.
*/
OplogEntry makeOplogEntry(OpTypeEnum opType,
NamespaceString nss,
const boost::optional<UUID>& uuid,
BSONObj o,
boost::optional<BSONObj> o2,
boost::optional<bool> fromMigrate) {
return makeOplogEntry({{1, 1}, 1}, opType, std::move(nss), uuid, o, o2, fromMigrate);
}
OplogEntry makeOplogEntry(OpTime opTime,
OpTypeEnum opType,
NamespaceString nss,
const boost::optional<UUID>& uuid,
BSONObj o,
boost::optional<BSONObj> o2,
boost::optional<bool> fromMigrate) {
return {DurableOplogEntry(opTime, // optime
opType, // opType
std::move(nss), // namespace
uuid, // uuid
fromMigrate, // fromMigrate
OplogEntry::kOplogVersion, // version
o, // o
o2, // o2
{}, // sessionInfo
boost::none, // upsert
Date_t(), // wall clock time
{}, // statement ids
boost::none, // optime of previous write within same transaction
boost::none, // pre-image optime
boost::none, // post-image optime
boost::none, // ShardId of resharding recipient
boost::none, // _id
boost::none)}; // needsRetryImage
}
OplogEntry makeOplogEntry(OpTypeEnum opType, NamespaceString nss, boost::optional<UUID> uuid) {
return makeOplogEntry(opType, nss, uuid, BSON("_id" << 0), boost::none);
}
CollectionOptions createOplogCollectionOptions() {
CollectionOptions options;
options.capped = true;
options.cappedSize = 64 * 1024 * 1024LL;
options.autoIndexId = CollectionOptions::NO;
return options;
}
CollectionOptions createRecordChangeStreamPreAndPostImagesCollectionOptions() {
CollectionOptions options;
options.changeStreamPreAndPostImagesOptions.setEnabled(true);
return options;
}
void createCollection(OperationContext* opCtx,
const NamespaceString& nss,
const CollectionOptions& options) {
writeConflictRetry(opCtx, "createCollection", nss.ns(), [&] {
Lock::DBLock dbLk(opCtx, nss.dbName(), MODE_IX);
Lock::CollectionLock collLk(opCtx, nss, MODE_X);
OldClientContext ctx(opCtx, nss);
auto db = ctx.db();
ASSERT_TRUE(db);
mongo::WriteUnitOfWork wuow(opCtx);
auto coll = db->createCollection(opCtx, nss, options);
ASSERT_TRUE(coll);
wuow.commit();
});
}
UUID createCollectionWithUuid(OperationContext* opCtx, const NamespaceString& nss) {
CollectionOptions options;
options.uuid = UUID::gen();
createCollection(opCtx, nss, options);
return options.uuid.value();
}
void createDatabase(OperationContext* opCtx, StringData dbName) {
Lock::GlobalWrite globalLock(opCtx);
bool justCreated;
auto databaseHolder = DatabaseHolder::get(opCtx);
const DatabaseName tenantDbName(boost::none, dbName);
auto db = databaseHolder->openDb(opCtx, tenantDbName, &justCreated);
ASSERT_TRUE(db);
ASSERT_TRUE(justCreated);
}
bool collectionExists(OperationContext* opCtx, const NamespaceString& nss) {
return static_cast<bool>(AutoGetCollectionForRead(opCtx, nss).getCollection());
}
void createIndex(OperationContext* opCtx,
const NamespaceString& nss,
const UUID collUUID,
const BSONObj& spec) {
Lock::DBLock dbLk(opCtx, nss.dbName(), MODE_IX);
Lock::CollectionLock collLk(opCtx, nss, MODE_X);
auto indexBuildsCoord = IndexBuildsCoordinator::get(opCtx);
// This fixture sets up some replication, but notably omits installing an OpObserverImpl. This
// state causes collection creation to timestamp catalog writes, but secondary index creation
// does not. We use an UnreplicatedWritesBlock to avoid timestamping any of the catalog setup.
repl::UnreplicatedWritesBlock noRep(opCtx);
indexBuildsCoord->createIndex(
opCtx, collUUID, spec, IndexBuildsManager::IndexConstraints::kEnforce, false);
}
} // namespace repl
} // namespace mongo
|