summaryrefslogtreecommitdiff
path: root/src/mongo/db/catalog/create_collection.cpp
blob: 461efd0bcb659970af465665f17e5233568e0c10 (plain)
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
/**
 *    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.
 */

#define MONGO_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kCommand

#include "mongo/platform/basic.h"

#include "mongo/db/catalog/create_collection.h"

#include <fmt/printf.h>

#include "mongo/bson/bsonobj.h"
#include "mongo/bson/json.h"
#include "mongo/db/catalog/collection_catalog.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/index_key_validate.h"
#include "mongo/db/commands.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/curop.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/index_builds_coordinator.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/op_observer.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/ops/insert.h"
#include "mongo/db/query/collation/collator_factory_interface.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/timeseries/timeseries_options.h"
#include "mongo/db/views/view_catalog.h"
#include "mongo/idl/command_generic_argument.h"
#include "mongo/logv2/log.h"
#include "mongo/util/fail_point.h"

namespace mongo {
namespace {

MONGO_FAIL_POINT_DEFINE(failTimeseriesViewCreation);

void _createSystemDotViewsIfNecessary(OperationContext* opCtx, const Database* db) {
    // Create 'system.views' in a separate WUOW if it does not exist.
    if (!CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx,
                                                                    db->getSystemViewsName())) {
        WriteUnitOfWork wuow(opCtx);
        invariant(db->createCollection(opCtx, db->getSystemViewsName()));
        wuow.commit();
    }
}

Status _createView(OperationContext* opCtx,
                   const NamespaceString& nss,
                   CollectionOptions&& collectionOptions) {
    // This must be checked before we take locks in order to avoid attempting to take multiple locks
    // on the <db>.system.views namespace: first a IX lock on 'ns' and then a X lock on the database
    // system.views collection.
    uassert(ErrorCodes::InvalidNamespace,
            str::stream() << "Cannot create a view called '" << nss.coll()
                          << "': this is a reserved system namespace",
            !nss.isSystemDotViews());

    return writeConflictRetry(opCtx, "create", nss.ns(), [&] {
        AutoGetDb autoDb(opCtx, nss.db(), MODE_IX);
        Lock::CollectionLock collLock(opCtx, nss, MODE_IX);
        // Operations all lock system.views in the end to prevent deadlock.
        Lock::CollectionLock systemViewsLock(
            opCtx,
            NamespaceString(nss.db(), NamespaceString::kSystemDotViewsCollectionName),
            MODE_X);

        auto db = autoDb.ensureDbExists();

        if (opCtx->writesAreReplicated() &&
            !repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, nss)) {
            return Status(ErrorCodes::NotWritablePrimary,
                          str::stream() << "Not primary while creating collection " << nss);
        }

        _createSystemDotViewsIfNecessary(opCtx, db);

        WriteUnitOfWork wunit(opCtx);

        AutoStatsTracker statsTracker(
            opCtx,
            nss,
            Top::LockType::NotLocked,
            AutoStatsTracker::LogMode::kUpdateTopAndCurOp,
            CollectionCatalog::get(opCtx)->getDatabaseProfileLevel(nss.db()));

        // If the view creation rolls back, ensure that the Top entry created for the view is
        // deleted.
        opCtx->recoveryUnit()->onRollback([nss, serviceContext = opCtx->getServiceContext()]() {
            Top::get(serviceContext).collectionDropped(nss);
        });

        // Even though 'collectionOptions' is passed by rvalue reference, it is not safe to move
        // because 'userCreateNS' may throw a WriteConflictException.
        Status status = db->userCreateNS(opCtx, nss, collectionOptions, /*createIdIndex=*/false);
        if (!status.isOK()) {
            return status;
        }
        wunit.commit();

        return Status::OK();
    });
}

Status _createTimeseries(OperationContext* opCtx,
                         const NamespaceString& ns,
                         const CollectionOptions& optionsArg) {
    // This path should only be taken when a user creates a new time-series collection on the
    // primary. Secondaries replicate individual oplog entries.
    invariant(!ns.isTimeseriesBucketsCollection());
    invariant(opCtx->writesAreReplicated());

    auto bucketsNs = ns.makeTimeseriesBucketsNamespace();

    CollectionOptions options = optionsArg;

    // Users may not pass a 'bucketMaxSpanSeconds' other than the default. Instead they should rely
    // on the default behavior from the 'granularity'.
    auto granularity = options.timeseries->getGranularity();
    auto maxSpanSeconds = timeseries::getMaxSpanSecondsFromGranularity(granularity);
    uassert(5510500,
            fmt::format("Timeseries 'bucketMaxSpanSeconds' is not configurable to a value other "
                        "than the default of {} for the provided granularity",
                        maxSpanSeconds),
            !options.timeseries->getBucketMaxSpanSeconds() ||
                maxSpanSeconds == options.timeseries->getBucketMaxSpanSeconds());
    options.timeseries->setBucketMaxSpanSeconds(maxSpanSeconds);

    // Set the validator option to a JSON schema enforcing constraints on bucket documents.
    // This validation is only structural to prevent accidental corruption by users and
    // cannot cover all constraints. Leave the validationLevel and validationAction to their
    // strict/error defaults.
    auto timeField = options.timeseries->getTimeField();
    auto validatorObj = fromjson(fmt::sprintf(R"(
{
'$jsonSchema' : {
    bsonType: 'object',
    required: ['_id', 'control', 'data'],
    properties: {
        _id: {bsonType: 'objectId'},
        control: {
            bsonType: 'object',
            required: ['version', 'min', 'max'],
            properties: {
                version: {bsonType: 'number'},
                min: {
                    bsonType: 'object',
                    required: ['%s'],
                    properties: {'%s': {bsonType: 'date'}}
                },
                max: {
                    bsonType: 'object',
                    required: ['%s'],
                    properties: {'%s': {bsonType: 'date'}}
                },
                closed: {bsonType: 'bool'}
            }
        },
        data: {bsonType: 'object'},
        meta: {}
    },
    additionalProperties: false
}
})",
                                              timeField,
                                              timeField,
                                              timeField,
                                              timeField));

    bool existingBucketCollectionIsCompatible = false;

    Status ret =
        writeConflictRetry(opCtx, "createBucketCollection", bucketsNs.ns(), [&]() -> Status {
            AutoGetDb autoDb(opCtx, bucketsNs.db(), MODE_IX);
            Lock::CollectionLock bucketsCollLock(opCtx, bucketsNs, MODE_IX);

            // Check if there already exist a Collection on the namespace we will later create a
            // view on. We're not holding a Collection lock for this Collection so we may only check
            // if the pointer is null or not. The answer may also change at any point after this
            // call which is fine as we properly handle an orphaned bucket collection. This check is
            // just here to prevent it from being created in the common case.
            if (CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, ns)) {
                return Status(ErrorCodes::NamespaceExists,
                              str::stream() << "Collection already exists. NS: " << ns);
            }

            auto db = autoDb.ensureDbExists();
            if (auto view = ViewCatalog::get(db)->lookup(opCtx, ns.ns()); view) {
                if (view->timeseries()) {
                    return Status(ErrorCodes::NamespaceExists,
                                  str::stream()
                                      << "A timeseries collection already exists. NS: " << ns);
                }
                return Status(ErrorCodes::NamespaceExists,
                              str::stream() << "A view already exists. NS: " << ns);
            }

            if (opCtx->writesAreReplicated() &&
                !repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, bucketsNs)) {
                // Report the error with the user provided namespace
                return Status(ErrorCodes::NotWritablePrimary,
                              str::stream() << "Not primary while creating collection " << ns);
            }

            WriteUnitOfWork wuow(opCtx);
            AutoStatsTracker bucketsStatsTracker(
                opCtx,
                bucketsNs,
                Top::LockType::NotLocked,
                AutoStatsTracker::LogMode::kUpdateTopAndCurOp,
                CollectionCatalog::get(opCtx)->getDatabaseProfileLevel(ns.db()));

            // If the buckets collection and time-series view creation roll back, ensure that their
            // Top entries are deleted.
            opCtx->recoveryUnit()->onRollback(
                [serviceContext = opCtx->getServiceContext(), bucketsNs]() {
                    Top::get(serviceContext).collectionDropped(bucketsNs);
                });


            // Prepare collection option and index spec using the provided options. In case the
            // collection already exist we use these to validate that they are the same as being
            // requested here.
            CollectionOptions bucketsOptions = options;
            bucketsOptions.validator = validatorObj;

            // If possible, cluster time-series buckets collections by _id.
            const bool useClusteredIdIndex = gTimeseriesBucketsCollectionClusterById;
            auto expireAfterSeconds = options.expireAfterSeconds;
            if (useClusteredIdIndex) {
                if (expireAfterSeconds) {
                    uassertStatusOK(
                        index_key_validate::validateExpireAfterSeconds(*expireAfterSeconds));
                    bucketsOptions.expireAfterSeconds = expireAfterSeconds;
                }
                bucketsOptions.clusteredIndex = true;
            }

            // Create a TTL index on 'control.min.[timeField]' if 'expireAfterSeconds' is provided
            // and the collection is not clustered by _id.
            BSONObj indexSpec;
            std::string indexName;
            if (expireAfterSeconds && !bucketsOptions.clusteredIndex) {
                const std::string controlMinTimeField = str::stream()
                    << "control.min." << options.timeseries->getTimeField();
                indexName = controlMinTimeField + "_1";
                indexSpec =
                    BSON(IndexDescriptor::kIndexVersionFieldName
                         << IndexDescriptor::kLatestIndexVersion
                         << IndexDescriptor::kKeyPatternFieldName << BSON(controlMinTimeField << 1)
                         << IndexDescriptor::kIndexNameFieldName << indexName
                         << IndexDescriptor::kExpireAfterSecondsFieldName << *expireAfterSeconds);
            }

            if (auto coll =
                    CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, bucketsNs)) {
                // Compare CollectionOptions and eventual TTL index to see if this bucket collection
                // may be reused for this request.
                existingBucketCollectionIsCompatible =
                    coll->getCollectionOptions().matchesStorageOptions(
                        bucketsOptions, CollatorFactoryInterface::get(opCtx->getServiceContext()));
                if (expireAfterSeconds && !bucketsOptions.clusteredIndex) {
                    auto indexDescriptor =
                        coll->getIndexCatalog()->findIndexByName(opCtx, indexName, true);
                    existingBucketCollectionIsCompatible &=
                        indexDescriptor && indexDescriptor->infoObj().woCompare(indexSpec) == 0;
                }

                return Status(ErrorCodes::NamespaceExists,
                              str::stream() << "Bucket Collection already exists. NS: " << bucketsNs
                                            << ". UUID: " << coll->uuid());
            }

            // Create the buckets collection that will back the view.
            const bool createIdIndex = !useClusteredIdIndex;
            uassertStatusOK(db->userCreateNS(opCtx, bucketsNs, bucketsOptions, createIdIndex));

            // Create a TTL index if 'expireAfterSeconds' is provided and the collection is not
            // clustered by _id.
            if (expireAfterSeconds && !useClusteredIdIndex) {
                CollectionWriter collectionWriter(opCtx, bucketsNs);
                auto indexBuildCoord = IndexBuildsCoordinator::get(opCtx);
                auto fromMigrate = false;
                try {
                    uassertStatusOK(index_key_validate::validateIndexSpecTTL(indexSpec));
                    indexBuildCoord->createIndexesOnEmptyCollection(
                        opCtx, collectionWriter, {indexSpec}, fromMigrate);
                } catch (DBException& ex) {
                    ex.addContext(str::stream()
                                  << "failed to create TTL index on bucket collection: "
                                  << bucketsNs << "; index spec: " << indexSpec);
                    return ex.toStatus();
                }
            }
            wuow.commit();
            return Status::OK();
        });

    // If compatible bucket collection already exists then proceed with creating view definition.
    if (!ret.isOK() && !existingBucketCollectionIsCompatible)
        return ret;

    ret = writeConflictRetry(opCtx, "create", ns.ns(), [&]() -> Status {
        AutoGetCollection autoColl(opCtx, ns, MODE_IX, AutoGetCollectionViewMode::kViewsPermitted);
        Lock::CollectionLock systemDotViewsLock(
            opCtx,
            NamespaceString(ns.db(), NamespaceString::kSystemDotViewsCollectionName),
            MODE_X);

        // This is a top-level handler for time-series creation name conflicts. New commands coming
        // in, or commands that generated a WriteConflict must return a NamespaceExists error here
        // on conflict.
        if (CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, ns)) {
            return Status(ErrorCodes::NamespaceExists,
                          str::stream() << "Collection already exists. NS: " << ns);
        }

        auto db = autoColl.ensureDbExists();
        if (auto view = ViewCatalog::get(db)->lookup(opCtx, ns.ns())) {
            if (view->timeseries()) {
                return {ErrorCodes::NamespaceExists,
                        str::stream() << "A timeseries collection already exists. NS: " << ns};
            }
            return {ErrorCodes::NamespaceExists,
                    str::stream() << "A view already exists. NS: " << ns};
        }

        if (opCtx->writesAreReplicated() &&
            !repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, ns)) {
            return {ErrorCodes::NotWritablePrimary,
                    str::stream() << "Not primary while creating collection " << ns};
        }

        _createSystemDotViewsIfNecessary(opCtx, db);

        auto catalog = CollectionCatalog::get(opCtx);
        WriteUnitOfWork wuow(opCtx);

        AutoStatsTracker statsTracker(opCtx,
                                      ns,
                                      Top::LockType::NotLocked,
                                      AutoStatsTracker::LogMode::kUpdateTopAndCurOp,
                                      catalog->getDatabaseProfileLevel(ns.db()));

        // If the buckets collection and time-series view creation roll back, ensure that their
        // Top entries are deleted.
        opCtx->recoveryUnit()->onRollback([serviceContext = opCtx->getServiceContext(), ns]() {
            Top::get(serviceContext).collectionDropped(ns);
        });

        if (MONGO_unlikely(failTimeseriesViewCreation.shouldFail(
                [&ns](const BSONObj& data) { return data["ns"_sd].String() == ns.ns(); }))) {
            LOGV2(5490200,
                  "failTimeseriesViewCreation fail point enabled. Failing creation of view "
                  "definition after bucket collection was created successfully.");
            return {ErrorCodes::OperationFailed,
                    str::stream() << "Timeseries view definition " << ns
                                  << " creation failed due to 'failTimeseriesViewCreation' "
                                     "fail point enabled."};
        }

        CollectionOptions viewOptions;
        viewOptions.viewOn = bucketsNs.coll().toString();
        viewOptions.collation = options.collation;
        constexpr bool asArray = true;
        viewOptions.pipeline = timeseries::generateViewPipeline(*options.timeseries, asArray);

        // Create the time-series view.
        auto status = db->userCreateNS(opCtx, ns, viewOptions);
        if (!status.isOK()) {
            return status.withContext(str::stream() << "Failed to create view on " << bucketsNs
                                                    << " for time-series collection " << ns
                                                    << " with options " << viewOptions.toBSON());
        }

        wuow.commit();
        return Status::OK();
    });

    return ret;
}

Status _createCollection(OperationContext* opCtx,
                         const NamespaceString& nss,
                         CollectionOptions&& collectionOptions,
                         boost::optional<BSONObj> idIndex) {
    return writeConflictRetry(opCtx, "create", nss.ns(), [&] {
        AutoGetDb autoDb(opCtx, nss.db(), MODE_IX);
        Lock::CollectionLock collLock(opCtx, nss, MODE_IX);
        // This is a top-level handler for collection creation name conflicts. New commands coming
        // in, or commands that generated a WriteConflict must return a NamespaceExists error here
        // on conflict.
        if (CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, nss)) {
            return Status(ErrorCodes::NamespaceExists,
                          str::stream() << "Collection already exists. NS: " << nss);
        }
        auto db = autoDb.ensureDbExists();
        if (auto view = ViewCatalog::get(db)->lookup(opCtx, nss.ns()); view) {
            if (view->timeseries()) {
                return Status(ErrorCodes::NamespaceExists,
                              str::stream()
                                  << "A timeseries collection already exists. NS: " << nss);
            }
            return Status(ErrorCodes::NamespaceExists,
                          str::stream() << "A view already exists. NS: " << nss);
        }

        if (collectionOptions.clusteredIndex && !nss.isTimeseriesBucketsCollection()) {
            return Status(
                ErrorCodes::InvalidOptions,
                "The 'clusteredIndex' option is only supported on time-series buckets collections");
        }

        if (collectionOptions.clusteredIndex && idIndex && !idIndex->isEmpty()) {
            return Status(ErrorCodes::InvalidOptions,
                          "The 'clusteredIndex' option is not supported with the 'idIndex' option");
        }


        if (opCtx->writesAreReplicated() &&
            !repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, nss)) {
            return Status(ErrorCodes::NotWritablePrimary,
                          str::stream() << "Not primary while creating collection " << nss);
        }

        WriteUnitOfWork wunit(opCtx);

        AutoStatsTracker statsTracker(
            opCtx,
            nss,
            Top::LockType::NotLocked,
            AutoStatsTracker::LogMode::kUpdateTopAndCurOp,
            CollectionCatalog::get(opCtx)->getDatabaseProfileLevel(nss.db()));

        // If the collection creation rolls back, ensure that the Top entry created for the
        // collection is deleted.
        opCtx->recoveryUnit()->onRollback([nss, serviceContext = opCtx->getServiceContext()]() {
            Top::get(serviceContext).collectionDropped(nss);
        });

        // Even though 'collectionOptions' is passed by rvalue reference, it is not safe to move
        // because 'userCreateNS' may throw a WriteConflictException.
        Status status = Status::OK();
        if (idIndex == boost::none || collectionOptions.clusteredIndex) {
            status = db->userCreateNS(opCtx, nss, collectionOptions, /*createIdIndex=*/false);
        } else {
            status =
                db->userCreateNS(opCtx, nss, collectionOptions, /*createIdIndex=*/true, *idIndex);
        }
        if (!status.isOK()) {
            return status;
        }
        wunit.commit();

        return Status::OK();
    });
}

/**
 * Creates the collection or the view as described by 'options'.
 */
Status createCollection(OperationContext* opCtx,
                        const NamespaceString& ns,
                        CollectionOptions&& options,
                        boost::optional<BSONObj> idIndex) {
    auto status = userAllowedCreateNS(opCtx, ns);
    if (!status.isOK()) {
        return status;
    }

    if (options.isView()) {
        uassert(ErrorCodes::OperationNotSupportedInTransaction,
                str::stream() << "Cannot create a view in a multi-document "
                                 "transaction.",
                !opCtx->inMultiDocumentTransaction());
        return _createView(opCtx, ns, std::move(options));
    } else if (options.timeseries && !ns.isTimeseriesBucketsCollection()) {
        // This helper is designed for user-created time-series collections on primaries. If a
        // time-series buckets collection is created explicitly or during replication, treat this as
        // a normal collection creation.
        uassert(ErrorCodes::OperationNotSupportedInTransaction,
                str::stream()
                    << "Cannot create a time-series collection in a multi-document transaction.",
                !opCtx->inMultiDocumentTransaction());
        return _createTimeseries(opCtx, ns, options);
    } else {
        uassert(ErrorCodes::OperationNotSupportedInTransaction,
                str::stream() << "Cannot create system collection " << ns
                              << " within a transaction.",
                !opCtx->inMultiDocumentTransaction() || !ns.isSystem());
        return _createCollection(opCtx, ns, std::move(options), idIndex);
    }
}

/**
 * Shared part of the implementation of the createCollection versions for replicated and regular
 * collection creation.
 */
Status createCollection(OperationContext* opCtx,
                        const NamespaceString& nss,
                        const BSONObj& cmdObj,
                        boost::optional<BSONObj> idIndex,
                        CollectionOptions::ParseKind kind) {
    BSONObjIterator it(cmdObj);

    // Skip the first cmdObj element.
    BSONElement firstElt = it.next();
    invariant(firstElt.fieldNameStringData() == "create");

    // Build options object from remaining cmdObj elements.
    BSONObjBuilder optionsBuilder;
    while (it.more()) {
        const auto elem = it.next();
        if (!isGenericArgument(elem.fieldNameStringData()))
            optionsBuilder.append(elem);
        if (elem.fieldNameStringData() == "viewOn") {
            // Views don't have UUIDs so it should always be parsed for command.
            kind = CollectionOptions::parseForCommand;
        }
    }

    BSONObj options = optionsBuilder.obj();
    uassert(14832,
            "specify size:<n> when capped is true",
            !options["capped"].trueValue() || options["size"].isNumber());

    CollectionOptions collectionOptions;
    {
        StatusWith<CollectionOptions> statusWith = CollectionOptions::parse(options, kind);
        if (!statusWith.isOK()) {
            return statusWith.getStatus();
        }
        collectionOptions = statusWith.getValue();
    }

    return createCollection(opCtx, nss, std::move(collectionOptions), idIndex);
}

}  // namespace

Status createCollection(OperationContext* opCtx,
                        const std::string& dbName,
                        const BSONObj& cmdObj,
                        const BSONObj& idIndex) {
    return createCollection(opCtx,
                            CommandHelpers::parseNsCollectionRequired(dbName, cmdObj),
                            cmdObj,
                            idIndex,
                            CollectionOptions::parseForCommand);
}

Status createCollection(OperationContext* opCtx,
                        const NamespaceString& ns,
                        const CreateCommand& cmd) {
    auto options = CollectionOptions::fromCreateCommand(cmd);
    auto idIndex = std::exchange(options.idIndex, {});
    return createCollection(opCtx, ns, std::move(options), idIndex);
}

Status createCollectionForApplyOps(OperationContext* opCtx,
                                   const std::string& dbName,
                                   const OptionalCollectionUUID& ui,
                                   const BSONObj& cmdObj,
                                   const bool allowRenameOutOfTheWay,
                                   boost::optional<BSONObj> idIndex) {
    invariant(opCtx->lockState()->isDbLockedForMode(dbName, MODE_IX));

    const NamespaceString newCollName(CommandHelpers::parseNsCollectionRequired(dbName, cmdObj));
    auto newCmd = cmdObj;

    auto databaseHolder = DatabaseHolder::get(opCtx);
    auto* const db = databaseHolder->getDb(opCtx, dbName);

    // If a UUID is given, see if we need to rename a collection out of the way, and whether the
    // collection already exists under a different name. If so, rename it into place. As this is
    // done during replay of the oplog, the operations do not need to be atomic, just idempotent.
    // We need to do the renaming part in a separate transaction, as we cannot transactionally
    // create a database, which could result in createCollection failing if the database
    // does not yet exist.
    if (ui) {
        auto uuid = ui.get();
        uassert(ErrorCodes::InvalidUUID,
                "Invalid UUID in applyOps create command: " + uuid.toString(),
                uuid.isRFC4122v4());

        auto catalog = CollectionCatalog::get(opCtx);
        const auto currentName = catalog->lookupNSSByUUID(opCtx, uuid);
        auto serviceContext = opCtx->getServiceContext();
        auto opObserver = serviceContext->getOpObserver();
        if (currentName && *currentName == newCollName)
            return Status::OK();

        if (currentName && currentName->isDropPendingNamespace()) {
            LOGV2(20308,
                  "CMD: create -- existing collection with conflicting UUID is in a drop-pending "
                  "state",
                  "newCollection"_attr = newCollName,
                  "conflictingUUID"_attr = uuid,
                  "existingCollection"_attr = *currentName);
            return Status(ErrorCodes::NamespaceExists,
                          str::stream() << "existing collection " << currentName->toString()
                                        << " with conflicting UUID " << uuid.toString()
                                        << " is in a drop-pending state.");
        }

        // In the case of oplog replay, a future command may have created or renamed a
        // collection with that same name. In that case, renaming this future collection to
        // a random temporary name is correct: once all entries are replayed no temporary
        // names will remain.
        const bool stayTemp = true;
        auto futureColl = db ? catalog->lookupCollectionByNamespace(opCtx, newCollName) : nullptr;
        bool needsRenaming = static_cast<bool>(futureColl);
        invariant(!needsRenaming || allowRenameOutOfTheWay,
                  str::stream() << "Current collection name: " << currentName << ", UUID: " << uuid
                                << ". Future collection name: " << newCollName);

        for (int tries = 0; needsRenaming && tries < 10; ++tries) {
            auto tmpNameResult = db->makeUniqueCollectionNamespace(opCtx, "tmp%%%%%.create");
            if (!tmpNameResult.isOK()) {
                return tmpNameResult.getStatus().withContext(str::stream()
                                                             << "Cannot generate temporary "
                                                                "collection namespace for applyOps "
                                                                "create command: collection: "
                                                             << newCollName);
            }

            const auto& tmpName = tmpNameResult.getValue();
            AutoGetCollection tmpCollLock(opCtx, tmpName, LockMode::MODE_X);
            if (tmpCollLock.getCollection()) {
                // Conflicting on generating a unique temp collection name. Try again.
                continue;
            }

            // It is ok to log this because this doesn't happen very frequently.
            LOGV2(20309,
                  "CMD: create -- renaming existing collection with conflicting UUID to "
                  "temporary collection",
                  "newCollection"_attr = newCollName,
                  "conflictingUUID"_attr = uuid,
                  "tempName"_attr = tmpName);
            Status status =
                writeConflictRetry(opCtx, "createCollectionForApplyOps", newCollName.ns(), [&] {
                    WriteUnitOfWork wuow(opCtx);
                    Status status = db->renameCollection(opCtx, newCollName, tmpName, stayTemp);
                    if (!status.isOK())
                        return status;
                    auto uuid = futureColl->uuid();
                    opObserver->onRenameCollection(opCtx,
                                                   newCollName,
                                                   tmpName,
                                                   uuid,
                                                   /*dropTargetUUID*/ {},
                                                   /*numRecords*/ 0U,
                                                   stayTemp);

                    wuow.commit();
                    // Re-fetch collection after commit to get a valid pointer
                    futureColl = CollectionCatalog::get(opCtx)->lookupCollectionByUUID(opCtx, uuid);
                    return Status::OK();
                });

            if (!status.isOK()) {
                return status;
            }

            // Abort any remaining index builds on the temporary collection.
            IndexBuildsCoordinator::get(opCtx)->abortCollectionIndexBuilds(
                opCtx,
                tmpName,
                futureColl->uuid(),
                "Aborting index builds on temporary collection");

            // The existing collection has been successfully moved out of the way.
            needsRenaming = false;
        }
        if (needsRenaming) {
            return Status(ErrorCodes::NamespaceExists,
                          str::stream() << "Cannot generate temporary "
                                           "collection namespace for applyOps "
                                           "create command: collection: "
                                        << newCollName);
        }

        // If the collection with the requested UUID already exists, but with a different
        // name, just rename it to 'newCollName'.
        if (catalog->lookupCollectionByUUID(opCtx, uuid)) {
            invariant(currentName);
            uassert(40655,
                    str::stream() << "Invalid name " << newCollName << " for UUID " << uuid,
                    currentName->db() == newCollName.db());
            return writeConflictRetry(opCtx, "createCollectionForApplyOps", newCollName.ns(), [&] {
                WriteUnitOfWork wuow(opCtx);
                Status status = db->renameCollection(opCtx, *currentName, newCollName, stayTemp);
                if (!status.isOK())
                    return status;
                opObserver->onRenameCollection(opCtx,
                                               *currentName,
                                               newCollName,
                                               uuid,
                                               /*dropTargetUUID*/ {},
                                               /*numRecords*/ 0U,
                                               stayTemp);

                wuow.commit();
                return Status::OK();
            });
        }

        // A new collection with the specific UUID must be created, so add the UUID to the
        // creation options. Regular user collection creation commands cannot do this.
        auto uuidObj = uuid.toBSON();
        newCmd = cmdObj.addField(uuidObj.firstElement());
    }

    return createCollection(
        opCtx, newCollName, newCmd, idIndex, CollectionOptions::parseForStorage);
}

}  // namespace mongo