summaryrefslogtreecommitdiff
path: root/src/mongo/db/cloner.cpp
blob: 69626a3e4b7faa4b0c35d39451063895bf589e62 (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
748
749
750
751
752
753
754
/**
 *    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_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kStorage

#include "mongo/platform/basic.h"

#include "mongo/db/cloner.h"
#include "mongo/db/cloner_gen.h"

#include <algorithm>

#include "mongo/base/status.h"
#include "mongo/bson/unordered_fields_bsonobj_comparator.h"
#include "mongo/bson/util/bson_extract.h"
#include "mongo/bson/util/builder.h"
#include "mongo/client/authenticate.h"
#include "mongo/db/catalog/collection.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/index_build_entry_gen.h"
#include "mongo/db/catalog/index_catalog.h"
#include "mongo/db/catalog/multi_index_block.h"
#include "mongo/db/commands.h"
#include "mongo/db/commands/list_collections_filter.h"
#include "mongo/db/commands/rename_collection.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/curop.h"
#include "mongo/db/dbdirectclient.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/index_build_entry_helpers.h"
#include "mongo/db/index_builds_coordinator.h"
#include "mongo/db/jsobj.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/op_observer.h"
#include "mongo/db/ops/insert.h"
#include "mongo/db/repl/isself.h"
#include "mongo/db/repl/read_concern_args.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/service_context.h"
#include "mongo/db/storage/durable_catalog.h"
#include "mongo/db/storage/storage_options.h"
#include "mongo/logv2/log.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/fail_point.h"
#include "mongo/util/str.h"

namespace mongo {

using std::string;
using std::unique_ptr;

using IndexVersion = IndexDescriptor::IndexVersion;

MONGO_FAIL_POINT_DEFINE(movePrimaryFailPoint);

BSONElement getErrField(const BSONObj& o);

BSONObj Cloner::_getIdIndexSpec(const std::list<BSONObj>& indexSpecs) {
    for (auto&& indexSpec : indexSpecs) {
        BSONElement indexName;
        uassertStatusOK(bsonExtractTypedField(
            indexSpec, IndexDescriptor::kIndexNameFieldName, String, &indexName));
        if (indexName.valueStringData() == "_id_"_sd) {
            return indexSpec;
        }
    }
    return BSONObj();
}

Cloner::Cloner() {}

struct Cloner::Fun {
    Fun(OperationContext* opCtx, const string& dbName)
        : lastLog(0), opCtx(opCtx), _dbName(dbName) {}

    void operator()(DBClientCursorBatchIterator& i) {
        boost::optional<Lock::DBLock> dbLock;
        dbLock.emplace(opCtx, _dbName, MODE_X);
        uassert(
            ErrorCodes::NotMaster,
            str::stream() << "Not primary while cloning collection " << from_collection.ns()
                          << " to " << to_collection.ns(),
            !opCtx->writesAreReplicated() ||
                repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, to_collection));

        // Make sure database still exists after we resume from the temp release
        auto databaseHolder = DatabaseHolder::get(opCtx);
        auto db = databaseHolder->openDb(opCtx, _dbName);

        bool createdCollection = false;
        Collection* collection = nullptr;

        collection =
            CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, to_collection);
        if (!collection) {
            massert(17321,
                    str::stream() << "collection dropped during clone [" << to_collection.ns()
                                  << "]",
                    !createdCollection);
            writeConflictRetry(opCtx, "createCollection", to_collection.ns(), [&] {
                opCtx->checkForInterrupt();

                WriteUnitOfWork wunit(opCtx);
                const bool createDefaultIndexes = true;
                CollectionOptions collectionOptions = uassertStatusOK(CollectionOptions::parse(
                    from_options, CollectionOptions::ParseKind::parseForCommand));
                invariant(db->userCreateNS(opCtx,
                                           to_collection,
                                           collectionOptions,
                                           createDefaultIndexes,
                                           from_id_index),
                          str::stream() << "collection creation failed during clone ["
                                        << to_collection.ns() << "]");
                wunit.commit();
                collection =
                    CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, to_collection);
                invariant(collection,
                          str::stream()
                              << "Missing collection during clone [" << to_collection.ns() << "]");
            });
        }

        const bool isSystemViewsClone = to_collection.isSystemDotViews();

        while (i.moreInCurrentBatch()) {
            if (numSeen % 128 == 127) {
                time_t now = time(nullptr);
                if (now - lastLog >= 60) {
                    // report progress
                    if (lastLog)
                        LOGV2(20412,
                              "clone {to_collection} {numSeen}",
                              "to_collection"_attr = to_collection,
                              "numSeen"_attr = numSeen);
                    lastLog = now;
                }
                opCtx->checkForInterrupt();

                dbLock.reset();

                CurOp::get(opCtx)->yielded();

                dbLock.emplace(opCtx, _dbName, MODE_X);

                // Check if everything is still all right.
                if (opCtx->writesAreReplicated()) {
                    uassert(ErrorCodes::PrimarySteppedDown,
                            str::stream() << "Cannot write to ns: " << to_collection.ns()
                                          << " after yielding",
                            repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(
                                opCtx, to_collection));
                }

                db = databaseHolder->getDb(opCtx, _dbName);
                uassert(28593,
                        str::stream() << "Database " << _dbName << " dropped while cloning",
                        db != nullptr);

                collection =
                    CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, to_collection);
                uassert(28594,
                        str::stream()
                            << "Collection " << to_collection.ns() << " dropped while cloning",
                        collection != nullptr);
            }

            BSONObj tmp = i.nextSafe();

            // If copying the system.views collection to a database with a different name, then any
            // view definitions must be modified to refer to the 'to' database.
            if (isSystemViewsClone && from_collection.db() != to_collection.db()) {
                BSONObjBuilder bob;
                for (auto&& item : tmp) {
                    if (item.fieldNameStringData() == "_id") {
                        auto viewNss = NamespaceString(item.checkAndGetStringData());

                        bob.append("_id",
                                   NamespaceString(to_collection.db(), viewNss.coll()).toString());
                    } else {
                        bob.append(item);
                    }
                }
                tmp = bob.obj();
            }

            /* assure object is valid.  note this will slow us down a little. */
            // Use the latest BSON validation version. We allow cloning of collections containing
            // decimal data even if decimal is disabled.
            const Status status = validateBSON(tmp.objdata(), tmp.objsize(), BSONVersion::kLatest);
            if (!status.isOK()) {
                str::stream ss;
                ss << "Cloner: found corrupt document in " << from_collection.toString() << ": "
                   << redact(status);
                if (gSkipCorruptDocumentsWhenCloning.load()) {
                    LOGV2_WARNING(20423, "{ss_ss_str}; skipping", "ss_ss_str"_attr = ss.ss.str());
                    continue;
                }
                msgasserted(28531, ss);
            }

            verify(collection);
            ++numSeen;

            writeConflictRetry(opCtx, "cloner insert", to_collection.ns(), [&] {
                opCtx->checkForInterrupt();

                WriteUnitOfWork wunit(opCtx);

                BSONObj doc = tmp;
                OpDebug* const nullOpDebug = nullptr;
                Status status =
                    collection->insertDocument(opCtx, InsertStatement(doc), nullOpDebug, true);
                if (!status.isOK() && status.code() != ErrorCodes::DuplicateKey) {
                    LOGV2_ERROR(
                        20424,
                        "error: exception cloning object in {from_collection} {status} obj:{doc}",
                        "from_collection"_attr = from_collection,
                        "status"_attr = redact(status),
                        "doc"_attr = redact(doc));
                    uassertStatusOK(status);
                }
                if (status.isOK()) {
                    wunit.commit();
                }
            });

            static Rarely sampler;
            if (sampler.tick() && (time(nullptr) - saveLast > 60)) {
                LOGV2(20413,
                      "{numSeen} objects cloned so far from collection {from_collection}",
                      "numSeen"_attr = numSeen,
                      "from_collection"_attr = from_collection);
                saveLast = time(nullptr);
            }
        }
    }

    time_t lastLog;
    OperationContext* opCtx;
    const string _dbName;

    int64_t numSeen;
    NamespaceString from_collection;
    BSONObj from_options;
    BSONObj from_id_index;
    NamespaceString to_collection;
    time_t saveLast;
};

/* copy the specified collection
 */
void Cloner::_copy(OperationContext* opCtx,
                   const string& toDBName,
                   const NamespaceString& from_collection,
                   const BSONObj& from_opts,
                   const BSONObj& from_id_index,
                   const NamespaceString& to_collection,
                   Query query) {
    LOGV2_DEBUG(20414,
                2,
                "\t\tcloning collection {from_collection} to {to_collection} on "
                "{conn_getServerAddress} with filter {query}",
                "from_collection"_attr = from_collection,
                "to_collection"_attr = to_collection,
                "conn_getServerAddress"_attr = _conn->getServerAddress(),
                "query"_attr = redact(query.toString()));

    Fun f(opCtx, toDBName);
    f.numSeen = 0;
    f.from_collection = from_collection;
    f.from_options = from_opts;
    f.from_id_index = from_id_index;
    f.to_collection = to_collection;
    f.saveLast = time(nullptr);

    int options = QueryOption_NoCursorTimeout | QueryOption_Exhaust;
    {
        Lock::TempRelease tempRelease(opCtx->lockState());
        _conn->query(std::function<void(DBClientCursorBatchIterator&)>(f),
                     from_collection,
                     query,
                     nullptr,
                     options,
                     0 /* batchSize */,
                     repl::ReadConcernArgs::kImplicitDefault);
    }

    uassert(ErrorCodes::PrimarySteppedDown,
            str::stream() << "Not primary while cloning collection " << from_collection.ns()
                          << " to " << to_collection.ns() << " with filter " << query.toString(),
            !opCtx->writesAreReplicated() ||
                repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, to_collection));
}

void Cloner::_copyIndexes(OperationContext* opCtx,
                          const string& toDBName,
                          const NamespaceString& from_collection,
                          const BSONObj& from_opts,
                          const std::list<BSONObj>& from_indexes,
                          const NamespaceString& to_collection) {
    LOGV2_DEBUG(20415,
                2,
                "\t\t copyIndexes {from_collection} to {to_collection} on {conn_getServerAddress}",
                "from_collection"_attr = from_collection,
                "to_collection"_attr = to_collection,
                "conn_getServerAddress"_attr = _conn->getServerAddress());

    uassert(ErrorCodes::PrimarySteppedDown,
            str::stream() << "Not primary while copying indexes from " << from_collection.ns()
                          << " to " << to_collection.ns() << " (Cloner)",
            !opCtx->writesAreReplicated() ||
                repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, to_collection));

    if (from_indexes.empty())
        return;

    // We are under lock here again, so reload the database in case it may have disappeared
    // during the temp release
    auto databaseHolder = DatabaseHolder::get(opCtx);
    auto db = databaseHolder->openDb(opCtx, toDBName);

    Collection* collection =
        CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, to_collection);
    if (!collection) {
        writeConflictRetry(opCtx, "createCollection", to_collection.ns(), [&] {
            opCtx->checkForInterrupt();

            WriteUnitOfWork wunit(opCtx);
            CollectionOptions collectionOptions = uassertStatusOK(
                CollectionOptions::parse(from_opts, CollectionOptions::ParseKind::parseForCommand));
            const bool createDefaultIndexes = true;
            invariant(db->userCreateNS(opCtx,
                                       to_collection,
                                       collectionOptions,
                                       createDefaultIndexes,
                                       _getIdIndexSpec(from_indexes)),
                      str::stream()
                          << "Collection creation failed while copying indexes from "
                          << from_collection.ns() << " to " << to_collection.ns() << " (Cloner)");
            wunit.commit();
            collection =
                CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, to_collection);
            invariant(collection,
                      str::stream() << "Missing collection " << to_collection.ns() << " (Cloner)");
        });
    }

    auto indexCatalog = collection->getIndexCatalog();
    auto indexesToBuild = indexCatalog->removeExistingIndexesNoChecks(
        opCtx, {std::begin(from_indexes), std::end(from_indexes)});
    if (indexesToBuild.empty()) {
        return;
    }

    // TODO pass the MultiIndexBlock when inserting into the collection rather than building the
    // indexes after the fact. This depends on holding a lock on the collection the whole time
    // from creation to completion without yielding to ensure the index and the collection
    // matches. It also wouldn't work on non-empty collections so we would need both
    // implementations anyway as long as that is supported.
    MultiIndexBlock indexer;

    // Emit startIndexBuild and commitIndexBuild oplog entries if supported by the current FCV.
    auto opObserver = opCtx->getServiceContext()->getOpObserver();
    auto fromMigrate = false;
    auto buildUUID = IndexBuildsCoordinator::supportsTwoPhaseIndexBuild()
        ? boost::make_optional(UUID::gen())
        : boost::none;

    MultiIndexBlock::OnInitFn onInitFn;
    if (opCtx->writesAreReplicated() && buildUUID) {
        onInitFn = [&](std::vector<BSONObj>& specs) {
            // TODO SERVER-47438: Should remove this onInitFn lambda function as we no longer
            // need to generate startIndexBuild and commitIndexBuild oplog entries.

            // Currently, primary doesn't wait for any votes from secondaries to commit
            // the index build. So, it's of no use to set the commit quorum option of any value
            // greater than 0. Disabling commit quorum is just an optimization to avoid secondaries
            // from trying to vote before committing index build.
            //
            // Persist the commit quorum value in the config.system.indexBuilds collection.
            IndexBuildEntry indexbuildEntry(*buildUUID,
                                            collection->uuid(),
                                            CommitQuorumOptions(CommitQuorumOptions::kDisabled),
                                            IndexBuildsCoordinator::extractIndexNames(specs));
            uassertStatusOK(indexbuildentryhelpers::addIndexBuildEntry(opCtx, indexbuildEntry));

            opObserver->onStartIndexBuild(
                opCtx, to_collection, collection->uuid(), *buildUUID, specs, fromMigrate);
            return Status::OK();
        };
    } else {
        onInitFn = MultiIndexBlock::kNoopOnInitFn;
    }

    auto indexInfoObjs = uassertStatusOK(indexer.init(opCtx, collection, indexesToBuild, onInitFn));

    // The code below throws, so ensure build cleanup occurs.
    auto abortOnExit = makeGuard(
        [&] { indexer.abortIndexBuild(opCtx, collection, MultiIndexBlock::kNoopOnCleanUpFn); });

    uassertStatusOK(indexer.insertAllDocumentsInCollection(opCtx, collection));
    uassertStatusOK(indexer.checkConstraints(opCtx));

    WriteUnitOfWork wunit(opCtx);
    uassertStatusOK(
        indexer.commit(opCtx,
                       collection,
                       [&](const BSONObj& spec) {
                           // If two phase index builds are enabled, the index build will be
                           // coordinated using startIndexBuild and commitIndexBuild oplog entries.
                           if (opCtx->writesAreReplicated() &&
                               !IndexBuildsCoordinator::supportsTwoPhaseIndexBuild()) {
                               opObserver->onCreateIndex(
                                   opCtx, collection->ns(), collection->uuid(), spec, fromMigrate);
                           }
                       },
                       [&] {
                           if (opCtx->writesAreReplicated() && buildUUID) {
                               opObserver->onCommitIndexBuild(opCtx,
                                                              collection->ns(),
                                                              collection->uuid(),
                                                              *buildUUID,
                                                              indexInfoObjs,
                                                              fromMigrate);
                           }
                       }));
    wunit.commit();
    abortOnExit.dismiss();
}

StatusWith<std::vector<BSONObj>> Cloner::_filterCollectionsForClone(
    const std::string& fromDBName, const std::list<BSONObj>& initialCollections) {
    std::vector<BSONObj> finalCollections;
    for (auto&& collection : initialCollections) {
        LOGV2_DEBUG(20418, 2, "\t cloner got {collection}", "collection"_attr = collection);

        BSONElement collectionOptions = collection["options"];
        if (collectionOptions.isABSONObj()) {
            auto statusWithCollectionOptions = CollectionOptions::parse(
                collectionOptions.Obj(), CollectionOptions::ParseKind::parseForCommand);
            if (!statusWithCollectionOptions.isOK()) {
                return statusWithCollectionOptions.getStatus();
            }
        }

        std::string collectionName;
        auto status = bsonExtractStringField(collection, "name", &collectionName);
        if (!status.isOK()) {
            return status;
        }

        const NamespaceString ns(fromDBName, collectionName.c_str());

        if (ns.isSystem()) {
            if (!ns.isLegalClientSystemNS()) {
                LOGV2_DEBUG(20419, 2, "\t\t not cloning because system collection");
                continue;
            }
        }

        finalCollections.push_back(collection.getOwned());
    }
    return finalCollections;
}

Status Cloner::_createCollectionsForDb(
    OperationContext* opCtx,
    const std::vector<CreateCollectionParams>& createCollectionParams,
    const std::string& dbName) {
    auto databaseHolder = DatabaseHolder::get(opCtx);
    auto db = databaseHolder->openDb(opCtx, dbName);
    invariant(opCtx->lockState()->isDbLockedForMode(dbName, MODE_X));

    auto collCount = 0;
    for (auto&& params : createCollectionParams) {
        if (MONGO_unlikely(movePrimaryFailPoint.shouldFail()) && collCount > 0) {
            return Status(ErrorCodes::CommandFailed, "movePrimary failed due to failpoint");
        }
        collCount++;

        BSONObjBuilder optionsBuilder;
        optionsBuilder.appendElements(params.collectionInfo["options"].Obj());

        const NamespaceString nss(dbName, params.collectionName);

        uassertStatusOK(userAllowedCreateNS(dbName, params.collectionName));
        Status status = writeConflictRetry(opCtx, "createCollection", nss.ns(), [&] {
            opCtx->checkForInterrupt();
            WriteUnitOfWork wunit(opCtx);

            Collection* collection =
                CollectionCatalog::get(opCtx).lookupCollectionByNamespace(opCtx, nss);
            if (collection) {
                if (!params.shardedColl) {
                    // If the collection is unsharded then we want to fail when a collection
                    // we're trying to create already exists.
                    return Status(ErrorCodes::NamespaceExists,
                                  str::stream() << "unsharded collection with same namespace "
                                                << nss.ns() << " already exists.");
                }

                // If the collection is sharded and a collection with the same name already
                // exists on the target, we check if the existing collection's UUID matches
                // that of the one we're trying to create. If it does, we treat the create
                // as a no-op; if it doesn't match, we return an error.
                auto existingOpts = DurableCatalog::get(opCtx)->getCollectionOptions(
                    opCtx, collection->getCatalogId());
                const UUID clonedUUID =
                    uassertStatusOK(UUID::parse(params.collectionInfo["info"]["uuid"]));

                if (clonedUUID == existingOpts.uuid)
                    return Status::OK();

                return Status(ErrorCodes::InvalidOptions,
                              str::stream()
                                  << "sharded collection with same namespace " << nss.ns()
                                  << " already exists, but UUIDs don't match. Existing UUID is "
                                  << existingOpts.uuid << " and new UUID is " << clonedUUID);
            }

            // If the collection does not already exist and is sharded, we create a new
            // collection on the target shard with the UUID of the original collection and
            // copy the options and secondary indexes. If the collection does not already
            // exist and is unsharded, we create a new collection with its own UUID and
            // copy the options and secondary indexes of the original collection.

            if (params.shardedColl) {
                optionsBuilder.append(params.collectionInfo["info"]["uuid"]);
            }

            const bool createDefaultIndexes = true;
            auto options = optionsBuilder.obj();

            CollectionOptions collectionOptions = uassertStatusOK(
                CollectionOptions::parse(options, CollectionOptions::ParseKind::parseForStorage));
            Status createStatus = db->userCreateNS(
                opCtx, nss, collectionOptions, createDefaultIndexes, params.idIndexSpec);
            if (!createStatus.isOK()) {
                return createStatus;
            }

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

        // Break early if one of the creations fails.
        if (!status.isOK()) {
            return status;
        }
    }

    return Status::OK();
}

Status Cloner::copyDb(OperationContext* opCtx,
                      const std::string& toDBName,
                      const std::string& masterHost,
                      const std::string& fromDBName,
                      const std::vector<NamespaceString>& shardedColls,
                      std::set<std::string>* clonedColls) {
    invariant(clonedColls, str::stream() << masterHost << ":" << toDBName);

    auto statusWithMasterHost = ConnectionString::parse(masterHost);
    if (!statusWithMasterHost.isOK()) {
        return statusWithMasterHost.getStatus();
    }

    const ConnectionString cs(statusWithMasterHost.getValue());

    bool masterSameProcess = false;
    std::vector<HostAndPort> csServers = cs.getServers();
    for (std::vector<HostAndPort>::const_iterator iter = csServers.begin(); iter != csServers.end();
         ++iter) {
        if (!repl::isSelf(*iter, opCtx->getServiceContext()))
            continue;

        masterSameProcess = true;
        break;
    }

    if (masterSameProcess) {
        if (fromDBName == toDBName) {
            // Guard against re-entrance
            return Status(ErrorCodes::IllegalOperation, "can't clone from self (localhost)");
        }
    }

    {
        // setup connection
        if (_conn.get()) {
            // nothing to do
        } else if (!masterSameProcess) {
            std::string errmsg;
            unique_ptr<DBClientBase> con(cs.connect(StringData(), errmsg));
            if (!con.get()) {
                return Status(ErrorCodes::HostUnreachable, errmsg);
            }

            if (auth::isInternalAuthSet()) {
                auto authStatus = con->authenticateInternalUser();
                if (!authStatus.isOK()) {
                    return authStatus;
                }
            }

            _conn = std::move(con);
        } else {
            _conn.reset(new DBDirectClient(opCtx));
        }
    }

    // Gather the list of collections to clone
    std::vector<BSONObj> toClone;
    clonedColls->clear();

    {
        // getCollectionInfos may make a remote call, which may block indefinitely, so release
        // the global lock that we are entering with.
        Lock::TempRelease tempRelease(opCtx->lockState());

        std::list<BSONObj> initialCollections = _conn->getCollectionInfos(
            fromDBName, ListCollectionsFilter::makeTypeCollectionFilter());

        auto status = _filterCollectionsForClone(fromDBName, initialCollections);
        if (!status.isOK()) {
            return status.getStatus();
        }
        toClone = status.getValue();
    }

    std::vector<CreateCollectionParams> createCollectionParams;
    for (auto&& collection : toClone) {
        CreateCollectionParams params;
        params.collectionName = collection["name"].String();
        params.collectionInfo = collection;
        if (auto idIndex = collection["idIndex"]) {
            params.idIndexSpec = idIndex.Obj();
        }

        const NamespaceString ns(fromDBName, params.collectionName);
        if (std::find(shardedColls.begin(), shardedColls.end(), ns) != shardedColls.end()) {
            params.shardedColl = true;
        }
        createCollectionParams.push_back(params);
    }

    // Get index specs for each collection.
    std::map<StringData, std::list<BSONObj>> collectionIndexSpecs;
    {
        Lock::TempRelease tempRelease(opCtx->lockState());
        for (auto&& params : createCollectionParams) {
            const NamespaceString nss(fromDBName, params.collectionName);
            auto indexSpecs = _conn->getIndexSpecs(nss);

            collectionIndexSpecs[params.collectionName] = indexSpecs;

            if (params.idIndexSpec.isEmpty()) {
                params.idIndexSpec = _getIdIndexSpec(indexSpecs);
            }
        }
    }

    uassert(
        ErrorCodes::NotMaster,
        str::stream() << "Not primary while cloning database " << fromDBName
                      << " (after getting list of collections to clone)",
        !opCtx->writesAreReplicated() ||
            repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesForDatabase(opCtx, toDBName));

    auto status = _createCollectionsForDb(opCtx, createCollectionParams, toDBName);
    if (!status.isOK()) {
        return status;
    }

    for (auto&& params : createCollectionParams) {
        if (params.shardedColl) {
            continue;
        }

        LOGV2_DEBUG(20420,
                    2,
                    "  really will clone: {params_collectionInfo}",
                    "params_collectionInfo"_attr = params.collectionInfo);

        const NamespaceString from_name(fromDBName, params.collectionName);
        const NamespaceString to_name(toDBName, params.collectionName);

        clonedColls->insert(from_name.ns());

        LOGV2_DEBUG(20421,
                    1,
                    "\t\t cloning {from_name} -> {to_name}",
                    "from_name"_attr = from_name,
                    "to_name"_attr = to_name);

        _copy(opCtx,
              toDBName,
              from_name,
              params.collectionInfo["options"].Obj(),
              params.idIndexSpec,
              to_name,
              Query());
    }

    // now build the secondary indexes
    for (auto&& params : createCollectionParams) {
        LOGV2(20422,
              "copying indexes for: {params_collectionInfo}",
              "params_collectionInfo"_attr = params.collectionInfo);

        const NamespaceString from_name(fromDBName, params.collectionName);
        const NamespaceString to_name(toDBName, params.collectionName);


        _copyIndexes(opCtx,
                     toDBName,
                     from_name,
                     params.collectionInfo["options"].Obj(),
                     collectionIndexSpecs[params.collectionName],
                     to_name);
    }

    return Status::OK();
}

}  // namespace mongo