summaryrefslogtreecommitdiff
path: root/src/mongo/db/repl/collection_cloner.cpp
blob: 85c9f39169d54ddfd6d1693f02bc12ea8139922e (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
/**
 *    Copyright (C) 2015 MongoDB Inc.
 *
 *    This program is free software: you can redistribute it and/or  modify
 *    it under the terms of the GNU Affero General Public License, version 3,
 *    as published by the Free Software Foundation.
 *
 *    This program is distributed in the hope that it will be useful,
 *    but WITHOUT ANY WARRANTY; without even the implied warranty of
 *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *    GNU Affero General Public License for more details.
 *
 *    You should have received a copy of the GNU Affero General Public License
 *    along with this program.  If not, see <http://www.gnu.org/licenses/>.
 *
 *    As a special exception, the copyright holders give permission to link the
 *    code of portions of this program with the OpenSSL library under certain
 *    conditions as described in each individual source file and distribute
 *    linked combinations including the program with the OpenSSL library. You
 *    must comply with the GNU Affero General Public License in all respects for
 *    all of the code used other than as permitted herein. If you modify file(s)
 *    with this exception, you may extend this exception to your version of the
 *    file(s), but you are not obligated to do so. If you do not wish to do so,
 *    delete this exception statement from your version. If you delete this
 *    exception statement from all source files in the program, then also delete
 *    it in the license file.
 */

#define MONGO_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kReplication

#include "mongo/platform/basic.h"

#include "mongo/db/repl/collection_cloner.h"

#include <utility>

#include "mongo/base/string_data.h"
#include "mongo/bson/util/bson_extract.h"
#include "mongo/client/remote_command_retry_scheduler.h"
#include "mongo/db/catalog/collection_options.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/repl/initial_sync_common.h"
#include "mongo/db/repl/storage_interface.h"
#include "mongo/db/repl/storage_interface_mock.h"
#include "mongo/db/server_parameters.h"
#include "mongo/rpc/get_status_from_command_result.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/destructor_guard.h"
#include "mongo/util/fail_point_service.h"
#include "mongo/util/log.h"
#include "mongo/util/mongoutils/str.h"

namespace mongo {
namespace repl {
namespace {

using LockGuard = stdx::lock_guard<stdx::mutex>;
using UniqueLock = stdx::unique_lock<stdx::mutex>;

constexpr auto kCountResponseDocumentCountFieldName = "n"_sd;

const int kProgressMeterSecondsBetween = 60;
const int kProgressMeterCheckInterval = 128;

// The batchSize to use for the query to get all documents from the collection.
// 16MB max batch size / 12 byte min doc size * 10 (for good measure) = batchSize to use.
const auto batchSize = (16 * 1024 * 1024) / 12 * 10;
// The number of attempts for the count command, which gets the document count.
MONGO_EXPORT_SERVER_PARAMETER(numInitialSyncCollectionCountAttempts, int, 3);
// The number of attempts for the listIndexes commands.
MONGO_EXPORT_SERVER_PARAMETER(numInitialSyncListIndexesAttempts, int, 3);
// The number of attempts for the find command, which gets the data.
MONGO_EXPORT_SERVER_PARAMETER(numInitialSyncCollectionFindAttempts, int, 3);
}  // namespace

CollectionCloner::CollectionCloner(executor::TaskExecutor* executor,
                                   OldThreadPool* dbWorkThreadPool,
                                   const HostAndPort& source,
                                   const NamespaceString& sourceNss,
                                   const CollectionOptions& options,
                                   const CallbackFn& onCompletion,
                                   StorageInterface* storageInterface)
    : _executor(executor),
      _dbWorkThreadPool(dbWorkThreadPool),
      _source(source),
      _sourceNss(sourceNss),
      _destNss(_sourceNss),
      _options(options),
      _onCompletion(onCompletion),
      _storageInterface(storageInterface),
      _countScheduler(_executor,
                      RemoteCommandRequest(_source,
                                           _sourceNss.db().toString(),
                                           BSON("count" << _sourceNss.coll()),
                                           rpc::ServerSelectionMetadata(true, boost::none).toBSON(),
                                           nullptr,
                                           RemoteCommandRequest::kNoTimeout),
                      stdx::bind(&CollectionCloner::_countCallback, this, stdx::placeholders::_1),
                      RemoteCommandRetryScheduler::makeRetryPolicy(
                          numInitialSyncCollectionCountAttempts.load(),
                          executor::RemoteCommandRequest::kNoTimeout,
                          RemoteCommandRetryScheduler::kAllRetriableErrors)),
      _listIndexesFetcher(_executor,
                          _source,
                          _sourceNss.db().toString(),
                          BSON("listIndexes" << _sourceNss.coll()),
                          stdx::bind(&CollectionCloner::_listIndexesCallback,
                                     this,
                                     stdx::placeholders::_1,
                                     stdx::placeholders::_2,
                                     stdx::placeholders::_3),
                          rpc::ServerSelectionMetadata(true, boost::none).toBSON(),
                          RemoteCommandRequest::kNoTimeout,
                          RemoteCommandRetryScheduler::makeRetryPolicy(
                              numInitialSyncListIndexesAttempts.load(),
                              executor::RemoteCommandRequest::kNoTimeout,
                              RemoteCommandRetryScheduler::kAllRetriableErrors)),
      _indexSpecs(),
      _documents(),
      _dbWorkTaskRunner(_dbWorkThreadPool),
      _scheduleDbWorkFn([this](const executor::TaskExecutor::CallbackFn& work) {
          auto task = [work](OperationContext* opCtx,
                             const Status& status) -> TaskRunner::NextAction {
              work(executor::TaskExecutor::CallbackArgs(nullptr, {}, status, opCtx));
              return TaskRunner::NextAction::kDisposeOperationContext;
          };
          _dbWorkTaskRunner.schedule(task);
          return executor::TaskExecutor::CallbackHandle();
      }),
      _progressMeter(1U,  // total will be replaced with count command result.
                     kProgressMeterSecondsBetween,
                     kProgressMeterCheckInterval,
                     "documents copied",
                     str::stream() << _sourceNss.toString() << " collection clone progress") {
    // Fetcher throws an exception on null executor.
    invariant(executor);
    uassert(ErrorCodes::BadValue,
            "invalid collection namespace: " + sourceNss.ns(),
            sourceNss.isValid());
    uassertStatusOK(options.validate());
    uassert(ErrorCodes::BadValue, "callback function cannot be null", onCompletion);
    uassert(ErrorCodes::BadValue, "storage interface cannot be null", storageInterface);
    _stats.ns = _sourceNss.ns();
}

CollectionCloner::~CollectionCloner() {
    DESTRUCTOR_GUARD(shutdown(); join(););
}

const NamespaceString& CollectionCloner::getSourceNamespace() const {
    return _sourceNss;
}

std::string CollectionCloner::getDiagnosticString() const {
    LockGuard lk(_mutex);
    str::stream output;
    output << "CollectionCloner";
    output << " executor: " << _executor->getDiagnosticString();
    output << " source: " << _source.toString();
    output << " source namespace: " << _sourceNss.toString();
    output << " destination namespace: " << _destNss.toString();
    output << " collection options: " << _options.toBSON();
    output << " active: " << _isActive_inlock();
    output << " listIndexes fetcher: " << _listIndexesFetcher.getDiagnosticString();
    output << " find fetcher: " << (_findFetcher ? _findFetcher->getDiagnosticString() : "");
    return output;
}

bool CollectionCloner::isActive() const {
    LockGuard lk(_mutex);
    return _isActive_inlock();
}

bool CollectionCloner::_isActive_inlock() const {
    return State::kRunning == _state || State::kShuttingDown == _state;
}

bool CollectionCloner::_isShuttingDown() const {
    stdx::lock_guard<stdx::mutex> lock(_mutex);
    return State::kShuttingDown == _state;
}

Status CollectionCloner::startup() noexcept {
    LockGuard lk(_mutex);
    LOG(0) << "CollectionCloner::start called, on ns:" << _destNss;

    switch (_state) {
        case State::kPreStart:
            _state = State::kRunning;
            break;
        case State::kRunning:
            return Status(ErrorCodes::InternalError, "collection cloner already started");
        case State::kShuttingDown:
            return Status(ErrorCodes::ShutdownInProgress, "collection cloner shutting down");
        case State::kComplete:
            return Status(ErrorCodes::ShutdownInProgress, "collection cloner completed");
    }

    _stats.start = _executor->now();
    Status scheduleResult = _countScheduler.startup();
    if (!scheduleResult.isOK()) {
        _state = State::kComplete;
        return scheduleResult;
    }

    return Status::OK();
}

void CollectionCloner::shutdown() {
    stdx::lock_guard<stdx::mutex> lock(_mutex);
    switch (_state) {
        case State::kPreStart:
            // Transition directly from PreStart to Complete if not started yet.
            _state = State::kComplete;
            return;
        case State::kRunning:
            _state = State::kShuttingDown;
            break;
        case State::kShuttingDown:
        case State::kComplete:
            // Nothing to do if we are already in ShuttingDown or Complete state.
            return;
    }

    _cancelRemainingWork_inlock();
}

void CollectionCloner::_cancelRemainingWork_inlock() {
    _countScheduler.shutdown();
    _listIndexesFetcher.shutdown();
    if (_findFetcher) {
        _findFetcher->shutdown();
    }
    _dbWorkTaskRunner.cancel();
}

CollectionCloner::Stats CollectionCloner::getStats() const {
    stdx::unique_lock<stdx::mutex> lk(_mutex);
    return _stats;
}

void CollectionCloner::join() {
    stdx::unique_lock<stdx::mutex> lk(_mutex);
    _condition.wait(lk, [this]() { return !_isActive_inlock(); });
}

void CollectionCloner::waitForDbWorker() {
    if (!isActive()) {
        return;
    }
    _dbWorkTaskRunner.join();
}

void CollectionCloner::setScheduleDbWorkFn_forTest(const ScheduleDbWorkFn& scheduleDbWorkFn) {
    LockGuard lk(_mutex);
    _scheduleDbWorkFn = scheduleDbWorkFn;
}

void CollectionCloner::_countCallback(
    const executor::TaskExecutor::RemoteCommandCallbackArgs& args) {

    // No need to reword status reason in the case of cancellation.
    if (ErrorCodes::CallbackCanceled == args.response.status) {
        _finishCallback(args.response.status);
        return;
    }

    if (!args.response.status.isOK()) {
        _finishCallback({args.response.status.code(),
                         str::stream() << "During count call on collection '" << _sourceNss.ns()
                                       << "' from "
                                       << _source.toString()
                                       << ", there was an error '"
                                       << args.response.status.reason()
                                       << "'"});
        return;
    }

    Status commandStatus = getStatusFromCommandResult(args.response.data);
    if (!commandStatus.isOK()) {
        _finishCallback({commandStatus.code(),
                         str::stream() << "During count call on collection '" << _sourceNss.ns()
                                       << "' from "
                                       << _source.toString()
                                       << ", there was a command error '"
                                       << commandStatus.reason()
                                       << "'"});
        return;
    }

    long long count = 0;
    auto countStatus =
        bsonExtractIntegerField(args.response.data, kCountResponseDocumentCountFieldName, &count);
    if (!countStatus.isOK()) {
        _finishCallback({countStatus.code(),
                         str::stream() << "There was an error parsing document count from count "
                                          "command result on collection "
                                       << _sourceNss.ns()
                                       << " from "
                                       << _source.toString()
                                       << ": "
                                       << countStatus.reason()});
        return;
    }

    if (count < 0) {
        _finishCallback({ErrorCodes::BadValue,
                         str::stream() << "Count call on collection " << _sourceNss.ns() << " from "
                                       << _source.toString()
                                       << " returned negative document count: "
                                       << count});
        return;
    }

    {
        LockGuard lk(_mutex);
        _stats.documentToCopy = count;
        _progressMeter.setTotalWhileRunning(static_cast<unsigned long long>(count));
    }

    auto scheduleStatus = _listIndexesFetcher.schedule();
    if (!scheduleStatus.isOK()) {
        _finishCallback(scheduleStatus);
        return;
    }
}

void CollectionCloner::_listIndexesCallback(const Fetcher::QueryResponseStatus& fetchResult,
                                            Fetcher::NextAction* nextAction,
                                            BSONObjBuilder* getMoreBob) {
    const bool collectionIsEmpty = fetchResult == ErrorCodes::NamespaceNotFound;
    if (collectionIsEmpty) {
        // Schedule collection creation and finish callback.
        auto&& scheduleResult =
            _scheduleDbWorkFn([this](const executor::TaskExecutor::CallbackArgs& cbd) {
                if (!cbd.status.isOK()) {
                    _finishCallback(cbd.status);
                    return;
                }
                auto opCtx = cbd.opCtx;
                opCtx->setReplicatedWrites(false);
                auto&& createStatus =
                    _storageInterface->createCollection(opCtx, _destNss, _options);
                _finishCallback(createStatus);
            });
        if (!scheduleResult.isOK()) {
            _finishCallback(scheduleResult.getStatus());
        }
        return;
    };
    if (!fetchResult.isOK()) {
        Status newStatus{fetchResult.getStatus().code(),
                         str::stream() << "During listIndexes call on collection '"
                                       << _sourceNss.ns()
                                       << "' there was an error '"
                                       << fetchResult.getStatus().reason()
                                       << "'"};

        _finishCallback(newStatus);
        return;
    }

    auto batchData(fetchResult.getValue());
    auto&& documents = batchData.documents;

    if (documents.empty()) {
        warning() << "No indexes found for collection " << _sourceNss.ns() << " while cloning from "
                  << _source;
    }

    UniqueLock lk(_mutex);
    // We may be called with multiple batches leading to a need to grow _indexSpecs.
    _indexSpecs.reserve(_indexSpecs.size() + documents.size());
    for (auto&& doc : documents) {
        if (StringData("_id_") == doc["name"].str()) {
            _idIndexSpec = doc;
            continue;
        }
        _indexSpecs.push_back(doc);
    }
    lk.unlock();

    // The fetcher will continue to call with kGetMore until an error or the last batch.
    if (*nextAction == Fetcher::NextAction::kGetMore) {
        invariant(getMoreBob);
        getMoreBob->append("getMore", batchData.cursorId);
        getMoreBob->append("collection", batchData.nss.coll());
        return;
    }

    // We have all of the indexes now, so we can start cloning the collection data.
    auto&& scheduleResult = _scheduleDbWorkFn(
        stdx::bind(&CollectionCloner::_beginCollectionCallback, this, stdx::placeholders::_1));
    if (!scheduleResult.isOK()) {
        _finishCallback(scheduleResult.getStatus());
        return;
    }
}

void CollectionCloner::_findCallback(const StatusWith<Fetcher::QueryResponse>& fetchResult,
                                     Fetcher::NextAction* nextAction,
                                     BSONObjBuilder* getMoreBob,
                                     std::shared_ptr<OnCompletionGuard> onCompletionGuard) {
    if (!fetchResult.isOK()) {
        // Wait for active inserts to complete.
        waitForDbWorker();

        Status newStatus{fetchResult.getStatus().code(),
                         str::stream() << "While querying collection '" << _sourceNss.ns()
                                       << "' there was an error '"
                                       << fetchResult.getStatus().reason()
                                       << "'"};

        stdx::lock_guard<stdx::mutex> lock(_mutex);
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lock, newStatus);
        return;
    }

    auto batchData(fetchResult.getValue());
    bool lastBatch = *nextAction == Fetcher::NextAction::kNoAction;
    if (batchData.documents.size() > 0) {
        LockGuard lk(_mutex);
        _documents.insert(_documents.end(), batchData.documents.begin(), batchData.documents.end());
    } else if (!batchData.first) {
        warning() << "No documents returned in batch; ns: " << _sourceNss
                  << ", cursorId:" << batchData.cursorId << ", isLastBatch:" << lastBatch;
    }

    auto&& scheduleResult =
        _scheduleDbWorkFn(stdx::bind(&CollectionCloner::_insertDocumentsCallback,
                                     this,
                                     stdx::placeholders::_1,
                                     lastBatch,
                                     onCompletionGuard));
    if (!scheduleResult.isOK()) {
        Status newStatus{scheduleResult.getStatus().code(),
                         str::stream() << "While cloning collection '" << _sourceNss.ns()
                                       << "' there was an error '"
                                       << scheduleResult.getStatus().reason()
                                       << "'"};

        stdx::lock_guard<stdx::mutex> lock(_mutex);
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lock, newStatus);
        return;
    }

    if (!lastBatch) {
        invariant(getMoreBob);
        getMoreBob->append("getMore", batchData.cursorId);
        getMoreBob->append("collection", batchData.nss.coll());
    }
}

void CollectionCloner::_beginCollectionCallback(const executor::TaskExecutor::CallbackArgs& cbd) {
    if (!cbd.status.isOK()) {
        _finishCallback(cbd.status);
        return;
    }

    // This completion guard invokes _finishCallback on destruction.
    auto cancelRemainingWorkInLock = [this]() { _cancelRemainingWork_inlock(); };
    auto finishCallbackFn = [this](const Status& status) { _finishCallback(status); };
    auto onCompletionGuard =
        std::make_shared<OnCompletionGuard>(cancelRemainingWorkInLock, finishCallbackFn);

    // Lock guard must be declared after completion guard. If there is an error in this function
    // that will cause the destructor of the completion guard to run, the destructor must be run
    // outside the mutex. This is a necessary condition to invoke _finishCallback.
    stdx::lock_guard<stdx::mutex> lock(_mutex);
    if (!_idIndexSpec.isEmpty() && _options.autoIndexId == CollectionOptions::NO) {
        warning()
            << "Found the _id_ index spec but the collection specified autoIndexId of false on ns:"
            << this->_sourceNss;
    }

    auto status = _storageInterface->createCollectionForBulkLoading(
        _destNss, _options, _idIndexSpec, _indexSpecs);

    if (!status.isOK()) {
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lock, status.getStatus());
        return;
    }

    _stats.indexes = _indexSpecs.size();
    if (!_idIndexSpec.isEmpty()) {
        ++_stats.indexes;
    }

    _collLoader = std::move(status.getValue());

    _findFetcher = stdx::make_unique<Fetcher>(
        _executor,
        _source,
        _sourceNss.db().toString(),
        // noCursorTimeout true, large batchSize (for older server versions to get larger batch)
        BSON("find" << _sourceNss.coll() << "noCursorTimeout" << true << "batchSize" << batchSize),
        stdx::bind(&CollectionCloner::_findCallback,
                   this,
                   stdx::placeholders::_1,
                   stdx::placeholders::_2,
                   stdx::placeholders::_3,
                   onCompletionGuard),
        rpc::ServerSelectionMetadata(true, boost::none).toBSON(),
        RemoteCommandRequest::kNoTimeout,
        RemoteCommandRetryScheduler::makeRetryPolicy(
            numInitialSyncCollectionFindAttempts.load(),
            executor::RemoteCommandRequest::kNoTimeout,
            RemoteCommandRetryScheduler::kAllRetriableErrors));

    Status scheduleStatus = _findFetcher->schedule();
    if (!scheduleStatus.isOK()) {
        _findFetcher.reset();
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lock, scheduleStatus);
        return;
    }
}

void CollectionCloner::_insertDocumentsCallback(
    const executor::TaskExecutor::CallbackArgs& cbd,
    bool lastBatch,
    std::shared_ptr<OnCompletionGuard> onCompletionGuard) {
    if (!cbd.status.isOK()) {
        stdx::lock_guard<stdx::mutex> lock(_mutex);
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lock, cbd.status);
        return;
    }

    std::vector<BSONObj> docs;
    UniqueLock lk(_mutex);
    if (_documents.size() == 0) {
        warning() << "_insertDocumentsCallback, but no documents to insert for ns:" << _destNss;

        if (lastBatch) {
            onCompletionGuard->setResultAndCancelRemainingWork_inlock(lk, Status::OK());
        }
        return;
    }

    _documents.swap(docs);
    _stats.documentsCopied += docs.size();
    ++_stats.fetchBatches;
    _progressMeter.hit(int(docs.size()));
    invariant(_collLoader);
    const auto status = _collLoader->insertDocuments(docs.cbegin(), docs.cend());

    if (!status.isOK()) {
        onCompletionGuard->setResultAndCancelRemainingWork_inlock(lk, status);
        return;
    }

    MONGO_FAIL_POINT_BLOCK(initialSyncHangDuringCollectionClone, options) {
        const BSONObj& data = options.getData();
        if (data["namespace"].String() == _destNss.ns() &&
            static_cast<int>(_stats.documentsCopied) >= data["numDocsToClone"].numberInt()) {
            lk.unlock();
            log() << "initial sync - initialSyncHangDuringCollectionClone fail point "
                     "enabled. Blocking until fail point is disabled.";
            while (MONGO_FAIL_POINT(initialSyncHangDuringCollectionClone) && !_isShuttingDown()) {
                mongo::sleepsecs(1);
            }
            lk.lock();
        }
    }

    if (!lastBatch) {
        return;
    }

    // Done with last batch and time to set result in completion guard to Status::OK().
    onCompletionGuard->setResultAndCancelRemainingWork_inlock(lk, Status::OK());
}

void CollectionCloner::_finishCallback(const Status& status) {
    log() << "CollectionCloner ns:" << _destNss
          << " finished cloning with status: " << redact(status);
    // Copy the status so we can change it below if needed.
    auto finalStatus = status;
    bool callCollectionLoader = false;

    decltype(_onCompletion) onCompletion;
    {
        LockGuard lk(_mutex);
        invariant(_state != State::kComplete);

        callCollectionLoader = _collLoader.operator bool();

        invariant(_onCompletion);
        std::swap(_onCompletion, onCompletion);
    }

    if (callCollectionLoader) {
        if (finalStatus.isOK()) {
            const auto loaderStatus = _collLoader->commit();
            if (!loaderStatus.isOK()) {
                warning() << "Failed to commit collection indexes " << _destNss.ns() << ": "
                          << redact(loaderStatus);
                finalStatus = loaderStatus;
            }
        }

        // This will release the resources held by the loader.
        _collLoader.reset();
    }

    onCompletion(finalStatus);

    // This will release the resources held by the callback function object. '_onCompletion' is
    // already cleared at this point and 'onCompletion' is the remaining reference to the callback
    // function (with any implicitly held resources). To avoid any issues with destruction logic
    // in the function object's resources accessing this CollectionCloner, we release this function
    // object outside the lock.
    onCompletion = {};

    LockGuard lk(_mutex);
    _stats.end = _executor->now();
    _progressMeter.finished();
    _state = State::kComplete;
    _condition.notify_all();
    LOG(1) << "    collection: " << _destNss << ", stats: " << _stats.toString();
}

constexpr StringData CollectionCloner::Stats::kDocumentsToCopyFieldName;
constexpr StringData CollectionCloner::Stats::kDocumentsCopiedFieldName;

std::string CollectionCloner::Stats::toString() const {
    return toBSON().toString();
}

BSONObj CollectionCloner::Stats::toBSON() const {
    BSONObjBuilder bob;
    bob.append("ns", ns);
    append(&bob);
    return bob.obj();
}

void CollectionCloner::Stats::append(BSONObjBuilder* builder) const {
    builder->appendNumber(kDocumentsToCopyFieldName, documentToCopy);
    builder->appendNumber(kDocumentsCopiedFieldName, documentsCopied);
    builder->appendNumber("indexes", indexes);
    builder->appendNumber("fetchedBatches", fetchBatches);
    if (start != Date_t()) {
        builder->appendDate("start", start);
        if (end != Date_t()) {
            builder->appendDate("end", end);
            auto elapsed = end - start;
            long long elapsedMillis = duration_cast<Milliseconds>(elapsed).count();
            builder->appendNumber("elapsedMillis", elapsedMillis);
        }
    }
}
}  // namespace repl
}  // namespace mongo