summaryrefslogtreecommitdiff
path: root/src/mongo/db/catalog/drop_collection.cpp
blob: bb087d0bf8d50a1a58bc972ad2ee3884658a86b6 (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
/**
 *    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/drop_collection.h"

#include "mongo/db/audit.h"
#include "mongo/db/catalog/index_catalog.h"
#include "mongo/db/catalog/uncommitted_collections.h"
#include "mongo/db/client.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/curop.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/index_builds_coordinator.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/s/collection_sharding_state.h"
#include "mongo/db/server_options.h"
#include "mongo/db/service_context.h"
#include "mongo/db/views/view_catalog.h"
#include "mongo/logv2/log.h"
#include "mongo/util/fail_point.h"

namespace mongo {

MONGO_FAIL_POINT_DEFINE(hangDropCollectionBeforeLockAcquisition);
MONGO_FAIL_POINT_DEFINE(hangDuringDropCollection);

Status _checkNssAndReplState(OperationContext* opCtx, const CollectionPtr& coll) {
    if (!coll) {
        return Status(ErrorCodes::NamespaceNotFound, "ns not found");
    }

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

    return Status::OK();
}

Status _dropView(OperationContext* opCtx,
                 Database* db,
                 const NamespaceString& collectionName,
                 DropReply* reply) {
    if (!db) {
        Status status = Status(ErrorCodes::NamespaceNotFound, "ns not found");
        audit::logDropView(opCtx->getClient(), collectionName, "", {}, status.code());
        return status;
    }
    auto view = ViewCatalog::get(db)->lookupWithoutValidatingDurableViews(opCtx, collectionName);
    if (!view) {
        Status status = Status(ErrorCodes::NamespaceNotFound, "ns not found");
        audit::logDropView(opCtx->getClient(), collectionName, "", {}, status.code());
        return status;
    }

    // Validates the view or throws an "invalid view" error.
    ViewCatalog::get(db)->lookup(opCtx, collectionName);

    // Take a MODE_X lock when dropping a view. This is to prevent a concurrent create
    // collection on the same namespace that will reserve an OpTime before this drop.
    Lock::CollectionLock collLock(opCtx, collectionName, MODE_X);
    // Operations all lock system.views in the end to prevent deadlock.
    Lock::CollectionLock systemViewsLock(opCtx, db->getSystemViewsName(), MODE_X);

    if (MONGO_unlikely(hangDuringDropCollection.shouldFail())) {
        LOGV2(20330,
              "hangDuringDropCollection fail point enabled. Blocking until fail point is "
              "disabled.");
        hangDuringDropCollection.pauseWhileSet();
    }

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

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

    WriteUnitOfWork wunit(opCtx);

    audit::logDropView(
        opCtx->getClient(), collectionName, view->viewOn().ns(), view->pipeline(), ErrorCodes::OK);

    Status status = db->dropView(opCtx, collectionName);
    if (!status.isOK()) {
        return status;
    }
    wunit.commit();

    reply->setNs(collectionName);
    return Status::OK();
}

Status _abortIndexBuildsAndDrop(OperationContext* opCtx,
                                AutoGetDb&& autoDb,
                                const NamespaceString& startingNss,
                                std::function<Status(Database*, const NamespaceString&)>&& dropFn,
                                DropReply* reply,
                                bool appendNs = true,
                                boost::optional<UUID> dropIfUUIDNotMatching = boost::none) {
    // We only need to hold an intent lock to send abort signals to the active index builder on this
    // collection.
    boost::optional<AutoGetDb> optionalAutoDb(std::move(autoDb));
    boost::optional<Lock::CollectionLock> collLock;
    collLock.emplace(opCtx, startingNss, MODE_IX);

    // Abandon the snapshot as the index catalog will compare the in-memory state to the disk state,
    // which may have changed when we released the collection lock temporarily.
    opCtx->recoveryUnit()->abandonSnapshot();

    CollectionPtr coll =
        CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, startingNss);
    Status status = _checkNssAndReplState(opCtx, coll);
    if (!status.isOK()) {
        return status;
    }

    if (MONGO_unlikely(hangDuringDropCollection.shouldFail())) {
        LOGV2(518090,
              "hangDuringDropCollection fail point enabled. Blocking until fail point is "
              "disabled.");
        hangDuringDropCollection.pauseWhileSet();
    }

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

    IndexBuildsCoordinator* indexBuildsCoord = IndexBuildsCoordinator::get(opCtx);
    const UUID collectionUUID = coll->uuid();
    if (dropIfUUIDNotMatching && collectionUUID == *dropIfUUIDNotMatching) {
        return Status::OK();
    }
    const NamespaceStringOrUUID dbAndUUID{coll->ns().db().toString(), coll->uuid()};
    const int numIndexes = coll->getIndexCatalog()->numIndexesTotal(opCtx);

    while (true) {
        // Save a copy of the namespace before yielding our locks.
        const NamespaceString collectionNs = coll->ns();

        // Release locks before aborting index builds. The helper will acquire locks on our behalf.
        collLock = boost::none;
        optionalAutoDb = boost::none;

        // Send the abort signal to any active index builds on the collection. This waits until all
        // aborted index builds complete.
        indexBuildsCoord->abortCollectionIndexBuilds(opCtx,
                                                     collectionNs,
                                                     collectionUUID,
                                                     str::stream()
                                                         << "Collection " << collectionNs << "("
                                                         << collectionUUID << ") is being dropped");

        // Take an exclusive lock to finish the collection drop.
        optionalAutoDb.emplace(opCtx, startingNss.db(), MODE_IX);
        collLock.emplace(opCtx, dbAndUUID, MODE_X);

        // Abandon the snapshot as the index catalog will compare the in-memory state to the
        // disk state, which may have changed when we released the collection lock temporarily.
        opCtx->recoveryUnit()->abandonSnapshot();

        coll = CollectionCatalog::get(opCtx)->lookupCollectionByUUID(opCtx, collectionUUID);
        status = _checkNssAndReplState(opCtx, coll);
        if (!status.isOK()) {
            return status;
        }

        // Check if any new index builds were started while releasing the collection lock
        // temporarily, if so, we need to abort the new index builders.
        const bool abortAgain = indexBuildsCoord->inProgForCollection(collectionUUID);
        if (!abortAgain) {
            break;
        }
    }

    // It's possible for the given collection to be drop pending after obtaining the locks again, if
    // that is the case, then the collection is already registered to be dropped. Return early.
    const NamespaceString resolvedNss = coll->ns();
    if (resolvedNss.isDropPendingNamespace()) {
        return Status::OK();
    }

    // Serialize the drop with refreshes to prevent dropping a collection and creating the same
    // nss as a view while refreshing.
    CollectionShardingState::get(opCtx, resolvedNss)->checkShardVersionOrThrow(opCtx);

    invariant(coll->getIndexCatalog()->numIndexesInProgress(opCtx) == 0);

    status = dropFn(optionalAutoDb->getDb(), resolvedNss);
    if (!status.isOK()) {
        return status;
    }

    reply->setNIndexesWas(numIndexes);
    if (appendNs) {
        reply->setNs(resolvedNss);
    }

    return Status::OK();
}

Status _dropCollectionForApplyOps(OperationContext* opCtx,
                                  Database* db,
                                  const NamespaceString& collectionName,
                                  const repl::OpTime& dropOpTime,
                                  DropCollectionSystemCollectionMode systemCollectionMode,
                                  DropReply* reply) {
    Lock::CollectionLock collLock(opCtx, collectionName, MODE_X);
    const CollectionPtr& coll =
        CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, collectionName);
    Status status = _checkNssAndReplState(opCtx, coll);
    if (!status.isOK()) {
        return status;
    }

    if (MONGO_unlikely(hangDuringDropCollection.shouldFail())) {
        LOGV2(20331,
              "hangDuringDropCollection fail point enabled. Blocking until fail point is "
              "disabled.");
        hangDuringDropCollection.pauseWhileSet();
    }

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

    WriteUnitOfWork wunit(opCtx);

    int numIndexes = coll->getIndexCatalog()->numIndexesTotal(opCtx);
    IndexBuildsCoordinator::get(opCtx)->assertNoIndexBuildInProgForCollection(coll->uuid());
    status =
        systemCollectionMode == DropCollectionSystemCollectionMode::kDisallowSystemCollectionDrops
        ? db->dropCollection(opCtx, collectionName, dropOpTime)
        : db->dropCollectionEvenIfSystem(opCtx, collectionName, dropOpTime);

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

    reply->setNIndexesWas(numIndexes);
    reply->setNs(collectionName);

    return Status::OK();
}

Status _dropCollection(OperationContext* opCtx,
                       const NamespaceString& collectionName,
                       DropReply* reply,
                       DropCollectionSystemCollectionMode systemCollectionMode,
                       boost::optional<UUID> dropIfUUIDNotMatching = boost::none) {

    try {
        return writeConflictRetry(opCtx, "drop", collectionName.ns(), [&] {
            AutoGetDb autoDb(opCtx, collectionName.db(), MODE_IX);
            auto db = autoDb.getDb();
            if (!db) {
                return Status(ErrorCodes::NamespaceNotFound, "ns not found");
            }

            if (CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, collectionName)) {
                return _abortIndexBuildsAndDrop(
                    opCtx,
                    std::move(autoDb),
                    collectionName,
                    [opCtx, systemCollectionMode](Database* db, const NamespaceString& resolvedNs) {
                        WriteUnitOfWork wuow(opCtx);

                        auto status = systemCollectionMode ==
                                DropCollectionSystemCollectionMode::kDisallowSystemCollectionDrops
                            ? db->dropCollection(opCtx, resolvedNs)
                            : db->dropCollectionEvenIfSystem(opCtx, resolvedNs);
                        if (!status.isOK()) {
                            return status;
                        }

                        wuow.commit();
                        return Status::OK();
                    },
                    reply,
                    true /* appendNs */,
                    dropIfUUIDNotMatching);
            }

            auto dropTimeseries = [opCtx, &autoDb, &collectionName, &reply](
                                      const NamespaceString& bucketNs, bool dropView) {
                return _abortIndexBuildsAndDrop(
                    opCtx,
                    std::move(autoDb),
                    bucketNs,
                    [opCtx, dropView, &collectionName, &reply](Database* db,
                                                               const NamespaceString& bucketsNs) {
                        if (dropView) {
                            // We already hold a MODE_X lock on the bucket collection inside
                            // '_abortIndexBuildsAndDrop' above. Taking that MODE_X lock on the
                            // view must happen after in order to prevent a deadlock.
                            auto status = _dropView(opCtx, db, collectionName, reply);
                            if (!status.isOK()) {
                                return status;
                            }
                        }

                        // Drop the buckets collection in its own writeConflictRetry so that if
                        // it throws a WCE, only the buckets collection drop is retried.
                        writeConflictRetry(opCtx, "drop", bucketsNs.ns(), [opCtx, db, &bucketsNs] {
                            WriteUnitOfWork wuow(opCtx);
                            db->dropCollectionEvenIfSystem(opCtx, bucketsNs).ignore();
                            wuow.commit();
                        });

                        return Status::OK();
                    },
                    reply,
                    false /* appendNs */);
            };

            auto view =
                ViewCatalog::get(db)->lookupWithoutValidatingDurableViews(opCtx, collectionName);
            if (!view) {
                // Timeseries bucket collection may exist even without the view. If that is the case
                // delete it.
                auto bucketsNs = collectionName.makeTimeseriesBucketsNamespace();
                if (CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, bucketsNs)) {
                    return dropTimeseries(bucketsNs, false);
                }

                Status status = Status(ErrorCodes::NamespaceNotFound, "ns not found");
                audit::logDropView(opCtx->getClient(), collectionName, "", {}, status.code());
                return status;
            }
            if (view->timeseries() &&
                CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, view->viewOn())) {
                return dropTimeseries(view->viewOn(), true);
            }

            return _dropView(opCtx, db, collectionName, reply);
        });
    } catch (ExceptionFor<ErrorCodes::NamespaceNotFound>&) {
        // The shell requires that NamespaceNotFound error codes return the "ns not found"
        // string.
        return Status(ErrorCodes::NamespaceNotFound, "ns not found");
    }
}

Status dropCollection(OperationContext* opCtx,
                      const NamespaceString& nss,
                      DropReply* reply,
                      DropCollectionSystemCollectionMode systemCollectionMode) {
    if (!serverGlobalParams.quiet.load()) {
        LOGV2(518070, "CMD: drop", logAttrs(nss));
    }

    if (MONGO_unlikely(hangDropCollectionBeforeLockAcquisition.shouldFail())) {
        LOGV2(518080, "Hanging drop collection before lock acquisition while fail point is set");
        hangDropCollectionBeforeLockAcquisition.pauseWhileSet();
    }

    // We rewrite drop of time-series buckets collection to drop of time-series view collection.
    // This ensures that such drop will delete both collections.
    const auto collectionName =
        nss.isTimeseriesBucketsCollection() ? nss.getTimeseriesViewNamespace() : nss;

    return _dropCollection(opCtx, collectionName, reply, systemCollectionMode);
}

Status dropCollectionIfUUIDNotMatching(OperationContext* opCtx,
                                       const NamespaceString& ns,
                                       const UUID& expectedUUID) {
    AutoGetDb autoDb(opCtx, ns.db(), MODE_IX);
    if (autoDb.getDb()) {
        {
            Lock::CollectionLock collLock(opCtx, ns, MODE_IS);
            const auto coll = CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, ns);
            if (!coll || coll->uuid() == expectedUUID) {
                return Status::OK();
            }
        }

        DropReply repl;
        return _dropCollection(opCtx,
                               ns,
                               &repl,
                               DropCollectionSystemCollectionMode::kDisallowSystemCollectionDrops,
                               expectedUUID);
    }

    return Status::OK();
}

Status dropCollectionForApplyOps(OperationContext* opCtx,
                                 const NamespaceString& collectionName,
                                 const repl::OpTime& dropOpTime,
                                 DropCollectionSystemCollectionMode systemCollectionMode) {
    if (!serverGlobalParams.quiet.load()) {
        LOGV2(20332, "CMD: drop", logAttrs(collectionName));
    }

    if (MONGO_unlikely(hangDropCollectionBeforeLockAcquisition.shouldFail())) {
        LOGV2(20333, "Hanging drop collection before lock acquisition while fail point is set");
        hangDropCollectionBeforeLockAcquisition.pauseWhileSet();
    }
    return writeConflictRetry(opCtx, "drop", collectionName.ns(), [&] {
        AutoGetDb autoDb(opCtx, collectionName.db(), MODE_IX);
        Database* db = autoDb.getDb();
        if (!db) {
            return Status(ErrorCodes::NamespaceNotFound, "ns not found");
        }

        const CollectionPtr& coll =
            CollectionCatalog::get(opCtx)->lookupCollectionByNamespace(opCtx, collectionName);

        DropReply unusedReply;
        if (!coll) {
            return _dropView(opCtx, db, collectionName, &unusedReply);
        } else {
            return _dropCollectionForApplyOps(
                opCtx, db, collectionName, dropOpTime, systemCollectionMode, &unusedReply);
        }
    });
}

}  // namespace mongo