summaryrefslogtreecommitdiff
path: root/src/mongo/db/repair_database.cpp
blob: 80963f123b7fb9649f8845a1f622db79d79c0069 (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
/**
*    Copyright (C) 2014 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::kStorage

#include "mongo/platform/basic.h"

#include "mongo/db/repair_database.h"


#include "mongo/db/background.h"
#include "mongo/base/status.h"
#include "mongo/base/string_data.h"
#include "mongo/bson/bson_validate.h"
#include "mongo/db/catalog/collection.h"
#include "mongo/db/catalog/collection_catalog_entry.h"
#include "mongo/db/catalog/database.h"
#include "mongo/db/catalog/database_catalog_entry.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/document_validation.h"
#include "mongo/db/catalog/index_create.h"
#include "mongo/db/catalog/index_key_validate.h"
#include "mongo/db/storage/mmap_v1/mmap_v1_engine.h"
#include "mongo/db/storage/storage_engine.h"
#include "mongo/util/log.h"

namespace mongo {

    using std::endl;
    using std::string;

namespace {
    Status rebuildIndexesOnCollection(OperationContext* txn,
                                      DatabaseCatalogEntry* dbce,
                                      const std::string& collectionName) {

        CollectionCatalogEntry* cce = dbce->getCollectionCatalogEntry(collectionName);

        std::vector<string> indexNames;
        std::vector<BSONObj> indexSpecs;
        {
            // Fetch all indexes
            cce->getAllIndexes( txn, &indexNames );
            for ( size_t i = 0; i < indexNames.size(); i++ ) {
                const string& name = indexNames[i];
                BSONObj spec = cce->getIndexSpec( txn, name );
                indexSpecs.push_back(spec.removeField("v").getOwned());

                const BSONObj key = spec.getObjectField("key");
                const Status keyStatus = validateKeyPattern(key);
                if (!keyStatus.isOK()) {
                    return Status(ErrorCodes::CannotCreateIndex, str::stream()
                        << "Cannot rebuild index " << spec << ": " << keyStatus.reason()
                        << " For more info see http://dochub.mongodb.org/core/index-validation");
                }
            }
        }

        // Skip the rest if there are no indexes to rebuild.
        if (indexSpecs.empty()) return Status::OK();

        std::unique_ptr<Collection> collection;
        std::unique_ptr<MultiIndexBlock> indexer;
        {
            // These steps are combined into a single WUOW to ensure there are no commits without
            // the indexes.
            // 1) Drop all indexes.
            // 2) Open the Collection
            // 3) Start the index build process.

            WriteUnitOfWork wuow(txn);

            { // 1
                for ( size_t i = 0; i < indexNames.size(); i++ ) {
                    Status s = cce->removeIndex(txn, indexNames[i]);
                    if (!s.isOK()) return s;
                }
            }

            // Indexes must be dropped before we open the Collection otherwise we could attempt to
            // open a bad index and fail.
            // TODO see if MultiIndexBlock can be made to work without a Collection.
            const StringData ns = cce->ns().ns();
            collection.reset(new Collection(txn, ns, cce, dbce->getRecordStore(ns), dbce));

            indexer.reset(new MultiIndexBlock(txn, collection.get()));
            Status status = indexer->init(indexSpecs);
            if (!status.isOK()) {
                // The WUOW will handle cleanup, so the indexer shouldn't do its own.
                indexer->abortWithoutCleanup();
                return status;
            }

            wuow.commit();
        }

        // Iterate all records in the collection. Delete them if they aren't valid BSON. Index them
        // if they are.

        long long numRecords = 0;
        long long dataSize = 0;

        RecordStore* rs = collection->getRecordStore();
        auto cursor = rs->getCursor(txn);
        while (auto record = cursor->next()) {
            RecordId id = record->id;
            RecordData& data = record->data;

            Status status = validateBSON(data.data(), data.size());
            if (!status.isOK()) {
                log() << "Invalid BSON detected at " << id << ": " << status << ". Deleting.";
                cursor->savePositioned(); // 'data' is no longer valid.
                {
                    WriteUnitOfWork wunit(txn);
                    rs->deleteRecord(txn, id);
                    wunit.commit();
                }
                cursor->restore(txn);
                continue;
            }

            numRecords++;
            dataSize += data.size();

            // Now index the record.
            // TODO SERVER-14812 add a mode that drops duplicates rather than failing
            WriteUnitOfWork wunit(txn);
            status = indexer->insert(data.releaseToBson(), id);
            if (!status.isOK()) return status;
            wunit.commit();
        }

        Status status = indexer->doneInserting();
        if (!status.isOK()) return status;

        {
            WriteUnitOfWork wunit(txn);
            indexer->commit();
            rs->updateStatsAfterRepair(txn, numRecords, dataSize);
            wunit.commit();
        }

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

    Status repairDatabase(OperationContext* txn,
                          StorageEngine* engine,
                          const std::string& dbName,
                          bool preserveClonedFilesOnFailure,
                          bool backupOriginalFiles) {

        DisableDocumentValidation validationDisabler(txn);

        // We must hold some form of lock here
        invariant(txn->lockState()->isLocked());
        invariant( dbName.find( '.' ) == string::npos );

        log() << "repairDatabase " << dbName << endl;

        BackgroundOperation::assertNoBgOpInProgForDb(dbName);

        txn->checkForInterrupt();

        if (engine->isMmapV1()) {
            // MMAPv1 is a layering violation so it implements its own repairDatabase.
            return static_cast<MMAPV1Engine*>(engine)->repairDatabase(txn,
                                                                      dbName,
                                                                      preserveClonedFilesOnFailure,
                                                                      backupOriginalFiles);
        }

        // These are MMAPv1 specific
        if ( preserveClonedFilesOnFailure ) {
            return Status( ErrorCodes::BadValue, "preserveClonedFilesOnFailure not supported" );
        }
        if ( backupOriginalFiles ) {
            return Status( ErrorCodes::BadValue, "backupOriginalFiles not supported" );
        }

        // Close the db to invalidate all current users and caches.
        dbHolder().close(txn, dbName);
        // Open the db after everything finishes
        class OpenDbInDestructor {
        public:
            OpenDbInDestructor(OperationContext* txn, const std::string& db) :
                  _dbName(db)
                , _txn(txn)
            {}
            ~OpenDbInDestructor() {
                dbHolder().openDb(_txn, _dbName);
            }
        private:
            const std::string& _dbName;
            OperationContext* _txn;
        } dbOpener(txn, dbName);
        DatabaseCatalogEntry* dbce = engine->getDatabaseCatalogEntry(txn, dbName);

        std::list<std::string> colls;
        dbce->getCollectionNamespaces(&colls);

        for (std::list<std::string>::const_iterator it = colls.begin(); it != colls.end(); ++it) {
            // Don't check for interrupt after starting to repair a collection otherwise we can
            // leave data in an inconsistent state. Interrupting between collections is ok, however.
            txn->checkForInterrupt();

            log() << "Repairing collection " << *it;

            Status status = engine->repairRecordStore(txn, *it);
            if (!status.isOK()) return status;

            status = rebuildIndexesOnCollection(txn, dbce, *it);
            if (!status.isOK()) return status;

            // TODO: uncomment once SERVER-16869
            // engine->flushAllFiles(true);
        }

        return Status::OK();
    }
}