summaryrefslogtreecommitdiff
path: root/src/mongo/db/ttl.cpp
blob: 0896b5708bdc19e52cea6764edbd7b90306ada4c (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
// ttl.cpp

/**
*    Copyright (C) 2008 10gen 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::kIndex

#include "mongo/platform/basic.h"

#include "mongo/db/ttl.h"

#include "mongo/base/counter.h"
#include "mongo/db/auth/authorization_session.h"
#include "mongo/db/auth/user_name.h"
#include "mongo/db/catalog/collection.h"
#include "mongo/db/catalog/collection_catalog_entry.h"
#include "mongo/db/catalog/database_catalog_entry.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/catalog/index_catalog.h"
#include "mongo/db/client.h"
#include "mongo/db/commands/fsync_locked.h"
#include "mongo/db/commands/server_status_metric.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/db_raii.h"
#include "mongo/db/exec/delete.h"
#include "mongo/db/index/index_descriptor.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/ops/insert.h"
#include "mongo/db/query/internal_plans.h"
#include "mongo/db/repl/replication_coordinator.h"
#include "mongo/db/server_parameters.h"
#include "mongo/db/ttl_collection_cache.h"
#include "mongo/util/background.h"
#include "mongo/util/concurrency/idle_thread_block.h"
#include "mongo/util/exit.h"
#include "mongo/util/log.h"

namespace mongo {

Counter64 ttlPasses;
Counter64 ttlDeletedDocuments;

ServerStatusMetricField<Counter64> ttlPassesDisplay("ttl.passes", &ttlPasses);
ServerStatusMetricField<Counter64> ttlDeletedDocumentsDisplay("ttl.deletedDocuments",
                                                              &ttlDeletedDocuments);

MONGO_EXPORT_SERVER_PARAMETER(ttlMonitorEnabled, bool, true);
MONGO_EXPORT_SERVER_PARAMETER(ttlMonitorSleepSecs, int, 60)
    ->withValidator([](const int& newVal) {
        if (newVal <= 0)
            return Status(ErrorCodes::BadValue, "ttlMonitorSleepSecs must be strictly positive");
        return Status::OK();
    });  // used for testing

class TTLMonitor : public BackgroundJob {
public:
    TTLMonitor() {}
    virtual ~TTLMonitor() {}

    virtual std::string name() const {
        return "TTLMonitor";
    }

    static std::string secondsExpireField;

    virtual void run() {
        Client::initThread(name().c_str());
        ON_BLOCK_EXIT([] { Client::destroy(); });
        AuthorizationSession::get(cc())->grantInternalAuthorization();

        while (!globalInShutdownDeprecated()) {
            {
                MONGO_IDLE_THREAD_BLOCK;
                sleepsecs(ttlMonitorSleepSecs.load());
            }

            LOG(3) << "thread awake";

            if (!ttlMonitorEnabled.load()) {
                LOG(1) << "disabled";
                continue;
            }

            if (lockedForWriting()) {
                // Note: this is not perfect as you can go into fsync+lock between this and actually
                // doing the delete later.
                LOG(3) << "locked for writing";
                continue;
            }

            try {
                doTTLPass();
            } catch (const WriteConflictException&) {
                LOG(1) << "got WriteConflictException";
            }
        }
    }

private:
    void doTTLPass() {
        const ServiceContext::UniqueOperationContext opCtxPtr = cc().makeOperationContext();
        OperationContext& opCtx = *opCtxPtr;

        // If part of replSet but not in a readable state (e.g. during initial sync), skip.
        if (repl::ReplicationCoordinator::get(&opCtx)->getReplicationMode() ==
                repl::ReplicationCoordinator::modeReplSet &&
            !repl::ReplicationCoordinator::get(&opCtx)->getMemberState().readable())
            return;

        TTLCollectionCache& ttlCollectionCache = TTLCollectionCache::get(getGlobalServiceContext());
        std::vector<std::string> ttlCollections = ttlCollectionCache.getCollections();
        std::vector<BSONObj> ttlIndexes;

        ttlPasses.increment();

        // Get all TTL indexes from every collection.
        for (const std::string& collectionNS : ttlCollections) {
            UninterruptibleLockGuard noInterrupt(opCtx.lockState());
            NamespaceString collectionNSS(collectionNS);
            AutoGetCollection autoGetCollection(&opCtx, collectionNSS, MODE_IS);
            Collection* coll = autoGetCollection.getCollection();
            if (!coll) {
                // Skip since collection has been dropped.
                continue;
            }

            CollectionCatalogEntry* collEntry = coll->getCatalogEntry();
            std::vector<std::string> indexNames;
            collEntry->getAllIndexes(&opCtx, &indexNames);
            for (const std::string& name : indexNames) {
                BSONObj spec = collEntry->getIndexSpec(&opCtx, name);
                if (spec.hasField(secondsExpireField)) {
                    ttlIndexes.push_back(spec.getOwned());
                }
            }
        }

        for (const BSONObj& idx : ttlIndexes) {
            try {
                doTTLForIndex(&opCtx, idx);
            } catch (const DBException& dbex) {
                error() << "Error processing ttl index: " << idx << " -- " << dbex.toString();
                // Continue on to the next index.
                continue;
            }
        }
    }

    /**
     * Remove documents from the collection using the specified TTL index after a sufficient amount
     * of time has passed according to its expiry specification.
     */
    void doTTLForIndex(OperationContext* opCtx, BSONObj idx) {
        const NamespaceString collectionNSS(idx["ns"].String());
        if (collectionNSS.isDropPendingNamespace()) {
            return;
        }
        if (!userAllowedWriteNS(collectionNSS).isOK()) {
            error() << "namespace '" << collectionNSS
                    << "' doesn't allow deletes, skipping ttl job for: " << idx;
            return;
        }

        const BSONObj key = idx["key"].Obj();
        const StringData name = idx["name"].valueStringData();
        if (key.nFields() != 1) {
            error() << "key for ttl index can only have 1 field, skipping ttl job for: " << idx;
            return;
        }

        LOG(1) << "ns: " << collectionNSS << " key: " << key << " name: " << name;

        AutoGetCollection autoGetCollection(opCtx, collectionNSS, MODE_IX);
        Collection* collection = autoGetCollection.getCollection();
        if (!collection) {
            // Collection was dropped.
            return;
        }

        if (!repl::ReplicationCoordinator::get(opCtx)->canAcceptWritesFor(opCtx, collectionNSS)) {
            return;
        }

        IndexDescriptor* desc = collection->getIndexCatalog()->findIndexByName(opCtx, name);
        if (!desc) {
            LOG(1) << "index not found (index build in progress? index dropped?), skipping "
                   << "ttl job for: " << idx;
            return;
        }

        // Re-read 'idx' from the descriptor, in case the collection or index definition changed
        // before we re-acquired the collection lock.
        idx = desc->infoObj();

        if (IndexType::INDEX_BTREE != IndexNames::nameToType(desc->getAccessMethodName())) {
            error() << "special index can't be used as a ttl index, skipping ttl job for: " << idx;
            return;
        }

        BSONElement secondsExpireElt = idx[secondsExpireField];
        if (!secondsExpireElt.isNumber()) {
            error() << "ttl indexes require the " << secondsExpireField << " field to be "
                    << "numeric but received a type of " << typeName(secondsExpireElt.type())
                    << ", skipping ttl job for: " << idx;
            return;
        }

        const Date_t kDawnOfTime =
            Date_t::fromMillisSinceEpoch(std::numeric_limits<long long>::min());
        const Date_t expirationTime = Date_t::now() - Seconds(secondsExpireElt.numberLong());
        const BSONObj startKey = BSON("" << kDawnOfTime);
        const BSONObj endKey = BSON("" << expirationTime);
        // The canonical check as to whether a key pattern element is "ascending" or
        // "descending" is (elt.number() >= 0).  This is defined by the Ordering class.
        const InternalPlanner::Direction direction = (key.firstElement().number() >= 0)
            ? InternalPlanner::Direction::FORWARD
            : InternalPlanner::Direction::BACKWARD;

        // We need to pass into the DeleteStageParams (below) a CanonicalQuery with a BSONObj that
        // queries for the expired documents correctly so that we do not delete documents that are
        // not actually expired when our snapshot changes during deletion.
        const char* keyFieldName = key.firstElement().fieldName();
        BSONObj query =
            BSON(keyFieldName << BSON("$gte" << kDawnOfTime << "$lte" << expirationTime));
        auto qr = stdx::make_unique<QueryRequest>(collectionNSS);
        qr->setFilter(query);
        auto canonicalQuery = CanonicalQuery::canonicalize(opCtx, std::move(qr));
        invariant(canonicalQuery.getStatus());

        DeleteStageParams params;
        params.isMulti = true;
        params.canonicalQuery = canonicalQuery.getValue().get();

        auto exec =
            InternalPlanner::deleteWithIndexScan(opCtx,
                                                 collection,
                                                 params,
                                                 desc,
                                                 startKey,
                                                 endKey,
                                                 BoundInclusion::kIncludeBothStartAndEndKeys,
                                                 PlanExecutor::YIELD_AUTO,
                                                 direction);

        Status result = exec->executePlan();
        if (!result.isOK()) {
            error() << "ttl query execution for index " << idx
                    << " failed with status: " << redact(result);
            return;
        }

        const long long numDeleted = DeleteStage::getNumDeleted(*exec);
        ttlDeletedDocuments.increment(numDeleted);
        LOG(1) << "deleted: " << numDeleted;
    }
};

namespace {
// The global TTLMonitor object is intentionally leaked.  Even though it is only used in one
// function, we declare it here to indicate to the leak sanitizer that the leak of this object
// should not be reported.
TTLMonitor* ttlMonitor = nullptr;
}  // namespace

void startTTLBackgroundJob() {
    ttlMonitor = new TTLMonitor();
    ttlMonitor->go();
}

std::string TTLMonitor::secondsExpireField = "expireAfterSeconds";
}  // namespace mongo