summaryrefslogtreecommitdiff
path: root/src/mongo/db/repl/replication_coordinator_external_state_impl.cpp
blob: 6f4a8330f34244fc4ea700a0ee70e7858220cec4 (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
/**
 *    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::kReplication

#include "mongo/platform/basic.h"

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

#include <sstream>
#include <string>

#include "mongo/base/status_with.h"
#include "mongo/bson/oid.h"
#include "mongo/db/catalog/database.h"
#include "mongo/db/catalog/database_holder.h"
#include "mongo/db/client.h"
#include "mongo/db/concurrency/d_concurrency.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/dbhelpers.h"
#include "mongo/db/jsobj.h"
#include "mongo/db/op_observer.h"
#include "mongo/db/operation_context_impl.h"
#include "mongo/db/repl/bgsync.h"
#include "mongo/db/repl/isself.h"
#include "mongo/db/repl/last_vote.h"
#include "mongo/db/repl/master_slave.h"
#include "mongo/db/repl/oplog.h"
#include "mongo/db/repl/replication_executor.h"
#include "mongo/db/repl/rs_sync.h"
#include "mongo/db/repl/snapshot_thread.h"
#include "mongo/db/server_parameters.h"
#include "mongo/db/service_context.h"
#include "mongo/db/storage/storage_engine.h"
#include "mongo/executor/network_interface.h"
#include "mongo/s/d_state.h"
#include "mongo/stdx/functional.h"
#include "mongo/stdx/thread.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/log.h"
#include "mongo/util/mongoutils/str.h"
#include "mongo/util/net/hostandport.h"
#include "mongo/util/net/message_port.h"
#include "mongo/util/net/sock.h"

namespace mongo {
namespace repl {

namespace {
const char configCollectionName[] = "local.system.replset";
const char configDatabaseName[] = "local";
const char lastVoteCollectionName[] = "local.replset.election";
const char lastVoteDatabaseName[] = "local";
const char meCollectionName[] = "local.me";
const char meDatabaseName[] = "local";
const char tsFieldName[] = "ts";

// Set this to false to disable the background creation of snapshots. This can be used for A-B
// benchmarking to find how much overhead repl::SnapshotThread introduces.
// TODO SERVER-19213 Make this the default once secondaries can advance their commit
//                   points (SERVER-19208) and we've validated the performance impact.
MONGO_EXPORT_STARTUP_SERVER_PARAMETER(enableReplSnapshotThread, bool, false);
}  // namespace

ReplicationCoordinatorExternalStateImpl::ReplicationCoordinatorExternalStateImpl()
    : _startedThreads(false), _nextThreadId(0) {}
ReplicationCoordinatorExternalStateImpl::~ReplicationCoordinatorExternalStateImpl() {}

void ReplicationCoordinatorExternalStateImpl::startThreads() {
    stdx::lock_guard<stdx::mutex> lk(_threadMutex);
    if (_startedThreads) {
        return;
    }
    log() << "Starting replication applier threads";
    _applierThread.reset(new stdx::thread(runSyncThread));
    BackgroundSync* bgsync = BackgroundSync::get();
    _producerThread.reset(new stdx::thread(stdx::bind(&BackgroundSync::producerThread, bgsync)));
    _syncSourceFeedbackThread.reset(
        new stdx::thread(stdx::bind(&SyncSourceFeedback::run, &_syncSourceFeedback)));
    if (enableReplSnapshotThread) {
        _snapshotThread = SnapshotThread::start(getGlobalServiceContext());
    }
    _startedThreads = true;
}

void ReplicationCoordinatorExternalStateImpl::startMasterSlave(OperationContext* txn) {
    repl::startMasterSlave(txn);
}

void ReplicationCoordinatorExternalStateImpl::shutdown() {
    stdx::lock_guard<stdx::mutex> lk(_threadMutex);
    if (_startedThreads) {
        log() << "Stopping replication applier threads";
        _syncSourceFeedback.shutdown();
        _syncSourceFeedbackThread->join();
        _applierThread->join();
        BackgroundSync* bgsync = BackgroundSync::get();
        bgsync->shutdown();
        _producerThread->join();

        if (_snapshotThread)
            _snapshotThread->shutdown();
    }
}

void ReplicationCoordinatorExternalStateImpl::initiateOplog(OperationContext* txn) {
    createOplog(txn);

    MONGO_WRITE_CONFLICT_RETRY_LOOP_BEGIN {
        ScopedTransaction scopedXact(txn, MODE_X);
        Lock::GlobalWrite globalWrite(txn->lockState());

        WriteUnitOfWork wuow(txn);
        getGlobalServiceContext()->getOpObserver()->onOpMessage(txn,
                                                                BSON("msg"
                                                                     << "initiating set"));
        wuow.commit();
    }
    MONGO_WRITE_CONFLICT_RETRY_LOOP_END(txn, "initiate oplog entry", "local.oplog.rs");
}

void ReplicationCoordinatorExternalStateImpl::forwardSlaveProgress() {
    _syncSourceFeedback.forwardSlaveProgress();
}

OID ReplicationCoordinatorExternalStateImpl::ensureMe(OperationContext* txn) {
    std::string myname = getHostName();
    OID myRID;
    {
        ScopedTransaction transaction(txn, MODE_IX);
        Lock::DBLock lock(txn->lockState(), meDatabaseName, MODE_X);

        BSONObj me;
        // local.me is an identifier for a server for getLastError w:2+
        // TODO: handle WriteConflictExceptions below
        if (!Helpers::getSingleton(txn, meCollectionName, me) || !me.hasField("host") ||
            me["host"].String() != myname) {
            myRID = OID::gen();

            // clean out local.me
            Helpers::emptyCollection(txn, meCollectionName);

            // repopulate
            BSONObjBuilder b;
            b.append("_id", myRID);
            b.append("host", myname);
            Helpers::putSingleton(txn, meCollectionName, b.done());
        } else {
            myRID = me["_id"].OID();
        }
    }
    return myRID;
}

StatusWith<BSONObj> ReplicationCoordinatorExternalStateImpl::loadLocalConfigDocument(
    OperationContext* txn) {
    try {
        MONGO_WRITE_CONFLICT_RETRY_LOOP_BEGIN {
            BSONObj config;
            if (!Helpers::getSingleton(txn, configCollectionName, config)) {
                return StatusWith<BSONObj>(
                    ErrorCodes::NoMatchingDocument,
                    str::stream() << "Did not find replica set configuration document in "
                                  << configCollectionName);
            }
            return StatusWith<BSONObj>(config);
        }
        MONGO_WRITE_CONFLICT_RETRY_LOOP_END(txn, "load replica set config", configCollectionName);
    } catch (const DBException& ex) {
        return StatusWith<BSONObj>(ex.toStatus());
    }
}

Status ReplicationCoordinatorExternalStateImpl::storeLocalConfigDocument(OperationContext* txn,
                                                                         const BSONObj& config) {
    try {
        MONGO_WRITE_CONFLICT_RETRY_LOOP_BEGIN {
            ScopedTransaction transaction(txn, MODE_IX);
            Lock::DBLock dbWriteLock(txn->lockState(), configDatabaseName, MODE_X);
            Helpers::putSingleton(txn, configCollectionName, config);
            return Status::OK();
        }
        MONGO_WRITE_CONFLICT_RETRY_LOOP_END(txn, "save replica set config", configCollectionName);
    } catch (const DBException& ex) {
        return ex.toStatus();
    }
}

StatusWith<LastVote> ReplicationCoordinatorExternalStateImpl::loadLocalLastVoteDocument(
    OperationContext* txn) {
    try {
        MONGO_WRITE_CONFLICT_RETRY_LOOP_BEGIN {
            BSONObj lastVoteObj;
            if (!Helpers::getSingleton(txn, lastVoteCollectionName, lastVoteObj)) {
                return StatusWith<LastVote>(ErrorCodes::NoMatchingDocument,
                                            str::stream()
                                                << "Did not find replica set lastVote document in "
                                                << lastVoteCollectionName);
            }
            LastVote lastVote;
            lastVote.initialize(lastVoteObj);
            return StatusWith<LastVote>(lastVote);
        }
        MONGO_WRITE_CONFLICT_RETRY_LOOP_END(
            txn, "load replica set lastVote", lastVoteCollectionName);
    } catch (const DBException& ex) {
        return StatusWith<LastVote>(ex.toStatus());
    }
}

Status ReplicationCoordinatorExternalStateImpl::storeLocalLastVoteDocument(
    OperationContext* txn, const LastVote& lastVote) {
    BSONObj lastVoteObj = lastVote.toBSON();
    try {
        MONGO_WRITE_CONFLICT_RETRY_LOOP_BEGIN {
            ScopedTransaction transaction(txn, MODE_IX);
            Lock::DBLock dbWriteLock(txn->lockState(), lastVoteDatabaseName, MODE_X);
            Helpers::putSingleton(txn, lastVoteCollectionName, lastVoteObj);
            return Status::OK();
        }
        MONGO_WRITE_CONFLICT_RETRY_LOOP_END(
            txn, "save replica set lastVote", lastVoteCollectionName);
        MONGO_UNREACHABLE;
    } catch (const DBException& ex) {
        return ex.toStatus();
    }
}

void ReplicationCoordinatorExternalStateImpl::setGlobalTimestamp(const Timestamp& newTime) {
    setNewTimestamp(newTime);
}

StatusWith<OpTime> ReplicationCoordinatorExternalStateImpl::loadLastOpTime(OperationContext* txn) {
    // TODO: handle WriteConflictExceptions below
    try {
        BSONObj oplogEntry;
        if (!Helpers::getLast(txn, rsOplogName.c_str(), oplogEntry)) {
            return StatusWith<OpTime>(ErrorCodes::NoMatchingDocument,
                                      str::stream() << "Did not find any entries in "
                                                    << rsOplogName);
        }
        BSONElement tsElement = oplogEntry[tsFieldName];
        if (tsElement.eoo()) {
            return StatusWith<OpTime>(ErrorCodes::NoSuchKey,
                                      str::stream() << "Most recent entry in " << rsOplogName
                                                    << " missing \"" << tsFieldName << "\" field");
        }
        if (tsElement.type() != bsonTimestamp) {
            return StatusWith<OpTime>(ErrorCodes::TypeMismatch,
                                      str::stream() << "Expected type of \"" << tsFieldName
                                                    << "\" in most recent " << rsOplogName
                                                    << " entry to have type Timestamp, but found "
                                                    << typeName(tsElement.type()));
        }
        return StatusWith<OpTime>(extractOpTime(oplogEntry));
    } catch (const DBException& ex) {
        return StatusWith<OpTime>(ex.toStatus());
    }
}

bool ReplicationCoordinatorExternalStateImpl::isSelf(const HostAndPort& host) {
    return repl::isSelf(host);
}

HostAndPort ReplicationCoordinatorExternalStateImpl::getClientHostAndPort(
    const OperationContext* txn) {
    return HostAndPort(txn->getClient()->clientAddress(true));
}

void ReplicationCoordinatorExternalStateImpl::closeConnections() {
    MessagingPort::closeAllSockets(executor::NetworkInterface::kMessagingPortKeepOpen);
}

void ReplicationCoordinatorExternalStateImpl::killAllUserOperations(OperationContext* txn) {
    ServiceContext* environment = getGlobalServiceContext();
    environment->killAllUserOperations(txn);
}

void ReplicationCoordinatorExternalStateImpl::clearShardingState() {
    shardingState.clearCollectionMetadata();
}

void ReplicationCoordinatorExternalStateImpl::signalApplierToChooseNewSyncSource() {
    BackgroundSync::get()->clearSyncTarget();
}

OperationContext* ReplicationCoordinatorExternalStateImpl::createOperationContext(
    const std::string& threadName) {
    Client::initThreadIfNotAlready(threadName.c_str());
    return new OperationContextImpl();
}

void ReplicationCoordinatorExternalStateImpl::dropAllTempCollections(OperationContext* txn) {
    std::vector<std::string> dbNames;
    StorageEngine* storageEngine = getGlobalServiceContext()->getGlobalStorageEngine();
    storageEngine->listDatabases(&dbNames);

    for (std::vector<std::string>::iterator it = dbNames.begin(); it != dbNames.end(); ++it) {
        // The local db is special because it isn't replicated. It is cleared at startup even on
        // replica set members.
        if (*it == "local")
            continue;
        LOG(2) << "Removing temporary collections from " << *it;
        Database* db = dbHolder().get(txn, *it);
        // Since we must be holding the global lock during this function, if listDatabases
        // returned this dbname, we should be able to get a reference to it - it can't have
        // been dropped.
        invariant(db);
        db->clearTmpCollections(txn);
    }
}

void ReplicationCoordinatorExternalStateImpl::dropAllSnapshots() {
    if (auto manager = getGlobalServiceContext()->getGlobalStorageEngine()->getSnapshotManager())
        manager->dropAllSnapshots();
}

void ReplicationCoordinatorExternalStateImpl::updateCommittedSnapshot(OpTime newCommitPoint) {
    auto manager = getGlobalServiceContext()->getGlobalStorageEngine()->getSnapshotManager();
    invariant(manager);  // This should never be called if there is no SnapshotManager.
    manager->setCommittedSnapshot(SnapshotName(newCommitPoint.getTimestamp()));
}

void ReplicationCoordinatorExternalStateImpl::forceSnapshotCreation() {
    if (_snapshotThread)
        _snapshotThread->forceSnapshot();
}
}  // namespace repl
}  // namespace mongo