summaryrefslogtreecommitdiff
path: root/src/mongo/db/write_concern.cpp
blob: ee09a8bd933cc9315a1e0d13bdbb616ccd9525b7 (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
/**
 *    Copyright (C) 2013 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.
 */

#include "mongo/base/counter.h"
#include "mongo/db/commands/server_status_metric.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/repl/is_master.h"
#include "mongo/db/repl/repl_settings.h"
#include "mongo/db/repl/write_concern.h"
#include "mongo/db/stats/timer_stats.h"
#include "mongo/db/write_concern.h"

namespace mongo {

    static TimerStats gleWtimeStats;
    static ServerStatusMetricField<TimerStats> displayGleLatency( "getLastError.wtime", &gleWtimeStats );

    static Counter64 gleWtimeouts;
    static ServerStatusMetricField<Counter64> gleWtimeoutsDisplay( "getLastError.wtimeouts", &gleWtimeouts );

    Status validateWriteConcern( const WriteConcernOptions& writeConcern ) {

        const bool isJournalEnabled = getDur().isDurable();

        if ( writeConcern.syncMode == WriteConcernOptions::JOURNAL && !isJournalEnabled ) {
            return Status( ErrorCodes::BadValue,
                           "cannot use 'j' option when a host does not have journaling enabled" );
        }

        const bool isConfigServer = serverGlobalParams.configsvr;
        const bool isMasterSlaveNode = anyReplEnabled() && !theReplSet;
        const bool isReplSetNode = anyReplEnabled() && theReplSet;

        if ( isConfigServer || ( !isMasterSlaveNode && !isReplSetNode ) ) {

            // Note that config servers can be replicated (have an oplog), but we still don't allow
            // w > 1

            if ( writeConcern.wNumNodes > 1 ) {
                return Status( ErrorCodes::BadValue,
                               string( "cannot use 'w' > 1 " ) +
                               ( isConfigServer ? "on a config server host" :
                                                  "when a host is not replicated" ) );
            }
        }

        if ( !isReplSetNode && !writeConcern.wMode.empty() && writeConcern.wMode != "majority" ) {
            return Status( ErrorCodes::BadValue,
                           string( "cannot use non-majority 'w' mode " ) + writeConcern.wMode
                           + " when a host is not a member of a replica set" );
        }

        return Status::OK();
    }

    void WriteConcernResult::appendTo( const WriteConcernOptions& writeConcern,
                                       BSONObjBuilder* result ) const {

        if ( syncMillis >= 0 )
            result->appendNumber( "syncMillis", syncMillis );

        if ( fsyncFiles >= 0 )
            result->appendNumber( "fsyncFiles", fsyncFiles );

        if ( wTime >= 0 ) {
            if ( wTimedOut )
                result->appendNumber( "waited", wTime );
            else
                result->appendNumber( "wtime", wTime );
        }

        if ( wTimedOut )
            result->appendBool( "wtimeout", true );

        if ( writtenTo.size() )
            result->append( "writtenTo", writtenTo );
        else
            result->appendNull( "writtenTo" );

        if ( err.empty() )
            result->appendNull( "err" );
        else
            result->append( "err", err );

        // *** 2.4 SyncClusterConnection compatibility ***
        // 2.4 expects either fsync'd files, or a "waited" field exist after running an fsync : true
        // GLE, but with journaling we don't actually need to run the fsync (fsync command is
        // preferred in 2.6).  So we add a "waited" field if one doesn't exist.

        if ( writeConcern.syncMode == WriteConcernOptions::FSYNC ) {

            if ( fsyncFiles < 0 && ( wTime < 0 || !wTimedOut ) ) {
                dassert( result->asTempObj()["waited"].eoo() );
                result->appendNumber( "waited", syncMillis );
            }

            dassert( result->asTempObj()["fsyncFiles"].numberInt() > 0 ||
                     !result->asTempObj()["waited"].eoo() );
        }
    }

    Status waitForWriteConcern( OperationContext* txn,
                                const WriteConcernOptions& writeConcern,
                                const OpTime& replOpTime,
                                WriteConcernResult* result ) {

        // We assume all options have been validated earlier, if not, programming error
        dassert( validateWriteConcern( writeConcern ).isOK() );

        // Next handle blocking on disk

        Timer syncTimer;

        switch( writeConcern.syncMode ) {
        case WriteConcernOptions::NONE:
            break;
        case WriteConcernOptions::FSYNC:
            if ( !getDur().isDurable() ) {
                result->fsyncFiles = MemoryMappedFile::flushAll( true );
            }
            else {
                // We only need to commit the journal if we're durable
                txn->recoveryUnit()->awaitCommit();
            }
            break;
        case WriteConcernOptions::JOURNAL:
            txn->recoveryUnit()->awaitCommit();
            break;
        }

        result->syncMillis = syncTimer.millis();

        // Now wait for replication

        if ( replOpTime.isNull() ) {
            // no write happened for this client yet
            return Status::OK();
        }

        if ( writeConcern.wNumNodes <= 1 && writeConcern.wMode.empty() ) {
            // no desired replication check
            return Status::OK();
        }

        if ( !anyReplEnabled() || serverGlobalParams.configsvr ) {
            // no replication check needed (validated above)
            return Status::OK();
        }

        const bool isMasterSlaveNode = anyReplEnabled() && !theReplSet;
        if ( writeConcern.wMode == "majority" && isMasterSlaveNode ) {
            // with master/slave, majority is equivalent to w=1
            return Status::OK();
        }

        // We're sure that replication is enabled and that we have more than one node or a wMode
        TimerHolder gleTimerHolder( &gleWtimeStats );

        // Now we wait for replication
        // Note that replica set stepdowns and gle mode changes are thrown as errors
        // TODO: Make this cleaner
        Status replStatus = Status::OK();
        try {
            while ( 1 ) {

                if ( writeConcern.wNumNodes > 0 ) {
                    if ( opReplicatedEnough( replOpTime, writeConcern.wNumNodes ) ) {
                        break;
                    }
                }
                else if ( opReplicatedEnough( replOpTime, writeConcern.wMode ) ) {
                    break;
                }

                if ( writeConcern.wTimeout > 0 &&
                     gleTimerHolder.millis() >= writeConcern.wTimeout ) {
                    gleWtimeouts.increment();
                    result->err = "timeout";
                    result->wTimedOut = true;
                    replStatus = Status( ErrorCodes::WriteConcernFailed,
                                         "waiting for replication timed out" );
                    break;
                }

                sleepmillis(1);
                txn->checkForInterrupt();
            }
        }
        catch( const AssertionException& ex ) {
            // Our replication state changed while enforcing write concern
            replStatus = ex.toStatus();
        }

        // Add stats
        result->writtenTo = getHostsWrittenTo( replOpTime );
        result->wTime = gleTimerHolder.recordMillis();

        return replStatus;
    }

} // namespace mongo