summaryrefslogtreecommitdiff
path: root/src/mongo/db/repl/replication_coordinator_external_state.h
blob: 3fc7913e3b4116d4da22eec937ce47d00ad2494f (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
/**
 *    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.
 */

#pragma once

#include <boost/optional.hpp>

#include "mongo/base/disallow_copying.h"
#include "mongo/bson/timestamp.h"
#include "mongo/db/repl/member_state.h"
#include "mongo/db/repl/multiapplier.h"
#include "mongo/db/repl/optime.h"
#include "mongo/stdx/functional.h"
#include "mongo/util/time_support.h"

namespace mongo {

class BSONObj;
class OID;
class OperationContext;
class SnapshotName;
class Status;
struct HostAndPort;
template <typename T>
class StatusWith;

namespace repl {

class LastVote;
class ReplSettings;

using OnInitialSyncFinishedFn = stdx::function<void()>;
using StartInitialSyncFn = stdx::function<void(OnInitialSyncFinishedFn callback)>;
using StartSteadyReplicationFn = stdx::function<void()>;
/**
 * This class represents the interface the ReplicationCoordinator uses to interact with the
 * rest of the system.  All functionality of the ReplicationCoordinatorImpl that would introduce
 * dependencies on large sections of the server code and thus break the unit testability of
 * ReplicationCoordinatorImpl should be moved here.
 */
class ReplicationCoordinatorExternalState {
    MONGO_DISALLOW_COPYING(ReplicationCoordinatorExternalState);

public:
    ReplicationCoordinatorExternalState();
    virtual ~ReplicationCoordinatorExternalState();

    /**
     * Starts the journal listener, and snapshot threads
     *
     * NOTE: Only starts threads if they are not already started,
     */
    virtual void startThreads(const ReplSettings& settings) = 0;

    /**
     * Starts an initial sync, and calls "finished" when done,
     * for replica set member -- legacy impl not in DataReplicator.
     *
     * NOTE: Use either this (and below function) or the Master/Slave version, but not both.
     */
    virtual void startInitialSync(OnInitialSyncFinishedFn finished) = 0;

    /**
     * Starts steady state sync for replica set member -- legacy impl not in DataReplicator.
     *
     * NOTE: Use either this or the Master/Slave version, but not both.
     */
    virtual void startSteadyStateReplication() = 0;

    /**
     * Starts the Master/Slave threads and sets up logOp
     */
    virtual void startMasterSlave(OperationContext* txn) = 0;

    /**
     * Performs any necessary external state specific shutdown tasks, such as cleaning up
     * the threads it started.
     */
    virtual void shutdown() = 0;

    /**
     * Creates the oplog, writes the first entry and stores the replica set config document.
     */
    virtual Status initializeReplSetStorage(OperationContext* txn, const BSONObj& config) = 0;

    /**
     * Writes a message about our transition to primary to the oplog.
     */
    virtual void logTransitionToPrimaryToOplog(OperationContext* txn) = 0;

    /**
     * Simple wrapper around SyncSourceFeedback::forwardSlaveProgress.  Signals to the
     * SyncSourceFeedback thread that it needs to wake up and send a replSetUpdatePosition
     * command upstream.
     */
    virtual void forwardSlaveProgress() = 0;

    /**
     * Queries the singleton document in local.me.  If it exists and our hostname has not
     * changed since we wrote, returns the RID stored in the object.  If the document does not
     * exist or our hostname doesn't match what was recorded in local.me, generates a new OID
     * to use as our RID, stores it in local.me, and returns it.
     */
    virtual OID ensureMe(OperationContext*) = 0;

    /**
     * Returns true if "host" is one of the network identities of this node.
     */
    virtual bool isSelf(const HostAndPort& host) = 0;

    /**
     * Gets the replica set config document from local storage, or returns an error.
     */
    virtual StatusWith<BSONObj> loadLocalConfigDocument(OperationContext* txn) = 0;

    /**
     * Stores the replica set config document in local storage, or returns an error.
     */
    virtual Status storeLocalConfigDocument(OperationContext* txn, const BSONObj& config) = 0;

    /**
     * Gets the replica set lastVote document from local storage, or returns an error.
     */
    virtual StatusWith<LastVote> loadLocalLastVoteDocument(OperationContext* txn) = 0;

    /**
     * Stores the replica set lastVote document in local storage, or returns an error.
     */
    virtual Status storeLocalLastVoteDocument(OperationContext* txn, const LastVote& lastVote) = 0;

    /**
     * Sets the global opTime to be 'newTime'.
     */
    virtual void setGlobalTimestamp(const Timestamp& newTime) = 0;

    /**
     * Gets the last optime of an operation performed on this host, from stable
     * storage.
     */
    virtual StatusWith<OpTime> loadLastOpTime(OperationContext* txn) = 0;

    /**
     * Cleaning up the oplog, by potentially truncating:
     * If we are recovering from a failed batch then minvalid.start though minvalid.end need
     * to be removed from the oplog before we can start applying operations.
     */
    virtual void cleanUpLastApplyBatch(OperationContext* txn) = 0;

    /**
     * Returns the HostAndPort of the remote client connected to us that initiated the operation
     * represented by "txn".
     */
    virtual HostAndPort getClientHostAndPort(const OperationContext* txn) = 0;

    /**
     * Closes all connections except those marked with the keepOpen property, which should
     * just be connections used for heartbeating.
     * This is used during stepdown, and transition out of primary.
     */
    virtual void closeConnections() = 0;

    /**
     * Kills all operations that have a Client that is associated with an incoming user
     * connection.  Used during stepdown.
     */
    virtual void killAllUserOperations(OperationContext* txn) = 0;

    /**
     * Clears all cached sharding metadata on this server.  This is called after stepDown to
     * ensure that if the node becomes primary again in the future it will reload an up-to-date
     * version of the sharding data.
     */
    virtual void clearShardingState() = 0;

    /**
     * Called when the instance transitions to primary in order to notify a potentially sharded
     * host to recover its sharding state.
     *
     * Throws on errors.
     */
    virtual void recoverShardingState(OperationContext* txn) = 0;

    /**
     * Called when the instance transitions to primary in order to update the config server
     * connection string of the shard identity document.
     */
    virtual void updateShardIdentityConfigString(OperationContext* txn) = 0;

    /**
     * Notifies the bgsync and syncSourceFeedback threads to choose a new sync source.
     */
    virtual void signalApplierToChooseNewSyncSource() = 0;

    /**
     * Notifies the bgsync to cancel the current oplog fetcher.
     */
    virtual void signalApplierToCancelFetcher() = 0;

    /**
     * Drops all temporary collections on all databases except "local".
     *
     * The implementation may assume that the caller has acquired the global exclusive lock
     * for "txn".
     */
    virtual void dropAllTempCollections(OperationContext* txn) = 0;

    /**
     * Drops all snapshots and clears the "committed" snapshot.
     */
    virtual void dropAllSnapshots() = 0;

    /**
     * Updates the committed snapshot to the newCommitPoint, and deletes older snapshots.
     *
     * It is illegal to call with a newCommitPoint that does not name an existing snapshot.
     */
    virtual void updateCommittedSnapshot(SnapshotName newCommitPoint) = 0;

    /**
     * Signals the SnapshotThread, if running, to take a forced snapshot even if the global
     * timestamp hasn't changed.
     *
     * Does not wait for the snapshot to be taken.
     */
    virtual void forceSnapshotCreation() = 0;

    /**
     * Returns whether or not the SnapshotThread is active.
     */
    virtual bool snapshotsEnabled() const = 0;

    virtual void notifyOplogMetadataWaiters() = 0;

    /**
     * Returns multiplier to apply to election timeout to obtain upper bound
     * on randomized offset.
     */
    virtual double getElectionTimeoutOffsetLimitFraction() const = 0;

    /**
     * Returns true if the current storage engine supports read committed.
     */
    virtual bool isReadCommittedSupportedByStorageEngine(OperationContext* txn) const = 0;

    /**
     * Applies the operations described in the oplog entries contained in "ops" using the
     * "applyOperation" function.
     */
    virtual StatusWith<OpTime> multiApply(OperationContext* txn,
                                          const MultiApplier::Operations& ops,
                                          MultiApplier::ApplyOperationFn applyOperation) = 0;

    /**
     * Used by multiApply() to writes operations to database during steady state replication.
     */
    virtual void multiSyncApply(const MultiApplier::Operations& ops) = 0;

    /**
     * Used by multiApply() to writes operations to database during initial sync.
     * Fetches missing documents from "source".
     */
    virtual void multiInitialSyncApply(const MultiApplier::Operations& ops,
                                       const HostAndPort& source) = 0;
};

}  // namespace repl
}  // namespace mongo