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
|
/**
* Copyright (C) 2018-present MongoDB, Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the Server Side Public License, version 1,
* as published by MongoDB, Inc.
*
* 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
* Server Side Public License for more details.
*
* You should have received a copy of the Server Side Public License
* along with this program. If not, see
* <http://www.mongodb.com/licensing/server-side-public-license>.
*
* 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 Server Side 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::kSharding
#include "mongo/platform/basic.h"
#include "mongo/db/s/sharding_state_recovery.h"
#include "mongo/bson/bsonobj.h"
#include "mongo/bson/bsonobjbuilder.h"
#include "mongo/bson/util/bson_extract.h"
#include "mongo/client/connection_string.h"
#include "mongo/db/catalog_raii.h"
#include "mongo/db/dbhelpers.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/ops/update.h"
#include "mongo/db/ops/update_request.h"
#include "mongo/db/repl/bson_extract_optime.h"
#include "mongo/db/repl/optime.h"
#include "mongo/db/repl/repl_client_info.h"
#include "mongo/db/s/sharding_logging.h"
#include "mongo/db/s/sharding_state.h"
#include "mongo/db/write_concern.h"
#include "mongo/db/write_concern_options.h"
#include "mongo/s/client/shard_registry.h"
#include "mongo/s/grid.h"
#include "mongo/util/log.h"
namespace mongo {
namespace {
const char kRecoveryDocumentId[] = "minOpTimeRecovery";
const char kMinOpTime[] = "minOpTime";
const char kMinOpTimeUpdaters[] = "minOpTimeUpdaters";
const WriteConcernOptions kMajorityWriteConcern(WriteConcernOptions::kMajority,
WriteConcernOptions::SyncMode::UNSET,
WriteConcernOptions::kWriteConcernTimeoutSharding);
const WriteConcernOptions kLocalWriteConcern(1,
WriteConcernOptions::SyncMode::UNSET,
Milliseconds(0));
/**
* Encapsulates the parsing and construction of the config server min opTime recovery document.
*/
class RecoveryDocument {
public:
enum ChangeType : int8_t { Increment = 1, Decrement = -1, Clear = 0 };
static StatusWith<RecoveryDocument> fromBSON(const BSONObj& obj) {
RecoveryDocument recDoc;
Status status = bsonExtractOpTimeField(obj, kMinOpTime, &recDoc._minOpTime);
if (!status.isOK())
return status;
status = bsonExtractIntegerField(obj, kMinOpTimeUpdaters, &recDoc._minOpTimeUpdaters);
if (!status.isOK())
return status;
return recDoc;
}
static BSONObj createChangeObj(repl::OpTime minOpTime, ChangeType change) {
BSONObjBuilder cmdBuilder;
{
BSONObjBuilder setBuilder(cmdBuilder.subobjStart("$set"));
minOpTime.append(&setBuilder, kMinOpTime);
}
if (change == Clear) {
cmdBuilder.append("$set", BSON(kMinOpTimeUpdaters << 0));
} else {
cmdBuilder.append("$inc", BSON(kMinOpTimeUpdaters << change));
}
return cmdBuilder.obj();
}
static BSONObj getQuery() {
return BSON("_id" << kRecoveryDocumentId);
}
BSONObj toBSON() const {
BSONObjBuilder builder;
builder.append("_id", kRecoveryDocumentId);
builder.append(kMinOpTime, _minOpTime.toBSON());
builder.append(kMinOpTimeUpdaters, _minOpTimeUpdaters);
return builder.obj();
}
repl::OpTime getMinOpTime() const {
return _minOpTime;
}
int64_t getMinOpTimeUpdaters() const {
return _minOpTimeUpdaters;
}
private:
RecoveryDocument() = default;
repl::OpTime _minOpTime;
long long _minOpTimeUpdaters;
};
/**
* This method is the main entry point for updating the sharding state recovery document. The goal
* it has is to always move the opTime forward for a currently running server. It achieves this by
* serializing the modify calls and reading the current opTime under X-lock on the admin database.
*/
Status modifyRecoveryDocument(OperationContext* opCtx,
RecoveryDocument::ChangeType change,
const WriteConcernOptions& writeConcern) {
try {
// Use boost::optional so we can release the locks early
boost::optional<AutoGetOrCreateDb> autoGetOrCreateDb;
autoGetOrCreateDb.emplace(
opCtx, NamespaceString::kServerConfigurationNamespace.db(), MODE_X);
// The config server connection string and shard name are no longer parsed in 4.0, but 3.6
// nodes still expect to find them, so we must include them until after 4.0 ships.
auto const grid = Grid::get(opCtx);
BSONObj updateObj = RecoveryDocument::createChangeObj(grid->configOpTime(), change);
LOG(1) << "Changing sharding recovery document " << redact(updateObj);
UpdateRequest updateReq(NamespaceString::kServerConfigurationNamespace);
updateReq.setQuery(RecoveryDocument::getQuery());
updateReq.setUpdateModification(updateObj);
updateReq.setUpsert();
UpdateResult result = update(opCtx, autoGetOrCreateDb->getDb(), updateReq);
invariant(result.numDocsModified == 1 || !result.upserted.isEmpty());
invariant(result.numMatched <= 1);
// Wait until the majority write concern has been satisfied, but do it outside of lock
autoGetOrCreateDb = boost::none;
WriteConcernResult writeConcernResult;
return waitForWriteConcern(opCtx,
repl::ReplClientInfo::forClient(opCtx->getClient()).getLastOp(),
writeConcern,
&writeConcernResult);
} catch (const DBException& ex) {
return ex.toStatus();
}
}
} // namespace
Status ShardingStateRecovery::startMetadataOp(OperationContext* opCtx) {
Status upsertStatus =
modifyRecoveryDocument(opCtx, RecoveryDocument::Increment, kMajorityWriteConcern);
if (upsertStatus == ErrorCodes::WriteConcernFailed) {
// Couldn't wait for the replication to complete, but the local write was performed. Clear
// it up fast (without any waiting for journal or replication) and still treat it as
// failure.
modifyRecoveryDocument(opCtx, RecoveryDocument::Decrement, WriteConcernOptions())
.transitional_ignore();
}
return upsertStatus;
}
void ShardingStateRecovery::endMetadataOp(OperationContext* opCtx) {
Status status =
modifyRecoveryDocument(opCtx, RecoveryDocument::Decrement, WriteConcernOptions());
if (!status.isOK()) {
warning() << "Failed to decrement minOpTimeUpdaters due to " << redact(status);
}
}
Status ShardingStateRecovery::recover(OperationContext* opCtx) {
Grid* const grid = Grid::get(opCtx);
ShardingState* const shardingState = ShardingState::get(opCtx);
invariant(shardingState->enabled());
BSONObj recoveryDocBSON;
try {
AutoGetCollection autoColl(opCtx, NamespaceString::kServerConfigurationNamespace, MODE_IS);
if (!Helpers::findOne(
opCtx, autoColl.getCollection(), RecoveryDocument::getQuery(), recoveryDocBSON)) {
return Status::OK();
}
} catch (const DBException& ex) {
return ex.toStatus();
}
const auto recoveryDocStatus = RecoveryDocument::fromBSON(recoveryDocBSON);
if (!recoveryDocStatus.isOK())
return recoveryDocStatus.getStatus();
const auto recoveryDoc = std::move(recoveryDocStatus.getValue());
log() << "Sharding state recovery process found document " << redact(recoveryDoc.toBSON());
if (!recoveryDoc.getMinOpTimeUpdaters()) {
// Treat the minOpTime as up-to-date
const auto prevOpTime = grid->advanceConfigOpTime(
opCtx, recoveryDoc.getMinOpTime(), "sharding state recovery document");
if (prevOpTime) {
log()
<< "No in flight metadata change operations, so config server optime updated from "
<< *prevOpTime << " to " << recoveryDoc.getMinOpTime();
}
return Status::OK();
}
log() << "Sharding state recovery document indicates there were "
<< recoveryDoc.getMinOpTimeUpdaters()
<< " metadata change operations in flight. Contacting the config server primary in order "
"to retrieve the most recent opTime.";
// Need to fetch the latest uptime from the config server, so do a logging write
Status status = ShardingLogging::get(opCtx)->logChangeChecked(
opCtx,
"Sharding minOpTime recovery",
NamespaceString::kServerConfigurationNamespace.ns(),
recoveryDocBSON,
ShardingCatalogClient::kMajorityWriteConcern);
if (!status.isOK())
return status;
log() << "Sharding state recovered. New config server opTime is " << grid->configOpTime();
// Finally, clear the recovery document so next time we don't need to recover
status = modifyRecoveryDocument(opCtx, RecoveryDocument::Clear, kLocalWriteConcern);
if (!status.isOK()) {
warning() << "Failed to reset sharding state recovery document due to " << redact(status);
}
return Status::OK();
}
} // namespace mongo
|