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
|
/**
* Copyright (C) 2018 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/oplog_applier.h"
#include "mongo/db/auth/authorization_session.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/repl/sync_tail.h"
#include "mongo/db/server_parameters.h"
#include "mongo/util/log.h"
namespace mongo {
namespace repl {
namespace {
/**
* This server parameter determines the number of writer threads OplogApplier will have.
*/
MONGO_EXPORT_STARTUP_SERVER_PARAMETER(replWriterThreadCount, int, 16)
->withValidator([](const int& newVal) {
if (newVal < 1 || newVal > 256) {
return Status(ErrorCodes::BadValue, "replWriterThreadCount must be between 1 and 256");
}
return Status::OK();
});
MONGO_EXPORT_SERVER_PARAMETER(replBatchLimitOperations, int, 5 * 1000)
->withValidator([](const int& newVal) {
if (newVal < 1 || newVal > (1000 * 1000)) {
return Status(ErrorCodes::BadValue,
"replBatchLimitOperations must be between 1 and 1 million, inclusive");
}
return Status::OK();
});
} // namespace
using CallbackArgs = executor::TaskExecutor::CallbackArgs;
// static
std::unique_ptr<ThreadPool> OplogApplier::makeWriterPool() {
return makeWriterPool(replWriterThreadCount);
}
// static
std::unique_ptr<ThreadPool> OplogApplier::makeWriterPool(int threadCount) {
ThreadPool::Options options;
options.threadNamePrefix = "repl writer worker ";
options.poolName = "repl writer worker Pool";
options.maxThreads = options.minThreads = static_cast<size_t>(threadCount);
options.onCreateThread = [](const std::string&) {
// Only do this once per thread
if (!Client::getCurrent()) {
Client::initThreadIfNotAlready();
AuthorizationSession::get(cc())->grantInternalAuthorization();
}
};
auto pool = stdx::make_unique<ThreadPool>(options);
pool->startup();
return pool;
}
// static
std::size_t OplogApplier::getBatchLimitOperations() {
return std::size_t(replBatchLimitOperations.load());
}
// static
std::size_t OplogApplier::calculateBatchLimitBytes(OperationContext* opCtx,
StorageInterface* storageInterface) {
auto oplogMaxSizeResult =
storageInterface->getOplogMaxSize(opCtx, NamespaceString::kRsOplogNamespace);
auto oplogMaxSize = fassert(40301, oplogMaxSizeResult);
return std::min(oplogMaxSize / 10, std::size_t(replBatchLimitBytes));
}
OplogApplier::OplogApplier(executor::TaskExecutor* executor,
OplogBuffer* oplogBuffer,
Observer* observer)
: _executor(executor), _oplogBuffer(oplogBuffer), _observer(observer) {}
OplogBuffer* OplogApplier::getBuffer() const {
return _oplogBuffer;
}
Future<void> OplogApplier::startup() {
auto pf = makePromiseFuture<void>();
auto callback =
[ this, promise = pf.promise.share() ](const CallbackArgs& args) mutable noexcept {
invariant(args.status);
log() << "Starting oplog application";
_run(_oplogBuffer);
log() << "Finished oplog application";
promise.setWith([] {});
};
invariant(_executor->scheduleWork(callback).getStatus());
return std::move(pf.future);
}
void OplogApplier::shutdown() {
_shutdown();
}
/**
* Pushes operations read from sync source into oplog buffer.
*/
void OplogApplier::enqueue(OperationContext* opCtx,
Operations::const_iterator begin,
Operations::const_iterator end) {
OplogBuffer::Batch batch;
for (auto i = begin; i != end; ++i) {
batch.push_back(i->raw);
}
enqueue(opCtx, batch.cbegin(), batch.cend());
}
void OplogApplier::enqueue(OperationContext* opCtx,
OplogBuffer::Batch::const_iterator begin,
OplogBuffer::Batch::const_iterator end) {
OCCASIONALLY {
LOG(2) << "oplog buffer has " << _oplogBuffer->getSize() << " bytes";
}
_oplogBuffer->pushAllNonBlocking(opCtx, begin, end);
}
StatusWith<OplogApplier::Operations> OplogApplier::getNextApplierBatch(
OperationContext* opCtx, const BatchLimits& batchLimits) {
if (batchLimits.ops == 0) {
return Status(ErrorCodes::InvalidOptions, "Batch size must be greater than 0.");
}
std::uint32_t totalBytes = 0;
Operations ops;
BSONObj op;
while (_oplogBuffer->peek(opCtx, &op)) {
auto entry = OplogEntry(op);
// Check for oplog version change. If it is absent, its value is one.
if (entry.getVersion() != OplogEntry::kOplogVersion) {
std::string message = str::stream()
<< "expected oplog version " << OplogEntry::kOplogVersion << " but found version "
<< entry.getVersion() << " in oplog entry: " << redact(entry.toBSON());
severe() << message;
return {ErrorCodes::BadValue, message};
}
// Commands must be processed one at a time. The only exception to this is applyOps because
// applyOps oplog entries are effectively containers for CRUD operations. Therefore, it is
// safe to batch applyOps commands with CRUD operations when reading from the oplog buffer.
if (entry.isCommand() && (entry.getCommandType() != OplogEntry::CommandType::kApplyOps ||
entry.shouldPrepare())) {
if (ops.empty()) {
// Apply commands one-at-a-time.
ops.push_back(std::move(entry));
BSONObj opToPopAndDiscard;
invariant(_oplogBuffer->tryPop(opCtx, &opToPopAndDiscard));
dassert(ops.back() == OplogEntry(opToPopAndDiscard));
}
// Otherwise, apply what we have so far and come back for the command.
return std::move(ops);
}
// Apply replication batch limits.
if (ops.size() >= batchLimits.ops) {
return std::move(ops);
}
// Never return an empty batch if there are operations left.
if ((totalBytes + entry.getRawObjSizeBytes() >= batchLimits.bytes) && (ops.size() > 0)) {
return std::move(ops);
}
// Add op to buffer.
totalBytes += entry.getRawObjSizeBytes();
ops.push_back(std::move(entry));
BSONObj opToPopAndDiscard;
invariant(_oplogBuffer->tryPop(opCtx, &opToPopAndDiscard));
dassert(ops.back() == OplogEntry(opToPopAndDiscard));
}
return std::move(ops);
}
StatusWith<OpTime> OplogApplier::multiApply(OperationContext* opCtx, Operations ops) {
_observer->onBatchBegin(ops);
auto lastApplied = _multiApply(opCtx, std::move(ops));
_observer->onBatchEnd(lastApplied, {});
return lastApplied;
}
} // namespace repl
} // namespace mongo
|
