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
|
/**
* 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::kTransaction
#include "mongo/platform/basic.h"
#include "mongo/db/s/transaction_coordinator_futures_util.h"
#include "mongo/client/remote_command_targeter.h"
#include "mongo/db/auth/authorization_session.h"
#include "mongo/db/s/sharding_state.h"
#include "mongo/rpc/get_status_from_command_result.h"
#include "mongo/s/grid.h"
#include "mongo/transport/service_entry_point.h"
#include "mongo/util/fail_point_service.h"
#include "mongo/util/log.h"
namespace mongo {
namespace txn {
namespace {
MONGO_FAIL_POINT_DEFINE(hangWhileTargetingRemoteHost);
MONGO_FAIL_POINT_DEFINE(hangWhileTargetingLocalHost);
using RemoteCommandCallbackArgs = executor::TaskExecutor::RemoteCommandCallbackArgs;
using ResponseStatus = executor::TaskExecutor::ResponseStatus;
} // namespace
AsyncWorkScheduler::AsyncWorkScheduler(ServiceContext* serviceContext)
: _serviceContext(serviceContext),
_executor(Grid::get(_serviceContext)->getExecutorPool()->getFixedExecutor()) {}
AsyncWorkScheduler::~AsyncWorkScheduler() {
{
stdx::lock_guard<stdx::mutex> lg(_mutex);
invariant(_quiesced(lg));
}
if (!_parent)
return;
stdx::lock_guard<stdx::mutex> lg(_parent->_mutex);
_parent->_childSchedulers.erase(_itToRemove);
_parent->_notifyAllTasksComplete(lg);
_parent = nullptr;
}
Future<executor::TaskExecutor::ResponseStatus> AsyncWorkScheduler::scheduleRemoteCommand(
const ShardId& shardId,
const ReadPreferenceSetting& readPref,
const BSONObj& commandObj,
OperationContextFn operationContextFn) {
const bool isSelfShard = (shardId == getLocalShardId(_serviceContext));
if (isSelfShard) {
// If sending a command to the same shard as this node is in, send it directly to this node
// rather than going through the host targeting below. This ensures that the state changes
// for the participant and coordinator occur sequentially on a single branch of replica set
// history. See SERVER-38142 for details.
return scheduleWork([this, shardId, operationContextFn, commandObj = commandObj.getOwned()](
OperationContext* opCtx) {
operationContextFn(opCtx);
// Note: This internal authorization is tied to the lifetime of the client, which will
// be destroyed by 'scheduleWork' immediately after this lambda ends
AuthorizationSession::get(opCtx->getClient())
->grantInternalAuthorization(opCtx->getClient());
if (MONGO_FAIL_POINT(hangWhileTargetingLocalHost)) {
LOG(0) << "Hit hangWhileTargetingLocalHost failpoint";
MONGO_FAIL_POINT_PAUSE_WHILE_SET_OR_INTERRUPTED(opCtx, hangWhileTargetingLocalHost);
}
const auto service = opCtx->getServiceContext();
auto start = _executor->now();
auto requestOpMsg =
OpMsgRequest::fromDBAndBody(NamespaceString::kAdminDb, commandObj).serialize();
const auto replyOpMsg = OpMsg::parseOwned(
service->getServiceEntryPoint()->handleRequest(opCtx, requestOpMsg).response);
// Document sequences are not yet being used for responses.
invariant(replyOpMsg.sequences.empty());
// 'ResponseStatus' is the response format of a remote request sent over the network
// so we simulate that format manually here, since we sent the request over the
// loopback.
return ResponseStatus{replyOpMsg.body.getOwned(), _executor->now() - start};
});
}
return _targetHostAsync(shardId, readPref, operationContextFn)
.then([this, shardId, commandObj = commandObj.getOwned(), readPref](
HostAndShard hostAndShard) mutable {
executor::RemoteCommandRequest request(hostAndShard.hostTargeted,
NamespaceString::kAdminDb.toString(),
commandObj,
readPref.toContainingBSON(),
nullptr);
auto pf = makePromiseFuture<ResponseStatus>();
stdx::unique_lock<stdx::mutex> ul(_mutex);
uassertStatusOK(_shutdownStatus);
auto scheduledCommandHandle =
uassertStatusOK(_executor->scheduleRemoteCommand(request, [
this,
commandObj = std::move(commandObj),
shardId = std::move(shardId),
hostTargeted = std::move(hostAndShard.hostTargeted),
shard = std::move(hostAndShard.shard),
promise = std::make_shared<Promise<ResponseStatus>>(std::move(pf.promise))
](const RemoteCommandCallbackArgs& args) mutable noexcept {
auto status = args.response.status;
shard->updateReplSetMonitor(hostTargeted, status);
// Only consider actual failures to send the command as errors.
if (status.isOK()) {
auto commandStatus = getStatusFromCommandResult(args.response.data);
shard->updateReplSetMonitor(hostTargeted, commandStatus);
auto writeConcernStatus =
getWriteConcernStatusFromCommandResult(args.response.data);
shard->updateReplSetMonitor(hostTargeted, writeConcernStatus);
promise->emplaceValue(std::move(args.response));
} else {
promise->setError([&] {
if (status == ErrorCodes::CallbackCanceled) {
stdx::unique_lock<stdx::mutex> ul(_mutex);
return _shutdownStatus.isOK() ? status : _shutdownStatus;
}
return status;
}());
}
}));
auto it =
_activeHandles.emplace(_activeHandles.begin(), std::move(scheduledCommandHandle));
ul.unlock();
return std::move(pf.future).tapAll(
[this, it = std::move(it)](StatusWith<ResponseStatus> s) {
stdx::lock_guard<stdx::mutex> lg(_mutex);
_activeHandles.erase(it);
_notifyAllTasksComplete(lg);
});
});
}
std::unique_ptr<AsyncWorkScheduler> AsyncWorkScheduler::makeChildScheduler() {
auto child = std::make_unique<AsyncWorkScheduler>(_serviceContext);
stdx::lock_guard<stdx::mutex> lg(_mutex);
if (!_shutdownStatus.isOK())
child->shutdown(_shutdownStatus);
child->_parent = this;
child->_itToRemove = _childSchedulers.emplace(_childSchedulers.begin(), child.get());
return child;
}
void AsyncWorkScheduler::shutdown(Status status) {
invariant(!status.isOK());
stdx::lock_guard<stdx::mutex> lg(_mutex);
if (!_shutdownStatus.isOK())
return;
_shutdownStatus = std::move(status);
for (const auto& it : _activeOpContexts) {
stdx::lock_guard<Client> clientLock(*it->getClient());
_serviceContext->killOperation(clientLock, it.get(), _shutdownStatus.code());
}
for (const auto& it : _activeHandles) {
_executor->cancel(it);
}
for (auto& child : _childSchedulers) {
child->shutdown(_shutdownStatus);
}
}
void AsyncWorkScheduler::join() {
stdx::unique_lock<stdx::mutex> ul(_mutex);
_allListsEmptyCV.wait(ul, [&] {
return _activeOpContexts.empty() && _activeHandles.empty() && _childSchedulers.empty();
});
}
Future<AsyncWorkScheduler::HostAndShard> AsyncWorkScheduler::_targetHostAsync(
const ShardId& shardId,
const ReadPreferenceSetting& readPref,
OperationContextFn operationContextFn) {
return scheduleWork([shardId, readPref, operationContextFn](OperationContext* opCtx) {
operationContextFn(opCtx);
const auto shardRegistry = Grid::get(opCtx)->shardRegistry();
const auto shard = uassertStatusOK(shardRegistry->getShard(opCtx, shardId));
if (MONGO_FAIL_POINT(hangWhileTargetingRemoteHost)) {
LOG(0) << "Hit hangWhileTargetingRemoteHost failpoint for shard " << shardId;
MONGO_FAIL_POINT_PAUSE_WHILE_SET_OR_INTERRUPTED(opCtx, hangWhileTargetingRemoteHost);
}
// TODO (SERVER-35678): Return a SemiFuture<HostAndShard> rather than using a blocking call
return HostAndShard{
shard->getTargeter()->findHostWithMaxWait(readPref, Seconds(20)).get(opCtx),
std::move(shard)};
});
}
bool AsyncWorkScheduler::_quiesced(WithLock) const {
return _activeOpContexts.empty() && _activeHandles.empty() && _childSchedulers.empty();
}
void AsyncWorkScheduler::_notifyAllTasksComplete(WithLock wl) {
if (_quiesced(wl))
_allListsEmptyCV.notify_all();
}
ShardId getLocalShardId(ServiceContext* service) {
if (serverGlobalParams.clusterRole == ClusterRole::ConfigServer) {
return ShardRegistry::kConfigServerShardId;
}
if (serverGlobalParams.clusterRole == ClusterRole::ShardServer) {
return ShardingState::get(service)->shardId();
}
// Only sharded systems should use the two-phase commit path
MONGO_UNREACHABLE;
}
Future<void> whenAll(std::vector<Future<void>>& futures) {
std::vector<Future<int>> dummyFutures;
for (auto&& f : futures) {
dummyFutures.push_back(std::move(f).then([]() { return 0; }));
}
return collect(
std::move(dummyFutures), 0, [](int, const int&) { return ShouldStopIteration::kNo; })
.ignoreValue();
}
} // namespace txn
} // namespace mongo
|