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
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
|
/**
* Copyright (C) 2020-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_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kResharding
#include "mongo/platform/basic.h"
#include "mongo/db/s/resharding/resharding_txn_cloner.h"
#include <vector>
#include "mongo/bson/bsonmisc.h"
#include "mongo/client/query.h"
#include "mongo/client/read_preference.h"
#include "mongo/db/client.h"
#include "mongo/db/concurrency/d_concurrency.h"
#include "mongo/db/concurrency/write_conflict_exception.h"
#include "mongo/db/curop.h"
#include "mongo/db/logical_session_id.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/persistent_task_store.h"
#include "mongo/db/pipeline/document_source_match.h"
#include "mongo/db/pipeline/document_source_sort.h"
#include "mongo/db/pipeline/process_interface/mongo_process_interface.h"
#include "mongo/db/pipeline/sharded_agg_helpers.h"
#include "mongo/db/repl/optime.h"
#include "mongo/db/repl/read_concern_args.h"
#include "mongo/db/repl/read_concern_level.h"
#include "mongo/db/s/resharding/resharding_data_copy_util.h"
#include "mongo/db/s/resharding/resharding_server_parameters_gen.h"
#include "mongo/db/s/resharding/resharding_txn_cloner_progress_gen.h"
#include "mongo/db/s/session_catalog_migration_destination.h"
#include "mongo/db/session_catalog_mongod.h"
#include "mongo/db/session_txn_record_gen.h"
#include "mongo/db/transaction_participant.h"
#include "mongo/logv2/log.h"
#include "mongo/logv2/redaction.h"
#include "mongo/util/future_util.h"
namespace mongo {
ReshardingTxnCloner::ReshardingTxnCloner(ReshardingSourceId sourceId, Timestamp fetchTimestamp)
: _sourceId(std::move(sourceId)), _fetchTimestamp(fetchTimestamp) {}
std::unique_ptr<Pipeline, PipelineDeleter> ReshardingTxnCloner::makePipeline(
OperationContext* opCtx,
std::shared_ptr<MongoProcessInterface> mongoProcessInterface,
const boost::optional<LogicalSessionId>& startAfter) {
const auto& sourceNss = NamespaceString::kSessionTransactionsTableNamespace;
StringMap<ExpressionContext::ResolvedNamespace> resolvedNamespaces;
resolvedNamespaces[sourceNss.coll()] = {sourceNss, std::vector<BSONObj>{}};
auto expCtx = make_intrusive<ExpressionContext>(opCtx,
boost::none, /* explain */
false, /* fromMongos */
false, /* needsMerge */
false, /* allowDiskUse */
false, /* bypassDocumentValidation */
false, /* isMapReduceCommand */
sourceNss,
boost::none, /* runtimeConstants */
nullptr, /* collator */
std::move(mongoProcessInterface),
std::move(resolvedNamespaces),
boost::none /* collUUID */);
Pipeline::SourceContainer stages;
if (startAfter) {
stages.emplace_back(DocumentSourceMatch::create(
BSON(SessionTxnRecord::kSessionIdFieldName << BSON("$gt" << startAfter->toBSON())),
expCtx));
}
stages.emplace_back(
DocumentSourceSort::create(expCtx, BSON(SessionTxnRecord::kSessionIdFieldName << 1)));
return Pipeline::create(std::move(stages), expCtx);
}
boost::optional<LogicalSessionId> ReshardingTxnCloner::_fetchProgressLsid(OperationContext* opCtx) {
PersistentTaskStore<ReshardingTxnClonerProgress> store(
NamespaceString::kReshardingTxnClonerProgressNamespace);
boost::optional<LogicalSessionId> progressLsid;
store.forEach(opCtx,
QUERY(ReshardingTxnClonerProgress::kSourceIdFieldName << _sourceId.toBSON()),
[&](const auto& doc) {
progressLsid = doc.getProgress();
return false;
});
return progressLsid;
}
std::unique_ptr<Pipeline, PipelineDeleter> ReshardingTxnCloner::_targetAggregationRequest(
OperationContext* opCtx, const Pipeline& pipeline) {
AggregateCommandRequest request(NamespaceString::kSessionTransactionsTableNamespace,
pipeline.serializeToBson());
request.setReadConcern(BSON(repl::ReadConcernArgs::kLevelFieldName
<< repl::readConcernLevels::kSnapshotName
<< repl::ReadConcernArgs::kAtClusterTimeFieldName
<< _fetchTimestamp));
request.setWriteConcern(WriteConcernOptions());
request.setHint(BSON(SessionTxnRecord::kSessionIdFieldName << 1));
request.setUnwrappedReadPref(ReadPreferenceSetting{ReadPreference::Nearest}.toContainingBSON());
return sharded_agg_helpers::runPipelineDirectlyOnSingleShard(
pipeline.getContext(), std::move(request), _sourceId.getShardId());
}
void ReshardingTxnCloner::_updateProgressDocument(OperationContext* opCtx,
const LogicalSessionId& progress) {
PersistentTaskStore<ReshardingTxnClonerProgress> store(
NamespaceString::kReshardingTxnClonerProgressNamespace);
store.upsert(
opCtx,
QUERY(ReshardingTxnClonerProgress::kSourceIdFieldName << _sourceId.toBSON()),
BSON("$set" << BSON(ReshardingTxnClonerProgress::kProgressFieldName << progress.toBSON())),
{1, WriteConcernOptions::SyncMode::UNSET, Seconds(0)});
}
/**
* Invokes the 'callable' function with a fresh OperationContext.
*
* The OperationContext is configured so the RstlKillOpThread would always interrupt the operation
* on step-up or stepdown, regardless of whether the operation has acquired any locks. This
* interruption is best-effort to stop doing wasteful work on stepdown as quickly as possible. It
* isn't required for the ReshardingTxnCloner's correctness. In particular, it is possible for an
* OperationContext to be constructed after stepdown has finished, for the ReshardingTxnCloner to
* run a getMore on the aggregation against the donor shards, and for the ReshardingTxnCloner to
* only discover afterwards the recipient had already stepped down from a NotPrimary error when
* updating a session record locally.
*
* Note that the recipient's primary-only service is responsible for managing the
* ReshardingTxnCloner and would shut down the ReshardingTxnCloner's task executor following the
* recipient stepping down.
*
* Also note that the ReshardingTxnCloner is only created after step-up as part of the recipient's
* primary-only service and therefore would never be interrupted by step-up.
*/
template <typename Callable>
auto ReshardingTxnCloner::_withTemporaryOperationContext(Callable&& callable) {
auto& client = cc();
{
stdx::lock_guard<Client> lk(client);
invariant(client.canKillSystemOperationInStepdown(lk));
}
auto opCtx = client.makeOperationContext();
opCtx->setAlwaysInterruptAtStepDownOrUp();
// The BlockingResultsMerger underlying by the $mergeCursors stage records how long the
// recipient spent waiting for documents from the donor shards. It doing so requires the CurOp
// to be marked as having started.
auto* curOp = CurOp::get(opCtx.get());
curOp->ensureStarted();
{
ON_BLOCK_EXIT([curOp] { curOp->done(); });
return callable(opCtx.get());
}
}
SemiFuture<void> ReshardingTxnCloner::run(
std::shared_ptr<executor::TaskExecutor> executor,
std::shared_ptr<executor::TaskExecutor> cleanupExecutor,
CancellationToken cancelToken,
std::shared_ptr<MongoProcessInterface> mongoProcessInterface_forTest) {
struct ChainContext {
std::unique_ptr<Pipeline, PipelineDeleter> pipeline;
boost::optional<SessionTxnRecord> donorRecord = boost::none;
bool moreToCome = true;
int progressCounter = 0;
};
auto chainCtx = std::make_shared<ChainContext>();
return AsyncTry([this, chainCtx, mongoProcessInterface_forTest] {
if (!chainCtx->pipeline) {
chainCtx->pipeline = _withTemporaryOperationContext([&](auto* opCtx) {
auto progressLsid = _fetchProgressLsid(opCtx);
auto mongoProcessInterface = MONGO_unlikely(mongoProcessInterface_forTest)
? mongoProcessInterface_forTest
: MongoProcessInterface::create(opCtx);
auto pipeline = _targetAggregationRequest(
opCtx,
*makePipeline(opCtx, std::move(mongoProcessInterface), progressLsid));
pipeline->detachFromOperationContext();
pipeline.get_deleter().dismissDisposal();
return pipeline;
});
chainCtx->donorRecord = boost::none;
}
// A donor record will have been stashed on the ChainContext if we are resuming due
// to a prepared transaction having been in progress.
if (!chainCtx->donorRecord) {
chainCtx->donorRecord = _withTemporaryOperationContext([&](auto* opCtx) {
chainCtx->pipeline->reattachToOperationContext(opCtx);
auto doc = chainCtx->pipeline->getNext();
chainCtx->pipeline->detachFromOperationContext();
return doc ? SessionTxnRecord::parse(
{"resharding config.transactions cloning"}, doc->toBson())
: boost::optional<SessionTxnRecord>{};
});
}
if (!chainCtx->donorRecord) {
chainCtx->moreToCome = false;
return makeReadyFutureWith([] {}).share();
}
auto hitPreparedTxn = _withTemporaryOperationContext([&](auto* opCtx) {
return resharding::data_copy::withSessionCheckedOut(
opCtx,
chainCtx->donorRecord->getSessionId(),
chainCtx->donorRecord->getTxnNum(),
boost::none /* stmtId */,
[&] {
resharding::data_copy::updateSessionRecord(
opCtx,
TransactionParticipant::kDeadEndSentinel,
{kIncompleteHistoryStmtId},
boost::none /* preImageOpTime */,
boost::none /* postImageOpTime */);
});
});
if (hitPreparedTxn) {
return *hitPreparedTxn;
}
chainCtx->progressCounter = (chainCtx->progressCounter + 1) %
resharding::gReshardingTxnClonerProgressBatchSize;
if (chainCtx->progressCounter == 0) {
_withTemporaryOperationContext([&](auto* opCtx) {
_updateProgressDocument(opCtx, chainCtx->donorRecord->getSessionId());
});
}
chainCtx->donorRecord = boost::none;
return makeReadyFutureWith([] {}).share();
})
.until([this, cancelToken, chainCtx](Status status) {
if (status.isOK() && chainCtx->moreToCome) {
return false;
}
if (chainCtx->pipeline) {
_withTemporaryOperationContext([&](auto* opCtx) {
chainCtx->pipeline->dispose(opCtx);
chainCtx->pipeline.reset();
});
}
if (status.isA<ErrorCategory::CancellationError>() ||
status.isA<ErrorCategory::NotPrimaryError>()) {
// Cancellation and NotPrimary errors indicate the primary-only service Instance
// will be shut down or is shutting down now - provided the cancelToken is also
// canceled. Otherwise, the errors may have originated from a remote response rather
// than the shard itself.
//
// Don't retry when primary-only service Instance is shutting down.
return !cancelToken.isCanceled();
}
if (status.isA<ErrorCategory::RetriableError>() ||
status.isA<ErrorCategory::CursorInvalidatedError>() ||
status == ErrorCodes::Interrupted) {
// Do retry on any other types of retryable errors though. Also retry on errors from
// stray killCursors and killOp commands being run.
LOGV2(5461600,
"Transient error while cloning config.transactions collection",
"sourceId"_attr = _sourceId,
"fetchTimestamp"_attr = _fetchTimestamp,
"error"_attr = redact(status));
return false;
}
if (!status.isOK()) {
LOGV2(5461601,
"Operation-fatal error for resharding while cloning config.transactions"
" collection",
"sourceId"_attr = _sourceId,
"fetchTimestamp"_attr = _fetchTimestamp,
"error"_attr = redact(status));
}
return true;
})
.on(executor, cancelToken)
.thenRunOn(cleanupExecutor)
.onCompletion([this, chainCtx](Status status) {
if (chainCtx->pipeline) {
// Guarantee the pipeline is always cleaned up - even upon cancellation.
_withTemporaryOperationContext([&](auto* opCtx) {
chainCtx->pipeline->dispose(opCtx);
chainCtx->pipeline.reset();
});
}
// Propagate the result of the AsyncTry.
return status;
})
.semi();
}
std::unique_ptr<Pipeline, PipelineDeleter> createConfigTxnCloningPipelineForResharding(
const boost::intrusive_ptr<ExpressionContext>& expCtx,
Timestamp fetchTimestamp,
boost::optional<LogicalSessionId> startAfter) {
invariant(!fetchTimestamp.isNull());
ReshardingSourceId sourceId(UUID::gen(), ShardId("dummyShardId"));
ReshardingTxnCloner cloner(std::move(sourceId), fetchTimestamp);
return cloner.makePipeline(expCtx->opCtx, expCtx->mongoProcessInterface, startAfter);
}
} // namespace mongo
|