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-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/collection_sharding_runtime.h"
#include "mongo/base/checked_cast.h"
#include "mongo/db/catalog_raii.h"
#include "mongo/db/server_parameters.h"
#include "mongo/util/duration.h"
#include "mongo/util/log.h"
namespace mongo {
MONGO_EXPORT_SERVER_PARAMETER(migrationLockAcquisitionMaxWaitMS, int, 500);
namespace {
// How long to wait before starting cleanup of an emigrated chunk range
MONGO_EXPORT_SERVER_PARAMETER(orphanCleanupDelaySecs, int, 900); // 900s = 15m
/**
* Returns whether the specified namespace is used for sharding-internal purposes only and can never
* be marked as anything other than UNSHARDED, because the call sites which reference these
* collections are not prepared to handle StaleConfig errors.
*/
bool isNamespaceAlwaysUnsharded(const NamespaceString& nss) {
// There should never be a case to mark as sharded collections which are on the config server
if (serverGlobalParams.clusterRole != ClusterRole::ShardServer)
return true;
// Local and admin never have sharded collections
if (nss.db() == NamespaceString::kLocalDb || nss.db() == NamespaceString::kAdminDb)
return true;
// Certain config collections can never be sharded
if (nss == NamespaceString::kSessionTransactionsTableNamespace)
return true;
if (nss.isSystemDotProfile())
return true;
return false;
}
} // namespace
CollectionShardingRuntime::CollectionShardingRuntime(ServiceContext* sc,
NamespaceString nss,
executor::TaskExecutor* rangeDeleterExecutor)
: CollectionShardingState(nss),
_nss(std::move(nss)),
_metadataManager(std::make_shared<MetadataManager>(sc, _nss, rangeDeleterExecutor)) {
if (isNamespaceAlwaysUnsharded(_nss)) {
_metadataManager->setFilteringMetadata(CollectionMetadata());
}
}
CollectionShardingRuntime* CollectionShardingRuntime::get(OperationContext* opCtx,
const NamespaceString& nss) {
auto* const css = CollectionShardingState::get(opCtx, nss);
return checked_cast<CollectionShardingRuntime*>(css);
}
void CollectionShardingRuntime::setFilteringMetadata(OperationContext* opCtx,
CollectionMetadata newMetadata) {
invariant(!newMetadata.isSharded() || !isNamespaceAlwaysUnsharded(_nss),
str::stream() << "Namespace " << _nss.ns() << " must never be sharded.");
invariant(opCtx->lockState()->isCollectionLockedForMode(_nss.ns(), MODE_X));
_metadataManager->setFilteringMetadata(std::move(newMetadata));
}
void CollectionShardingRuntime::clearFilteringMetadata() {
if (!isNamespaceAlwaysUnsharded(_nss)) {
_metadataManager->clearFilteringMetadata();
}
}
auto CollectionShardingRuntime::beginReceive(ChunkRange const& range) -> CleanupNotification {
return _metadataManager->beginReceive(range);
}
void CollectionShardingRuntime::forgetReceive(const ChunkRange& range) {
_metadataManager->forgetReceive(range);
}
auto CollectionShardingRuntime::cleanUpRange(ChunkRange const& range, CleanWhen when)
-> CleanupNotification {
Date_t time =
(when == kNow) ? Date_t{} : Date_t::now() + Seconds(orphanCleanupDelaySecs.load());
return _metadataManager->cleanUpRange(range, time);
}
Status CollectionShardingRuntime::waitForClean(OperationContext* opCtx,
const NamespaceString& nss,
OID const& epoch,
ChunkRange orphanRange) {
while (true) {
boost::optional<CleanupNotification> stillScheduled;
{
AutoGetCollection autoColl(opCtx, nss, MODE_IX);
auto* const self = CollectionShardingRuntime::get(opCtx, nss);
{
// First, see if collection was dropped, but do it in a separate scope in order to
// not hold reference on it, which would make it appear in use
const auto optMetadata =
self->_metadataManager->getActiveMetadata(self->_metadataManager, boost::none);
if (!optMetadata)
return {ErrorCodes::ConflictingOperationInProgress,
"Collection being migrated had its metadata reset"};
const auto& metadata = *optMetadata;
if (!metadata->isSharded() || metadata->getCollVersion().epoch() != epoch) {
return {ErrorCodes::ConflictingOperationInProgress,
"Collection being migrated was dropped"};
}
}
stillScheduled = self->trackOrphanedDataCleanup(orphanRange);
if (!stillScheduled) {
log() << "Finished deleting " << nss.ns() << " range "
<< redact(orphanRange.toString());
return Status::OK();
}
}
log() << "Waiting for deletion of " << nss.ns() << " range " << orphanRange;
Status result = stillScheduled->waitStatus(opCtx);
if (!result.isOK()) {
return result.withContext(str::stream() << "Failed to delete orphaned " << nss.ns()
<< " range "
<< orphanRange.toString());
}
}
MONGO_UNREACHABLE;
}
auto CollectionShardingRuntime::trackOrphanedDataCleanup(ChunkRange const& range)
-> boost::optional<CleanupNotification> {
return _metadataManager->trackOrphanedDataCleanup(range);
}
boost::optional<ChunkRange> CollectionShardingRuntime::getNextOrphanRange(BSONObj const& from) {
return _metadataManager->getNextOrphanRange(from);
}
boost::optional<ScopedCollectionMetadata> CollectionShardingRuntime::_getMetadata(
const boost::optional<mongo::LogicalTime>& atClusterTime) {
return _metadataManager->getActiveMetadata(_metadataManager, atClusterTime);
}
CollectionCriticalSection::CollectionCriticalSection(OperationContext* opCtx, NamespaceString ns)
: _nss(std::move(ns)), _opCtx(opCtx) {
AutoGetCollection autoColl(_opCtx,
_nss,
MODE_IX,
MODE_X,
AutoGetCollection::ViewMode::kViewsForbidden,
opCtx->getServiceContext()->getPreciseClockSource()->now() +
Milliseconds(migrationLockAcquisitionMaxWaitMS.load()));
auto* const csr = CollectionShardingRuntime::get(_opCtx, _nss);
auto csrLock = CollectionShardingState::CSRLock::lockExclusive(_opCtx, csr);
csr->enterCriticalSectionCatchUpPhase(_opCtx, csrLock);
}
CollectionCriticalSection::~CollectionCriticalSection() {
UninterruptibleLockGuard noInterrupt(_opCtx->lockState());
AutoGetCollection autoColl(_opCtx, _nss, MODE_IX, MODE_IX);
auto* const csr = CollectionShardingRuntime::get(_opCtx, _nss);
auto csrLock = CollectionShardingRuntime::CSRLock::lockExclusive(_opCtx, csr);
csr->exitCriticalSection(_opCtx, csrLock);
}
void CollectionCriticalSection::enterCommitPhase() {
AutoGetCollection autoColl(_opCtx,
_nss,
MODE_IX,
MODE_X,
AutoGetCollection::ViewMode::kViewsForbidden,
_opCtx->getServiceContext()->getPreciseClockSource()->now() +
Milliseconds(migrationLockAcquisitionMaxWaitMS.load()));
auto* const csr = CollectionShardingRuntime::get(_opCtx, _nss);
auto csrLock = CollectionShardingRuntime::CSRLock::lockExclusive(_opCtx, csr);
csr->enterCriticalSectionCommitPhase(_opCtx, csrLock);
}
} // namespace mongo
|