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
|
/**
* 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.
*/
#include "mongo/platform/basic.h"
#include "mongo/db/s/periodic_sharded_index_consistency_checker.h"
#include "mongo/db/auth/privilege.h"
#include "mongo/db/curop.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/s/sharding_runtime_d_params_gen.h"
#include "mongo/db/service_context.h"
#include "mongo/logv2/log.h"
#include "mongo/s/grid.h"
#include "mongo/s/query/cluster_aggregate.h"
#include "mongo/s/stale_shard_version_helpers.h"
#define MONGO_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kSharding
namespace mongo {
namespace {
const auto getPeriodicShardedIndexConsistencyChecker =
ServiceContext::declareDecoration<PeriodicShardedIndexConsistencyChecker>();
} // namespace
PeriodicShardedIndexConsistencyChecker& PeriodicShardedIndexConsistencyChecker::get(
OperationContext* opCtx) {
return get(opCtx->getServiceContext());
}
PeriodicShardedIndexConsistencyChecker& PeriodicShardedIndexConsistencyChecker::get(
ServiceContext* serviceContext) {
return getPeriodicShardedIndexConsistencyChecker(serviceContext);
}
long long PeriodicShardedIndexConsistencyChecker::getNumShardedCollsWithInconsistentIndexes()
const {
stdx::lock_guard<Latch> lk(_mutex);
return _numShardedCollsWithInconsistentIndexes;
}
void PeriodicShardedIndexConsistencyChecker::_launchShardedIndexConsistencyChecker(
WithLock, ServiceContext* serviceContext) {
auto periodicRunner = serviceContext->getPeriodicRunner();
invariant(periodicRunner);
PeriodicRunner::PeriodicJob job(
"PeriodicShardedIndexConsistencyChecker",
[this](Client* client) {
if (!enableShardedIndexConsistencyCheck.load()) {
return;
}
LOGV2(22049, "Checking consistency of sharded collection indexes across the cluster");
const auto aggRequestBSON = fromjson(
"{pipeline: [{$indexStats: {}},"
"{$group:"
"{_id: null, indexDoc: {$push: \"$$ROOT\"}, allShards: {$addToSet: \"$shard\"}}},"
"{$unwind: \"$indexDoc\"},"
"{$group: "
"{\"_id\": \"$indexDoc.name\","
"\"shards\": {$push: \"$indexDoc.shard\"},"
"\"specs\": {$push: {$objectToArray: {$ifNull: [\"$indexDoc.spec\", {}]}}},"
"\"allShards\": {$first: \"$allShards\"}}},"
"{$project:"
" {missingFromShards: {$setDifference: [\"$allShards\", \"$shards\"]},"
" inconsistentProperties: {"
" $setDifference: ["
" {$reduce: {input: \"$specs\", initialValue: {$arrayElemAt: [\"$specs\", 0]},"
"in: {$setUnion: [\"$$value\", \"$$this\"]}}},"
" {$reduce: {input: \"$specs\", initialValue: {$arrayElemAt: [\"$specs\", 0]},"
"in: {$setIntersection: [\"$$value\", \"$$this\"]}}}]}}},"
"{$match:"
"{$expr: {$or: ["
" {$gt: [{$size: \"$missingFromShards\"}, 0]},"
" {$gt: [{$size: \"$inconsistentProperties\"}, 0]}]}}},"
"{$project:"
"{_id: 0, indexName: \"$$ROOT._id\", inconsistentProperties: 1, missingFromShards:"
"1}},"
"{$limit: 1}], cursor: {}}");
auto uniqueOpCtx = client->makeOperationContext();
auto opCtx = uniqueOpCtx.get();
auto curOp = CurOp::get(opCtx);
curOp->ensureStarted();
ON_BLOCK_EXIT([&] { curOp->done(); });
try {
long long numShardedCollsWithInconsistentIndexes = 0;
auto collections = Grid::get(opCtx)->catalogClient()->getCollections(
opCtx, {}, repl::ReadConcernLevel::kLocalReadConcern);
for (const auto& coll : collections) {
auto nss = coll.getNss();
// The only sharded collection in the config database with indexes is
// config.system.sessions. Unfortunately, the code path to run aggregation
// below would currently invariant if one of the targeted shards was the config
// server itself.
if (nss.isConfigDB()) {
continue;
}
auto request = aggregation_request_helper::parseFromBSON(
opCtx, nss, aggRequestBSON, boost::none, false);
auto catalogCache = Grid::get(opCtx)->catalogCache();
shardVersionRetry(
opCtx,
catalogCache,
nss,
"pipeline to detect inconsistent sharded indexes"_sd,
[&] {
BSONObjBuilder responseBuilder;
auto status = ClusterAggregate::runAggregate(
opCtx,
ClusterAggregate::Namespaces{nss, nss},
request,
LiteParsedPipeline{request},
PrivilegeVector(),
&responseBuilder);
// Stop counting if the agg command failed for one of the collections
// to avoid recording a false count.
uassertStatusOKWithContext(status, str::stream() << "nss " << nss);
if (!responseBuilder.obj()["cursor"]["firstBatch"].Array().empty()) {
numShardedCollsWithInconsistentIndexes++;
}
});
}
if (numShardedCollsWithInconsistentIndexes) {
LOGV2_WARNING(22051,
"Found {numShardedCollectionsWithInconsistentIndexes} sharded "
"collection(s) with inconsistent indexes",
"Found sharded collections with inconsistent indexes",
"numShardedCollectionsWithInconsistentIndexes"_attr =
numShardedCollsWithInconsistentIndexes);
}
// Update the count if this node is still primary. This is necessary because a
// stepdown may complete while this job is running and the count should always be
// zero on a non-primary node.
stdx::lock_guard<Latch> lk(_mutex);
if (_isPrimary) {
_numShardedCollsWithInconsistentIndexes =
numShardedCollsWithInconsistentIndexes;
}
} catch (DBException& ex) {
LOGV2(22052,
"Checking sharded index consistency failed with {error}",
"Error while checking sharded index consistency",
"error"_attr = ex.toStatus());
}
},
Milliseconds(shardedIndexConsistencyCheckIntervalMS));
_shardedIndexConsistencyChecker = periodicRunner->makeJob(std::move(job));
_shardedIndexConsistencyChecker.start();
}
void PeriodicShardedIndexConsistencyChecker::onStepUp(ServiceContext* serviceContext) {
stdx::lock_guard<Latch> lk(_mutex);
if (!_isPrimary) {
_isPrimary = true;
if (!_shardedIndexConsistencyChecker.isValid()) {
// If this is the first time we're stepping up, start a thread to periodically check
// index consistency.
_launchShardedIndexConsistencyChecker(lk, serviceContext);
} else {
// If we're stepping up again after having stepped down, just resume the existing task.
_shardedIndexConsistencyChecker.resume();
}
}
}
void PeriodicShardedIndexConsistencyChecker::onStepDown() {
stdx::lock_guard<Latch> lk(_mutex);
if (_isPrimary) {
_isPrimary = false;
invariant(_shardedIndexConsistencyChecker.isValid());
// Note pausing a periodic job does not wait for the job to complete if it is concurrently
// running, otherwise this would deadlock when the index check tries to lock _mutex when
// updating the inconsistent index count.
_shardedIndexConsistencyChecker.pause();
// Clear the counter to prevent a secondary from reporting an out-of-date count.
_numShardedCollsWithInconsistentIndexes = 0;
}
}
void PeriodicShardedIndexConsistencyChecker::onShutDown() {
if (_shardedIndexConsistencyChecker.isValid()) {
_shardedIndexConsistencyChecker.stop();
}
}
} // namespace mongo
|
