-
Notifications
You must be signed in to change notification settings - Fork 5.6k
/
sharding_catalog_client_impl.cpp
1882 lines (1649 loc) · 79.5 KB
/
sharding_catalog_client_impl.cpp
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
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/**
* 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.
*/
#include "mongo/s/catalog/sharding_catalog_client_impl.h"
#include <absl/container/node_hash_map.h>
#include <algorithm>
#include <boost/cstdint.hpp>
#include <boost/move/utility_core.hpp>
#include <boost/smart_ptr/intrusive_ptr.hpp>
#include <cstdint>
#include <fmt/format.h>
#include <iterator>
#include <type_traits>
#include <variant>
#include <boost/none.hpp>
#include <boost/optional/optional.hpp>
#include "mongo/base/error_codes.h"
#include "mongo/bson/bson_field.h"
#include "mongo/bson/bsonelement.h"
#include "mongo/bson/bsonmisc.h"
#include "mongo/bson/bsonobjbuilder.h"
#include "mongo/bson/util/bson_extract.h"
#include "mongo/client/read_preference.h"
#include "mongo/db/cluster_role.h"
#include "mongo/db/commands.h"
#include "mongo/db/database_name.h"
#include "mongo/db/exec/document_value/document.h"
#include "mongo/db/exec/document_value/value.h"
#include "mongo/db/feature_flag.h"
#include "mongo/db/namespace_string.h"
#include "mongo/db/operation_context.h"
#include "mongo/db/pipeline/aggregate_command_gen.h"
#include "mongo/db/pipeline/document_source_facet.h"
#include "mongo/db/pipeline/document_source_group.h"
#include "mongo/db/pipeline/document_source_limit.h"
#include "mongo/db/pipeline/document_source_lookup.h"
#include "mongo/db/pipeline/document_source_match.h"
#include "mongo/db/pipeline/document_source_project.h"
#include "mongo/db/pipeline/document_source_sort.h"
#include "mongo/db/pipeline/document_source_union_with.h"
#include "mongo/db/pipeline/expression_context.h"
#include "mongo/db/pipeline/pipeline.h"
#include "mongo/db/query/write_ops/write_ops_gen.h"
#include "mongo/db/query/write_ops/write_ops_parsers.h"
#include "mongo/db/repl/optime.h"
#include "mongo/db/repl/read_concern_args.h"
#include "mongo/db/server_options.h"
#include "mongo/db/vector_clock.h"
#include "mongo/idl/idl_parser.h"
#include "mongo/logv2/log.h"
#include "mongo/logv2/log_component.h"
#include "mongo/platform/atomic_word.h"
#include "mongo/s/catalog/type_chunk.h"
#include "mongo/s/catalog/type_collection.h"
#include "mongo/s/catalog/type_collection_gen.h"
#include "mongo/s/catalog/type_config_version.h"
#include "mongo/s/catalog/type_database_gen.h"
#include "mongo/s/catalog/type_namespace_placement_gen.h"
#include "mongo/s/catalog/type_remove_shard_event_gen.h"
#include "mongo/s/catalog/type_shard.h"
#include "mongo/s/catalog/type_tags.h"
#include "mongo/s/client/shard.h"
#include "mongo/s/client/shard_registry.h"
#include "mongo/s/database_version.h"
#include "mongo/s/grid.h"
#include "mongo/s/sharding_cluster_parameters_gen.h"
#include "mongo/s/write_ops/batched_command_request.h"
#include "mongo/s/write_ops/batched_command_response.h"
#include "mongo/util/assert_util.h"
#include "mongo/util/intrusive_counter.h"
#include "mongo/util/namespace_string_util.h"
#include "mongo/util/pcre_util.h"
#include "mongo/util/str.h"
#include "mongo/util/string_map.h"
#define MONGO_LOGV2_DEFAULT_COMPONENT ::mongo::logv2::LogComponent::kSharding
namespace mongo {
namespace {
using namespace fmt::literals;
const ReadPreferenceSetting kConfigNearestReadPreference(ReadPreference::Nearest, TagSet{});
const ReadPreferenceSetting kConfigPrimaryPreferredReadPreference(ReadPreference::PrimaryPreferred,
TagSet{});
const int kMaxReadRetry = 3;
const int kMaxWriteRetry = 3;
/**
* This function returns `true` if the `configServerReadPreferenceForCatalogQueries`
* cluster server parameter exists and is set to indicate that the nearest read preference
* must always be used. If the parameter is not known or is set to false, it returns `false`.
*/
bool forceUsageOfNearestReadPreference() {
auto* configServerReadPreferenceForCatalogQueriesParam =
ServerParameterSet::getClusterParameterSet()
->getIfExists<
ClusterParameterWithStorage<ConfigServerReadPreferenceForCatalogQueriesParam>>(
"configServerReadPreferenceForCatalogQueries");
if (!configServerReadPreferenceForCatalogQueriesParam) {
return false;
}
return configServerReadPreferenceForCatalogQueriesParam->getValue(boost::none)
.getMustAlwaysUseNearest();
}
/**
* Returns the appropriate read preference for queries targeting the config server:
* - If the read preference cluster server parameter is set, it returns 'nearest'.
* - For clusters with a config shard, it returns 'primaryPreferred'.
* - For clusters with a dedicated config server, it returns 'nearest'.
*
* To check if the cluster has a config shard, the cached data is consulted without causal
* consistency. This is because we don't want to introduce extra latency to queries targeting the
* config server to determine a read preference.
*
* Note: The selection of the read preference only makes sense when using the remote catalog client
* and not the local one.
* TODO (SERVER-91526): Make an early exit when using the local catalog client.
*/
ReadPreferenceSetting getConfigReadPreference(OperationContext* opCtx) {
if (forceUsageOfNearestReadPreference()) {
return kConfigNearestReadPreference;
}
const auto shardRegistry = Grid::get(opCtx)->shardRegistry();
const auto optClusterHasConfigShard = shardRegistry->cachedClusterHasConfigShard();
if (!optClusterHasConfigShard) {
return kConfigPrimaryPreferredReadPreference;
}
return *optClusterHasConfigShard ? kConfigPrimaryPreferredReadPreference
: kConfigNearestReadPreference;
}
void toBatchError(const Status& status, BatchedCommandResponse* response) {
response->clear();
response->setStatus(status);
}
AggregateCommandRequest makeCollectionAndChunksAggregation(OperationContext* opCtx,
const NamespaceString& nss,
const ChunkVersion& sinceVersion) {
auto expCtx = make_intrusive<ExpressionContext>(opCtx, nullptr, CollectionType::ConfigNS);
StringMap<ExpressionContext::ResolvedNamespace> resolvedNamespaces;
resolvedNamespaces[CollectionType::ConfigNS.coll()] = {CollectionType::ConfigNS,
std::vector<BSONObj>()};
resolvedNamespaces[ChunkType::ConfigNS.coll()] = {ChunkType::ConfigNS, std::vector<BSONObj>()};
expCtx->setResolvedNamespaces(resolvedNamespaces);
using Doc = Document;
using Arr = std::vector<Value>;
Pipeline::SourceContainer stages;
// 1. Match config.collections entries with {_id: nss}. This stage will produce, at most, one
// config.collections document.
// {
// $match: {
// _id: <nss>
// }
// }
stages.emplace_back(DocumentSourceMatch::create(
Doc{{CollectionType::kNssFieldName,
NamespaceStringUtil::serialize(nss, SerializationContext::stateDefault())}}
.toBson(),
expCtx));
// 2. Two $unionWith stages guarded by a mutually exclusive condition on whether the refresh is
// incremental ('lastmodEpoch' matches sinceVersion.epoch), so that only a single one of them
// will possibly execute their $lookup stage. This is necessary because the query optimizer is
// not able to use indexes when a $match inside a $lookup includes a $cond operator.
//
// The $lookup stages get the config.chunks documents according to the
// type of refresh (incremental or full), sorted by ascending 'lastmod'. The $lookup is
// immediately followed by $unwind to take advantage of the $lookup + $unwind coalescence
// optimization which avoids creating large intermediate documents.
//
// This $unionWith stage will produce one result document for each config.chunks document
// matching the refresh query.
// Note that we must not make any assumption on where the document produced by stage 1 will be
// placed in the response in relation with the documents produced by stage 2. The
// config.collections document produced in stage 1 could be interleaved between the
// config.chunks documents produced by stage 2.
//
// {
// $unionWith: {
// coll: "collections",
// pipeline: [
// { $match: { _id: <nss> } },
// { $match: { lastmodEpoch: <sinceVersion.epoch> } },
// {
// $lookup: {
// from: "chunks",
// as: "chunks",
// let: { local_uuid: "$uuid" },
// pipeline: [
// {
// $match: {
// $expr: {
// $eq: ["$uuid", "$$local_uuid"],
// },
// }
// },
// { $match: { lastmod: { $gte: <sinceVersion> } } },
// {
// $sort: {
// lastmod: 1
// }
// }
// ]
// }
// },
// {
// $unwind: {
// path: "$chunks"
// }
// },
// {
// $project: { _id: false, chunks: true }
// },
// ]
// }
// },
// {
// $unionWith: {
// coll: "collections",
// pipeline: [
// { $match: { _id: <nss> } },
// { $match: { lastmodEpoch: { $ne: <sinceVersion.epoch> } } },
// {
// $lookup: {
// from: "chunks",
// as: "chunks",
// let: { local_uuid: "$uuid" },
// pipeline: [
// {
// $match: {
// $expr: {
// $eq: ["$uuid", "$$local_uuid"],
// },
// }
// },
// {
// $sort: {
// lastmod: 1
// }
// }
// ]
// }
// },
// {
// $unwind: {
// path: "$chunks"
// }
// },
// {
// $project: { _id: false, chunks: true }
// },
// ]
// }
// }
const auto buildUnionWithFn = [&](bool incremental) {
const auto lastmodEpochMatch = Doc{{incremental ? "$eq" : "$ne", sinceVersion.epoch()}};
const auto letExpr = Doc{{"local_uuid", "$" + CollectionType::kUuidFieldName}};
const auto uuidExpr =
Arr{Value{"$" + ChunkType::collectionUUID.name()}, Value{"$$local_uuid"_sd}};
constexpr auto chunksLookupOutputFieldName = "chunks"_sd;
const auto lookupPipeline = [&]() {
return Doc{
{"from", ChunkType::ConfigNS.coll()},
{"as", chunksLookupOutputFieldName},
{"let", letExpr},
{"pipeline",
Arr{Value{Doc{{"$match", Doc{{"$expr", Doc{{"$eq", uuidExpr}}}}}}},
incremental
? Value{Doc{{"$match",
Doc{{ChunkType::lastmod.name(),
Doc{{"$gte", Timestamp(sinceVersion.toLong())}}}}}}}
: Value{/*noop*/},
Value{Doc{{"$sort", Doc{{ChunkType::lastmod.name(), 1}}}}}}}};
}();
return Doc{
{"coll", CollectionType::ConfigNS.coll()},
{"pipeline",
Arr{Value{Doc{{"$match",
Doc{{CollectionType::kNssFieldName,
NamespaceStringUtil::serialize(
nss, SerializationContext::stateDefault())}}}}},
Value{Doc{{"$match", Doc{{CollectionType::kEpochFieldName, lastmodEpochMatch}}}}},
Value{Doc{{"$lookup", lookupPipeline}}},
Value{Doc{{"$unwind", Doc{{"path", "$" + chunksLookupOutputFieldName}}}}},
Value{Doc{
{"$project", Doc{{"_id", false}, {chunksLookupOutputFieldName, true}}}}}}}};
};
stages.emplace_back(DocumentSourceUnionWith::createFromBson(
Doc{{"$unionWith", buildUnionWithFn(true /* incremental */)}}.toBson().firstElement(),
expCtx));
stages.emplace_back(DocumentSourceUnionWith::createFromBson(
Doc{{"$unionWith", buildUnionWithFn(false /* incremental */)}}.toBson().firstElement(),
expCtx));
auto pipeline = Pipeline::create(std::move(stages), expCtx);
auto serializedPipeline = pipeline->serializeToBson();
return AggregateCommandRequest(CollectionType::ConfigNS, std::move(serializedPipeline));
}
AggregateCommandRequest makeCollectionAndIndexesAggregation(OperationContext* opCtx,
const NamespaceString& nss) {
auto expCtx = make_intrusive<ExpressionContext>(opCtx, nullptr, CollectionType::ConfigNS);
StringMap<ExpressionContext::ResolvedNamespace> resolvedNamespaces;
resolvedNamespaces[CollectionType::ConfigNS.coll()] = {CollectionType::ConfigNS,
std::vector<BSONObj>()};
resolvedNamespaces[NamespaceString::kConfigsvrIndexCatalogNamespace.coll()] = {
NamespaceString::kConfigsvrIndexCatalogNamespace, std::vector<BSONObj>()};
expCtx->setResolvedNamespaces(resolvedNamespaces);
using Doc = Document;
using Arr = std::vector<Value>;
Pipeline::SourceContainer stages;
// 1. Match config.collections entries with {_id: nss}. This stage will produce, at most, one
// config.collections document.
// {
// $match: {
// _id: <nss>
// }
// }
stages.emplace_back(DocumentSourceMatch::create(
Doc{{CollectionType::kNssFieldName,
NamespaceStringUtil::serialize(nss, SerializationContext::stateDefault())}}
.toBson(),
expCtx));
// 2. Retrieve config.csrs.indexes entries with the same uuid as the one from the
// config.collections document.
//
// The $lookup stage gets the config.csrs.indexes documents and puts them in a field called
// "indexes" in the document produced during stage 1.
//
// {
// $lookup: {
// from: "csrs.indexes",
// as: "indexes",
// localField: "uuid",
// foreignField: "collectionUUID"
// }
// }
const Doc lookupPipeline{{"from", NamespaceString::kConfigsvrIndexCatalogNamespace.coll()},
{"as", "indexes"_sd},
{"localField", CollectionType::kUuidFieldName},
{"foreignField", IndexCatalogType::kCollectionUUIDFieldName}};
stages.emplace_back(DocumentSourceLookUp::createFromBson(
Doc{{"$lookup", lookupPipeline}}.toBson().firstElement(), expCtx));
auto pipeline = Pipeline::create(std::move(stages), expCtx);
auto serializedPipeline = pipeline->serializeToBson();
return AggregateCommandRequest(CollectionType::ConfigNS, std::move(serializedPipeline));
}
AggregateCommandRequest makeUnsplittableCollectionsDataShardAggregation(
OperationContext* opCtx,
const DatabaseName& dbName,
const std::vector<ShardId>& excludedShards) {
auto expCtx = make_intrusive<ExpressionContext>(opCtx, nullptr, CollectionType::ConfigNS);
StringMap<ExpressionContext::ResolvedNamespace> resolvedNamespaces;
resolvedNamespaces[CollectionType::ConfigNS.coll()] = {CollectionType::ConfigNS,
std::vector<BSONObj>()};
resolvedNamespaces[NamespaceString::kConfigsvrChunksNamespace.coll()] = {
NamespaceString::kConfigsvrChunksNamespace, std::vector<BSONObj>()};
expCtx->setResolvedNamespaces(resolvedNamespaces);
using Doc = Document;
using Arr = std::vector<Value>;
Pipeline::SourceContainer stages;
// 1. Match config.collections entries with database name = dbName
// {
// $match: {
// _id: {$regex: dbName.*, unsplittable: true}
// }
// }
const auto db =
DatabaseNameUtil::serialize(dbName, SerializationContext::stateCommandRequest());
stages.emplace_back(DocumentSourceMatch::create(
BSON(CollectionType::kNssFieldName << BSON("$regex"
<< "^{}\\."_format(pcre_util::quoteMeta(db)))
<< CollectionType::kUnsplittableFieldName << true),
expCtx));
// 2. Retrieve config.chunks entries with the same uuid as the one from the
// config.collections document.
//
// The $lookup stage gets the config.chunks documents and puts them in a field called
// "chunks" in the document produced during stage 1.
//
// {
// $lookup: {
// from: "chunks",
// as: "chunks",
// localField: "uuid",
// foreignField: "uuid"
// }
// }
const Doc lookupPipeline{{"from", NamespaceString::kConfigsvrChunksNamespace.coll()},
{"as", "chunks"_sd},
{"localField", CollectionType::kUuidFieldName},
{"foreignField", CollectionType::kUuidFieldName}};
stages.emplace_back(DocumentSourceLookUp::createFromBson(
Doc{{"$lookup", lookupPipeline}}.toBson().firstElement(), expCtx));
// 3. Filter only the collection entries where the chunk has the shard field equal to shardId.
// {
// $match: {
// chunks.shard: {$nin: <excludedShards>}
// }
// }
BSONObjBuilder ninBuilder;
ninBuilder.append("$nin", excludedShards);
stages.emplace_back(
DocumentSourceMatch::create(Doc{{"chunks.shard", ninBuilder.obj()}}.toBson(), expCtx));
auto pipeline = Pipeline::create(std::move(stages), expCtx);
auto serializedPipeline = pipeline->serializeToBson();
return AggregateCommandRequest(CollectionType::ConfigNS, std::move(serializedPipeline));
}
/**
* Returns keys for the given purpose and have an expiresAt value greater than newerThanThis on the
* given shard.
*/
template <typename KeyDocumentType>
StatusWith<std::vector<KeyDocumentType>> _getNewKeys(OperationContext* opCtx,
std::shared_ptr<Shard> shard,
const NamespaceString& nss,
StringData purpose,
const LogicalTime& newerThanThis,
repl::ReadConcernLevel readConcernLevel) {
BSONObjBuilder queryBuilder;
queryBuilder.append("purpose", purpose);
queryBuilder.append("expiresAt", BSON("$gt" << newerThanThis.asTimestamp()));
auto findStatus = shard->exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcernLevel,
nss,
queryBuilder.obj(),
BSON("expiresAt" << 1),
boost::none);
if (!findStatus.isOK()) {
return findStatus.getStatus();
}
const auto& objs = findStatus.getValue().docs;
std::vector<KeyDocumentType> keyDocs;
keyDocs.reserve(objs.size());
for (auto&& obj : objs) {
try {
keyDocs.push_back(KeyDocumentType::parse(IDLParserContext("keyDoc"), obj));
} catch (...) {
return exceptionToStatus();
}
}
return keyDocs;
}
} // namespace
ShardingCatalogClientImpl::ShardingCatalogClientImpl(std::shared_ptr<Shard> overrideConfigShard)
: _overrideConfigShard(std::move(overrideConfigShard)) {}
ShardingCatalogClientImpl::~ShardingCatalogClientImpl() = default;
DatabaseType ShardingCatalogClientImpl::getDatabase(OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcernLevel) {
uassert(ErrorCodes::InvalidNamespace,
str::stream() << dbName.toStringForErrorMsg() << " is not a valid db name",
DatabaseName::isValid(dbName, DatabaseName::DollarInDbNameBehavior::Allow));
// The admin database is always hosted on the config server.
if (dbName.isAdminDB()) {
return DatabaseType(dbName, ShardId::kConfigServerId, DatabaseVersion::makeFixed());
}
// The config database's primary shard is always config, and it is always sharded.
if (dbName.isConfigDB()) {
return DatabaseType(dbName, ShardId::kConfigServerId, DatabaseVersion::makeFixed());
}
auto result =
_fetchDatabaseMetadata(opCtx, dbName, getConfigReadPreference(opCtx), readConcernLevel);
if (result == ErrorCodes::NamespaceNotFound) {
// If we failed to find the database metadata on the 'nearest' config server, try again
// against the primary, in case the database was recently created.
return uassertStatusOK(
_fetchDatabaseMetadata(opCtx,
dbName,
ReadPreferenceSetting{ReadPreference::PrimaryOnly},
readConcernLevel))
.value;
}
return uassertStatusOK(std::move(result)).value;
}
std::vector<DatabaseType> ShardingCatalogClientImpl::getAllDBs(OperationContext* opCtx,
repl::ReadConcernLevel readConcern) {
auto dbs = uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcern,
NamespaceString::kConfigDatabasesNamespace,
BSONObj(),
BSONObj(),
boost::none))
.value;
std::vector<DatabaseType> databases;
databases.reserve(dbs.size());
for (const BSONObj& doc : dbs) {
databases.emplace_back(DatabaseType::parse(IDLParserContext("DatabaseType"), doc));
}
return databases;
}
StatusWith<repl::OpTimeWith<DatabaseType>> ShardingCatalogClientImpl::_fetchDatabaseMetadata(
OperationContext* opCtx,
const DatabaseName& dbName,
const ReadPreferenceSetting& readPref,
repl::ReadConcernLevel readConcernLevel) {
invariant(!dbName.isAdminDB() && !dbName.isConfigDB());
auto findStatus =
_exhaustiveFindOnConfig(opCtx,
readPref,
readConcernLevel,
NamespaceString::kConfigDatabasesNamespace,
BSON(DatabaseType::kDbNameFieldName << DatabaseNameUtil::serialize(
dbName, SerializationContext::stateCommandRequest())),
BSONObj(),
boost::none);
if (!findStatus.isOK()) {
return findStatus.getStatus();
}
const auto& docsWithOpTime = findStatus.getValue();
if (docsWithOpTime.value.empty()) {
return {ErrorCodes::NamespaceNotFound,
str::stream() << "database " << dbName.toStringForErrorMsg() << " not found"};
}
invariant(docsWithOpTime.value.size() == 1);
try {
auto db =
DatabaseType::parse(IDLParserContext("DatabaseType"), docsWithOpTime.value.front());
return repl::OpTimeWith<DatabaseType>(db, docsWithOpTime.opTime);
} catch (const DBException& e) {
return e.toStatus("Failed to parse DatabaseType");
}
}
HistoricalPlacement ShardingCatalogClientImpl::_fetchPlacementMetadata(
OperationContext* opCtx, ConfigsvrGetHistoricalPlacement&& request) {
auto remoteResponse = uassertStatusOK(_getConfigShard(opCtx)->runCommandWithFixedRetryAttempts(
opCtx,
ReadPreferenceSetting{ReadPreference::PrimaryOnly},
DatabaseName::kAdmin,
request.toBSON(),
Milliseconds(defaultConfigCommandTimeoutMS.load()),
Shard::RetryPolicy::kIdempotentOrCursorInvalidated));
uassertStatusOK(remoteResponse.commandStatus);
auto placementDetails = ConfigsvrGetHistoricalPlacementResponse::parse(
IDLParserContext("ShardingCatalogClient"), remoteResponse.response);
return placementDetails.getHistoricalPlacement();
}
std::vector<BSONObj> ShardingCatalogClientImpl::runCatalogAggregation(
OperationContext* opCtx,
AggregateCommandRequest& aggRequest,
const repl::ReadConcernArgs& readConcern,
const Milliseconds& maxTimeout) {
// Reads on the config server may run on any node in its replica set. Such reads use the config
// time as an afterClusterTime token, but config time is only inclusive of majority committed
// data, so we should not use a weaker read concern. Note if the local node is a config server,
// it can use these concerns safely with a ShardLocal, which would require relaxing this
// invariant.
invariant(readConcern.getLevel() == repl::ReadConcernLevel::kMajorityReadConcern ||
readConcern.getLevel() == repl::ReadConcernLevel::kSnapshotReadConcern ||
readConcern.getLevel() == repl::ReadConcernLevel::kLinearizableReadConcern,
str::stream() << "Disallowed read concern: " << readConcern.toBSONInner());
aggRequest.setReadConcern(readConcern);
aggRequest.setWriteConcern(WriteConcernOptions());
const auto readPref = [&]() -> ReadPreferenceSetting {
const auto vcTime = VectorClock::get(opCtx)->getTime();
ReadPreferenceSetting readPref = getConfigReadPreference(opCtx);
readPref.minClusterTime = vcTime.configTime().asTimestamp();
return readPref;
}();
aggRequest.setUnwrappedReadPref(readPref.toContainingBSON());
if (!serverGlobalParams.clusterRole.has(ClusterRole::ConfigServer)) {
// Don't use a timeout on the config server to guarantee it can always refresh.
const Milliseconds maxTimeMS = std::min(opCtx->getRemainingMaxTimeMillis(), maxTimeout);
aggRequest.setMaxTimeMS(durationCount<Milliseconds>(maxTimeMS));
}
// Run the aggregation
std::vector<BSONObj> aggResult;
auto callback = [&aggResult](const std::vector<BSONObj>& batch,
const boost::optional<BSONObj>& postBatchResumeToken) {
aggResult.insert(aggResult.end(),
std::make_move_iterator(batch.begin()),
std::make_move_iterator(batch.end()));
return true;
};
const auto configShard = _getConfigShard(opCtx);
for (int retry = 1; retry <= kMaxWriteRetry; retry++) {
const Status status = configShard->runAggregation(opCtx, aggRequest, callback);
if (retry < kMaxWriteRetry &&
configShard->isRetriableError(status.code(), Shard::RetryPolicy::kIdempotent)) {
aggResult.clear();
continue;
}
uassertStatusOK(status);
break;
}
return aggResult;
}
CollectionType ShardingCatalogClientImpl::getCollection(OperationContext* opCtx,
const NamespaceString& nss,
repl::ReadConcernLevel readConcernLevel) {
auto collDoc =
uassertStatusOK(_exhaustiveFindOnConfig(
opCtx,
getConfigReadPreference(opCtx),
readConcernLevel,
CollectionType::ConfigNS,
BSON(CollectionType::kNssFieldName << NamespaceStringUtil::serialize(
nss, SerializationContext::stateDefault())),
BSONObj(),
1))
.value;
uassert(ErrorCodes::NamespaceNotFound,
str::stream() << "collection " << nss.toStringForErrorMsg() << " not found",
!collDoc.empty());
return CollectionType(collDoc[0]);
}
CollectionType ShardingCatalogClientImpl::getCollection(OperationContext* opCtx,
const UUID& uuid,
repl::ReadConcernLevel readConcernLevel) {
auto collDoc =
uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcernLevel,
CollectionType::ConfigNS,
BSON(CollectionType::kUuidFieldName << uuid),
BSONObj(),
1))
.value;
uassert(ErrorCodes::NamespaceNotFound,
str::stream() << "Collection with UUID '" << uuid << "' not found",
!collDoc.empty());
return CollectionType(collDoc[0]);
}
std::vector<CollectionType> ShardingCatalogClientImpl::getShardedCollections(
OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcernLevel,
const BSONObj& sort) {
BSONObjBuilder b;
if (!dbName.isEmpty()) {
const auto db =
DatabaseNameUtil::serialize(dbName, SerializationContext::stateCommandRequest());
b.appendRegex(CollectionType::kNssFieldName, "^{}\\."_format(pcre_util::quoteMeta(db)));
}
b.append(CollectionType::kUnsplittableFieldName, BSON("$ne" << true));
auto collDocs = uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcernLevel,
CollectionType::ConfigNS,
b.obj(),
sort,
boost::none))
.value;
std::vector<CollectionType> collections;
collections.reserve(collDocs.size());
for (const BSONObj& obj : collDocs)
collections.emplace_back(obj);
return collections;
}
std::vector<CollectionType> ShardingCatalogClientImpl::getCollections(
OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcernLevel,
const BSONObj& sort) {
BSONObjBuilder b;
if (!dbName.isEmpty()) {
const auto db =
DatabaseNameUtil::serialize(dbName, SerializationContext::stateCommandRequest());
b.appendRegex(CollectionType::kNssFieldName, "^{}\\."_format(pcre_util::quoteMeta(db)));
}
auto collDocs = uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcernLevel,
CollectionType::ConfigNS,
b.obj(),
sort,
boost::none))
.value;
std::vector<CollectionType> collections;
collections.reserve(collDocs.size());
for (const BSONObj& obj : collDocs)
collections.emplace_back(obj);
return collections;
}
std::vector<NamespaceString> ShardingCatalogClientImpl::getShardedCollectionNamespacesForDb(
OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcern,
const BSONObj& sort) {
BSONObjBuilder b;
const auto db =
DatabaseNameUtil::serialize(dbName, SerializationContext::stateCommandRequest());
b.appendRegex(CollectionType::kNssFieldName, "^{}\\."_format(pcre_util::quoteMeta(db)));
b.append(CollectionTypeBase::kUnsplittableFieldName, BSON("$ne" << true));
auto collDocs = uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcern,
CollectionType::ConfigNS,
b.obj(),
sort,
boost::none))
.value;
std::vector<NamespaceString> collections;
collections.reserve(collDocs.size());
for (const BSONObj& obj : collDocs) {
auto coll = CollectionTypeBase::parse(IDLParserContext("getShardedCollectionsForDb"), obj);
collections.emplace_back(coll.getNss());
}
return collections;
}
std::vector<NamespaceString> ShardingCatalogClientImpl::getCollectionNamespacesForDb(
OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcern,
const BSONObj& sort) {
auto collectionsOnConfig = getCollections(opCtx, dbName, readConcern, sort);
std::vector<NamespaceString> collectionsToReturn;
collectionsToReturn.reserve(collectionsOnConfig.size());
for (const auto& coll : collectionsOnConfig) {
collectionsToReturn.push_back(coll.getNss());
}
return collectionsToReturn;
}
std::vector<NamespaceString> ShardingCatalogClientImpl::getUnsplittableCollectionNamespacesForDb(
OperationContext* opCtx,
const DatabaseName& dbName,
repl::ReadConcernLevel readConcern,
const BSONObj& sort) {
BSONObjBuilder b;
const auto db =
DatabaseNameUtil::serialize(dbName, SerializationContext::stateCommandRequest());
b.appendRegex(CollectionType::kNssFieldName, "^{}\\."_format(pcre_util::quoteMeta(db)));
b.append(CollectionTypeBase::kUnsplittableFieldName, true);
auto collDocs = uassertStatusOK(_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcern,
CollectionType::ConfigNS,
b.obj(),
sort,
boost::none))
.value;
std::vector<NamespaceString> collections;
collections.reserve(collDocs.size());
for (const BSONObj& obj : collDocs) {
auto coll =
CollectionTypeBase::parse(IDLParserContext("getAllUnsplittableCollectionsForDb"), obj);
collections.emplace_back(coll.getNss());
}
return collections;
}
std::vector<NamespaceString>
ShardingCatalogClientImpl::getUnsplittableCollectionNamespacesForDbOutsideOfShards(
OperationContext* opCtx,
const DatabaseName& dbName,
const std::vector<ShardId>& excludedShards,
repl::ReadConcernLevel readConcern) {
auto aggRequest =
makeUnsplittableCollectionsDataShardAggregation(opCtx, dbName, excludedShards);
std::vector<BSONObj> collectionEntries =
Grid::get(opCtx)->catalogClient()->runCatalogAggregation(
opCtx, aggRequest, repl::ReadConcernArgs(readConcern));
std::vector<NamespaceString> collectionNames;
collectionNames.reserve(collectionEntries.size());
for (const auto& coll : collectionEntries) {
auto nssField = coll.getField(CollectionType::kNssFieldName);
collectionNames.push_back(NamespaceStringUtil::deserialize(
boost::none, nssField.String(), SerializationContext::stateDefault()));
}
return collectionNames;
}
StatusWith<BSONObj> ShardingCatalogClientImpl::getGlobalSettings(OperationContext* opCtx,
StringData key) {
auto findStatus = _exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
repl::ReadConcernLevel::kMajorityReadConcern,
NamespaceString::kConfigSettingsNamespace,
BSON("_id" << key),
BSONObj(),
1);
if (!findStatus.isOK()) {
return findStatus.getStatus();
}
const auto& docs = findStatus.getValue().value;
if (docs.empty()) {
return {ErrorCodes::NoMatchingDocument,
str::stream() << "can't find settings document with key: " << key};
}
invariant(docs.size() == 1);
return docs.front();
}
StatusWith<VersionType> ShardingCatalogClientImpl::getConfigVersion(
OperationContext* opCtx, repl::ReadConcernLevel readConcern) {
auto findStatus = _getConfigShard(opCtx)->exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcern,
VersionType::ConfigNS,
BSONObj(),
BSONObj(),
boost::none /* no limit */);
if (!findStatus.isOK()) {
return findStatus.getStatus();
}
auto queryResults = findStatus.getValue().docs;
if (queryResults.size() > 1) {
return {ErrorCodes::TooManyMatchingDocuments,
str::stream() << "should only have 1 document in "
<< VersionType::ConfigNS.toStringForErrorMsg()};
}
if (queryResults.empty()) {
return {ErrorCodes::NoMatchingDocument,
str::stream() << "No documents found in "
<< VersionType::ConfigNS.toStringForErrorMsg()};
}
BSONObj versionDoc = queryResults.front();
auto versionTypeResult = VersionType::fromBSON(versionDoc);
if (!versionTypeResult.isOK()) {
return versionTypeResult.getStatus().withContext(
str::stream() << "Unable to parse config.version document " << versionDoc);
}
auto validationStatus = versionTypeResult.getValue().validate();
if (!validationStatus.isOK()) {
return Status(validationStatus.withContext(
str::stream() << "Unable to validate config.version document " << versionDoc));
}
return versionTypeResult.getValue();
}
StatusWith<std::vector<DatabaseName>> ShardingCatalogClientImpl::getDatabasesForShard(
OperationContext* opCtx, const ShardId& shardId) {
auto findStatus =
_exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
repl::ReadConcernLevel::kMajorityReadConcern,
NamespaceString::kConfigDatabasesNamespace,
BSON(DatabaseType::kPrimaryFieldName << shardId.toString()),
BSONObj(),
boost::none); // no limit
if (!findStatus.isOK()) {
return findStatus.getStatus();
}
const std::vector<BSONObj>& values = findStatus.getValue().value;
std::vector<DatabaseName> dbs;
dbs.reserve(values.size());
for (const BSONObj& obj : values) {
std::string dbName;
Status status = bsonExtractStringField(obj, DatabaseType::kDbNameFieldName, &dbName);
if (!status.isOK()) {
return status;
}
dbs.push_back(DatabaseNameUtil::deserialize(
boost::none, std::move(dbName), SerializationContext::stateDefault()));
}
return dbs;
}
StatusWith<std::vector<ChunkType>> ShardingCatalogClientImpl::getChunks(
OperationContext* opCtx,
const BSONObj& query,
const BSONObj& sort,
boost::optional<int> limit,
repl::OpTime* opTime,
const OID& epoch,
const Timestamp& timestamp,
repl::ReadConcernLevel readConcern,
const boost::optional<BSONObj>& hint) {
invariant(serverGlobalParams.clusterRole.has(ClusterRole::ConfigServer) ||
readConcern == repl::ReadConcernLevel::kMajorityReadConcern);
// Convert boost::optional<int> to boost::optional<long long>.
auto longLimit = limit ? boost::optional<long long>(*limit) : boost::none;
auto findStatus = _exhaustiveFindOnConfig(opCtx,
getConfigReadPreference(opCtx),
readConcern,
ChunkType::ConfigNS,
query,
sort,
longLimit,
hint);
if (!findStatus.isOK()) {
return findStatus.getStatus().withContext("Failed to load chunks");