Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-134531ClosedFixed
timeseries_create_indexes.js FSM workload does not tolerate ConflictingOperationInProgress from createIndexes
[BF-46154|https://jira.mongodb.org/browse/BF-46154]: the [createIndex state|https://github.com/mongodb/mongo/blob/67ee0fa0fed0263185b4d2eed3fbfbe61cb763eb/jstests/concurrency/fsm_workloads/timeseries/timeseries_create_indexes.js#L123] of {{timeseries_create_indexes.js}} fails with {{createIndexes ... conflicted with another operation after 10 retries}} (code 117, ConflictingOperationInProgress) in concurrency_simultaneous suites.
Since [SERVER-132822|https://jira.mongodb.org/browse/SERVER-132
Affected Versions
9.0.0-rc1
Fixed Versions
9.1.0-rc09.1.0-rc1022
SERVER-138178Needs Scheduling
Write conflict can escape split command
It appears that the {{split}} command can return a write conflict error rather than retrying it internally.
SERVER-138177Needs Scheduling
Write conflict can escape shardCollection command
It appears that the {{shardCollection}} command can return a write conflict error rather than retrying it internally.
SERVER-139073Needs Scheduling
Stabilize serverStatus metrics-filtering tests by removing cross-fixture schema assumptions
h2. Problem
BF-47116 exposed a fragile assumption in the metrics-filtering JS tests. They use the complete metrics response from an auth-disabled deployment as the expected schema for a separate auth-enabled deployment. These processes can legitimately report different fields because of sampling history, node role, or dynamically keyed metrics.
Consequently, a timing-dependent schema difference can fail the test even when filtering is correct. Maintaining exceptions for each such metric is
SERVER-138176Needs Scheduling
Write conflict can escape addShard command
It appears that the {{addShard}} command can return a write conflict error rather than retrying it internally.
SERVER-138624ClosedFixed
Implement delta syntax for collMod persistentStatsConfig
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138000ClosedFixed
$setWindowFields with sortBy after a join-optimized $lookup+$unwind hits tassert
$setWindowFields desugars to \{$sort: {..., $_internalOutputSortKeyMetadata: true}} plus $_internalSetWindowFields. The sort-key metadata dependency reaches the base collection CanonicalQuery, the planner puts a SORT_KEY_GENERATOR on top of the access plan, and CBR has no estimate for that node in single_table_access.cpp. A plain \{$sort: {_id: 1}} suffix does not trigger it.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-132676ClosedFixed
Support $_internalConstructNdvSketch in SBE
{{$_internalConstructNdvSketch}} is implemented only as a classic accumulator. As a result, any {{$group}} containing it is ineligible for SBE pushdown and executes in the classic engine.
A lot of future QE performance work (e.g. scan performance) will focus on SBE. Providing an SBE implementation (presumably wiring up the exact same hyperloglog implementation to SBE hash agg / the VM) seems like a good piece of future work.
Fixed Versions
9.1.0-rc0
SERVER-138987ClosedFixed
Disable the creation of minidump on sharding_metadata_indexes_missing_during_step_up.js
Test like sharding_metadata_indexes_missing_during_step_up.js can explicitly crash the server to verify that the proper assertions fire.
On a windows platform, a crash causes the creation of a mini-dump which can sometime hang (see [SERVER-124513]).
The solution to this is to disable the creation of mini-dumps for such tests.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-131440Backlog
Simplify checkCollectionMetadataInShardCatalog in CMC by reading at a fixed snapshot
checkCollectionMetadataInShardCatalog does some [juggling of CSR acquisitions|https://github.com/10gen/mongo/blob/5e55bcb582df7119fae59050952df4f9893c35f5/src/mongo/db/global_catalog/metadata_consistency_validation/metadata_consistency_util.cpp#L1094-L1104] to try and defend itself from concurrent migration commits (or any chunk operation commit after auth shards). Instead, it should be cleaner to just establish a consistent snapshot and read the local and remote catalogs at the same timestamp,
SERVER-135788ClosedFixed
Enable featureFlagChangeStreamOptimizedUpdateLookup
The flag is enabled on master (SERVER-117888) and all fixes are backported. Enable it on the v9.0 branch: flip {{default: true}} + {{incremental_rollout_phase: release}} in {{query_feature_flags.idl}} and regenerate the IFR registry.
Fixed Versions
9.1.0-rc09.0.9
SERVER-137651ClosedFixed
Reenable featureFlagChangeStreamReaderV2 in 9.0.9
Reenable the feature {{featureFlagChangeStreamReaderV2}} in 9.0.9
The feature was enabled for 9.0.0 and 9.0.1, but was disabled in 9.0.2 after bugs were found with the initial implementation.
Our goal is to reenable the flag so that 9.0 users will ultimately be able to get the benefits of the change stream improvements from SPM-1941.
A requirement for this is a successful rollout of 9.0.7/9.0.8 on Atlas with the flag reenabled (just on Atlas).
Affected Versions
9.0.6
Fixed Versions
9.0.9
SERVER-138937ClosedFixed
Cover change stream recovery from a rename invalidate by reopening the stream on the renamed collection
h2. Background
A collection-level change stream is invalidated when its watched collection is renamed (A→B):
the stream delivers a {{rename}} event followed by an {{invalidate}}. To keep consuming, the
client must reopen a stream — the natural recovery is to open it over the RENAMED collection B,
resuming from the invalidate token. That recovery path is untested today: existing coverage only
resumes on the original namespace.
h2. Behavior (pinned by the new test)
* {{resumeAfter}} with an inval
Fixed Versions
9.1.0-rc09.1.0-rc1025
JAVA-6369Needs Triage
Command tracing span scope leaks when CommandListener throws
In InternalStreamConnection, the one-way command path (responseExpected=false) dispatches commandSucceeded before closing the tracing scope and ending the span.
If a CommandListener throws an Error, tracing cleanup is skipped, leaving the scope active and the span unended and unexported.
The same issue behaviour in the asynchronous path.
*Acceptance Criteria*
* Ensure tracing cleanup executes via try/finally for both sync and async one-way commands.
JAVA-5068Needs Review
Simplify Iterable API interface hierarchy to eliminate method duplication
The Iterable API currently contains duplicative methods in various interfaces, including FindIterable, AggregateIterable, ListIndexesIterable, DistinctIterable, DatabaseIterable, and ChangeStreamIterable. These duplications impact code maintenance and can lead to inconsistencies.
Introduce internal interfaces and abstract classes, if necessary, to consolidate the duplicative methods and achieve a cleaner design without compromising backward and binary compatibility.
The objective is to i
Components
Internal
SERVER-139064Needs Scheduling
Complete TODO listed in SERVER-136465
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* buildscripts/modules/atlas/suites/no_passthrough_disagg_override_sharded.yml:112
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-136465&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow the steps in the TODO if it is correct. If the latter, please update the summary and description of th
SERVER-139061Needs Scheduling
Complete TODO listed in SERVER-117506
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* src/mongo/db/query/compiler/optimizer/join/join_cost_estimator.h:42
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-117506&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow the steps in the TODO if it is correct. If the latter, please update the summary and description of this ticket to
SERVER-117506ClosedWon't Do
Support all QSN stages in join cost model
Currently the join cost model only supports QSNs with the following stages: Fetch, CollScan and IxScan. As we expand the join optimization module to support more complex queries, we may find that we end up generating single table plan fragments with other operators like projections, index intersection, sort, etc.
This ticket represents the work to properly incorporate these costs into the join cost model.
SERVER-139060Needs Scheduling
Complete TODO listed in SERVER-115490
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* buildscripts/modules/atlas/suites/no_passthrough_disagg_override_sharded.yml:30
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-115490&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow the steps in the TODO if it is correct. If the latter, please update the summary and description of thi
SERVER-133447ClosedFixed
Ensure incorrect order records are covered across slice boundaries in parallel validate
under SERVER-128372, we made traverseRecordStore a 'pure-ish' function to support parallelism. while the base validate tests pass, check if we're missing any coverage gaps around the validate code.
Fixed Versions
9.1.0-rc09.1.0-rc1025