Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-127787ClosedFixed
Retry on NetworkInterfaceExceededTimeLimit in blockUntilIndexQueryable
Stacked on top of the CommandFailed retry fix (SERVER-127065).
`blockUntilIndexQueryable` sends `_shardsvrRunSearchIndexCommand` to each
shard with a 60 s `kRemoteCommandTimeout`. On slow aubsan builds the shard
can exceed that deadline, returning `NetworkInterfaceExceededTimeLimit`
(code 89) at the network layer. `Shard::shouldErrorBePropagated(89)` returns
`false` (hardcoded exclusion), so the old code wrapped the error in
`OperationFailed` and threw immediately — causing a spurious BF
(`sear
Fixed Versions
9.0.0-rc08.3.16
SERVER-131792ClosedFixed
Fix initialization order problem with numCommandsTargetingSystemBuckets
The `numCommandsTargetingSystemBuckets` metric is initialized differently to most metrics. While most metrics are registered at file scope, this metric is initialized during the constructor of the SystemBucketsMetricsCommandHooks. That constructor runs when initializeCommandHooks() builds MongodCommandInvocationHooks which is called from _initAndListen().
This triggers [an invariant|https://github.com/mongodb/mongo/blob/be79e8062344f8290b81757c42403312b380630d/src/mongo/db/commands/server_sta
Affected Versions
9.0.0-rc08.3.0
Fixed Versions
9.1.0-rc09.0.0-rc28.3.16
SERVER-129939ClosedFixed
Make sure $idLookup is rejected if provided by user in pipeline
$idLookup is registered with the non-internal macro, this should probably be changed.
Fixed Versions
9.0.0-rc08.3.16
SERVER-135507ClosedFixed
Exclude agg_lookup_in_txn from running in step-down suites
Fixed Versions
9.0.48.3.168.0.37
SERVER-129716ClosedFixed
Multikey transition race during CBR sampling results in tassert
samplingCE's sample generation can yield, which results in the following race:
* a collection has an index that is initially not multikey
* the query planner gets called on it. it reads the catalog, and fetches the index as not multikey
* after that, but before sampling ce generates the samples, a document is inserted to the collection that makes the index mulktikey
* during index seek estimation, [CardinalityEstimator|https://github.com/10gen/mongo/blob/f4015024ff68fdbb90a8e74762a5d0494
Fixed Versions
9.1.0-rc09.0.0-rc38.3.16
SERVER-136515ClosedFixed
Update PGO profiles
Updated PGO profile numbers for performance.
Fixed Versions
8.3.16
SERVER-137058ClosedFixed
[v8.3] Assign remaining crypt release builds to release distros
h2. Problem
The v8.3-staging backport of SERVER-136184 added release distro enforcement, but four enterprise variants still select regular distros for {{crypt_create_lib}}. Waterfall builds fail before compilation.
[Failing Amazon Linux 2 task|https://spruce.corp.mongodb.com/task/mongodb_mongo_v8.3_staging_enterprise_amazon2_crypt_create_lib_5461db789d88352c62bccb55a55716b1fa576b76_26_10_14_18_32_30/logs?execution=0]
h2. Affected variants
* {{enterprise-amazon2}}: {{amazon2-latest-large}}
* {
Fixed Versions
8.3.16
SERVER-133231ClosedFixed
Fix read_pref_cmd.js to wait for replica set before running profile
BF-45277 — read_pref_cmd.js: profile entry lost on a lagged secondary after dropDatabase/re-create
Mechanism. The test loops: setUp() → testCursorReadPreference() / testConnReadPreference() → tearDown(). tearDown() drops read_pref_cmd and waits for replication; the next setUp() re-creates the database via enableSharding + shardCollection on the primary. Those are majority writes, so in a 3-node set they return as soon as the primary and one secondary ack — the third node can still be behind.
Fixed Versions
9.1.0-rc09.0.48.3.168.0.37
SERVER-133278ClosedFixed
Disable fsyncLock tests in sharding CSRS stepdown suites
FsyncLock interferes with RSTL acquisition during stepdown. This makes tests which run fsyncLock incompatible with stepdowns. There is also an additional issue caused by SERVER-133098 in sharded clusters with CSRS stepdowns and fsyncLock, but even if that issue were fixed thus allowing another node to step up and run fsyncUnlock, the fsyncUnlock command must run against shard primaries and so this test would leave the initial primary still fscynLocked and trying to step down.
Adding the requi
Fixed Versions
9.1.0-rc09.0.48.3.168.0.37
SERVER-125410ClosedFixed
Fix query benchmarks to set IndexCatalogEntry
Fixed Versions
9.0.0-rc08.3.16
SERVER-131327ClosedFixed
checkMetadataConsistency may return MissingShardKeyIndex false positive when migrating the first chunk to a shard
We only [check that a sharded collection is supported by a valid shard key index under the optimistic check|https://github.com/mongodb/mongo/blob/37ff58cbd999dd136497bb28e4ab9a6d96e83b65/src/mongo/db/global_catalog/metadata_consistency_validation/metadata_consistency_util.cpp#L1237-L1245] (only holding the DDL lock, but not collection locks). A chunk migration does not acquire the DDL lock, so checkMetadataConsistency may observes the intermediate state where the collection has been cloned by th
Components
Sharding
Affected Versions
8.3.0-rc08.2.09.0.0-rc09.1.0-rc0
Fixed Versions
9.1.0-rc09.0.48.3.16
SERVER-130694ClosedFixed
Re-enable MalformedTimeseriesBucketsCollection check in CheckMetadataConsistencyInBackground hook
After SERVER-130586, re-enabling the {{CheckMetadataConsistencyInBackground}} reports {{MalformedTimeseriesBucketsCollection}} inconsistencies (suspected false positives).
This ticket should investigate those failures and remove the associated TODO ticket from the codebase.
Components
Catalog
Affected Versions
8.0.08.2.09.0.0-rc08.3.0
Fixed Versions
9.0.0-rc18.3.16
SERVER-122126ClosedFixed
Fix race in InitialSyncerTransitionsToCompleteWhenFinishCallbackThrowsException
The test hangs because {{shutdown()}} causes {{_finishCallback}} to be scheduled on {{{}_exec->scheduleWork(){}}}, which enqueues a task into the {{{}ThreadPoolMock{}}}. But the mock's single worker thread is blocked in {{NetworkInterfaceMock::waitForWork()}} and nobody signals it to wake up. The task sits in the queue forever, so {{_finishCallback}} never runs and the failpoint is never hit.
The sibling test ({{{}InitialSyncerResetsOnCompletionCallbackFunctionPointerUponCompletion{}}}) handl
Fixed Versions
9.0.0-rc08.3.16
SERVER-136976ClosedFixed
Update PGO profiles
Updated PGO profile numbers for performance.
Fixed Versions
8.3.16
SERVER-131526ClosedFixed
Remove duplicate attribute from log line in metadata consistency util
[This|https://github.com/mongodb/mongo/blob/295c3d2ac202b76071df4012bcdee9bf27bdf8dd/src/mongo/db/global_catalog/metadata_consistency_validation/metadata_consistency_util.cpp#L1305-L1310] log line has the namespace logged twice which triggers an assertion in the logv2 library.
Fixed Versions
9.1.0-rc09.0.48.3.16
SERVER-131262ClosedFixed
Populate Locker debugInfo with OperationContext state in getGlobalLockManager()->dump()
At the RSTL timeout in {{replication_coordinator_impl.cpp}} (v7.0), {{getGlobalLockManager()->dump()}} emits log id {{20521}} with per-holder {{{}desc{}}}, {{{}connectionId{}}}, {{client}} IP, and {{opid}} but the {{debugInfo}} field is empty in practice. {{Locker::setDebugInfo}} is called at {{commands.cpp:914}} on the sync run path only, so it does not cover async paths, transactions, and background contexts. Attribution still requires cross-referencing slow-op logs by {{ctx}} after the fact.
Fixed Versions
9.1.0-rc09.0.48.3.168.0.37
SERVER-133198ClosedFixed
Handle stale primary member data case in sync source selection
In SERVER-96495, we [added logic|https://github.com/mongodb/mongo/blob/6efa48969872ee9a29ed082b26e63c83ab6a6fe6/src/mongo/db/repl/topology_coordinator.cpp#L395-L429] to prefer syncing from the primary if we are choosing between the primary and another node in a datacenter.
The check is done using [member data of each node|https://github.com/mongodb/mongo/blob/6efa48969872ee9a29ed082b26e63c83ab6a6fe6/src/mongo/db/repl/topology_coordinator.cpp#L395-L396]. However, this data may be stale, since
Components
Replication
Fixed Versions
9.1.0-rc09.0.48.3.168.0.37
SERVER-134453ClosedFixed
The transition to CSRS test hook can fail to step up a new primary due to starvation from test writes
*Problem*
In the replicaset_transition_to_and_from_csrs.py test hook we do a rolling restart which involves stepping up a new primary. This is done via replSetStepUp which will fail to elect the new primary if the running jstest is doing writes during the election (see related tickets).
The hook gives each secondary a single chance to step up, so with heavy write activity all attempts fail.
*Solution*
Do more stepUp attempts. We expect that eventually we will win the race o
Components
ReplicationSharding
Affected Versions
8.3.09.1.0-rc09.0.0
Fixed Versions
9.1.0-rc09.1.0-rc10229.0.48.3.16
SERVER-134768ClosedFixed
Reduce buildSort stack frame size to reduce likelihood of stack overflow with some PGO profiles
Fixed Versions
9.1.0-rc09.1.0-rc10249.0.48.3.16
SERVER-131802ClosedFixed
Query optimizer can drop a disjunct when converting a | ~a to a bitset tree during boolean simplification
Description info removed to limit exposure of customer details.
Affected Versions
8.0.89.0.0-rc08.0.298.3.8
Fixed Versions
9.1.0-rc09.0.0-rc48.3.168.0.37