Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-139072ClosedFixed
Re-enable shard_identity_rollback.js in no_passthrough_disagg_override_sharded
{{shard_identity_rollback.js}} verifies that rolling back a shard identity causes the affected node to terminate, preventing it from continuing with stale initialized sharding state. The test assumes ASC replication and rollback behavior. Under DSC, the replication-stop override pauses the primary’s SLS shipping loop, and the subsequent graceful shutdown stalls.
As part of this ticket, we should adapt and re-enable the test for DSC.
h3.
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-137188ClosedFixed
Run the custom build's standalone compile tasks on release distros (v9.0-staging)
The [SERVER-136184|https://jira.mongodb.org/browse/SERVER-136184] backport on v9.0-staging (512f0436) requires release configurations to run on release-* distros, and the custom builds' MONGO_VERSION_OVERRIDE makes every compile in the variant resolve public-release, so every compile outside the release-pool task groups has failed on plain rhel8.8-large since 2026-10-14 (tracked in [BF-46787|https://jira.mongodb.org/browse/BF-46787]).
Fix: [PR 64532|https://github.com/10gen/mongo/pull/64532]
Fixed Versions
9.0.7
SERVER-138765Backlog
Integrate ConflictDetector into the staged-write commit path
h2. Overview
Wire the conflict detector into commit for staged writes, following design note section 7, so that write-write conflicts are decided by the detector rather than WiredTiger.
h2. Scope of Work
* Commit ordering: take the {{WriteSet}} accumulated by the write helpers; call {{ConflictDetector::tryCommit(writeSet, readTs, reserveSlot)}}; on conflict throw {{WriteConflictException}} (non-transactional writes retry, transactions abort); apply staged writes to WiredTiger at the reserved ti
SERVER-138762Backlog
Remove apply-at-commit so only the oplog entry commits on the primary
h2. Overview
Remove the intermediate apply-at-commit step from the staged-write commit path, so the primary's WiredTiger transaction commits only the oplog entry. Collection and index writes are applied later by primary oplog application.
h2. Background
The commit path initially replays the staged buffer into WiredTiger inside the commit critical section (design note section 7, step 3). That keeps storage writes on the critical path. The oplog entry is already built from the {{OpObserver}}'s lo
SERVER-138768Backlog
Prototype startup recovery and failover with staged writes and primary application
h2. Overview
Prototype correct behavior across step-up, step-down, restart, and rollback for the new write path.
h2. Scope of Work
* Step-up: drain unapplied oplog, initialize the detector floor to lastApplied, then accept writes.
* Step-down: stop accepting writes, stop/drain the primary applier, abort staged transactions.
* Restart: replay through existing startup recovery.
* Rollback of an unapplied oplog tail on a former primary.
* One jstest per scenario.
h2. Acceptance Criteria
* Each sc
SERVER-138766Backlog
Verify unique index constraint enforcement with staged writes
h2. Overview
Verify that unique index constraints hold under staged writes without special handling, as described in design note section 3.4.
h2. Background
{{WiredTigerIndex::_checkDups}} inserts and removes a prefix key to conflict with concurrent insertions of the same value. Under staging this key enters the {{WriteSet}}, so the conflict detector rejects the later committer. The {{_id}} index uses one table key per value, so duplicate {{_id}} inserts conflict directly. The bounded duplicate
SERVER-138921ClosedGone away
Complete TODO listed in SERVER-136298
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* src/mongo/db/repl/oplog.cpp:258
* src/mongo/db/repl/oplog.cpp:2206
* src/mongo/db/repl/oplog.cpp:2416
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-136298&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
SERVER-137326ClosedFixed
Hook the conversion stage into the data-node and router pipelines
h2. Deliverable
Make the option user-facing and install the stage.
h3. User-facing option
* Add a {{showErrorsAsEvents}} boolean to the {{$changeStream}} IDL spec.
* Gate it behind {{featureFlagChangeStreamEmitErrorEvents}} (FCV-gated and checked at parse time).
* Normalize an explicit {{false}} out of the query shape.
* Forward the option from shell {{{}watch(){}}}.
h3. Stage injection
Add {{change_stream::pipeline_helpers::addErrorConversionStageIfNeeded(Pipeline&)}} wi
Fixed Versions
9.1.0-rc0
SERVER-139277Open
Ensure streaming group takes precedence over the batched ixscan+fetch rewrite
SERVER-136952Open
SBE $lookup using a partial foreign index drops foreign documents that the index does not contain
Found by an AI bug hunt and confirmed by a differential oracle (SBE vs classic).
h2. Problem
When the foreign collection has a partial index and $lookup runs in SBE with the dynamic indexed loop join, the join probes the index whenever the local key passes the index's partial filter. It does not check that the index holds every foreign document the key matches. Foreign documents left out of the index are lost from the result. Classic returns them.
h2. Repro
The test starts its own mongod with
SERVER-138466Backlog
Propagate isExpectedToExecuteQueries into the extension parse path
Expose the host's execution intent to the extension stage parse path, so an extension stage can tell a real execution request apart from analysis-only parsing. This lets the extension $listSearchIndexes counterpart skip the search management-endpoint availability [check|https://github.com/10gen/mongo/blob/6453da0686ca7ff56cae4c161c81b7bead5076e8/src/mongo/db/pipeline/search/document_source_list_search_indexes.cpp#L30-L41] for analysis-only parsing while still raising it before syntax validation
SERVER-139246In Code Review
Fix the semantics for resume gap in case of shard retargetting.
* sharded cluster with 2 shards A and B
* open a stream over a collection which lives on shard A only
* move collection to another shard B
* oplog roll over (we won’t receive control event for retargeting)
* getMore → error event resume gap
* we still target shard A and never retarget to shard B and never receive new events
SERVER-132685ClosedDuplicate
Incrementally spill sorter data when using container-based spilling
Currently all of the sorted data is spilled in one burst of work, which can overwhelm the system and cause replication lag spikes that interfere with the user's foreground workload. We should instead incrementally spill the sorter data over the course of the collection scan.
SERVER-132430ClosedFixed
Bound memory usage in $_internalKeyStringValue
Currently, {$_internalKeyStringValue} has no input size validation or memory limits and is available to the public, despite the {_internal} prefix. We cannot restrict access beyond setting `apiStrict` to false (which should be done in a follow-up ticket), so we must add memory limits. Add a new parameter internalQueryKeyStringValueMaxCollationBytes settable at startup and runtime that limits how much memory _internalKeyStringValue can use.
Fixed Versions
9.1.0-rc08.3.169.0.78.0.377.0.489.1.0-rc1026
SERVER-122256Backlog
Add stricter requirements about abortIndexBuild and commitIndexBuild oplog entries always have index idents
SERVER-138004ClosedFixed
suppress LSAN failure due to mongo client shutdown order
<https://jira.mongodb.org/browse/BF-46959> so far looks like it's spurious, due to the order in which threads terminate in the mongo test driver. At least according to AI.
This ticket is the work of looking into more closely, and either adding a LeakSanitizer (LSAN) exception to silence the failure, or making a code change to fix a real underlying leak or shutdown race.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139275Open
Expose sharding.lastSeenConfigServerOpTime and sharding.lastSeenTopologyOpTime in mms
PYTHON-6261ClosedDone
Opt into ruff D rules for whitespace and escapes
h3. *Context*
{color:#505f79}Describe the background behind the problem.{color}
PYTHON-6249 enabled the small number of {{D}} rules on by default. Opting into the full suite of {{D}} rules will improve docstring consistency and make enforcing our coding standards easier and more automated.
This ticket tracks the enablement of ruff rules D202, D209, D301, and D403, which deal with docstring whitespace and escape characters.
h3. *Definition of done*
{color:#505f79}What must be done to con
SERVER-139276In Progress
Run candidate mongotune policies against various connection storm workflows