Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
PYTHON-6233Backlog
Remove pymongocrypt version test helpers
h3. *Context*
PYTHON-5959 raises the pymongocrypt floor for PyMongo 4.20 to {{pymongocrypt>=1.21.0,<2.0.0}}. The test suite does not explicitly need version gates: tests should run against what CI installs or fail, not silently skip.
h3. *Definition of done*
- The test-side version gates are removed.
- The runtime check stays: {{check_min_pymongocrypt()}} in {{pymongo/encryption_options.py}}. It is the only guard for pip resolver bypass (an old pymongocrypt installed alongside the ne
SERVER-136912In Code Review
TLA+ proposed changes to prevent a replaced node from affecting the majority commit point
In WRITING-40450 we came up with proposed changes in order to avoid a replaced/restarted node from going back in time and affecting the majority commit point of the replset.
Changes taken from the [decision doc|https://docs.google.com/document/d/1dGRjirN43xWQv-XlXqekWeMo77oJkc0zzSKlJhH6R2M/edit?tab=t.0#heading=h.6vzuqy1ox6ng]:
* Server changes so that a node in STARTUP2 cannot vote.
* Server changes so that initial sync (LIS or FCBIS) is not considered completed until it is caught up to t
SERVER-134248ClosedFixed
Make an error upon the xxh3 hash generated by validate and continuous validation's hash mismatch
Fixed Versions
9.1.0-rc09.1.0-rc1026
PYTHON-6275ClosedDone
Enable all ruff PERF rules
h3. *Context*
{color:#505f79}Describe the background behind the problem.{color}
Only a portion of ruff's {{PERF}} linting rules are enabled by default. Enabling all of them improves performance and code consistency.
h3. *Definition of done*
{color:#505f79}What must be done to consider the task complete?{color}
Enable the entire suite of ruff {{PERF}} linting rules and make the required changes to pass them.
h3. *Pitfalls*
{color:#505f79}What should the implementer watch out for? What
SERVER-137887In Progress
Bound interrupt-check gaps while serializing and deserializing large $group spill partials
A single large $group spill record can take longer than the interrupt-check interval to serialize or deserialize. SERVER-136021 checks between records and merge partials; it does not bound work inside a record. SERVER-136213 adds classic spill-sort/$push checks and removes an SBE recovered-array deep copy, but large-record spans remain. These can delay killOp/maxTimeMS handling and produce interrupt-check delinquency.
SERVER-139181ClosedFixed
Move transactions task ownership to Storage Execution
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139089ClosedFixed
Transfer bulkWrite task ownership to query
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-137443In Code Review
Remove Parallel Collection Validation Feature Flag
Enable namespace level parallelism in Offline Validate by default.
SERVER-139014Blocked
Support accumulators with variable initializers in streaming group
When an accumulator has a variable initializer, we add a project to rewrite the group key into an object form, and use that as the group key slot as long as the group key is also an object. Currently we don't use streaming group in this case.
To support, need to correctly build a cluster key representation of the rewritten grouping key.
Example rewrites:
Grouping key is an object. We build an object form of the key (s6) and use that as the grouping key and the slot used by the init
SERVER-139355Needs Scheduling
Support for parallel scans when query settings/index hints are applied
Spun out from [https://github.com/10gen/mongo/pull/65920] where we conservatively bail out of parallelism on _any_ index hint being applied. At the time, this meant $natural (which arguably defines an order and thus parallelism would have to sort-merge) but once index scans are supported we should revisit this check.
SERVER-138778ClosedFixed
Add incremental spilling support to Spiller and ContainerBasedSpiller
When a primary-driven index build spills its sorter, every batch is written back to back while the collection scan is suspended, which engages flow control against concurrent user writes. Incremental spilling, one batch per call paced to the collection scan, avoids the spike, but {{Spiller::spill()}} has no way to express it.
This adds {{beginSpill()}} and {{spillIncremental()}} to the Spiller interface (the writer is created by {{beginSpill()}} and held open across {{spillIncremental()}} calls
Fixed Versions
9.1.0-rc09.1.0-rc1026
GODRIVER-4258ClosedFixed
Consistent failure: TestFLE2CreateCollection panics on server 7.0.48 / 9.0.8+ and fails all 7.0+ integration tasks
h3. *Name of Failure:*
TestFLE2CreateCollection panics on server 7.0.48 / 9.0.8+ and fails all 7.0+ integration tasks
h3. *Link to task:*
https://spruce.corp.mongodb.com/version/6af3bc9a97d58300076ac0d3
7.0.45 and 9.0.4 accepted the deviating name, and 7.0.48 / 9.0.8 rejected it.
h3. *Context of when and why the failure occurred:*
The test creates a data collection named {{coll}} but passes an
{{escCollection}} name that belongs to a different collection:
internal/integration/clien
Fixed Versions
2.10.1
SERVER-139349In Progress
Enable stepdowns in disagg_concurrency_sharded_terminate_kill_primary_with_balancer
* Enable stepdowns in the suite (resolve this [TODO|https://github.com/10gen/mongo/blob/master/buildscripts/modules/atlas/matrix_suites/overrides/disagg.yml#L819-L820])
* Add "_stepdown" to the suite name
** disagg_concurrency_sharded_stepdown_terminate_kill_primary_with_balancer
Affected Versions
9.1.0-rc0
SERVER-139348Needs Scheduling
replSetGetStatus reports pingMs=0 for members that have a priorityPort configured
h3. Summary
When a replica set member has {{members[].priorityPort}} set (and {{disableReplicationUsageOfPriorityPort}} is false), {{replSetGetStatus}} always reports {{members[].pingMs: 0}} for that member. The heartbeat RTT is recorded, but under a different key than the one {{replSetGetStatus}} reads.
h3. Cause
* Heartbeats are sent to {{getHostAndPortPriority()}} ({{replication_coordinator_impl_heartbeat.cpp}}, {{auto target = rsc.getMemberAt(i).getHostAndPortPriority();}}). The ping stats
SERVER-138619In Code Review
Refactor to harden in-flight OpCtx logic
Coverity issue SERVER-138382, pointed to the logic around managing in-flight OpCtx values as a potential use-after-free case. The analysis turned out to be a false positive and I don't see any risks in the current logic, but the current implementation is fragile and future refactors could break the assumptions currently in place.
Here are a few possible strategies:
* CancelableOperationContext
** This is seemingly not viable, or any sort of cancellation-token-style approach, as a hard {{mar
SERVER-135894Backlog
checkReferentialIntegrity(): single-session validation support for inserts
h2. Overview
Implement the core {{checkReferentialIntegrity\(\)}} mechanism that validates foreign keys on insert within a single session/transaction context, as prototyped in the POC. Includes detecting violations against the referenced \(parent\) collection.h1. Dev\-only: POC branch — [https://github.com/10gen/mongo/tree/mjrb/ri\-lockmgr2|https://github.com/10gen/mongo/tree/mjrb/ri-lockmgr2]
SERVER-121413ClosedFixed
Track the number of times the KEK has been used to encrypt DEKs
From a security perspective, we need to guarantee that a KEK has not been used 2^32-1 times to perform encryption. This follows a security guideline outlined by section 8.3 of [NIST GCM recommendations|section 8.3 of https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-38d.pdf].
This will likely have to be persisted in the keystore itself.
Fixed Versions
9.1.0-rc1026
SERVER-136296ClosedFixed
Enable standby collection validation fassert server parameter
Set the default value of {{fassertOnPerCollectionContinuousInternodeValidationFailure}} to {{{}true{}}}. When this parameter is set to true, fassert on a validation failure. Otherwise, log an error.
{color:#ff0000}This change enables checkpoint validation on all clusters, so checkpoint validation must be fully correct.{color}
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-136891Open
Update the OTEL prod allowlist with the combined step-up metric
SERVER-136676 combined the lower/upper bound metrics for failover unavailability, and added a label for planned/unplanned failover so we can differentiate them. Now that it's merged, we should update the [OTEL prod allowlist|https://github.com/10gen/mms/blob/5829760bdb34098dcd40afff36720802ae276c81/server/src/main/com/xgen/cloud/services/metrics/resource/usecase/privateConfig/prod/dedicated-mongo-otel-metrics.yaml#L29-L30] to use the [new combined step up metric|https://github.com/10gen/mongo/bl