Known Issues
5166 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-139287ClosedFixed
Ignore $clusterTime gossip that cannot be validated when the key cache has never loaded
The receiving node must verify the HMAC signature against its key cache before adopting the gossiped {{{}$clusterTime{}}}. However, if the node has not loaded its HMAC keys yet, then it can't validate the {{{}$clusterTime{}}}. As written today, the node fails the request rather than continuing to process the command. (Note: the {{$clusterTime}} is just metadata, not the command itself.)
Why aren't the keys loaded? That requires a majority read. If a PSA primary is starting up and its secondar
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139452Needs Scheduling
Add CAR ownership to jstests/sharding/lagged_config_secondary.js
With SERVER-139331 we enabled the test in {{{}no_passthrough_disagg_override_sharded{}}}, and during the code changes review, we spot the file does not have ownership.
As part of this ticket, we should analyze whether {{jstests/sharding/lagged_config_secondary.js}} has to fall under the ownership of CAR and to eventually add the ownership.
SERVER-112520ClosedFixed
Failure to write to geo index when inserting a document to timeseries existing bucket
{panel:title=Issue Status as of November 20, 2025|borderColor=#cccccc|titleBGColor=#6cb33f|bgColor=#eeeeee}
*ISSUE DESCRIPTION AND IMPACT*
2dsphere indexes on metric data in time series collections are not correctly maintained when inserting data. Index keys may not be generated and inserted into the index, causing geospatial queries optimized by it to return incomplete results.
2dsphere indexes on the metadata value in time series collections are not affected.
*DIAGNOSIS AND AFFECTED
Fixed Versions
8.3.0-rc08.2.28.0.16
SERVER-139329ClosedFixed
Re-enable startup_with_all_configs_down.js in no_passthrough_disagg_override_sharded
The ASC path verifies that a lagged secondary can catch up from another secondary without a primary or config servers. DSC standbys instead read from SLS, so its path should verify SLS catch-up under those conditions.
As part of this ticket, we should 1) Adapt the test for DSC, and 2) Add the test to {{no_passthrough_disagg_override_sharded}} suite.
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-137303Needs Merge
Select extension or built-in search index commands at startup
Mark the built-in create, update, and dropSearchIndex commands {{replaceableByExtension}}. When the built-ins are in use, the extension's registration of those names is skipped. Add the mock search-index-commands test extension with correct auth.
SERVER-138371Needs Scheduling
Get concurrent resharding coverage on resharding tests that use ShardingTest
Many tests in jstests/sharding/* make their own cluster with ShardingTest, not ReshardingTest fixture. start by seeing how to use a shell override for resharding tests like change streams, index builds, internal transactions, analyze_shard_key, and timeseries. look at the existing sharding_change_streams_* suites as a starting point.
SERVER-139449Open
Re-enable and rework replicate_record_ids_dbhash.js to groupId membership semantics
[{{replicate_record_ids_dbhash.js}}|https://github.com/10gen/mongo/blob/20753a5d812/jstests/noPassthrough/replication/replicate_record_ids_dbhash.js] was disabled as it [asserts|https://github.com/10gen/mongo/blob/20753a5d812/jstests/noPassthrough/replication/replicate_record_ids_dbhash.js#L86] the old per-collection semantics: dbHash catches injected {{rid}} divergence only when the collection has {{recordIdsReplicated: true}}, and ignores it otherwise. Under the set-wide {{groupId}} design, {{
PYTHON-6185ClosedFixed
Migrate macOS 14 tests to macOS 15
This ticket was split from {{DRIVERS-3691}}, please see that ticket for a detailed description.
Components
Evergreen
Fixed Versions
4.20.1
SERVER-112895Backlog
MemoryUsageTokenWith utility has problems
The memory usage tracking provides a few utilities to make memory tracking easier:
* a [token class|https://github.com/10gen/mongo/blob/580332adc410237402b54ffe42c16d7a53987661/src/mongo/db/memory_tracking/memory_usage_tracker.h#L306] that increments memory in the constructor and decrements in the destructor
* a ["token with" helper class|https://github.com/10gen/mongo/blob/580332adc410237402b54ffe42c16d7a53987661/src/mongo/db/memory_tracking/memory_usage_tracker.h#L367] that bundles a type in
SERVER-139354Open
Respect collMod sample size setting in read path
SERVER-137869 will update the write path (`analyze`) to choose sample sizes set in collMod over the knob configuration if set.
The read path needs to mirror that logic and respect samples set via collMod as well. If not, samples that are smaller than the knob configuration will be rejected even when they are the correct size (set by collMod). This means the read path will silently fall back to on-the-fly sampling, wasting resources.
For this ticket, we want to update the read path to respe
SERVER-137656Open
shutdownPinnedExecutors() can skip a PCTE whose destructor is still in flight
h3. Context
During search executor shutdown, {{shutdownTaskExecutor()}} drains the registered PinnedConnectionTaskExecutors (PCTEs) before shutting down the underlying executor:
{{shutdownPinnedExecutors()}} ({{{}src/mongo/executor/pinned_connection_task_executor_registry.cpp{}}}) scans the registry under the registry lock and skips entries whose {{pinned}} weak_ptr is expired:
h3. Problem (confirmed)
An expired weak_ptr only means the strong ref count reached zero; it does not mean
SERVER-108191ClosedWon't Do
Enable multiversion tests on Extensions-enabled variant
In SPM-4234, we created an Evergreen variant that runs extensions-agnostic tasks on nodes with extensions loaded. We won't be able to run multiversion tests until the "previous" versions have extensions enabled.
SERVER-138940Open
Potential null pointer dereference in getDBClientWithAutoEncryptionRef()
I found a potential null pointer dereference in {{getDBClientWithAutoEncryptionRef()}} in MongoDB 8.0.9:
[https://github.com/mongodb/mongo/blob/r8.0.9/src/mongo/scripting/mozjs/mongo.cpp#L148-L156]
In this method, the return value of {{JS::GetMaybePtrFromReservedSlot()}} is dereferenced without checking it for nullptr:
{{auto ret = JS::GetMaybePtrFromReservedSlot<std::shared_ptr<DBClientWithAutoEncryption>>(
args.thisv().toObjectOrNull(), MongoBase::DBClientWithAutoEncryptionSlot);
Affected Versions
8.0.9
SERVER-139140In Code Review
Add transaction test model and TxnCommandGenerator to the change stream FSM framework
Add the transaction\-centric test model and its walk generator to the change stream FSM framework, for endOfTransaction \(SPM\-4826\) testing.
h2. Scope of Work
\* Create {{jstests/libs/util/change\_stream/change\_stream\_txn\_test\_model.js}} — {{TxnAction}}/{{TxnState}} enums \(kept separate from the sharding {{Action}}/{{State}} so each model's all\-edges coverage check stays scoped\) and {{TransactionTestModel}}: RS graph \(3 in\-txn states\) vs sharded graph \(7 states, sharded/unsharded s
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139397In Code Review
Gate monguard-baseline-scan Evergreen variant to master only
h2. Overview
Run the {{monguard\-baseline\-scan}} Evergreen build variant only on {{mongodb\-mongo\-master}}, not on the release staging branches.
h2. Background
The variant runs an Endor Labs baseline scan of the {{monguard}} Rust crate. It is defined in {{etc/evergreen\_yml\_components/variants/misc/misc.yml}}, which the staging branches also pick up, so the Rust scan runs on branches where it is not wanted.
h2. Scope of Work
\* {{etc/evergreen\_yml\_components/variants/misc/misc.yml}} — rem
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-138754ClosedFixed
Move heap profiler and jsCore_heap_profiling to WR
This change moves ownership of the heap profiler and associated tests to workload resilience.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-113064ClosedFixed
Complete TODO listed in SERVER-74926
There is a TODO in the codebase referencing a resolved ticket which is assigned to you.
Please follow this link to see the lines of code referencing this resolved ticket:
https://github.com/10gen/mongo/search?q=SERVER-74926&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 represent the work you're actually doing.
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139446Needs Scheduling
Absorb relevant $project/$addFields fields into $lookup
We could do the following rewrite:
+ similar with renames and dotted paths where possible.
We already have a similar rewrite on $match, to absorb the $match into the $lookup subpipeline.
This $project rewrite will reduce the size of the returned data which has a positive performance impact, but also enable other optimisations, such as the projection pushdown, $match pushdown (including with computed fields), etc.
A simple local test showed potential for 35% improvement when the forei