Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-130965ClosedFixed
Handle unsigned overflow in FLE2 binary search
In {{binarySearchCommon()}} at [src/mongo/crypto/fle_crypto.cpp|https://github.com/10gen/mongo/blob/85387fec361fa96d0ef9ba7efb3fcf4e86d068a1/src/mongo/crypto/fle_crypto.cpp#L2293], the variable rho (uint64_t) is doubled via rho = 2 * rho inside a galloping loop with no overflow guard. If this continuous doubling eventually becomes 0, the {{ceil(log2(rho))}} would return -inf and converting that to uint64_t is undefined behavior.
Components
Queryable Encryption
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-96495ClosedFixed
Choosing sync source across regions doesn't take into account which node is primary
The topology of sync sources has a large impact on majority write acknowledgement. Since chaining is [enabled by default|https://github.com/mongodb/mongo/blob/4f244903852237fe3681cef5949c74a366c39d19/src/mongo/db/repl/repl_set_config.idl#L90-L92] in replica sets, we should make sure that replication automatically trends towards sync source topologies that promote lower majority write acknowledgement latencies.
After SPM-1409, we try to optimize sync source topologies to make it so that per r
Components
Replication
Fixed Versions
8.1.0-rc08.0.20
SERVER-135756In Code Review
Replace bare wait() with opCtx->waitForInterrupt() in exchange_stage.cpp to fix delinquency in exchange getMores.
SERVER-137986ClosedFixed
Define pre transition checks for resharding coordinator
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139013ClosedFixed
Make $cursor batch memory accounting aware of Document shared BSON backing
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139160ClosedFixed
Make detailed_join_explain.js and join_explain.js pass under the disaggregated storage override
Two tests fail in {{{}no_passthrough_disagg_override_replsets{}}}:
* *{{detailed_join_explain.js}}* (no-feature-flags disagg variants): the suite starts replica sets at last stable FCV (9.0), as {{getExpectedSuiteFCV()}} expects. The {{enableDetailedJoinExplain}} query knob needs FCV 9.1, so setting it via query settings fails with {{query knob not settable via QuerySettings: enableDetailedJoinExplain}} (12955401).
* *{{join_explain.js}}* (all-feature-flags disagg variants)
Related: SER
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-136959Open
Add histograms to serverStatus/Otel for the latency distributions of write ops
We know operations like findOneByID/Index, updateOneByID/Index, insertOne, deleteOneById/Index should have very low latency. We should create histograms showing how long they actually take to show how stalls can affect user operations beyond the canary.
SERVER-127567 already added {{queryLatencies.express}} to cover the read queries.
The goal is to provide health metrics for the host itself, so we should exclude time spent waiting for write concern (on ASC) or for a log server (on DSC).
SERVER-133468ClosedWon't Do
Make FCV checks for per collection hash validation the same as the replicated size and count
SERVER-134116Backlog
MigrationChunkClonerSourceOpObserver holds CSS mutex while acquiring DSS mutex
{{MigrationChunkClonerSourceOpObserver}} acquires the DSR mutex while still holding the CSR mutex when checking {{movePrimary}} on untracked collections. If other sharding paths take these locks in the opposite order, it would deadlock.
SERVER-133732Needs Scheduling
onStepDown starts an async job that interrupts and joins leftovers PrimaryOnlyService instances
SERVER-133518 implemented a workaround to make sharding coordinators release locks during stepdown.
Ideally, cleanup work should not be inline in the stepdown path, and should instead be executed asynchronously, or during a potential next stepup.
This ticket tracks the work to improve the cleanup of sharding coordinators, making stepdown start an asynchronous job that will interrupt and join leftovers (including releasing the locks) initiated via stepdown. That way, the SERVER-133518 worka
SERVER-133668Backlog
Split checkMetadataConsistency across catalog modules
{{checkMetadataConsistency}} used to live as a single unit under the global catalog.
New checks — including authoritative shard catalog — have grown that logic well beyond the global catalog. It is hard to review, and it no longer matches module or ownership boundaries.
We need to split this work into layers. The commands and the dispatcher should live in one place. And then, each catalog (global catalog, shard catalog, and so on) exposes APIs to check its own consistency, and the dispatch
SERVER-138861ClosedFixed
Update timezone data to 2026f
Updates to the latest timezone data 2026f: Cuba will not fall back 2026-11-01 and stays on -04 through 2027-11-07.
The patch contains the new timezone data, generated files for timelib, and updated golden tests.
Furthermore, we adjusted the {{import.sh}} script to facilitate future automated updates:
* Clone over HTTPS instead of SSH as CI has now SSH key.
* Record the tzdata version in file {{version-info.txt}} to simplify PR generation in the future. Line 1 is the IANA tzdata release version
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139048In Code Review
Add a trySbeEngine all-feature-flags Evergreen variant
Add an Evergreen variant that runs the existing trySbeEngine variant's tasks with {{--runAllFeatureFlagTests}}.
* New variant {{amazon-linux2023-arm64-try-sbe-engine-all-feature-flags}}, copied from {{amazon-linux2023-arm64-try-sbe-engine}}
* Also runs {{query_golden_classic}}, {{query_golden_sharding}} and {{query_golden_join_optimization}}
Fixed Versions
9.1.0-rc09.1.0-rc1025
NODE-6517ClosedFixed
dropIndexes swallows all errors to and returns false
h3. Use Case
As a... driver user
I want... important errors to be surfaced from dropIndexes
So that... I am alerted if something unexpected happens while using this API
[https://github.com/mongodb/specifications/blob/master/source/index-management/index-management.md#namespacenotfound-errors]
h3. User Experience
* What is the desired/expected outcome for the user once this ticket is implemented?
** If the operation succeeds, resolve true. If the operation fails due to name space not
Components
Index Management
Fixed Versions
7.9.0
SERVER-127567ClosedFixed
Implement additional metrics for CBR
We are missing a few metrics to effectively measure the success of the CBR metrics. Details in the internal comments.
Fixed Versions
9.1.0-rc0
SERVER-138291ClosedFixed
Add microbenchmark for replica set transaction commit with allowEndOfTransaction enabled
h2. Goal
Add a C++ microbenchmark that measures the cost allowEndOfTransaction adds to the commit path on a replica set: building and writing the final applyOps entry with the allowEndOfTransaction marker.
h2. Proposal
* Benchmark the unprepared transaction commit path (OpObserverImpl::onUnpreparedTransactionCommit through to the oplog write) with allowEndOfTransaction off and on.
* Vary:
** Collections per transaction: 2, 10, 50
** Operations per transaction: 1, 100
* Report ti
Fixed Versions
9.1.0-rc09.1.0-rc1025
NODE-7998ClosedGone away
Performance drop in largeDocBulkInsert
2 tasks:
Project: mongo-node-driver-next
Variant: performance-tests
Task: run-spec-benchmark-tests-node-server-logging
Test: largeDocBulkInsert
Measurement: normalized_throughput
Percent change: -24.719190188897866
Project: mongo-node-driver-next
Variant: performance-tests
Task: run-spec-benchmark-tests-node-server
Test: largeDocInsertOne
Measurement: normalized_throughput
Percent change: -16.163382346825074
[Link1|https://spruce.mongodb.com/task/mongo_node_driver_n
SERVER-139159ClosedFixed
Remove samplingMethod from analyze calls in analyze_sample_persist_unreplicated.js
SERVER-137870 removed the {{samplingMethod}} and {{numChunks}} fields from the {{analyze}} command and updated the existing tests. {{jstests/noPassthrough/query/query-optimization/analyze_sample_persist_unreplicated.js}} was split out by SERVER-138718 after that PR was branched, and still passes {{{}samplingMethod: "random"{}}}.
Both cases now fail with {{IDLUnknownField}} (40415) before reaching the unreplicated-collection check.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139158ClosedFixed
Fix join_plan_cache_key.js replica-set tests reading the removed joinPlanCacheKey explain field
SERVER-138849 moved the join plan cache key in explain from {{queryPlanner.joinPlanCacheKey}} to {{{}queryPlanner.joinOptimizer.planCacheKey{}}}. SERVER-137120 was branched before that change and merged afterwards. Its new "replica set with cluster-wide query settings" block in {{jstests/noPassthrough/query/joins/join_plan_cache_key.js}} still reads the old field, so {{getKey()}} always returns {{{}undefined{}}}.
Fixed Versions
9.1.0-rc09.1.0-rc1025