Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-138679Backlog
Ensure spilling sort yields resources
This is the task to cover spilling sort induced resource contention. Description in linked ticket.
SERVER-138681Backlog
Ensure spilling hash join/lookup yields resources
This is the task to cover spilling hash join/lookup induced resource contention. Description in linked ticket.
SERVER-123532Backlog
Extend join cost estimation for multikey indexes
SERVER-122379 added support for multikey index NDV estimation.
Address this [todo|https://github.com/mongodb/mongo/blob/63f1205f84f24e0c3c3e45e38cbcab1b2063b326/src/mongo/db/query/compiler/optimizer/join/join_cost_estimator_impl.cpp#L84] and extend {{JoinCostEstimatorImpl::costIndexScanFragment}} to use this functionality.
SERVER-139007Needs Scheduling
Update Homebrew Formulae for New Major Patch Release (4.4.32)
Update Homebrew Formulae to support new Major Patch Version 4.4.32
SERVER-138340Backlog
Health observers create child CancellationSources off a long living parent source
Recently in SERVER-135020, we discovered that a long lived CancellationSource will hold a strong reference to the child cancellation states of every child CancellationSource until the parent source is canceled or destroyed. This can lead to steadily increasing memory usage until then (see HELP-102354).
It looks like we also use this pattern in the health observer code where we have [one global CancellationSource|https://github.com/10gen/mongo/blob/2de947c5f21dd4b1e676823d92f065b1bd3e2c33/src/
NODE-7994Blocked
node-mongodb-native - PR #5110: chore(main): release 7.9.0
[github-actions[bot]|https://github.com/apps/github-actions] has created [PR #5110: chore(main): release 7.8.1|https://github.com/mongodb/node-mongodb-native/pull/5110] in [node-mongodb-native|https://github.com/mongodb/node-mongodb-native]
NODE-7988Investigating
Add a supported isConnectionError helper
Could the Node.js driver expose a supported {{isConnectionError(err)}} helper? We're classifying Atlas Remote MCP failures as expected vs. alertable, and maintaining our own list of driver error types is brittle. It'd be useful if it covered network error subclasses, SRV/TXT DNS failures, and the underlying causes in {{{}MongoServerSelectionError{}}}.
Context: [Slack thread|https://mongodb.slack.com/archives/C03MPA2FY6S/p1793696623790289]
Components
Error
SERVER-134234ClosedDone
Investigate improvements to OplogWriter for large oplog entries
The thread which takes fetched entries and writes them to the oplog collection before sending them to the apply buffer is single threaded, and only grabs oplog entries up to 16MB before writing them and sending to the apply buffer.
If we have a scenario (resharding is a common case) where we have a large amount of big oplog entries (~5MB each) then we are only able to write a couple of entries into the oplog at a time, which turns the writer thread into the bottleneck, which can cause Replica
SERVER-135501Open
Add observability for collection hash repair
h2. Summary
Add metrics and structured logs for collection hash repair so operators can determine whether repairs are being requested, making progress, successfully incorporated into replicated metadata, or repeatedly failing.
h2. Proposed Metrics
* Repair requests and outcomes: counters for requests, committed repair oplog entries, failures, and no-ops. Distinguish no-op reasons such as collection no longer existing or empty repair metadata.
* Repair mode: distinguish hash replacement (
NODE-6910Investigating
Add test for SRV hostname validation when resolver and resolved hostnames are identical with three domain levels
This ticket was split from {{DRIVERS-3162}}, please see that ticket for a detailed description.
Components
Initial DNS Seedlist Discovery
NODE-7812Backlog
Benchmark OpenTelemetry support
This ticket was split from {{DRIVERS-3620}}, please see that ticket for a detailed description.
Components
Observability
SERVER-138843ClosedFixed
Fix metrics filtering test failure on missing disagg.pali.read moving average
h2. Problem
{{filtering_sharded_cluster.js}} in the {{disagg_storage}} suite intermittently fails on the all-feature-flags TSAN variant with a schema mismatch in {{getDiagnosticData}} (also reachable via {{serverStatus}}) metrics:
BF: https://jira.mongodb.org/browse/BF-47116
Failing test: {{src/mongo/db/modules/atlas/jstests/disagg_storage/metrics_filtering/filtering_sharded_cluster.js}}
h2. Root cause
The fixture compares the metrics schema of an auth-enabled cluster against an auth-disab
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-134568Backlog
See if we can quantify the perf impacts of SERVER-133331 from the internode hash validation stress test
SERVER-138604Backlog
Use the collection's collator when building field stats
If a collation is specified for the collection the analyze command should pass the collator to the stats builder.
NODE-7987Backlog
Connection pool can check out a socket from the generation cleared before it finished handshaking
h3. Current behavior
ConnectionPool.createConnection copies the pool generation into the connection options when the attempt starts. The success path then checks only that the pool state is ready. clear() increments the generation and pauses a non-load-balanced pool. ready() sets the pool ready again without moving the generation back. A handshake that was in flight across that sequence is checked out, or pushed into the available list, with the pre-clear generation. The pool spec does not allow
Components
CMAP
NODE-7984Backlog
Server session pool reuses an lsid after logicalSessionTimeoutMinutes when idle time is a multiple of one hour
h3. Current behavior
ServerSession.hasTimedOut computes idle minutes as round(((elapsedMs % 86400000) % 3600000) / 60000) and returns idleTimeMinutes > sessionTimeoutMinutes - 1. The modulo folds idle time into minutes past the hour. A session idle for 60 minutes with logicalSessionTimeoutMinutes of 30 is compared as 0 > 29 and is reused.
h3. Expected behavior
Idle time is the elapsed minutes since lastUse. The session is not reused when that value is greater than logicalSessionTimeoutMinutes -
Components
Sessions
NODE-7986Backlog
Auto-encryption internal client keeps command listeners after the application removes them
h3. Current behavior
Encrypter.getInternalClient copies listeners onto the internal key-vault client, then listens for newListener and forwards every later subscription. There is no removeListener forwarder, and close() does not remove the newListener hook. After the application client removes a commandStarted listener, the same function still runs for key-vault commands and remains reachable from the internal client. The forwarder itself stays on the application client after Encrypter.close().
Components
Client Side Encryption
SERVER-135027Open
Existing time-series remediation scripts will not work with viewless time-series
The existing time-series remediation scripts, like [this one|https://github.com/mongodb/support-tools/blob/21920b7ee0b4696efd6062e9609e7927732a3638/timeseries-bucket-repair/timeseries_bucket_repair.js#L123], operate directly on the buckets collection. For customers running 9.0+, time-series collections will be viewless and will not have the buckets collection; the equivalent interaction can be achieved by using the rawData parameter.
We should either create another set of scripts that work on
PYTHON-6209Investigating
[Rust core] Audit how one shared Rust library serves PyMongo's sync and async clients: runtime, cancellation, TLS, fork
h3. *Context*
{color:#505f79}Describe the background behind the problem.{color}
h3. *Definition of done*
{color:#505f79}What must be done to consider the task complete?{color}
h3. *Pitfalls*
{color:#505f79}What should the implementer watch out for? What are the risks?{color}
SERVER-138871ClosedDuplicate
Ordinary $lookup inherits isHybridSearch from enclosing pipeline and enters hybrid-search-only serialization
h2. Problem
{{ExpressionContext::isHybridSearch\(\)}} is inherited by sub\-pipeline expression contexts \({{expression\_context\_builder.cpp}}\). Any code that routes on it cannot distinguish "my own pipeline is hybrid search" from "I am merely nested inside a hybrid search pipeline."
The clearest instance: {{DocumentSourceLookUp}}'s serialization swap \({{document\_source\_lookup.cpp}}, gated on {{\_fromNsIsAView && \_fromExpCtx\->isHybridSearch\(\)}}\) replaces the lookup's resolved pipeline