Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-138403ClosedFixed
Reset SizeCountCheckpointValidator on step up
Today, the lifetime of the {{SizeCountCheckpointValidator}} is equal to lifetime of the {{OplogApplier}} which is actually equal to the process lifetime. If a node creates an {{OplogApplier}} at any point, steps up, and then steps back down, it continues using the state in the pre-existing {{{}SizeCountCheckpointValidator{}}}. Oplog application stops during the node's primary tenure, so the {{SizeCountCheckpointValidator}} state is stale, and subsequent checkpoint validation is fails.
Change
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138560ClosedFixed
Refactor expectValidatorMetrics() to test_helpers
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138539ClosedFixed
Delete outdated _lastGetNext in `PlanExecutorSBE`
This logic is outdated as of [SERVER-86887|https://jira.mongodb.org/browse/SERVER-86887], since {{PlanExecutorSBE}} never moves out of the result slot
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-86887ClosedFixed
Remove cross-stage use of SlotAccessor::copyOrMoveValue
Remove cross-stage use of SlotAccessor::copyOrMoveValue.
copyOrMoveValue is unsafe when combined with loopy or unwindy stages and can lead to segfaults:
buffering and keep the value in scope until EOF or child EOF is insufficient, and will segfault when combined with correlated variables:
Similar problem exists in PlanExecutor use of copyOrMoveValue. This optimization targets the mkObj + PlanExecutor combination, where we want to avoid making extra copy when converting resul
Fixed Versions
8.0.0-rc0
SERVER-138472ClosedFixed
Log continuous internode validation failure by default
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138399ClosedDuplicate
Set fassert parameter off explicitly in checkpoint_validation_steady_state.js
The test "does not validate checkpoints while the server parameter is disabled" does not work when \{[enableContinuousInternodePerCollectionValidation}} defaults to true, so we need to explicitly set enableContinuousInternodePerCollectionValidation to false in that test.
Also, checkpointsValidated == 0 must become == validatedBefore because the counter is cumulative since process start, so when the parameter defaults to true, the standby could validate checkpoints before we turn the paramater
GODRIVER-4224Backlog
mongo-go-driver - PR #2696: Merge release/2.9 into master
[mongodb-drivers-pr-bot[bot]|https://github.com/apps/mongodb-drivers-pr-bot] has created [PR #2696: Merge release/2.9 into master|https://github.com/mongodb/mongo-go-driver/pull/2696] in [mongo-go-driver|https://github.com/mongodb/mongo-go-driver]
SERVER-132246ClosedFixed
minWaitForStreamingHelloMillis default breaks spec-compliant drivers
h3. Context
SERVER-128517 added minWaitForStreamingHelloMillis, which enforces a server-side floor on how long a pre-auth streamable hello is held, the drivers team understands that the effective await is
The default value of this floor exceeds the minimum {{maxAwaitTimeMS}} a spec-compliant driver is allowed to send, which is [{{500ms}}|https://github.com/mongodb/specifications/blob/master/source/server-discovery-and-monitoring/server-monitoring.md#heartbeatfrequencyms].
The default wa
Fixed Versions
9.1.0-rc08.3.98.0.309.0.0-rc27.0.45
SERVER-138184ClosedWorks as Designed
Use stdx::thread in place of std::async in IntentRegistry::killConflictingOperations
h4. Background
{{IntentRegistry::killConflictingOperations}} (src/mongo/db/repl/intent_registry.cpp) runs the kill-and-drain work on a {{std::async(std::launch::async, ...)}} thread and returns a {{{}std::future<ReplicationStateTransitionGuard>{}}}. Evaluate replacing {{std::async}} with our internal library {{{}stdx::thread{}}}, and decide whether the differences in behavior are acceptable.
h4. Proposed change (to be evaluated)
Replace {{std::async}} with a {{stdx::thread}} plus a {
SERVER-138021ClosedFixed
Harden num_orphan_docs.js against stale secondary reads
In chunkMigration, there is a document for the donor and recipient in {{{}config.rangeDeletions{}}}. As documents are copied to the recipient, the recipient document increments its {{numOrphanDocs}} . For example, 2 documents are copied and the recipient document has a {{numOrphanDocs}} count of 2 (the purpose of the recipient document is to delete these orphans on abort), BUT {{pending: true}} is set, which means it's not finalized yet.
So then what's happening is {{analyzeShardKey}} queries
Fixed Versions
9.1.0-rc09.1.0-rc1025
JAVA-5850Needs Triage
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
PYTHON-5327ClosedDuplicate
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-7855Backlog
[Claude Agent SDK] Package and publish
After completing the implementation of a Claude SDK adapter for TypeScript (NODE-7846), ABX will require support/guidance to package and publish to npm.
Please reach out to NODE-7846's implementer directly for any questions/information.
Components
ABX
NODE-7846Ready for Work
[Claude Agent SDK] SessionStorage adapter (TypeScript)
h2. TL;DR
Create a standalone TypeScript package that provides the same user-facing behavior as the Python MongoDB session store (INTPYTHON-1109), using the official MongoDB Node.js driver.
h2. Context
The Claude Agent SDK runs agent conversations. A *session* contains the main conversation and any conversations started by subagents; each conversation's recorded messages and metadata form a {*}transcript{*}.
The Python implementation persists these transcripts in MongoDB so application
Components
ABX
SERVER-138304ClosedDuplicate
[POA] Make internode validation work with POA
Just a placeholder ticket.
SERVER-138158ClosedDeclined
Survey storex tests to see if we skip testing coverage with otel test skips
We can just not check otel metrics instead of completely skipping tests when there is no otel. See if we potentially miss test coverage due to skipping when otel isn't supported.