Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
NODE-7877Needs Verification
[Mastra] Auto Embeddings for Observational Memory
h2. TL;DR
Allow Observational Memory's optional semantic retrieval to use MongoDB Automated Embedding without a client-side embedder. MongoDB should embed observation text and search queries server-side, while existing client-side retrieval remains unchanged.
h2. Context
Mastra is a TypeScript framework for building AI agents. Observational Memory compresses conversations into summaries called {*}observations{*}. Its optional semantic retrieval searches those summaries by meaning and link
Components
ABX
NODE-7930ClosedDone
kerberos - PR #371: chore(deps-dev): bump markdown-it from 14.2.0 to 14.3.3
[dependabot[bot]|https://github.com/apps/dependabot] has created [PR #371: chore(deps-dev): bump markdown-it from 14.2.0 to 14.3.3|https://github.com/mongodb-js/kerberos/pull/371] in [kerberos|https://github.com/mongodb-js/kerberos]
NODE-7915ClosedDone
kerberos - PR #370: chore(deps-dev): bump follow-redirects from 1.16.0 to 1.16.1
[dependabot[bot]|https://github.com/apps/dependabot] has created [PR #370: chore(deps-dev): bump follow-redirects from 1.16.0 to 1.16.1|https://github.com/mongodb-js/kerberos/pull/370] in [kerberos|https://github.com/mongodb-js/kerberos]
NODE-7913ClosedGone away
js-bson - PR #945: chore(deps-dev): bump the development-dependencies group across 1 directory with 8 updates
[dependabot[bot]|https://github.com/apps/dependabot] has created [PR #945: chore(deps-dev): bump the development-dependencies group across 1 directory with 8 updates|https://github.com/mongodb/js-bson/pull/945] in [js-bson|https://github.com/mongodb/js-bson]
JAVA-5934Backlog
Add contextual serializer for kotlin.date.Instant
The MongoDB Kotlin driver supports contextual serializers for the kotlinx.datetime library using the @Contextual annotation from kotlinx-serialization. Since Kotlin version 2.1 these types have moved from kotlinx-datetime to the Kotlin stdlib and are thus deprecated.
It would be nice to have @Contextual support for the new stdlib classes such as kotlin.date.Instant out of the box.
Workaround in the meantime: Create your custom serializer.
Components
Kotlin
NODE-7610Development CompleteFixed
Configurable DNS domain validation for SRV records
This ticket was split from {{DRIVERS-3329}}, please see that ticket for a detailed description.
Components
DocsInitial DNS Seedlist Discovery
Fixed Versions
7.8.0
SERVER-135699ClosedDuplicate
Absolute waitForTimesEntered(1) barrier is a no-op in StepUpReconfigConcurrentWithForceHeartbeatReconfig reconfig test
h2. Origin
Adjacent-risk hunt from the *BF-46466 postmortem* — SUSPECTED, NOT YET CONFIRMED. Confidence: highest of the sibling risks. Related: [BF-46466], [SERVER-133148] (which established the correct relative-count pattern).
h2. Summary
In {{src/mongo/db/repl/replication_coordinator_impl_reconfig_test.cpp}}, the test {{StepUpReconfigConcurrentWithForceHeartbeatReconfig}} (line 2478) calls {{fpb->waitForTimesEntered(1)}} at *line 2554* to assert that a heartbeat reconfig is "in progress, st
SERVER-135707ClosedWon't Fix
FP Safety 5/5: Restrict absolute failpoint wait API
h3. Background
After T1-T4, no test waits on the absolute failpoint counter by accident. But nothing stops someone from writing {{fpb->waitForTimesEntered(1)}} again in the future.
h3. What to do
Change {{FailPoint::waitForTimesEntered}} so that absolute waits can't be written by accident. Evaluate three options and pick one:
* (a) require an explicit \{{AbsoluteEntryCount\{n\}}} wrapper type
* (b) make the method private, reachable only through {{FailPointEnableBlock}}
* (c) rena
SERVER-135706ClosedFixed
FP Safety 4/5: Migrate initialTimesEntered()+N pattern to relative-wait API
h3. Background
About 76 waits in 47 test files already avoid the counter bug by adding the starting count by hand: {{fpb->waitForTimesEntered(fpb.initialTimesEntered() + 1)}}. These waits are correct but wordy. Copying the pattern makes it easy to drop the {{initialTimesEntered()}} part, and that brings the bug back.
h3. What to do
* Rewrite these waits with the T1 API ({{waitForOneNewEntry()}} / {{waitForNNewEntries(N)}}). This is a mechanical rewrite: test behavior does not change.
* Update t
Fixed Versions
9.1.0-rc09.1.0-rc1024
SERVER-135705ClosedFixed
FP Safety 3/5: Sweep absolute waitForTimesEntered literal call sites
h3. Background
A failpoint's hit counter is process-wide and is never reset between tests in the same unit-test binary. So {{waitForTimesEntered(1)}} returns immediately if an _earlier_ test already hit it, and the wait silently does nothing.
h3. What to do
Two other tests have the same bug as T2. Convert both to count only new hits:
* {{async_rpc_test.cpp}} ({{CancelAfterNetworkResponse}})
* {{version_context_decoration_test.cpp}} ({{FixedOperationFCVRegionSetDuringFcvTransition}})
Both enabl
Fixed Versions
9.1.0-rc09.1.0-rc1024
SERVER-135703ClosedFixed
FP Safety 1/5: Add relative-wait API to FailPointEnableBlock
h3. Background
A failpoint's hit counter is process-wide and is never reset between tests in the same unit-test binary. So {{fpb->waitForTimesEntered(1)}} ("wait until the failpoint has been hit at least once") returns immediately if an _earlier_ test already hit it, and the wait silently does nothing.
h3. What to do
Add methods to {{FailPointEnableBlock}} that count only hits since the block enabled the failpoint:
* {{waitForOneNewEntry()}} and {{waitForNNewEntries\(n)}}
* {{timesEntered
Fixed Versions
9.1.0-rc09.1.0-rc1024
SERVER-135704ClosedWon't Fix
FP Safety 2/5: Fix no-op failpoint wait in reconfig test
h3. Background
A failpoint's hit counter is process-wide and is never reset between tests in the same unit-test binary. So {{fpb->waitForTimesEntered(1)}} returns immediately if an _earlier_ test already hit it, and the wait silently does nothing.
h3. What to do
In {{replication_coordinator_impl_reconfig_test.cpp}}, test {{StepUpReconfigConcurrentWithForceHeartbeatReconfig}} has exactly this bug. An earlier test in the same binary hits {{blockHeartbeatReconfigFinish}}, so the wait passes immedi
SERVER-108693ClosedWon't Do
Expose `supportsWriteConcern` via the listCommands command
Every command includes a function `[supportsWriteConcern|https://github.com/mongodb/mongo/blob/ec48d07f1e3b402785d78bf8bfacb7e5258858be/src/mongo/db/commands.h#L862-L866]` which, if exposed via the [listCommands|https://github.com/mongodb/mongo/blob/ec48d07f1e3b402785d78bf8bfacb7e5258858be/src/mongo/db/commands/generic.cpp#L226-L250] command would allow us to greatly simplify both the [write_commands_all_commands.js|https://github.com/mongodb/mongo/blob/master/jstests/libs/write_concern_all_comm
SERVER-138425ClosedFixed
Revert PQS settability of internalQueryStatsSampleRate
Reverts the PQS-settable {{internalQueryStatsSampleRate}} knob added in SERVER-134528.
Fixed Versions
9.1.0-rc09.1.0-rc1025
NODE-2624Backlog
BSON Timestamp should wrap its value and present as an unsigned long
{{Timestamp}} extends {{Long}}, so it exposes the whole 64-bit integer API: arithmetic, bitwise ops, signedness, raw {{low}}/{{high}}. A BSON Timestamp is just two uint32s, {{t}} (seconds since epoch) and {{i}} (increment), and most of that API makes no sense for it. It also gives odd results, e.g. {{ts.add(...)}} returns a {{Long}} (NODE-4098).
NODE-7069 (bson 7.3.0, js-bson #887) did the groundwork with no runtime changes. It split the inherited {{Long}} members into kept and deprecated (th
Components
BSON
SERVER-137664ClosedFixed
Implement control event type
Make change stream control events distinguishable by their type.
The type should be a new metadata field exclusively set on change stream control events.
For the project, add the new metadata field and give every existing control event the type {{topology}}.
Once we add the control event for {{endOfTransaction}} accounting events, they will get the type {{accounting}}.
This work only introduces the new type metadata field for control events and unit-tests them. The control event type will
Fixed Versions
9.1.0-rc09.1.0-rc1025
CSHARP-6081In Progress
Estabilish codestyle for driver
{panel:bgColor=#fafbfc}
h3. *Summary*
Our codebase has accumulated different various code styles over many years with no single agreed reference, so the same construct is often written several different ways depending on when it was added.
The target of this epic is creating a single human-readable style guide for the driver, and we want to enforce as many of its rules as possible automatically through tooling, so that violations break the build rather than depending on a reviewer noticin
Fixed Versions
4.0.0
SERVER-138492ClosedWorks as Designed
BackfillCoordinator sleepFor onError cancels the current State instead of its own
h2. Problem
The {{sleepFor}} {{.onError}} lambda in {{BackfillCoordinator::markForBackfillAndScheduleIfNeeded()}} calls {{cancel()}} unconditionally. {{cancel()}} consumes whichever {{State}} is current, not the one the sleep was scheduled for.
When a State's cancellation token is cancelled (for example by {{cancel()}} in the {{catch}} block of the same function), the cancelled sleep's {{onError}} runs later on an executor thread. By then a newer State is installed, so the stale callback:
* dro
SERVER-135415ClosedFixed
Modify read path to read all pages for all gens and get the highest complete one, if any. Assume no gen = 0.
Read every page of the stat in a single snapshot and use the highest gen_id whose pages 0 to num_pages-1 are all present. Fall back if no generation is complete.
* Read with the {{INTERRUPT_ONLY}} yield policy and exhaust the cursor in one go (no getMores), so the read never abandons its snapshot.
* Assume gen_id = 0 for documents written before gen_ids existed.
See WRITING-11150, ANNEX II.
Fixed Versions
9.1.0-rc09.1.0-rc1025