Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
NODE-7900In Progress
Update project dependencies
h3. Use Case
As a... Node driver engineer
I want... to have up to date dependencies
So that... we can be more efficient by using the best available tools
h3. User Impact
* N/A
h3. Dependencies
* N/A
h3. Unknowns
* Do we want to update any previous majors?
h3. Acceptance Criteria
h3. Implementation Requirements
* Review and update the dev and prod dependencies in the following libraries:
** driver
** js-bson
** mongo-client-encryption
** kerberos
** zstd
** dbx-js-tools
** mongodb-
Components
Dependencies
SERVER-139295ClosedFixed
Retry WriteConflict when acquiring/promoting/releasing the critical section
*Issue*
During a DSC step down, DDLs can bubble up a WriteConflict error to the user (instead of the critical section code internally retrying it).
*Analysis*
[Acquiring, promoting or releasing the critical section writes via DBDirectClient with locks held|https://github.com/mongodb/mongo/blob/b2b94807a4d84159532979b97387cc806f691e0e/src/mongo/db/global_catalog/ddl/sharding_recovery_service.cpp#L282-L364].
In a DSC cluster, during stepdown we may hit a WriteConflict there due to
Affected Versions
9.1.0-rc0
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-137182In Progress
Parent task tracking work for exposing query knobs for PQS
See the syntax doc for the list of knobs. This should involve:
* updating each existing knob to be visible via query settings
* updating sites where we check the knob to use the QKC context/ not use the knob directly
* adding tests validating each knob does the right thing for all its values
* adding tests to verify interactions between related knobs work as expected
SERVER-139082ClosedFixed
Enable multikey_timestamp_mixed_transaction_large_oplog_batches.js in no_passthrough_disagg_override_replsets
The test pauses secondary replication using {{stopReplProducer}} before executing its mixed workload. DSC bypasses the conventional oplog fetcher that checks this failpoint, so the test waits 10 minutes and fails with {{MaxTimeMSExpired}} before exercising its multikey timestamp assertions.
As part of this ticket, we have to adapt this test by using {{pauseOplogApplication}} on the secondary when running under DSC. We have then enable the test in {{no_passthrough_disagg_override_replsets.}}
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-138782In Code Review
Implement kickback and backfill on the router
Make the router throw a {{ChangeStreamEndOfTransactionBackfillRequired}} error if it encounters a transaction with a commit floor that is beyond the stream's starting point.
The router will catch the error, close the current stream and open a new one at the backfill point. The backfill scan will emit accounting control events for {{endOfTransaction}} and nothing else, until the scan reaches the backfill resume position. Then the stream will continue normally.
SERVER-82520Backlog
Remove TSAN suppressions for WiredTiger
There is a TODO in the codebase referencing a resolved ticket to remove TSAN suppressions for WiredTiger [here|https://github.com/mongodb/mongo/blob/0348832e0d5796ff6f70f84827e832b691027f26/etc/tsan.suppressions#L8-L22].
We should evaluate if we can remove these suppressions after modernizing the WT memory model.
SERVER-137180ClosedFixed
Add a per attempt timeout to CMS calls
Calls from mongod to CMS currently use a heartbeat to detect dead servers and network errors. There is still a possible gap where a CMS connection is accepted and established but takes a long time to service the request (the connection gets stuck and theres no response) In that case calls like GetLogSegmentList can leave startup waiting on the stuck CMS instead of trying another cell. If we had something like a 30s timeout for each attempt we could fallback
The ORP change in SERVER-135177 gav
Fixed Versions
9.1.0-rc1026
GODRIVER-4213Backlog
Return an error message that describes all retry attempts
h3. *Context*
Currently, there's a [bunch of logic|https://github.com/mongodb/mongo-go-driver/blob/6598290d5e92edc6de9428c4bf109decdd33ba50/x/mongo/driver/operation.go#L579-L597] that tries to pick which error to return after a failed retry. That logic frequently discards important error information and makes it difficult or impossible to troubleshoot problems without a bunch of additional event monitoring or logging enabled. It would be much more useful to collect all error information for all
SERVER-138252Backlog
Add WT database wide check to test sweep and validate command
In WT-18582 WiredTiger made changes to run unique btree ID check separately from verifying each table. Running verify with the database: namespace will trigger the unique btree ID check and this is something worth adding to the testing sweep. This ticket will make that implementation and decide whether this call should be added to the non-modal validate command.
SERVER-122285Backlog
Add e2e functional testing of interesting scenarios
In addition to the unit testing of individual components, we need some integration testing of the scenarios which we expect may have problems. Some ideas for things to test:
* Create several collections outside of a multi-document transaction. Advance the stable timestamp incrementally to each of the timestamps, take a checkpoint, and verify that the checkpoint contains only the expected tables.
* Create several collections inside of a multi-document transactions. Take a checkpoint without a
SERVER-122277Open
Enable strict validation when installing checkpoints when possible
Once validation of metadata and graceful stepdown have been implemented, the following sequence of events becomes invalid:
# Node is primary.
# Concurrent conflicting createCollection calls happen, resulting in one of them creating a table and then rolling back. Table is added to the reaper.
# Reaper gets EBUSY when it tries to drop the table.
# Node steps down.
# Node installs a checkpoint from the new primary
Here the validation will fail because the node has a table that it shouldn'
SERVER-138332Needs Scheduling
Build the MetricsRegistry for materialization, validation, and unmaterialize()
h2. Overview
Build the process-wide MetricsRegistry for the unified metrics framework (epic SPM-4962), plus the registration codegen that feeds it. This ticket absorbs SERVER-138333 (registration codegen): the registration spec, the macro, and the generated initializers are one contract with the registry that consumes them.
h2. Scope of Work
* One registry per process, created explicitly via {{materialize(ServiceContext*)}} and accessed via {{MetricsRegistry::get()}}.
* Members: {{_sections}}
SERVER-138333ClosedDuplicate
Registration codegen: per-file MONGO_INITIALIZER metrics registration
h2. Overview
Build the registration codegen for the unified metrics framework (epic SPM-4962). Design revised 2026-11-11: registration is a per-file MONGO_INITIALIZER with no aggregate anchor.
h2. Scope of Work
* Emit one MONGO_INITIALIZER per *_metrics.idl file into the generated _gen.cpp, named RegisterMetrics_<file_path_underscores>, whose body calls addMetricsRegistration(MetricsRegistrationSpec{parent, name, resolve, leaves, appender, otel}) directly. No wrapper registration function, no
SERVER-135689In Code Review
Wire fallback triggers and the run loop
When the decision function says kFallbackToORP, we should set some fallback requested flag and return a normal connection failure error. The decision function site itself doesn't run the ORP, we need to bubble it back up so the run loop does the ORP read and the flag says that this wasn't a network problem.
For this ticket, we let the run loop see the flag and perform the fetch and reconnect. Fetch our missing range from ORP, apply it, then call the normal connect path to cut back to the LS.
SERVER-138995In Code Review
Remove BSONElementValue
This class cannot be used safely has it does not have type information. We should simply switch usage to plain BSONElement
SERVER-139202In Code Review
Make SBE empty TypeSignature safer
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-137173In Progress
Avoid exposing the entire config manager via pali runtime
This inlines the API used by non-pali components that (transitively) depend on the config manager. Which avoids exposing a lifetime-extending handle to the config manager via the runtime interface.
This exposes an odd piece of previously untracked techdebt where pali used repl to get a pali component back from itself (the rate limiter). We are concurrently figuring out what the internal-facing API considerations for pali look like, so i am leaving this oddity as future work. (note: the state
SERVER-136063In Progress
control lifetimes of pali components
We make PALI's runtime interface return references rather than shared ptrs, since it should control their lifetime. This primarily impacts the state_machine, which does the same thing, and the replication coordinator, which we modify to use a state machine interface as well.
Components
Replication
SERVER-90264ClosedFixed
Remove slow step from compile
Fixed Versions
5.0.278.1.0-rc06.0.167.3.38.0.0-rc57.0.11