Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-136778ClosedFixed
Complete TODO listed in SERVER-134550
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* src/mongo/db/query/explain.cpp:425
* src/mongo/db/query/plan_explainer_test.cpp:703
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-134550&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow the steps in the TODO if it is correct. If the latter, please update the summary and description of
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-119745ClosedWon't Do
Implement all plans execution explain for join-optimizable queries
As of work for SERVER-119736, this doesn't execute rejected plans.
SERVER-138360ClosedFixed
Make test less flaky
The underlying test currently records the current time and opens a change stream at that time using {{startAtOperation´. }}startAtOperationTime{{is inclusive, so the change stream can observe an unrelated event. If the test is run in a loop inside the same fixture (e.g. during burn-in), it may observe the collection drop from the previous test run's}}afterEach()` cleanup. This will immediately invalidate the change stream when it is opened, which can cause flakiness.
Components
Change streams
Fixed Versions
9.1.0-rc09.1.0-rc1025
CSHARP-6018Backlog
SRV connection times out on VPN due to random DNS nameserver order
h4. Summary
When I'm connected to a VPN (WireGuard) on Windows 11, trying to create a Mongo connection using the "mongodb+srv" protocol is very slow, and sometimes throws a timeout exception after 30 seconds.
This is because WireGuard sets a DNS server for the VPN interface, and blocks requests to outside DNS servers. This is what I want, and what many VPNs do, and it works fine with other applications (nslookup, browsers, etc.) because they use the OS's DNS resolution, which tries the DNS
Affected Versions
3.8.0
NODE-7341ClosedDone
LangGraphJS Long Term Memory Store
h4. {*}(i) Target date: April 15h, 2026; This feature will be presented at the London .local on May 7th.{*}{*}{*}
h2. Context
[LangGraph|https://www.langchain.com/langgraph] is a framework, part of the LangChain ecosystem, that allows the programmatic creation of "Graphs" to represent the flow of the interaction with AI agents. Each "Node" represents a function that alters the data "State" and connect to other Nodes through "Edges". LangGraph provides short-term memory ("[Checkpoints|https:/
SERVER-138361ClosedFixed
Remove remaining invariant() calls in change stream pipeline code
Remove all remaining {{invariant()}} calls in change stream pipeline code, outside of tests and benchmarks.
Components
Change streams
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-137777ClosedFixed
Tolerate node shutdown races in CheckMultikeyConsistencyInBackground hook
Just like SERVER-136637 added tolerance for QueryPlanKilled, the hook should tolerate errors coming from a node shutting down.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135986ClosedFixed
Fix unique document count for projected samples
[countUniqueDocuments|[https://github.com/10gen/mongo/blob/f10c8e92fc81557bbe021594e044d11bd3a4b02c/src/mongo/db/query/compiler/ce/ce_common.cpp#L338])] counts distinct _id values, but CBR samples that are projected without _id, result in the count of 1.
In SamplingEstimatorImpl::[estimateNDV|https://github.com/10gen/mongo/blob/27378125adb5b6e4e80f8dfcf4ffa4d34ea265b2/src/mongo/db/query/compiler/ce/sampling/sampling_estimator_impl.cpp#L1421], a sample NDV of 1 combined with 1 unique document
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135980ClosedFixed
Refactor memory tracker to have chunkSize in both Simple memory usage tracker and Operation memory usage tracker
Operation Memory Usage tracker depends on SimpleMemory Usage tracker to have a chunk size. In SPM-4781 we want to use the operation memory usage tracker directly, without going through a simple memory usage tracker. Thus, it needs to have its own chunk size. SimpleMemoryUsageTracker will use its chunkSize to send data to Operation Usage Tracker and Operation Usage Tracker will use its chunk size to send data to curOp.
Fixed Versions
9.1.0-rc09.1.0-rc1024
CSHARP-6083In Progress
Discuss and agree the style rules
The target of this task is to agree on the various rules that have been added to the code style guide. Most of the guide will capture what we already do, but some rules are contentious and others are ones the codebase is inconsistent about. We will possibly spend some time on those, but all the rules are up for debate, including their classification (tier/severity).
The likely format will be a shared document (for example a Google Doc) where each rule is laid out with its options, the current
CSHARP-6085In Progress
Configure the tooling to enforce the agreed rules
The aim of this task is to take the rules the team agreed on and encode them in the modernized tooling, so the guide and what the build enforces are the same thing. For each deterministic rule, set its severity in {{.editorconfig}} to the agreed level (the tiers).
We could add also a check that the format has been done as a git-hook and/or on CI.
CSHARP-6189ClosedDone
Make the license headers, encoding and line endings uniform and enforce them in CI
This ticket normalizes every header to a single banner with a uniform 2010 year, updates the Rider file templates to match, and adds a CI check so they cannot drift apart again. It also sets `charset = utf-8` and `end_of_line = lf` for `.cs` files and normalizes the tree to match, so encoding and line endings are enforced by the `dotnet format` check we already run.
SERVER-121901ClosedFixed
Remove unneeded internalValidateFeaturesAsPrimary flag
Secondaries used to set {{internalValidateFeaturesAsPrimary}} to false to opt out of FCV validation during oplog application. That functionality has moved to the {{OperationContext::_enforceConstraints}} flag, however.
From my reading, there is no longer any code that changes its behavior based on {{internalValidateFeaturesAsPrimary}}, and who doesn't love a good opportunity to delete some code?
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138373ClosedFixed
Metric for frontier advancement while log-entry blocked
The value of SERVER-135319 is the ability to advance the mat frontier while log entry processing is otherwise stalled.
we should add a metric to directly count the number of times that happens.
Something like: every time the entry loop completes, set a flag to 1, every time the frontier stream tries to apply updates, decrement the flag or bump the metric if the flag is already zero.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-110427Backlog
Update Profile Data Files
Fixed Versions
8.3.0-rc09.1.0-rc09.0.0-rc19.1.0-rc1025
SERVER-138273ClosedFixed
Explicitly discard [[nodiscard]] value
{{std::exchange}} is marked {{[[nodiscard]]}} in libc++ 23. Discarding the return value triggers a {{Wunused-result}} warning.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-137508ClosedDuplicate
Complete TODO listed in SERVER-121845
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* jstests/disk/wt_size_storer_cleanup_replica_set.js:60
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-121845&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow the steps in the TODO if it is correct. If the latter, please update the summary and description of this ticket to represent the
SERVER-137355ClosedWon't Fix
Suppress prefetch WT_PANIC from when verifying table in WT_VERIFY during validate
SERVER-135015ClosedFixed
Link up WT metric registry and ensure that the otel metric exceptions are correctly handled
Warning: An uncaught exception on a {{PeriodicRunner}} thread is real bad. Things that can throw:
* {{Counter::add()}} on a negative (guarded by the clamp)
* malformed WT stat name
* metric creation with a mismatched definition.
Wrap the cycle and count failures in try catch, theres a metric for this {{{}recordWTCollectError{}}}.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-138153ClosedFixed
Move the previous-segment check from LSM to MaterializedTracker
The last step in the pre-refactoring is to move the CSCT dependency from LSM to mat-tracker. This way, the LSM acts purely as a provider of mat updates, and the tracker is responsible for sorting out what updates to forward to its observers.
Importantly, we modify the {{update()}} to take the segment id. That is the critical part of this PR chain, as it is a blocker for allowing updates about historic segments independently of the LSM.
Fixed Versions
9.1.0-rc1025