Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-137331ClosedFixed
Encapsulate join plan cache logic and move it out of join/executor.cpp
Currently, a lot of the core logic for join plan caching, e.g. baseline trial (after SERVER-136737), index fingerprint invalidation logic, etc are handled in join/executor.cpp.
For better modularity, we should move the join plan cache logic out and encapsulate it in a class that abstracts away the necessary state and logic.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-137395Backlog
Update variants used on sys-perf for release branches
Review the sys-perf variants on the 9.1 branch and delete any that are no longer required. Make updates to older release branches as needed.
SERVER-139001Needs Scheduling
Update Homebrew Formulae for New Major Patch Release (5.0.35)
Update Homebrew Formulae to support new Major Patch Version 5.0.35
SERVER-138874ClosedFixed
Transient negative metric via disabling load shedding during a serverStatus read
{{serverStatus().queryMemory.loadShedding.lowMarkBytes}} can briefly report a negative value when load shedding is disabled at runtime.
{{appendQueryMemoryLoadSheddingStats()}} decides {{active}} from the RSS monitor state ([L588|https://github.com/10gen/mongo/blob/8fc0d56da14997f6a5c6b02d20ce941c3cd55ccd/src/mongo/db/memory_tracking/query_memory_load_shedding.cpp#L588]), then separately reads {{queryMemoryLoadSheddingLowMarkPercent}} ([L600|https://github.com/10gen/mongo/blob/8fc0d56da14997f
Affected Versions
9.1 Required
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135582Open
Add an architecture guide for server backpressure and queueing mechanisms
NODE-7471Backlog
Flaky: ObjectId > range query based on objectId timestamp [5]
https://parsley.corp.mongodb.com/evergreen/mongo_node_driver_next_windows_2022_latest_large_Node24_test_7.0_sharded_cluster_4157b2600dda54ee14693ee951abd72e4da2b07e_26_03_13_00_30_08/0/task?bookmarks=0,9903
Components
Flaky Test
SERVER-138392In Progress
Test DDL operations in checkpoint_validation_standby_restart.js
This is follow-up work from SERVER-138287. Test at least the following cases:
# Create and drop before adoption
## Both entries are adopted without validation.
# Create before adoption, drop post-adoption
## The drop is validated as a container delete because the create is in the baseline.
# Create + drop post-adoption
## Validated as normal. If within a batch, nothing is validated. If they are across batches, it's the same expected value as 2.
Use a separate {{it("")}} for legibi
SERVER-138872In Code Review
Create a jstest to ensure the standby skips adopting a checkpoint when we don't see the corresponding watermark
SERVER-136319Open
Clean up options for Validate Command
Most validation options pass through an enum `ValidateMode` which generates a combination of options that are fed forward into the validate command.
This enum should be replaced with a set of options that can be precisely switched to enable specific checks as needed.
This ticket will most likely require updates to the public-facing docs if these options need to change.
SERVER-132996Open
Re-enable tests disabled due to FCV upgrade to 9.1
*Team:* Validate
A number of tests have been disabled as they were failing after the FCV version is bumped to 9.1.
All affected tests are labelled with this ticket number.
Review the tests and evaluate whether they need to be updated to reflect the new FCV version or can be removed.
SERVER-137812In Progress
SeekableRecordCursor API for partitioning a RecordStore
Given a range between minRecordId and maxRecordId, and a number of partitions N, aka Degree of Parallelism (DOP), returns a sorted vector of unique RecordIds that partition the range, with the array size <= N -1. (Question for : or would you prefer N + 1 entries, with the first and last RecordId returned as well?)
The bounds are optional: a missing lower bound means the beginning of a collection, and a missing upper bound means the end of a collection.
The implementation will make use of
SERVER-137517In Code Review
Complete TODO listed in SERVER-76346
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* etc/evergreen_yml_components/variants/rhel/test_dev_master_branch_only.yml:276
* jstests/noPassthrough/validate/modal_validate_specify.js:42
* src/mongo/db/server_feature_flags.idl:365
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-76346&type=Code
The next steps for this ticket are to either remove the outdated TODO or follow
SERVER-125934In Code Review
Timeseries collections created with default parameters should be eligible for the fixed bucketing optimization
SERVER-129611In Progress
Conditionally enable instantiating the oplog on each use
SERVER-122142 fixed one of the problems that forces CollectionCatalog::_instantiateCollectionIfTesting() to exclude the oplog, but it has turned out that even with it not starting/stopping a thread every time it's still quite slow and makes some tests time out, especially when debug logging is involved. As such we still need to conditionally exclude the oplog for some tests.
SERVER-132665Open
Limit the number of indexes per index build on disaggregated storage clusters
Option 1 of the side-write drain decision document. Cap the number of index specs accepted by a single {{createIndexes}} command on a disaggregated storage cluster (proposed limit: 3), so that the per-index share of the serial drain budget stays large enough for the index build to converge.
This is a stopgap intended to make multi-index builds usable for Atlas Infinite public preview, while the longer-term options (replicated container truncates, parallelizing the drain, backpressure on side w
SERVER-118795Open
Update HELP Playbook for HELP-86102
The information and context for HELP-86102 was flagged for adding to the HELP Playbook. As owner of the original HELP issue (Assigned Teams), please review the linked issue and update the playbook accordingly
SERVER-133580Open
Complete TODO listed in SERVER-131379
There is a TODO in the codebase referencing a resolved ticket which is assigned to you.
Please follow this link to see the lines of code referencing this resolved ticket:
https://github.com/10gen/mongo/search?q=SERVER-131379&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 work you're actually doing.
SERVER-133740In Progress
Log the number of documents modified by sub-operations in a bulk write
When a slow query log is emitted for a operation within a bulkWrite, it appears that the number of documents modified by that operation is not logged (see [this comment|https://jira.mongodb.org/browse/HELP-98158?focusedCommentId=8779035&focusedId=8779035&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-8779035] for an example of a batched update that is not reporting nModified. This could be because each statement in the bulkwrite command has its own nested CurOp.