Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-137961In Progress
Support MCV and NDV in in-memory stats cache
* combine MCV and NDV and histograms in to idl file
* migrate in-memory-cache to read from the persistent field stats collection
* add reading API functions for single field stats and compound field stats
* the functions should support collator as a parameter
* add the collator to the key of the field stats collection
SERVER-135359Backlog
Fix the recovery duration metrics reporting after node restarts in Authoritative Shards Dashboard Recovery
After investigating several spikes of recovery times from the dashboard as part of [HELP-100264|https://jira.mongodb.org/browse/HELP-100264], it was concluded that all of the spikes that exceed 1sec follow the same pattern.
There was a recovery initiated, but it was interrupted by the user/automation in Atlas, which was observed through the following log in the mongod:
For the next 15 seconds, the cluster wasn't accepting any requests and doing any work, reporting that it entered the quies
SERVER-134054Backlog
Improve authoritative shard recovery time
The authoritative shards metadata recovery process has to navigate a lot of race conditions successfully.
One of these involves the way it waits until it's safe to read at a given timestamp. As explained in [this comment|https://github.com/mongodb/mongo/blob/c68e0440e5e887abebde44003cbd2392a1a900af/src/mongo/db/shard_role/shard_catalog/collection_metadata_synchronizer.cpp#L156-L172] the recoverer has to do the following series of steps in a deliberate order:
# Install the recoverer so that
SERVER-132590Backlog
Prove and test that clearing of the CSR at rollback is not really needed
Authoritative shards cause all the CSRs to be cleared at live recovery rollback (when the process doesn't restart). This is a regression from FCV 8, where we left CSRs intact on live rollback.
We actually think that even in authoritative shards, clearing the CSR is not really needed. The protocol ensures that CSRs are updated under a critical section, and once a critical section is taken, DDLs always retry until successful. The steps they try are also idempotent. In other words, the system wi
SERVER-132573Backlog
Reduce impact of clearing CSR for all namespace at rollback
Shards clear the CSRs for all the namespaces when recovering from a live rollback if the reported rolled-back namespaces or operations involve the auth shards DDLs.
SERVER-130703 prepared for any adverse impact by introducing a server parameter to disable CSR clearing. It also added some telemetry to ensure that we capture the issue in logs and FTDC.
We should address this long-term. There are several possible approaches, such as tracking exactly which namespaces could be impacted and only
SERVER-138636ClosedFixed
Add microbenchmark for sharded cluster transaction commit with allowEndOfTransaction enabled
Fixed Versions
9.1.0-rc0
SERVER-138949ClosedFixed
Reconstruct the prefix type for collection-origin fields
When we ask for a type of a dotted path defined in the graph, we currently use createDocumentFromLeafMatch.
For example, if we have \{$set: {'a.b': '$c'}}, where we know '$c' is a 'number', we nest the 'number' type into the prefix to construct:
Root = \{a: array|scalar|{b: array|scalar|number}}
This is not pointless, once we have this Root type, we can intersect it with collection schema and $match type inference to potentially remove the prefixes.
We need to do the same for fields
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-139426ClosedFixed
Fix mirrored-read race in change_stream_read_preference.js test
jstests/sharding/query/change\_streams/change\_stream\_read\_preference.js samples $indexStats accesses.ops for the _id_ index on each shard secondary around the change\-stream getMore and asserts the delta is exactly 1 — the optimized post\-image update lookup.
Mirrored reads can break that assertion: an update served by the shard primary is mirrored to a secondary as a synthesized find of its filter \(CmdUpdate::supportsReadMirroring, write\_commands.cpp\), recording an extra _id_ index acces
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139437In Progress
Reads' operationTime must cover data read by parallel query workers
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-134367ClosedFixed
Integration of Scan stages with transaction processing
Enhance OperationContext to ensure the transactional correctness of parallel scan stages
# scan stages should see only the snapshot in which the original query was placed when it was started
# is it possible to resynchronize after yielding of single threads?
# coordinate detaching of OperationContext from the main query to the parallel threads
Include writing unit tests stressing the locking part; e.g.
# parallel scan with DOP > number_of_cores reading a collection that is being comp
Fixed Versions
9.1.0-rc09.1.0-rc1026
NODE-7902Needs Review
Update dependencies for zstd
h3. AC
* Merge dependabot PRs that are safe to merge (semver-compatible updates, including any standalone package-lock-only PRs)
* When updating the dev dependencies to latest available semver compatible versions ANY reformatting MUST be done in a separate pull request and should ONLY be done if the team agrees that it is an important improvement
* Check dev dependency major version updates
** Update the major version if and ONLY if this can be done with minimal code changes; otherwise, file
SERVER-137759ClosedFixed
Complete TODO listed in SERVER-117506
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* src/mongo/db/query/compiler/optimizer/join/join_cost_estimator.h:42
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-117506&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
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-136150ClosedFixed
Require querySettings action privilege for $_internalListQuerySettings
*$_internalListQuerySettings* lacks the required querySettings action privilege allowing unintended information disclosure.
Affected Versions
9.0.3
Fixed Versions
9.1.0-rc09.0.79.1.0-rc1026
PYTHON-6216ClosedGone away
GridFS fails to read files with integral BSON Decimal128 length or chunkSize metadata
h3. Detailed steps to reproduce the problem?
_Include full traceback, if possible_
h3. Definition of done: what must be done to consider the task complete?
h3. The exact Python version used, with patch level:
{{$ python -c "import sys; print(sys.version)"}}
h3. The exact version of PyMongo used, with patch level:
{{$ python -c "import pymongo; print(pymongo.version); print(pymongo.has_c())"}}
h3. Describe how MongoDB is set up. Local vs Hosted, version, topology, load balanced, etc.
Components
GridFS
PYTHON-6217ClosedGone away
GridFS upload_from_stream does not delete already-uploaded chunks when the source stream raises
h3. Detailed steps to reproduce the problem?
_Include full traceback, if possible_
h3. Definition of done: what must be done to consider the task complete?
h3. The exact Python version used, with patch level:
{{$ python -c "import sys; print(sys.version)"}}
h3. The exact version of PyMongo used, with patch level:
{{$ python -c "import pymongo; print(pymongo.version); print(pymongo.has_c())"}}
h3. Describe how MongoDB is set up. Local vs Hosted, version, topology, load balanced, etc.
SERVER-138064ClosedFixed
Get types out of subpipeline in the dependency graph
canPathBeArrayFromFieldMetadata checks the subpipeline with getSubpipelineForPath.
We should reuse that logic for subpipeline checks and use it in getType too.
We should also take this opportunity to make getConstant also check the subpipeline, such that the three metadata accessors (arrayness, type, constants) operate the same.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-137195In Progress
block drop command with inbound fk constraints
Ensure collections with FK outbound constraints can be dropped.
Block collection drop of parent collections (when inbound constraints exist).
Also test $out / rename with drop (or similar).
SERVER-107842In Code Review
Improve error message emitted by mongos when 'QueryExceededMemoryLimitNoDiskUseAllowed' is thrown, but 'allowDiskUseByDefault' is set to 'true'
As the title describes, his ticket tracks the work to improve error message emitted by mongos when 'QueryExceededMemoryLimitNoDiskUseAllowed' is thrown, but 'allowDiskUseByDefault' is set to 'true'.
Following the decision to not pursue SERVER-87735 given that the performance hit of this change (from executing merging $groups/spilling stages on the shards instead of on the router) is too significant, we should improve the error mentioned above to clarify that 'allowDiskUseByDefault' is ignore
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-138856ClosedFixed
Change streams maxTimeMS metric should be maxAwaitTimeMS
The change stream {{aggregate}} command’s {{maxTimeMS}} and the cursor’s {{getMore}} {{maxAwaitTimeMS}} represent different options. The metric is intended to measure the cursor wait option ({{{}maxAwaitTimeMS{}}}), but its current name and test behavior can imply that it measures {{maxTimeMS}} on {{{}aggregate{}}}.
Update the metric and its coverage so they reflect the intended option and match customer-facing terminology.
We should do the following;
* Update the Extend Change Streams M
Components
Change streams
Fixed Versions
9.1.0-rc09.1.0-rc1026