Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-131938ClosedFixed
Skip WiredTiger extended HS key validation
BF-44595 identified a regression added by WT-10779. WT-18094 added a flag to bypass the (potentially time-consuming) extended per-key HS validation. While this extra validation is something we should eventually use, the short-term fix to work around the validation performance loss is to simply opt out of the new behaviour.
Fixed Versions
9.1.0-rc09.0.0-rc2
SERVER-137044ClosedDuplicate
Re-land: Reject $_internalKeyStringValue in collection validators and profile filters
SERVER-137446ClosedFixed
Fix PeriodicKillSecondaries background hook date parsing
Datetime conversion should fall back to raw {{ms}} when the value is an invalid {{{}datetime.datetime{}}}.
Fixed Versions
9.1.0-rc09.1.0-rc1025
JAVA-5125Backlog
Some write concern errors are not retried according to spec
JAVA-5124 was opened because the legacy retryable writes test suite was not being run against sharded clusters. After that issue was addressed it was observed that tests of write concern error retryability fail on 4.2 and 4.4 sharded clusters. The tests that fail mostly have description ending with "succeeds after WriteConcernError ShutdownInProgress", although there are a few other failures as well.
While attempting to debug these test failures, the following observations were made:
1.
Components
Retryability
JAVA-5226Backlog
Clarify exactly what code/label fields drivers should inspect to determine retryability
This ticket was split from {{DRIVERS-1641}}, please see that ticket for a detailed description.
Components
Retryability
JAVA-6087In Code Review
RetryableWriteError label not preserved for retryable writeConcernError
In MixedBulkWriteOperation, when we decide to attach RetryableWriteError for a retryable write concern failure (e.g., last-attempt path), the label is currently written to the {*}writeConcernErro{*}r subdocument. However, when building MongoBulkWriteException, BulkWriteBatch.getBulkWriteException() reads labels from the top-level {*}result.errorLabels{*}. Labels added under *writeConcernError* are not parsed, so they are lost.
*Impact*
Applications checking e.hasErrorLabel("RetryableWriteErr
JAVA-6111In Code Review
MongoCollection single write requests write concern errors not always handled correctly
Both sync and async exception handling for single inserts, updates, deletes can lose write concern exception / write results information.
It is possible to have both a write concern exception and a bulk write exception. Currently, in sync there are any bulk write exceptions, the write concern exception is not propogated. In async a write concern exception is returned but no write results information.
Components
CRUD
SERVER-136567ClosedFixed
compile_commands.json omits mongo_cc_fuzztest targets from the compilation database
Slack Thread Capture
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-137139ClosedDone
Create test cluster with root privileges to prototype SECL ideas
One avenue I wanted to briefly explore is using cgroups to limit CPU used during session establishment. For that I need custom VMs.
SECL = session establishment concurrency limiter
SERVER-121845ClosedFixed
Remove requires_replicated_fast_count_recovery tag
Reenable the excluded suites for the project test variant. If some suites are still failing for other reasons, like initial sync, upgrade/downgrade, etc., fix or handle them separately.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135054ClosedFixed
Create feature flag
Create the feature flag for this project.
Fixed Versions
9.1.0-rc0
SERVER-137816ClosedFixed
Write `allowEndOfTransaction` flag into applyOps
This change writes an {{allowEndOfTransaction}} flag into the final applyOps oplog entry of replica set transactions if the server parameter {{allowEndOfTransactions}} is enabled.
This allows change streams to unambiguously determine the final applyOps entry in a non-sharded transaction _and_ if the parameter was enabled at commit time.
Sharded transaction oplog augmentation will follow in a separate commit.
This work is part of SPM-4826.
Components
Change streams
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135061ClosedDuplicate
Create feature flag
Create the feature flag for this project.
SERVER-135611ClosedFixed
Flatten Inner Worker Pool for Parallel Record Store Validation
Currently, each collection validation will spin up a pool of new worker threads to operate on "slices" of the record store. This can result in n^2 active threads at a time with nested parallelism.
The inner pools can be flattened into one shared pool, this would further support adding more parallelism for tracking indexes and better manages the division of work.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-134528ClosedFixed
Make internalQueryStatsSampleRate settable via (P)QS queryKnobs
h2. Overview
Make the query stats sampling server parameters overridable per-query via query settings ({{queryKnobs}}):
* {{internalQueryStatsSampleRate}} - sampling-based rate limiting for reads
* {{internalQueryStatsWriteCmdSampleRate}} - sampling-based rate limiting for write commands
h2. Background
SPM-4364 added support for overriding query knobs via (P)QS, and SERVER-123044 allows passing {{querySettings}} directly as part of command invocation. Query knob exposure is opt-in, not opt-o
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-135908ClosedWon't Do
Immutable shard key enforcement for FK-referenced fields
no description yet
PYTHON-6196ClosedDone
Automated spec sync job should succeed instead of failing when there are no changes to commit
h3. *Context*
{color:#505f79}Describe the background behind the problem.{color}
h3. *Definition of done*
{color:#505f79}What must be done to consider the task complete?{color}
h3. *Pitfalls*
{color:#505f79}What should the implementer watch out for? What are the risks?{color}
SERVER-90448ClosedFixed
Support resharding collections where the local catalogs and the sharding catalog do not agree on the collection's UUID
[The ReshardingCollectionCloner includes the {{collectionUUID}} parameter to the aggregate command it runs|https://github.com/mongodb/mongo/blob/r7.0.7/src/mongo/db/s/resharding/resharding_collection_cloner.cpp#L184]. [The CommonReshardingMetadata::getSourceUUID() value comes from the sharding catalog's belief of the source collection's UUID|https://github.com/mongodb/mongo/blob/r7.0.7/src/mongo/db/s/resharding/resharding_donor_recipient_common.cpp#L269-L273]. We have seen a small number of inst
Components
Sharding
Fixed Versions
8.1.0-rc0
SERVER-92607ClosedFixed
Eager reaping should not reap a session with a yielded TransactionRouter
Eager reaping is a mechanism to reap internal transactions spawned for retryable writes (ie session ids with txnNumbers in them) before their parent session expires to avoid unbounded memory growth. This works by [triggering a callback when a new "client" transaction number is seen|https://github.com/10gen/mongo/blob/0ed820f428e08a19a4f01b34833395f772ea274f/src/mongo/db/session/session_catalog.cpp#L372-L386], ie the top level txnNumber in a retryable write or client started transaction, and will
Fixed Versions
8.1.0-rc0