Known Issues
5166 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-139339Open
Add zone observability for resharding operations
Expose zone configuration for resharding operations that reports:
* Whether the request supplied explicit zone definitions.
* Whether the source collection had pre-existing zone ranges.
SERVER-139507Needs Scheduling
Allow option to select which donor node the resharding recipients clone from
Resharding recipients clone from donor shards using a hardcoded \{mode: "nearest"} read preference. The member within each donor shard is chosen by standard server selection: all members whose RTT from the recipient is within fastest RTT + localThresholdMS (default 15ms) are eligible, one is picked at random, and it is then pinned for the entire clone phase. There is no way to influence this pick.
Add a supported way to select which donor member recipients clone from. For example: a tag set a
SERVER-138274Backlog
Add option to make resharding always clone from secondaries
Nowadays resharding uses {{nearest}} read preference for cloning, which could read from a secondary or a primary. In general primaries are more overloaded, so in some situations reading from a secondary might be preferable. We should have an option to use {{secondary}} read preference in those cases.
SERVER-139509In Progress
Explicitly requested resharding validations can be silently downgraded
Today, a user could request performValidations: true in a resharding command, and have that request be [downgraded with only a warning|https://github.com/10gen/mongo/blob/b725078378cfd5e52670e307306594ad739ac846/src/mongo/db/s/resharding/resharding_util.cpp#L870-L888] if the size of the collection is beyond the threshold or if its size cannot be determined.
After SERVER-138245, and part of SERVER-135224's related work generally, some top level resharding parameters are going to be deprecated
SERVER-136679ClosedFixed
Idle $search cursors on views are not disposed at shutdown and stall mongod for the executor timeout
h3. Summary
An idle {{$search}} cursor opened with {{batchSize: 0}} against a *view* (view definition contains {{$search}}) is not marked as holding the mongot task executor, so it is not disposed during shutdown. mongod then stalls for the full {{searchTaskExecutorShutdownTimeoutMS}} backstop (default 60s), or indefinitely if the timeout is disabled ({{0}}).
h3. Background
SERVER-129849 added {{ClientCursorParams::mayHoldMongotTaskExecutor}} and disposes idle cursors holding mongot executo
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139141ClosedFixed
Add CommittedTxnMatcher pool matching and stream-level txn deduplication to the change stream matchers
Add the matching machinery for transaction event sets: one committed transaction is matched by a pool \(order\-free inside the txn\), with the endOfTransaction event asserted last, and re\-emitted txn events are deduplicated at the stream level.
h2. Scope of Work
\* Extend {{jstests/libs/util/change\_stream/change\_stream\_event.js}} — {{CommittedTxnMatcher}}: pool of the txn's statement matchers \+ a gated EOT arm \(mandatory/optional/absent, derived from the reader config\); position\-locked
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-139278Backlog
[POA] Preserve split prepared transaction ordering across live worker resize
Make changes to {{replWriterThreadCount}} safe while a split transaction remains prepared. This expands the ticket from test coverage alone to the ordering coordination needed for [SERVER-138432|https://jira.mongodb.org/browse/SERVER-138432] alongside live worker resizing from [SERVER-134783|https://jira.mongodb.org/browse/SERVER-134783].
Split commit/abort dispatch uses the original split's requester ID modulo the current worker count. Subsequent CRUD uses the document hash modulo the current
SERVER-138432Backlog
Pipeline split prepared batches in POA
Let POA dispatch the next batch without waiting for an eligible split prepare or commit/abort batch to finish. Workers still apply their queues in order, and {{lastApplied}} must wait for all transaction work and session updates in earlier batches. Keep the serialized path for non-split commit/abort and transactions requiring serialization (ones containing DDL operations).
Wait for pending split-session cleanup before preparing another transaction on the same logical session, and make that wait
SERVER-139352Needs Scheduling
Refactor SampleSpec to better describe different spec types
Currently we have one SampleSpec class that describes all sample requests, of which there are 3 types:
* request to generate an OTF sample
* request to load a persistent sample
* request to refresh a persistent sample
These types all have different parameter requirements, and it's messy to manage which fields should be filled out for which types. It would be better to handle in an object oriented way by defining each of the 3 spec types as subclasses.
SERVER-139501In Code Review
Adopt corrected IBM cross GCC toolchains
Release-branch PPC64LE and s390x cross-GCC pins predate two toolchain configuration fixes. Adopt the rebuilt v4/v5 GCC archives at toolchain-builder revision 6313c32e971dc7f77be261669a72f80a11ce3346 on v7.0, v8.0, v8.3, and v9.0.
h2. Problem and correction
* Cross GCC lacked {{--enable-initfini-array}}. In 7.0 PPC server binaries this caused application constructors to run in the reverse of the native build's order. The rebuilt v4/v5 archives use modern {{.init_array}}/{{.fini_array}} handling.
SERVER-134712Needs Scheduling
Add server metric for total space used by stats vs its limit
SERVER-139460In Code Review
sharding_jscore_passthrough_with_balancer_rate_limited lead to OOM and system stall
in [https://jira.mongodb.org/browse/BF-47272] 6 parallel task exhausted system memory for 20 minutes and lead to tasks timeout.
Try reduce parallelism allowed for this task
SERVER-138647Backlog
Waive quarantined test failure results from resmoke
Resmoke already skips tests that have been quarantined.
However, there is a TOCTOU window where something could be quarantined during the run, and the result should be appropriately marked as "waived".
This will also allow us to upload test results, specifically to trunk.io for further study of flaky signals.
Components
Testing Infrastructure
SERVER-137510ClosedFixed
Complete TODO listed in SERVER-123384
There is a TODO in code owned by your team which references a resolved ticket.
The TODO appears in:
* src/mongo/db/replicated_fast_count/replicated_fast_count_test_helpers.cpp:483
You can also search for all references to the resolved ticket here:
https://github.com/10gen/mongo/search?q=SERVER-123384&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
Fixed Versions
9.1.0-rc09.1.0-rc1026
SERVER-131864Needs Scheduling
Revisit persisted samples validation
This ticket should address the issue of how we raise errors when we fail to find a persistent sample. There are a few possible cases, for example:
# no sample exists
# sample was found but format is invalid
# the samples collection is not clustered
Currently we return a Status when checking for a persistent sample. If it's an error status, we fall back to OTF sampling. For case 1 this is clearly correct, but cases 2 and 3 are situations where something has gone badly wrong with sample p
SERVER-137144Needs Scheduling
Add persistent samples collection to collection validation
There is as validation process that runs on demand and while testing that validates contents, schema, etc... in each collection. We should add persisted samples validation to your workstream, something similar to how we validate timeseries buckets.
Readme: [https://github.com/10gen/mongo/blob/fd376da77021ba8a16a668ab26a8f41fe0f65bb5/src/mongo/db/validate/README.md]
Eg: [https://github.com/10gen/mongo/blob/fd376da77021ba8a16a668ab26a8f41fe0f65bb5/src/mongo/db/validate/validate_adaptor.cpp#
SERVER-127371Needs Scheduling
Consolidate system.stats (persistent samples) and system.statistics (histograms) under one namespace
The analyze command previously existed as a test only command generating histograms and storing them in the system.statistics namespace. Now analyze has sample mode that stores sample docs in system.stats.samples. We should consolidate these both under system.statistics, something like system.statistics.histograms and system.statistics.samples
SERVER-134893ClosedFixed
Fix BSON generation advancement in WASM bridge
Fixed Versions
9.1.0-rc09.0.79.1.0-rc1026
SERVER-139309ClosedFixed
make the CMS channel manager concurrency-safe
We use an internal synchronized state which is copied when callers make a CMS RPC.
Channel rotations are guarded so that if two concurrent RPCs fail and want to rotate the channels, only one of them will do so and the other will detect the stale rotate and bail out.
Channels are initialized lazily. Aside from being a useful way to spread out connections, this is the same behaviour as before and several tests depend on it.
Components
Concurrency
Fixed Versions
9.1.0-rc1026