Known Issues
5279 tracked MongoDB server issues from Jira with status, affected versions, and resolution details.
SERVER-108019Blocked
Fix how voters are counted in _validate in reconfig
Add the bug fix from SERVER-69418 back in to [ReplSetConfig::_validate|https://github.com/10gen/mongo/blob/d848742a00a71932d92e3042b65e4f3fabb0bf05/src/mongo/db/repl/repl_set_config.cpp#L304-L308].
This ticket is blocked by SERVER-106079, see [this comment|https://jira.mongodb.org/browse/SERVER-106079?focusedCommentId=7518616&focusedId=7518616&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-7518616] for further explanation
Fixed Versions
9.1 Required
SERVER-15157Open
Controlled primary step-down with no loss of write availability
Stepping down a primary, to allow another replica member to become primary, incurs a brief period where no writes can be accepted. This can lead to application errors when not handled correctly and is a deterrent for some when considering machine maintenance or database upgrade.
Please allow for a [controlled step down|http://docs.mongodb.org/manual/tutorial/force-member-to-be-primary/#force-a-member-to-be-primary-using-database-commands] where there is no loss of write availability.
Components
Replication
Fixed Versions
Needs Further Definition
SERVER-13244Open
Secondary that enters RECOVERY mode does not reset connections
When a node changes state normally (most notably from an election) it disconnects all sockets partly to ensure the change of state is noticed. However, when a Secondary enters RECOVERY mode due to a compact command, it does not close existing sockets.
Clients using secondaryPreferred continue to utilize the node as a normal secondary (operations either get blocked or return a failure due to "not primary or secondary") but new clients do not attempt to issue queries because the invalid state is
Components
Replication
Fixed Versions
Needs Further Definition
SERVER-123169Open
Antithesis: make composer tests work
We have two composer tests tasks to run with antithesis:
* [antithesis_replica_sets_basic_js_commands|https://github.com/10gen/mongo/blob/ffa2a9d8559f4bdbd673f0eed04b046951a40b0d/etc/evergreen_yml_components/tasks/misc_tasks.yml#L368]
* [antithesis_sharding_basic_js_commands|https://github.com/10gen/mongo/blob/ffa2a9d8559f4bdbd673f0eed04b046951a40b0d/etc/evergreen_yml_components/tasks/misc_tasks.yml#L390]
Currently, these tasks are not being run in our daily antithesis runs as they were a
Fixed Versions
9.0.0-rc0
SERVER-17127Open
Make accounting for upserts consistent with oplog entries on Primaries
There are three scenarios for an upsert. On the Primary all three will be reported as an update, but the actual oplog entry will not necessarily reflect this. When those oplog entries are replicated to Secondaries, the correct counters will be increased, but this makes the counts inconsistent with those on the Primary.
The three scenarios are as follows
# The document with a given key does not exist. Insert occurs on the Primary and an "i" record written to oplog. This is still recorded in s
Components
Replication
Affected Versions
2.6.6
Fixed Versions
Needs Further Definition
SERVER-18918Open
Specify replica level storage engine settings
Given that MongoDB has storage engines with options, and these options should greatly increase and be very impactful on performance with btree vs LSM, column vs row, etc, it would be helpful is the default storage engine options were stored in the replica set and by default MongoDB used that. This will provide unexpected performance problems when resyncing etc.
Components
Admin
Fixed Versions
features we're not sure of
SERVER-16087Open
Add support for bulk loading (with WiredTiger)
Since some storage engines, like WiredTiger, have special bulk loading options it would be good if the server supported them.
http://source.wiredtiger.com/2.4.1/tune_bulk_load.html
*Note*:
This may be more complicated in sharding and with replication but in the standalone case it should be straight-forward. Even through that is true, unless support exists for all deployment types it is probably best not to support any.
Components
ReplicationShardingStorage
Fixed Versions
Needs Further Definition
SERVER-106810Open
Oplog Visibility Manager - Creation & Lifetime
The new {{OplogVisibilityManager}} needs to be created & owned somewhere with process lifetime, similar to other services that run on the {{{}ServiceContext{}}}. A pointer to this manager can be placed on the {{OperationContext}} as a decorable, or as a service and a decorable on the parent {{{}ServiceContext{}}}.
Once SERVER-106673 is merged, we can create the pointers/references but defer instantiation until the feature flag is checked - we only want to enable the new behavior when instruc
Fixed Versions
8.3.0-rc0
SERVER-14157Open
rs.status() shell helper reports repllag
The result returned by replSetGetStatus contains enough information to compute a "repllag" value for each member with an optime. Having the server actually perform this computation would avoid the need for it to be done (potentially poorly) client-side.
Presumably the most sensible definition of repllag would be {{latest_optime - optime}}, where {{latest_optime}} is the largest {{optime.getTime()}} value of any replset member (that has an optime), but anything that is self-consistent would b
Components
Replication
Fixed Versions
Needs Further Definition
SERVER-2402Open
Allow for replica set initialization via CLI options and/or config file
It would be very nice if there were a way to tell the mongo server to initialize a replica set automatically after starting up. For automated sysadmin setups (like Chef), it can be very difficult to fully bootstrap several mongo nodes for a replica set. This is due to the fact that you have to wait for the server to come up, then call rs.initialize() in the shell. If there were some way via the CLI options (and config file) to instruct the server to do this automatically, it would make automated
Components
Admin
Fixed Versions
features we're not sure of
SERVER-5917Open
More advanced availability checks required for Arbiter
We have a replica set setup with 2 nodes + 1 arbiter. The SAN storage in our primary's DC went down(ish), resulting in timeouts on I/O operations. Secondary node had detected that the primary was unreachable, but arbiter still marked the primary as available, since the TCP connection to the primary was still active. Failover to the secondary node did not work, causing downtime.
The arbiter should periodically query the different nodes with an I/O operation to detect whether the underlying I/O s
Components
Replication
Affected Versions
2.0.5
Fixed Versions
Needs Further Definition
NODE-7661ClosedFixed
count empty-scope Code as code_with_scope in calculateObjectSize
h3. Expected Behavior
As a... developer
I expect... alignment between the predicted/precalculated size of a BSON object and the size of a BSON object after it is written
When... invoked through the parallel `calculate_size.ts` and `serializer.ts` paths
h3. Actual Behavior and Impact
* _What is the actual behavior?_
** In the specific case of `code_with_scope`, *serializer.ts* and the *calculate_size.ts* have subtly different conditionals before emitting {*}code_with_scope{*}:
*** {*}
Fixed Versions
bson-7.3.4
SERVER-135300ClosedFixed
Record multikey updates in their own side write entries
With SERVER-134912, PDIB piggybacks multikey updates in "i" side writes, rather than having them in their own "m" entries. We should consider making it so that all hybrid builds (PDIB or not) write multikey updates into dedicated "m" entries. This would allow us to avoid much of the complexity of SERVER-134912 where we have to have special handling for the case where an update operation changes multikey while not contributing any new keys.
Fixed Versions
9.1.0-rc09.1.0-rc1025
SERVER-132294ClosedWon't Do
[Join Optimization] Eligibility for join optimization should not depend featureFlagCostBasedRanker
Currently we [consult|https://github.com/10gen/mongo/blob/621283435ecf1259e26cfb3352a92572bc8a9c4a/src/mongo/db/query/compiler/optimizer/join/executor.cpp#L138-L143] featureFlagCostBasedRanker as part of join eligibility checks. The team discussed that we the join flag implies CBR so we should avoid the intertwining of these flags (don't bail out of join is the CBR feature flag is disabled).
SERVER-135403ClosedGone away
Complete TODO listed in SERVER-134058
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-134058&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-134260ClosedGone away
M4b: Export snapshots from mciuploads to provisioned S3 bucket
To work around the time-consuming S3 provisioning process, we'll temporarily store snapshots in the general-purpose S3 bucket {{{}mciuploads{}}}. Once the correct bucket is provisioned, we'll point the Evergreen tasks to it and move the already generated snapshots over to the permanent bucket.
SERVER-111199ClosedFixed
Remove noisy startup warnings about missing ethtool stats
When I start mongod I see the following warnings:
We should remove these if they are not actionable
Fixed Versions
9.1.0-rc09.1.0-rc1024
SERVER-79298ClosedFixed
FCBIS should succeed in unhealthy set
File Copy Based Initial Sync requires the majority point of the set to advance in order to start an extension round. This means if the set is unhealthy the sync will not succeed. We should fix this, either by
1) If extension fails before we've copied anything, logging an error and continuing with the sync. This may leave the node more lagged than desired, but it will catch up with normal replication (and perhaps make the set healthy again, if this was a re-sync)
and/or
2) Remove the
Components
Replication
Fixed Versions
8.1.0-rc08.0.207.0.32
SERVER-138711ClosedFixed
Make test multiversion-safe
Fixes test flakiness in a new multiversion test that was introduced via SERVER-138293 and caused BF-47119.
The test starts a sharded cluster on mixed versions, which led to the old versions starting up with an unexpected 8.0 FCV. This made the automatic FCV upgrade race with the race of the test, so it could fail on slower build variants.
This change also refactors the test so it reuses the same {{ShardingTest}} fixture across multiple test cases, so that overall test runtime is greatly reduce
Components
Build
Fixed Versions
9.1.0-rc09.1.0-rc1025