[
https://issues.apache.org/jira/browse/KAFKA-15366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17793355#comment-17793355
]
Igor Soarez commented on KAFKA-15366:
-------------------------------------
[~viktorsomogyi] thanks for helping with this! Yes, thats seems to me like a
pretty good place to start. If that test supports KRaft the main cases should
be covered.
It would be nice to also have tests covering:
* admin.alterReplicaLogDirs() moving a replica between log directories in the
same broker, with the source/destination directory failing
* dir failure during zk->kraft migration
* moving partitions across directories while a broker is offline, checking
upon restart, the partition isn't recreated, but rather the assignment is
corrected in the cluster metadata
* using alterReplicaLogDirs to set preferred log dir (see
LogManager.maybeUpdatePreferredLogDir) and action it with a following
reassignmentÂ
There might be more, this is what I can think of now. And we can look into
these later, the main thing to test is leadership failover as currently covered
in LogDirFailureTest.
> Log directory failure integration test
> --------------------------------------
>
> Key: KAFKA-15366
> URL: https://issues.apache.org/jira/browse/KAFKA-15366
> Project: Kafka
> Issue Type: Sub-task
> Reporter: Igor Soarez
> Assignee: Viktor Somogyi-Vass
> Priority: Major
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
