[jira] [Updated] (KAFKA-840) Controller tries to perform preferred replica election on failover before state machines have started up

["Swapnil Ghike (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/KAFKA-840?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Swapnil Ghike updated KAFKA-840:
--------------------------------

    Attachment: kafka-840.patch

I think the right solution is to move   initializeAndMaybeTriggerPartitionReassignment() and initializeAndMaybeTriggerPreferredReplicaElection() after the controller mbean is registered and the controller logs that it is ready to serve. Attached a patch.
                
> Controller tries to perform preferred replica election on failover before state machines have started up
> --------------------------------------------------------------------------------------------------------
>
>                 Key: KAFKA-840
>                 URL: https://issues.apache.org/jira/browse/KAFKA-840
>             Project: Kafka
>          Issue Type: Bug
>    Affects Versions: 0.8
>            Reporter: Swapnil Ghike
>            Assignee: Swapnil Ghike
>            Priority: Blocker
>              Labels: bugs
>             Fix For: 0.8
>
>         Attachments: kafka-840.patch
>
>
> If the admin/preferred_replica_election path is non-empty when a new controller starts, the controller attempts to perform preferred replica election before the partition and replica state machine have been initialized. In this case, the controller will try to make invalid state transitions. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira