[
https://issues.apache.org/jira/browse/KAFKA-3083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15098607#comment-15098607
]
Mayuresh Gharat edited comment on KAFKA-3083 at 1/14/16 6:34 PM:
-----------------------------------------------------------------
That's a very good point, I will verify if this can happen.
Moreover, I think the behavior should be :
1) Broker A was the controller.
2) Broker A faces a session expiration, invokes the controllerResignation and
clears all its caches and also stops all the ongoing controller work.
3) Broker B becomes the controller and proceeds.
what do you think?
was (Author: mgharat):
That's a very good point, I will verify if this can happen.
Moreover, I think the behavior should be :
1) Broker A was the controller.
2) Broker A faces a session expiration, invokes the controllerResignation and
clears all its caches and also stops all the ongoing controller work.
3) Broker B becomes the controller and proceeds.
> a soft failure in controller may leader a topic partition in an inconsistent
> state
> ----------------------------------------------------------------------------------
>
> Key: KAFKA-3083
> URL: https://issues.apache.org/jira/browse/KAFKA-3083
> Project: Kafka
> Issue Type: Bug
> Components: core
> Affects Versions: 0.9.0.0
> Reporter: Jun Rao
> Assignee: Mayuresh Gharat
>
> The following sequence can happen.
> 1. Broker A is the controller and is in the middle of processing a broker
> change event. As part of this process, let's say it's about to shrink the isr
> of a partition.
> 2. Then broker A's session expires and broker B takes over as the new
> controller. Broker B sends the initial leaderAndIsr request to all brokers.
> 3. Broker A continues by shrinking the isr of the partition in ZK and sends
> the new leaderAndIsr request to the broker (say C) that leads the partition.
> Broker C will reject this leaderAndIsr since the request comes from a
> controller with an older epoch. Now we could be in a situation that Broker C
> thinks the isr has all replicas, but the isr stored in ZK is different.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
