[ 
https://issues.apache.org/jira/browse/KAFKA-13406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17435239#comment-17435239
 ] 

Luke Chen commented on KAFKA-13406:
-----------------------------------

[~Andy_Dufresne], I understand, but I'm afraid we have to fix it in the 
`internals/ConsumerCoordinator` file, because this issue is happened after the 
assignor completed its job. And we had already fixed the issue in assignor in  
KAFKA-12984 , and the `internals/ConsumerCoordinator` is what we missed before. 
We have to make sure when this issue (out-of-date `ownedPartition`) happened, 
there's no other places missing and cause the rebalance stuck issue happen 
again.

I created another PR based on V2.8: 
[https://github.com/apache/kafka/pull/11444] . I think it should be easier for 
you if you are running on v2.8.1. Or, please let me know how I can help you 
with it.

Thanks.

 

[~ableegoldman], I agree with you for fixing it in V3.1.0, and have another 
ticket for long-term fix.

> Cooperative sticky assignor got stuck due to assignment validation failed
> -------------------------------------------------------------------------
>
>                 Key: KAFKA-13406
>                 URL: https://issues.apache.org/jira/browse/KAFKA-13406
>             Project: Kafka
>          Issue Type: Bug
>          Components: clients
>    Affects Versions: 3.0.0
>            Reporter: Luke Chen
>            Assignee: Luke Chen
>            Priority: Major
>             Fix For: 3.1.0, 3.0.1
>
>
> We'll do validateCooperativeAssignment for cooperative assignor, where we 
> validate if there are previously owned partitions directly transfer to other 
> consumers without "revoke" step. However, the "ownedPartition" in 
> subscription might contain out-of-dated data, which might cause the 
> validation always failure.
> We should consider the short-term fix it by disabling 
> validateCooperationAssignment for built-in cooperativeStickyAssignor because 
> we've already consider the generation in the assignor, and discard the old 
> generation ownedPartition if any.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to