You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "Philip Nee (Jira)" <ji...@apache.org> on 2022/04/04 18:41:00 UTC

[jira] [Assigned] (KAFKA-13668) Failed cluster authorization should not be fatal for producer

     [ https://issues.apache.org/jira/browse/KAFKA-13668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Philip Nee reassigned KAFKA-13668:
----------------------------------

    Assignee: Philip Nee  (was: Kirk True)

> Failed cluster authorization should not be fatal for producer
> -------------------------------------------------------------
>
>                 Key: KAFKA-13668
>                 URL: https://issues.apache.org/jira/browse/KAFKA-13668
>             Project: Kafka
>          Issue Type: Bug
>            Reporter: Jason Gustafson
>            Assignee: Philip Nee
>            Priority: Major
>
> The idempotent producer fails fatally if the initial `InitProducerId` returns CLUSTER_AUTHORIZATION_FAILED. This makes the producer unusable until a new instance is constructed. For some applications, it is more convenient to keep the producer instance active and let the administrator fix the permission problem instead of going into a crash loop. Additionally, most applications will probably not be smart enough to reconstruct the producer instance, so if the application does not handle the error by failing, users will have to restart the application manually. 
> I think it would be better to let the producer retry the `InitProducerId` request as long as the user keeps trying to use the producer. 



--
This message was sent by Atlassian Jira
(v8.20.1#820001)