[jira] [Commented] (KAFKA-3879) KafkaConsumer with auto commit enabled gets stuck when killed after broker is dead

["Ashish K Singh (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/KAFKA-3879?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15340283#comment-15340283 ] 

Ashish K Singh commented on KAFKA-3879:
---------------------------------------

[~hachikuji] did you guys ever hit this? I am thinking of adding timeout to {{ConsumerCoordinator.maybeAutoCommitOffsetsSync}}. Wondering if we need a mechanism to make a retriable error, non-retriable after certain number of failures.

> KafkaConsumer with auto commit enabled gets stuck when killed after broker is dead
> ----------------------------------------------------------------------------------
>
>                 Key: KAFKA-3879
>                 URL: https://issues.apache.org/jira/browse/KAFKA-3879
>             Project: Kafka
>          Issue Type: Bug
>    Affects Versions: 0.10.0.0
>            Reporter: Ashish K Singh
>            Assignee: Ashish K Singh
>             Fix For: 0.10.0.1
>
>
> KafkaConsumer with auto commit enabled gets stuck when killed after broker is dead.
> * KafkaConsumer on close tries to close coordinator.
> * Coordinator, if auto commit is enabled, tries to commit offsets synchronously before closing.
> * While trying to synchronously commit offsets, coordinator checks if coordinator is alive by sending {{GroupCoordinatorRequest}}. As brokers are dead, this returns {{NoAvailableBrokersException}}, which is a retriable exception.
> * Coordinator ready check enters into an infinite loop as it keeps retrying to discover group coordinator.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)