You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@kafka.apache.org by "Jun Rao (JIRA)" <ji...@apache.org> on 2013/02/20 19:19:12 UTC

[jira] [Commented] (KAFKA-765) Corrupted messages in produce request could shutdown the broker

    [ https://issues.apache.org/jira/browse/KAFKA-765?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13582396#comment-13582396 ] 

Jun Rao commented on KAFKA-765:
-------------------------------

One way to fix this is to only convert IOException from segment.append and maybeFlush to KafkaStorageException.
                
> Corrupted messages in produce request could shutdown the broker
> ---------------------------------------------------------------
>
>                 Key: KAFKA-765
>                 URL: https://issues.apache.org/jira/browse/KAFKA-765
>             Project: Kafka
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 0.8
>            Reporter: Jun Rao
>            Priority: Blocker
>
> In kafka.log.append(), we convert all IOException to KafkaStorageException. This will cause the caller to shutdown the broker. However, if there is a corrupted compressed message, validMessages.assignOffsets() in append() could also throw an IOException when decompressing the message. In this case, we shouldn't shut down the broker and should just fail this particular produce request.

--
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