[
https://issues.apache.org/jira/browse/KAFKA-6834?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16458810#comment-16458810
]
Colin P. McCabe commented on KAFKA-6834:
----------------------------------------
Good catch!
Maybe we should validate the CRC of the overlarge message batch before
enlarging the buffer, just to make sure we're not allocating memory based on
corrupt data.
> log cleaner should handle the case when the size of a message set is larger
> than the max message size
> -----------------------------------------------------------------------------------------------------
>
> Key: KAFKA-6834
> URL: https://issues.apache.org/jira/browse/KAFKA-6834
> Project: Kafka
> Issue Type: Bug
> Reporter: Jun Rao
> Assignee: Rajini Sivaram
> Priority: Major
>
> In KAFKA-5316, we added the logic to allow a message (set) larger than the
> per topic message size to be written to the log during log cleaning. However,
> the buffer size in the log cleaner is still bounded by the per topic message
> size. This can cause the log cleaner to die and cause the broker to run out
> of disk space.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
