[
https://issues.apache.org/jira/browse/KAFKA-3413?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15197861#comment-15197861
]
Gwen Shapira commented on KAFKA-3413:
-------------------------------------
This is an error due to the severe implications - on a replica, this will cause
a re-replication of an entire partition. Sysadmins will want to know that their
network is about to get really really busy, so this is logged at a level where
they are likely to catch it.
If you see it frequently (to the point where log level matters), you must be
hitting a pretty serious issue and we should discuss that :)
> Load Error Message should be a Warning
> --------------------------------------
>
> Key: KAFKA-3413
> URL: https://issues.apache.org/jira/browse/KAFKA-3413
> Project: Kafka
> Issue Type: Improvement
> Components: consumer, offset manager
> Affects Versions: 0.9.0.0, 0.9.0.1
> Reporter: Scott Reynolds
> Assignee: Neha Narkhede
>
> There is a Error message from AbstractReplicaFetcherThread that isn't really
> an error.
> Each implementation
> of this thread can logs out when an error or fatal error occurs.
> ReplicaFetcherThread, has both warn, error and fatal in the
> handleOffsetOutOfRange method.
> ConsumerFetcherThread seems to reset itself without logging out an error.
> Seems that the Reset message isn't shouldn't be an error level as it
> doesn't indicate any real error.
> This patch makes it a warning:
> https://github.com/apache/kafka/compare/trunk...SupermanScott:offset-reset-to-warn?diff=split&name=offset-reset-to-warn
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
