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

Posted to dev@kafka.apache.org by "Alexey Ozeritskiy (JIRA)" <ji...@apache.org> on 2016/12/23 12:29:58 UTC

[jira] [Commented] (KAFKA-1923) Negative offsets in replication-offset-checkpoint file

    [ https://issues.apache.org/jira/browse/KAFKA-1923?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15772787#comment-15772787 ] 

Alexey Ozeritskiy commented on KAFKA-1923:
------------------------------------------

Problem stil exists (v0.10.1.1). Trying to reproduce it. I've read the code and found that the only place with negative offsets is Replica.logEndOffset (for remote Replica), but I dont understand how it can get into highwatermark.

> Negative offsets in replication-offset-checkpoint file
> ------------------------------------------------------
>
>                 Key: KAFKA-1923
>                 URL: https://issues.apache.org/jira/browse/KAFKA-1923
>             Project: Kafka
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 0.8.2.0
>            Reporter: Oleg Golovin
>
> Today was the second time we witnessed negative offsets in replication-offset-checkpoint file. After restart the node stops replicating some of its partitions.
> Unfortunately we can't reproduce it yet. But the two cases we encountered indicate a bug which should be addressed.



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