hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Fengdong Yu (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-4631) Support customized call back method during failover automatically.
Date Fri, 12 Apr 2013 02:01:16 GMT

    [ https://issues.apache.org/jira/browse/HDFS-4631?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13629689#comment-13629689

Fengdong Yu commented on HDFS-4631:


I do think we need this patch for customized callback during state changed.(maybe do failover
during state change, maybe not)

but in the prod cluster, we still need get informed when there is health state changed.
> Support customized call back method during failover automatically.
> ------------------------------------------------------------------
>                 Key: HDFS-4631
>                 URL: https://issues.apache.org/jira/browse/HDFS-4631
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: ha
>    Affects Versions: 3.0.0
>            Reporter: Fengdong Yu
>              Labels: features, ha, hadoop
>             Fix For: 3.0.0
>         Attachments: HDFS-4631.patch
>   Original Estimate: 0.5m
>  Remaining Estimate: 0.5m
> ZKFC add HealthCallbacks by default, this can do quiteElection at least. but we often
want to be alerted if there is fail over occurring(such as send email, short messages), especially
for prod cluster.
> There is a configured fence script. maybe we can put all these logics in the script.
but in reasonable, fence script does only one thing: fence :)
> So I added this patch, we can configure customized HM callback method, if there is no
configuration, then only HealthCallbacks is added.

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

View raw message