[jira] Updated: (HBASE-1667) hbase-daemon.sh stop master should only stop the master, not the cluster

["Jean-Daniel Cryans (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jean-Daniel Cryans updated HBASE-1667:
--------------------------------------

    Summary: hbase-daemon.sh stop master should only stop the master, not the cluster  (was: hbase-daemon.sh stop master for standby masters should not stop the whole cluster)

Changing the title to reflect a broader scope, stop master should only stop the master, not the cluster.

> hbase-daemon.sh stop master should only stop the master, not the cluster
> ------------------------------------------------------------------------
>
>                 Key: HBASE-1667
>                 URL: https://issues.apache.org/jira/browse/HBASE-1667
>             Project: Hadoop HBase
>          Issue Type: Improvement
>          Components: master, scripts
>    Affects Versions: 0.20.0
>            Reporter: Rong-En Fan
>             Fix For: 0.21.0
>
>
> 0.20 supports multi masters. However,
> bin/hbase-daemon.sh stop master
> on backup masters will bring the whole cluster down.
> Per rolling upgrade wiki that stack pointed out, kill -9 for backup master is the only way to go currently.
> I think it's better to make some sort of magic that we can use something like
> bin/hbase-daemon.sh stop master
> to properly stop either the backup master or the whole cluster.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.