hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pankaj Kumar (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-14499) Master coprocessors shutdown will not happen on master abort
Date Sat, 03 Oct 2015 03:42:26 GMT

    [ https://issues.apache.org/jira/browse/HBASE-14499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14942084#comment-14942084
] 

Pankaj Kumar commented on HBASE-14499:
--------------------------------------

Thanks [~tedyu] for reviewing the patch. Yeah locally those test cases are successful. 

> Master coprocessors shutdown will not happen on master abort
> ------------------------------------------------------------
>
>                 Key: HBASE-14499
>                 URL: https://issues.apache.org/jira/browse/HBASE-14499
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>            Reporter: Pankaj Kumar
>            Assignee: Pankaj Kumar
>             Fix For: 2.0.0
>
>         Attachments: HBASE-14499.patch, HBASE-14499.patch
>
>
> In HMaster,
> {code}
> @Override
>   public void abort(final String msg, final Throwable t) {
>     if (isAborted() || isStopped()) {
>       return;
>     }
>     if (cpHost != null) {
>       // HBASE-4014: dump a list of loaded coprocessors.
>       LOG.fatal("Master server abort: loaded coprocessors are: " +
>           getLoadedCoprocessors());
>     }
>     if (t != null) LOG.fatal(msg, t);
>     stop(msg);
>   }
> {code}
> Here we are invoking stop(...) of HRegionServer, which will try to stop RS coprocessors
if rsHost is not NULL.
> So Master coprocessors will not be stopped. We should invoke stopMaster() instead of
stop(...).



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

Mime
View raw message