You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "nijel (JIRA)" <ji...@apache.org> on 2015/09/29 06:12:04 UTC

[jira] [Created] (HBASE-14504) Improve the master exception handling in Region server using zookeeper watcher

nijel created HBASE-14504:
-----------------------------

             Summary: Improve the master exception handling in Region server using zookeeper watcher
                 Key: HBASE-14504
                 URL: https://issues.apache.org/jira/browse/HBASE-14504
             Project: HBase
          Issue Type: Improvement
            Reporter: nijel
            Assignee: Pankaj Kumar


As per the current design, region server will fail over only if the connection got some exception.
As described in HBASE-14498, if there is any issue in master instance related to zookeeper connection, the standBy master will become active. But the region servers will not fail over
This particular issue can be fixed in master side.

But as a generic solution, i think region server needs to depend on zookeeper watch to identify the master change.
Assigning to [~pankaj2461] since this work is related to HBASE-14498

Any thoughts ? 





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