[
https://issues.apache.org/jira/browse/HBASE-13559?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14512029#comment-14512029
]
Elliott Clark commented on HBASE-13559:
---------------------------------------
bq.These are supposed to get cleaned up... we going a code path not considered
previously?
Not completely sure. This was a a freak accident on a production cluster. I'll
try and reproduce on a test cluster.
> Region stuck in PENDING_OPEN on repeated master failure
> -------------------------------------------------------
>
> Key: HBASE-13559
> URL: https://issues.apache.org/jira/browse/HBASE-13559
> Project: HBase
> Issue Type: Bug
> Reporter: Elliott Clark
>
> We had some regions stay un-assigned. They were in pending open on a server
> that was dead. The cluster was restarted as a whole a few times and then the
> regions were stuck until a manual assign was performed on them.
> I know timing things out has been a pain. Could we have a timeout on servers
> that are not alive?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
