hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-3113) Don't reassign regions if cluster is being shutdown
Date Fri, 15 Oct 2010 17:58:32 GMT

    [ https://issues.apache.org/jira/browse/HBASE-3113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12921472#action_12921472

stack commented on HBASE-3113:

This patch needs a part two.  While we don't assign any more if cluster is going down, I saw
that #1552 build up on hudson failed because load balancer gave a region to a server that
we'd asked to shutdown.  The RS proceeded to open the region and since it had come in under
the wire, the RS stayed up holding up test completion.

A RS that has been asked to stop -- by user via HBaseAdmin or in a unit test testing balancing
-- should reject requests to open regions.  Master should then catch the rejection exception
and reassign the region elsewhere.

> Don't reassign regions if cluster is being shutdown
> ---------------------------------------------------
>                 Key: HBASE-3113
>                 URL: https://issues.apache.org/jira/browse/HBASE-3113
>             Project: HBase
>          Issue Type: Bug
>            Reporter: stack
>            Assignee: stack
>             Fix For: 0.90.0
>         Attachments: 3113.txt
> On stop of a cluster, handling a close message, we'll go ahead and reassign regions as
per normal though the cluster up flag is false.  This is what cause the TestRegionRebalancing
test to fail up on hudson just now, #1546.

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

View raw message