[
https://issues.apache.org/jira/browse/HBASE-17580?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16227730#comment-16227730
]
Andrew Purtell commented on HBASE-17580:
----------------------------------------
Hey [~stack] I'm looking at handling this automatically on HBASE-19144.
And maybe more generally, we can take a look at kicking regions in FAILED_OPEN
state whenever a server joins the cluster (with hysteresis so if a few join in
a short while we don't pile all regions on the first one to show up)?
> Tooling to move table out of FAILED_OPEN state
> ----------------------------------------------
>
> Key: HBASE-17580
> URL: https://issues.apache.org/jira/browse/HBASE-17580
> Project: HBase
> Issue Type: Sub-task
> Components: Region Assignment
> Affects Versions: 2.0.0
> Reporter: stack
> Priority: Critical
> Fix For: 2.0.0
>
>
> Playing with rsgroup feature, I was able to manufacture a state where
> assignment landed regions into a FAILED_OPEN state; i.e. assignment could not
> complete (in this case because no server for the region to land on inside its
> rsgroup). Once in this FAILED_OPEN state, the only way currently (seemingly)
> to move it on from this state is to manually assign the region. If table is
> large, this is burdensome. FAILED_OPEN seems a legitmate endpoint for AM to
> dump regions into if it can't do anything with them. We need tooling for
> 'fix'. One thought is that assign take a tableName.
> Disable/Enable does not work.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
