[
https://issues.apache.org/jira/browse/HBASE-24759?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17163396#comment-17163396
]
Pankaj Kumar commented on HBASE-24759:
--------------------------------------
Ok, got it..
Recently we met a blocking issue (in HBase-2.2.3) where all system tables were
assigned to default rsgroup and due to some reason all RSs (only 1 RS is in
default rsgroup) of default rsgroup went down. Hmaster stuck while persisting
default rsgroup info to rsgroup table (since no RS in deafult rsgroup, so meta
and other system region were not assigned). Even new RS join the default
rsgroup, AM couldn't pick them as candidate since rsGroupMap will be updated
after persisting default rsgroup to rsgroup table. HBASE-22820 addressed the
problem.
> Persisting configuration of default rsgroup
> -------------------------------------------
>
> Key: HBASE-24759
> URL: https://issues.apache.org/jira/browse/HBASE-24759
> Project: HBase
> Issue Type: Bug
> Components: rsgroup
> Affects Versions: 3.0.0-alpha-1
> Reporter: Sun Xin
> Assignee: Sun Xin
> Priority: Major
> Fix For: 3.0.0-alpha-1
>
>
> In the current scenario, we didn't store the default rsgroup information. But
> after HBASE-24431 , we have added a config map, which need to be persisted to
> avoid lossing config of default rsgroup.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
