[jira] Updated: (HBASE-3337) Restore HBCK fix of unassignment and dupe assignment for new master

["Jonathan Gray (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-3337?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Gray updated HBASE-3337:
---------------------------------

    Attachment: HBASE-3337-v2.patch

This is the patch from over in HBASE-3332 plus an additional fix to modernize disabled table handling in HBCK.  Currently it considers a table disabled if it's in DISABLED state.  It ignores other states like DISABLING (for now), not really sure if we should do something there but perhaps for HBCK, disabled and disabling should both be treated as equivalent (and thus, region should not be deployed).

> Restore HBCK fix of unassignment and dupe assignment for new master
> -------------------------------------------------------------------
>
>                 Key: HBASE-3337
>                 URL: https://issues.apache.org/jira/browse/HBASE-3337
>             Project: HBase
>          Issue Type: Bug
>          Components: master, util
>            Reporter: Jonathan Gray
>            Assignee: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0, 0.92.0
>
>         Attachments: HBASE-3337-v2.patch
>
>
> HBCK fixing of region unassignment and duplicate assignment was broken with the move to the new master.
> We've seen it happen doing testing of 0.90 RCs like that over in HBASE-3332.
> Rather than the old "clear everything out approach" which relied on the BaseScanner, in the new master we should just manipulate unassigned ZK nodes and let the master handle the transition.

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