[ 
https://issues.apache.org/jira/browse/HBASE-22537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16862286#comment-16862286
 ] 

Wellington Chevreuil commented on HBASE-22537:
----------------------------------------------

Hi [~sreenivasulureddy], are you able to consistently reproduce this problem? 
Tried to go through the described steps, but couldn't manage to reproduce it. 
Also, in the proposed patch, I think the given change would cause secondary 
replicas to never be removed from Master's cache, which I don't think is 
desirable. Is your added test constantly passing within this change, and 
failing if you run it against original code version? Unfortunately, I had not 
been able to run any of these TestAsyncNonMetaRegionLocator tests, locally, 
keeps giving me timeouts.

> Split happened Replica region can not be deleted after deleting table 
> successfully and restarting RegionServer
> --------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-22537
>                 URL: https://issues.apache.org/jira/browse/HBASE-22537
>             Project: HBase
>          Issue Type: Bug
>          Components: Region Assignment
>    Affects Versions: 2.1.1
>            Reporter: Y. SREENIVASULU REDDY
>            Priority: Minor
>             Fix For: 2.1.6
>
>         Attachments: HBASE-22537.branch-2.1.002.patch, 
> HBASE-22537.branch-2.1.patch
>
>
> [Test step]
> 1.create a table (set RegionReplication=2).
> 2.insert data to the table utill region be splitted.
> 3.Disable and Drop the table.
> 4.Parent replica region holding Regionserver, Kill forcefully 
> 5.HBase WebUI will show that the replica regions will be in RIT.
> [Expect Output]
> Parent replica region should be deleted.
> [Actual Output]
> Parent replica region still exists.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to