Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Message-ID: <17073863.32951276710986025.JavaMail.jira@thor>
Date: Wed, 16 Jun 2010 13:56:26 -0400 (EDT)
From: "Andrei Dragomir (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Subject: [jira] Commented: (HBASE-2736) Disable table does not work anymore
In-Reply-To: <28872228.29981276697843717.JavaMail.jira@thor>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-2736?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12879419#action_12879419 ] 

Andrei Dragomir commented on HBASE-2736:
----------------------------------------

I don't know exactly what the logic is, but debugging the Master process yields the following (I looked where I expect the problem is):

1. In ChangeTableState::postProcessMeta, there is an iteration on this.unservedRegions, which has size 0. This loop seems to persist the changes to the meta table that say the region is offline
2. However, this.unservedRegions has size 0, and this.servedRegions has size 1. The table has only 1 region, so I suspect the issue is that somehow, the region should be in unservedRegions, but it isn't.

> Disable table does not work anymore
> -----------------------------------
>
>                 Key: HBASE-2736
>                 URL: https://issues.apache.org/jira/browse/HBASE-2736
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.20.5, 0.20.6
>            Reporter: Andrei Dragomir
>
> When trying to disable a table, the shell exhausts the retries, then throws the following: 
> {noformat}
> ERROR: org.apache.hadoop.hbase.RegionException: Retries exhausted, it took too long to wait for the table t1 to be disabled.
> {noformat}
> in HBaseAdmin.java:disableTable, after disabling, the code checks again, using isTableDisabled, which always returns false. 
> Also, after running a disable, in Zookeeper, in "/hbase/UNASSIGNED", I get an entry, which is exactly the name of the only region of the table I am trying to disable. This leads me to believe this bug is somehow linked to HBASE-2694.
> This reproduces on my machine, our cluster, with freshly installed HBase, with new data, as well as with old one.

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