Mailing-List: contact hbase-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hbase-dev@hadoop.apache.org
Message-ID: <1077394669.1240299587616.JavaMail.jira@brutus>
Date: Tue, 21 Apr 2009 00:39:47 -0700 (PDT)
From: "ryan rawson (JIRA)" <jira@apache.org>
To: hbase-dev@hadoop.apache.org
Subject: [jira] Updated: (HBASE-1332) regionserver carrying .META. starts
 sucking all cpu, drives load up - infinite loop?
In-Reply-To: <1920910069.1240299467423.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/HBASE-1332?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

ryan rawson updated HBASE-1332:
-------------------------------

    Attachment: HBASE-1332.patch

It turns out that if the .META. has a hfile that doesnt start at ',,' (aka: first row in the table), we end up with an infinite loop.  As you can see, in this function, we 'continue' and never finish this function.

> regionserver carrying .META. starts sucking all cpu, drives load up - infinite loop?
> ------------------------------------------------------------------------------------
>
>                 Key: HBASE-1332
>                 URL: https://issues.apache.org/jira/browse/HBASE-1332
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0
>            Reporter: ryan rawson
>            Assignee: ryan rawson
>             Fix For: 0.20.0
>
>         Attachments: HBASE-1332.patch
>
>
> the symptom is the cluster comes to a dead halt.  Lookups on meta don't seem to work, and the regionserver carrying .META. goes hot - using 800% CPU or more, driving system LA up really really high (I've seen it as high as 26).  Thread dumps seem to indicate every IPC handler is stuck in Bytes.binarySearch().

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