hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jean-Daniel Cryans (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1601) Region server going down as a result of expiration by ZK
Date Fri, 03 Jul 2009 18:15:47 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1601?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12727075#action_12727075
] 

Jean-Daniel Cryans commented on HBASE-1601:
-------------------------------------------

Amandeep,

Did you consult this page? http://wiki.apache.org/hadoop/PerformanceTuning
Did you use a ZK quorum during your upload?
What else can you tell us about your setup?

FYI, one reason a RS can get expired by ZK is because its JVM was stalled by something big
like GC. See HBASE-1316.

> Region server going down as a result of expiration by ZK
> --------------------------------------------------------
>
>                 Key: HBASE-1601
>                 URL: https://issues.apache.org/jira/browse/HBASE-1601
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0
>            Reporter: Amandeep Khurana
>             Fix For: 0.20.0
>
>         Attachments: master.log, regionserver2.log, regionserver_on_zk_failure_node.log,
web_ui.log, web_ui2.log, web_ui3.log, zk.log
>
>
> During an MR job, ZK expires one of the nodes. This results in losing contact with a
region server and eventually the map task timing out and the job failing.
> Attaching the following logs:
> 1. HBase master logs
> 2. Logs from the web UI
> 3. Logs from the particular RS that went down.
> 4. Logs from the particular node that ZK expired.

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


Mime
View raw message