hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gopal V (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-564) Job history logs do not log anything when JVM fails to start
Date Wed, 17 Apr 2013 08:09:16 GMT

    [ https://issues.apache.org/jira/browse/YARN-564?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13633884#comment-13633884

Gopal V commented on YARN-564:

[~shekhar.kotekar]: Add a non-sensical option to mapred.map.child.java.opts in hadoop branch-2
(mapred-site.xml or in command line args), then the error indication does not show the command
line args in YARN or anything beyond KILLED as the status.

I would suggest picking up YARN-307 first (because it would solve 50% of this problem) - I
debugged this issue using debug.delay options, which works for dev deployments, but not practical
for prod.
> Job history logs do not log anything when JVM fails to start
> ------------------------------------------------------------
>                 Key: YARN-564
>                 URL: https://issues.apache.org/jira/browse/YARN-564
>             Project: Hadoop YARN
>          Issue Type: Bug
>    Affects Versions: 3.0.0
>         Environment: Ubuntu 64bit 
>            Reporter: Gopal V
>            Priority: Minor
>              Labels: usability
>         Attachments: yarn-error.png
> If the -Xmx line in a java invocation has errors such as is possible with quoting issues
as would be possible with hive
> hive> set mapred.map.child.java.opts="-server -Xmx2248m -Djava.net.preferIPv4Stack=true";
> The diagnostic error message says
> {code}
> FAILED: Execution Error, return code 2 from org.apache.hadoop.hive.ql.exec.MapRedTask
> {code}
> The job history UI is rather unhelpful and the UI says 
> {code}
> Attmpt state missing from History : marked as KILLED
> {code}
> And the log files are not available and instead show
> {code}
> Logs not available for attempt_1365673149565_0002_m_000000_3. Aggregation may not be
complete, Check back later or try the nodemanager at :-1
> {code}

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message