hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benjamin Francisoud (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HADOOP-2517) mapred/local/jobTracker is not cleanup correctly
Date Wed, 09 Jan 2008 15:36:34 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-2517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Benjamin Francisoud updated HADOOP-2517:

    Attachment: patch-JobTracker-cleanup-submission-dir-0.14.0.diff

JobTracker.java patch to delete those files after a some time (configurable using jobtracker.cleanup.submission.dir.delay
if no value provided: default is 24 hours

You can remove some/all the LOG.info 

> mapred/local/jobTracker is not cleanup correctly
> ------------------------------------------------
>                 Key: HADOOP-2517
>                 URL: https://issues.apache.org/jira/browse/HADOOP-2517
>             Project: Hadoop
>          Issue Type: Bug
>          Components: mapred
>    Affects Versions: 0.14.0
>            Reporter: Benjamin Francisoud
>         Attachments: patch-JobTracker-cleanup-submission-dir-0.14.0.diff
> Jobtracker store 1.jar file and 1 .xml per job in /srv/hadoop/var/mapred/local/jobTracker.
> They seems to be correctly delete in most cases BUT not when a job is launch with 
> Map total: 0 and Reduce total: 0
> This happen when the job doesn't find any input file to process.
> After few months, lots of disk space is lost because of that.
> This issue is different from HADOOP-2427 since it's on the jobtracker, not on the nodes,
but it may be related in some way in the code (my 2 cents).

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

View raw message