[jira] Updated: (HADOOP-4060) [HOD] Make HOD to roll log files on the client

["Vinod Kumar Vavilapalli (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HADOOP-4060?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Vinod Kumar Vavilapalli updated HADOOP-4060:
--------------------------------------------

    Attachment: HOD-LOG-ROTATION.txt

Attaching a patch. This introduces a config parameter hod.log-rollover-count that takes the number of rolled-over logs that HOD client has to keep. It has to be a positive integer, defaults to 5. A zero value disables the roll-over and the log grows indefinitely (current behaviour). Rolled over logs have file-names of the form hod-%user.log.1, hod-%user.log.2 etc.

> [HOD] Make HOD to roll log files on the client 
> -----------------------------------------------
>
>                 Key: HADOOP-4060
>                 URL: https://issues.apache.org/jira/browse/HADOOP-4060
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: contrib/hod
>            Reporter: Hemanth Yamijala
>            Assignee: Vinod Kumar Vavilapalli
>            Priority: Blocker
>             Fix For: 0.18.1, 0.19.0
>
>         Attachments: HOD-LOG-ROTATION.txt
>
>
> Currently HOD writes a log file on the client in the cluster directory, named hod-<username>.log. This file is appended to for each run of hod allocation that runs on the same cluster directory. Thus, the file can become quite large - particularly if the job is queued for a long time. If there are problems with the HOD client subsequently, this could result in stale file handles pointing to this large file, and cause disk space to fill up. Another problem is that a large log file is usually unusable for a user.

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