You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Arun C Murthy (JIRA)" <ji...@apache.org> on 2007/05/04 10:24:15 UTC
[jira] Updated: (HADOOP-1276) TaskTracker expiry interval is not
configurable
[ https://issues.apache.org/jira/browse/HADOOP-1276?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1276:
----------------------------------
Attachment: HADOOP-1276_20070504_1.patch
Here is a straight-forward patch...
Personally I find this feature quite useful, especially when dealing with issues relating to lost trackers... I'm not very sure about the original constraints which led to this being hard-coded and to acquiesce to the same.
> TaskTracker expiry interval is not configurable
> -----------------------------------------------
>
> Key: HADOOP-1276
> URL: https://issues.apache.org/jira/browse/HADOOP-1276
> Project: Hadoop
> Issue Type: Improvement
> Components: mapred
> Environment: all
> Reporter: Alejandro Abdelnur
> Assigned To: Arun C Murthy
> Fix For: 0.13.0
>
> Attachments: HADOOP-1276_20070504_1.patch
>
>
> The tasktracker expiry interval is hardcoded to 10 mins with the MRConstants.TASKTRACKER_EXPIRY_INTERVAL constant.
> For small clusters, running small jobs, this interval is too high.
> Making it configurable it would require:
> * Introducing a 'tasktracker.expiry.interval' property with default value of 10 mins.
> * Load the property in the JobTracker
> * Change the 5 usages of the constant by a property in the JobTracker
> * Remove the constant from MRConstants
> * add default value to the hadoop-default.xml file
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.