hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sanjay Dahiya (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-263) task status should include timestamps for when a job transitions
Date Thu, 03 Aug 2006 06:01:16 GMT
    [ http://issues.apache.org/jira/browse/HADOOP-263?page=comments#action_12425441 ] 
Sanjay Dahiya commented on HADOOP-263:

Here is what I am thinking for this. 

We add an extra field (timestamp) in org.apache.hadoop.mapred.TaskStatus. The timestamp gets
TaskTracker.TaskInprogress.reportProgress(), when a task reports progress. As part of Task
the timestamp is updated in Progress.set(), Progress.setStatus() and complete(). The timestamp
is sent over to TaskTracker over RPC along with status.

Another option is to update the timestamp when TaskTracker gets the progress report, without
making any change to the Task. But sending timestamp over RPC gives a more accurate timestamp
of the last progress update. 

> task status should include timestamps for when a job transitions
> ----------------------------------------------------------------
>                 Key: HADOOP-263
>                 URL: http://issues.apache.org/jira/browse/HADOOP-263
>             Project: Hadoop
>          Issue Type: Improvement
>          Components: mapred
>    Affects Versions: 0.2.1
>            Reporter: Owen O'Malley
>         Assigned To: Sanjay Dahiya
>             Fix For: 0.5.0
> It would help users to understand what happened if the task status included information
about when the task transitioned:
> Map:
>    started
>    finished
> Reduce:
>    started
>    shuffle finished
>    sort finished
>    finished

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


View raw message