[ 
http://issues.apache.org/jira/browse/HADOOP-39?page=comments#action_12366674 ] 

Doug Cutting commented on HADOOP-39:
------------------------------------

The point is to try to get map tasks with side effects to sometimes succeed, 
even with speculative execution?  That sounds like it could be a bad idea.  
Wouldn't it be better to have map tasks with side effects fail more frequently 
with speculative execution, so that you find such problems sooner, with smaller 
datasets on a smaller cluster, before you try a big run?  Or am I 
misunderstanding you?

> Job killed when backup tasks fail
> ---------------------------------
>
>          Key: HADOOP-39
>          URL: http://issues.apache.org/jira/browse/HADOOP-39
>      Project: Hadoop
>         Type: Bug
>   Components: mapred
>     Reporter: Owen O'Malley

>
> I had a map job with side effects that meant that any speculative tasks would 
> fail.
> Currently, the job tracker kills the job when the speculative task fails 4 
> times.
> It would be better to stop scheduling speculative tasks for that fragment, 
> but let the job continue as long as one of the the instances of that fragment 
> continue to run.

-- 
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

Reply via email to