[jira] [Commented] (PIG-3254) Fail a failed Pig script quicker

["Dmitriy V. Ryaboy (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PIG-3254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13608070#comment-13608070 ] 

Dmitriy V. Ryaboy commented on PIG-3254:
----------------------------------------

Can I add a request for whoever will work on this ticket?

Right now we die with "MR Job Failed" but don't say which job. In cases when multiple jobs are launched, one of them fails, the other ones are killed, and users find it hard to figure out which job was the cause of all badness. It would be nice to print out the job id of the failed job.
                
> Fail a failed Pig script quicker
> --------------------------------
>
>                 Key: PIG-3254
>                 URL: https://issues.apache.org/jira/browse/PIG-3254
>             Project: Pig
>          Issue Type: Improvement
>            Reporter: Daniel Dai
>             Fix For: 0.12
>
>
> Credit to [~asitecn]. Currently Pig could launch several mapreduce job simultaneously. When one mapreduce job fail, we need to wait for simultaneous mapreduce job finish. In addition, we could potentially launch additional jobs which is doomed to fail. However, this is unnecessary in some cases:
> * If "stop.on.failure==true", we can kill parallel jobs, and fail the whole script
> * If "stop.on.failure==false", and no "store" could success, we can also kill parallel jobs, and fail the whole script
> Consider simultaneous jobs may take a long time to finish, this could significantly improve the turn around in some cases.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira