[jira] [Commented] (OOZIE-1700) Oozie job fails because it doesnt see even though result is SUCCESS

["jay vyas (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OOZIE-1700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904618#comment-13904618 ] 

jay vyas commented on OOZIE-1700:
---------------------------------

(Original bug report from the mailing list) 

I've run into quite cryptic error when submitting a very simple java action (DemoJavaMain, just prints args out). 

*** Oozie is failing out with JA017: ***

"Oozie cant find the hadoop job associated with it. "

It seems to be available in YARN.... Here are some more details:

- I've tried increasing user-retry to 3, that did not work.
- YARN doesnt seem to report any failures: Yarn somehow "thinks" the job succeeded.
- I dont see any system outs from the job (which is all the DemoJavaMain does).

*** Ive seen this unresolved in a few places online... any thoughts on how to fix it? ***

http://hortonworks.com/community/forums/topic/oozie-unknown-hadoop-job/

http://mail-archives.apache.org/mod_mbox/oozie-user/201211.mbox/%3C19AD42E3F64F0F468A305399D0DF39D9335DAC6F34@winops07.win.compete.com%3E


*** my specifc  exception is below **** 
{noformat}
org.apache.oozie.action.ActionExecutorException: JA017: Unknown hadoop job [job_1392665114408_0007] associated with action [0000037-140217072450087-oozie-oozi-W@java-node].  Failing this action!
    at org.apache.oozie.action.hadoop.JavaActionExecutor.check(JavaActionExecutor.java:1009)
    at org.apache.oozie.command.wf.ActionCheckXCommand.execute(ActionCheckXCommand.java:177)
    at org.apache.oozie.command.wf.ActionCheckXCommand.execute(ActionCheckXCommand.java:56)
    at org.apache.oozie.command.XCommand.call(XCommand.java:280)
    at org.apache.oozie.service.CallableQueueService$CallableWrapper.run(CallableQueueService.java:175)
    at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:662)
{noformat}

> Oozie job fails because it doesnt see even though result is SUCCESS
> -------------------------------------------------------------------
>
>                 Key: OOZIE-1700
>                 URL: https://issues.apache.org/jira/browse/OOZIE-1700
>             Project: Oozie
>          Issue Type: Bug
>            Reporter: jay vyas
>
> This is a very difficult bug to describe, so ive made a video of it.  
> It seems that others have seen this as well: Sometimes , when you submit a job, oozie cant find a hadoop job matching it, so it fails out.  
> But in this video, as shown, the job indeed succeeds on the hadoop cluster and ouputput + SUCCESS file is created. 
> http://www.youtube.com/watch?v=EPdkJ3zMp0c



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)