[ 
https://issues.apache.org/jira/browse/OOZIE-3294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16519209#comment-16519209
 ] 

Ahmed edited comment on OOZIE-3294 at 6/25/18 5:20 AM:
-------------------------------------------------------

[~andras.piros] Thanks for the quick response,

Yes we are doing oozie shell actions. post your comment i evaluated all the 
servers with the spark-submit available, i see 1 node is *not installed* with 
the client(suspecting that node is causing intermittent issue sometimes if the 
oozie calls its action and YARN runs on that node & the failure Error is 
noticed.) 

Rephrasing the question:

How do i exactly identifiy node without the programme spark-submit is causing 
failure as the oozie error logs says "*no such file+(but does not specify the 
hostname+)*". is there a way to find the exact failure node  ? currently we did 
manual evaluation on each host (what if in large production setup of thousand 
of nodes,[manual validation would not be possible,]) How do we identify the 
host.

Regards

-Ahmed 


was (Author: big-d):
[~andras.piros] Thanks for the quick response,

Yes we are doing oozie shell actions. post your comment i evaluated all the 
servers with the spark-submit available, i see 1 node is not installed with the 
client(suspecting that node is causing intermittent issue sometimes if the 
oozie calls its action and YARN runs on that node & the failure Error is 
noticed.) 

How do i exactly identifiy node without the programme spark-submit is causing 
failure as the error logs says no such file. is there a way to find the exact 
failure cause on that specific node ?

How do we fix this going forward? does installing the spark-client resolves 
this or we stop the nodemanager on this host and enable on other node?

 

Regards

-Ahmed 

> Launcher exception: java.io.IOException
> ---------------------------------------
>
>                 Key: OOZIE-3294
>                 URL: https://issues.apache.org/jira/browse/OOZIE-3294
>             Project: Oozie
>          Issue Type: Bug
>          Components: action
>    Affects Versions: 4.2.0
>            Reporter: Ahmed
>            Priority: Major
>
> Hi ,
> There is an intermittent issue in oozie workflow, where the oozie action 
> fails sometimes with file not found error for some actions of spark-submit or 
> sqoop. (The issue is intermittent.)
>  
> I am using hadoop - 2.7.3
> oozie version-4.2.0
>  
> "Launcher exception: java.io.IOException Cannot run program "./spark-submit" 
> (in directory "/usr/hdp/current/spark2-client/bin"): error=2, No such file or 
> directory"
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to