> On 2012-03-23 01:14:02, Mohammad Islam wrote:
> > +1.
> > The question is: how it is working now without this extra settings.

it my case I was always testing with YARN in same host and local port, that it 
is why i didn't see it. I you had Hadoop site.xml injected in your Oozie setup, 
the value was picked up from there.

Our QA folks found this out testing in a cluster.


- Alejandro


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/4455/#review6262
-----------------------------------------------------------


On 2012-03-22 21:27:55, Alejandro Abdelnur wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/4455/
> -----------------------------------------------------------
> 
> (Updated 2012-03-22 21:27:55)
> 
> 
> Review request for oozie.
> 
> 
> Summary
> -------
> 
> A bug in Hadoop MAPREDUCE-4044 prevents the JobClient to process correctly 
> 'mapred.job.tracker'. 
> 
> To avoid this, until Hadoop fixes it, we should inject the jobtracker 
> host:port using the 3 possible properties for it: 
> 
> * mapred.job.tracker 
> * mapreduce.jobtracker.address 
> * yarn.resourcemanager.address 
> 
> 
> This addresses bug OOZIE-779.
>     https://issues.apache.org/jira/browse/OOZIE-779
> 
> 
> Diffs
> -----
> 
>   
> trunk/core/src/main/java/org/apache/oozie/action/hadoop/JavaActionExecutor.java
>  1304075 
> 
> Diff: https://reviews.apache.org/r/4455/diff
> 
> 
> Testing
> -------
> 
> built/deployed and submitted jobs to a hadoop 0.23 cluster running in a 
> non-default port
> 
> 
> Thanks,
> 
> Alejandro
> 
>

Reply via email to