[ 
https://issues.apache.org/jira/browse/MAPREDUCE-2880?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13109610#comment-13109610
 ] 

Vinod Kumar Vavilapalli commented on MAPREDUCE-2880:
----------------------------------------------------

I did some debugging with the latest patch and cornered the test failures to 
this: The symlinks that are supposed to be created in container-dir are instead 
getting created in hadoop-mapreduce-jobclient dir.

*Expected*: 
_hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/target/org.apache.hadoop.mapreduce.v2.TestMRJobs/org.apache.hadoop.mapreduce.v2.TestMRJobs-localDir/usercache/vinodkv/appcache/application_1316619583041_0001/container_1316619583041_0001_01_000001/MRAppJar.jar_
*Found*: 
_hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/MRAppJar.jar_

This is because of the change in {{DefaultContainerExecutor.launchContainer()}} 
which changed the currentWorkingDirectory of the shell to be null from new 
{{File(containerWorkDir.toUri().getPath())}}. We need to restore that, this PWD 
makes sense for both tests and the real world in case of 
_DefaultContainerExecutor_. For _LinuxContainerExecutor_, null makes sense for 
the real-world but not tests, but we don't have tests for LCE, so we are fine.

> Fix classpath construction for MRv2
> -----------------------------------
>
>                 Key: MAPREDUCE-2880
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2880
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: mrv2
>    Affects Versions: 0.23.0
>            Reporter: Luke Lu
>            Assignee: Arun C Murthy
>            Priority: Blocker
>             Fix For: 0.23.0
>
>         Attachments: MAPREDUCE-2880.patch, MAPREDUCE-2880.patch, 
> MAPREDUCE-2880.patch, MAPREDUCE-2880.patch, MAPREDUCE-2880.patch
>
>
> MRConstants.java refers a hard-coded version of MR AM jar. The build config 
> works around with a symlink. The deployment currently needs symlink 
> workaround as well. We need to fix this so that we can actually launch 
> arbitrary versions of AMs.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to