[ 
https://issues.apache.org/jira/browse/MAPREDUCE-2943?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13098965#comment-13098965
 ] 

Vinod Kumar Vavilapalli commented on MAPREDUCE-2943:
----------------------------------------------------

I just looked at {{LocalJobRunner}} assuming it can be taken out of MRV1 into 
one of the MRV2 modules. But turns out it is pretty wired into MRV1 stuff, 
DistributedCache and all.

So for now, I guess we should just wait for HADOOP-7590 and use the MRV1 
artifact.

It is a pity that DistCpV2 has to depend on the old artifacts just for this. 
Guess we can use this ticket for making a version of local-job-runner that 
works with YARN.

> Test-failures when using local-job-runner.
> ------------------------------------------
>
>                 Key: MAPREDUCE-2943
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2943
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: distcp, mrv2
>    Affects Versions: 0.24.0
>            Reporter: Mithun Radhakrishnan
>            Priority: Minor
>
> Here's an example of a failure when a test attempts to run using the 
> local-job-runner:
> [ERROR] TestIntegration - Exception encountered  
> <java.lang.NullPointerException>java.lang.NullPointerException
>     at org.apache.hadoop.mapreduce.Cluster.getStagingAreaDir(Cluster.java:263)
>     at 
> org.apache.hadoop.mapreduce.JobSubmissionFiles.getStagingDir(JobSubmissionFiles.java:95)
>     at org.apache.hadoop.tools.DistCp.createMetaFolderPath(DistCp.java:337)
>     at org.apache.hadoop.tools.DistCp.<init>(DistCp.java:77)
>     at 
> org.apache.hadoop.tools.TestIntegration.runTest(TestIntegration.java:449)
>     at 
> org.apache.hadoop.tools.TestIntegration.caseSingleFileMissingTarget(TestIntegration.java:79)
>     at 
> org.apache.hadoop.tools.TestIntegration.testSingleFileMissingTarget(TestIntegration.java:69)
>     at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
>     at 
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
>     at 
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
>     at java.lang.reflect.Method.invoke(Method.java:597)
>     at 
> org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:44)
>     at 
> org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:15)
>     at 
> org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:41)
>     ...
> This is a DistCp(v2) test, built and run using Maven. DistCp's tests depend 
> on the following maven artifacts:
> 1. hadoop-mapreduce-client-jobclient
> 2. hadoop-hdfs
> 3. hadoop-common
> The failure is in spite of doing a " conf.set(MRConfig.FRAMEWORK_NAME, 
> "local");". Vinod (KV)'s in the loop on this. It appears that the root-cause 
> lies in org/apache/hadoop/mapred/LocalClientProtocolProvider.class not being 
> available in class-path. This might have to do with MR1 not having been 
> maven-ized, or some such.
> I haven't a clean way around this for the moment, so I'll temporarily disable 
> this distcp-test from running. It would be nice, however, to have maven-tests 
> that use local-job-runner to run through.

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

        

Reply via email to