[jira] [Commented] (GIRAPH-391) Zookeeper setup in GraphMapper should check for existence of the jar file

["Avery Ching (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GIRAPH-391?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486581#comment-13486581 ] 

Avery Ching commented on GIRAPH-391:
------------------------------------

Thanks for taking a look at this Ravi.  Unfortunately, your patch breaks unittests (try running 'mvn clean install').  Can you please rework it?
                
> Zookeeper setup in GraphMapper should check for existence of the jar file
> -------------------------------------------------------------------------
>
>                 Key: GIRAPH-391
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-391
>             Project: Giraph
>          Issue Type: Bug
>    Affects Versions: 0.2.0
>            Reporter: Ravi Prakash
>         Attachments: giraph-391.patch
>
>
> Distributed Cache behaviour has changed in Hadoop 0.23. 
> When running on the PageRankBenchmark on a pseudo-distributed setup, GraphMapper.java:303 returns null
> bq. Path[] fileClassPaths = DistributedCache.getLocalCacheArchives(conf);
> Further, context.getJar() returns the location of the job jar in *HDFS* . Obviously this can't be added to the classpath. If we only checked for the existence of this jar file, it would follow on to using the findContainingJar() method which returns the right thing.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira