[
https://issues.apache.org/jira/browse/OOZIE-2547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15317486#comment-15317486
]
Satish Subhashrao Saley commented on OOZIE-2547:
------------------------------------------------
Hello [~rkanter], Could you please review the patch?
I have removed the logic behind populating {{spark.executor.extraClassPath}},
{{spark.driver.extraClassPath}}, {{-- jars}} and {{spark.yarn.dist.files}}.
Instead of that, now we are adding distributed cached files in {{-- files}}.
While doing so, I also make sure that hdfs paths to those files are formulated
such that spark won't make another copy.
I have tested the patch locally as well as in clusters, it seems working fine
with {{-- master}} as local,yarn-client and yarn-cluster.
> Add mapreduce.job.cache.files to spark action
> ---------------------------------------------
>
> Key: OOZIE-2547
> URL: https://issues.apache.org/jira/browse/OOZIE-2547
> Project: Oozie
> Issue Type: Bug
> Reporter: Satish Subhashrao Saley
> Assignee: Satish Subhashrao Saley
> Priority: Minor
> Attachments: OOZIE-2547-1.patch
>
>
> Currently, we pass jars using --jars option while submitting spark job. Also,
> we add spark.yarn.dist.files option in case of yarn-client mode.
> Instead of that, we can have only --files option and pass on the files which
> are present in mapreduce.job.cache.files. While doing so, we make sure that
> spark won't make another copy of the files if files exist on the hdfs. We saw
> the issues when files are getting copied multiple times and causing
> exceptions such as :
> {code}
> Diagnostics: Resource
> hdfs://localhost/user/saley/.sparkStaging/application_1234_123/oozie-examples.jar
> changed on src filesystem
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
