[
https://issues.apache.org/jira/browse/SPARK-1441?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13963021#comment-13963021
]
Thomas Graves commented on SPARK-1441:
--------------------------------------
The yarn-alpha profile is tied to meaning included the spark on yarn support
for hadoop 0.23. If you don't need the spark on yarn support you should be
able to just compile spark for hadoop 0.23 so that you can talk to HDFS on a
hadoop 0.23 cluster, but you are running spark separately (standalone cluster
or mesos). I think the latter is what is trying to be done here. It
doesn't hurt anything to included the yarn pieces (other then slightly larger
assembly jar) so that is atleast a work around for now.
> Compile Spark Core error with Hadoop 0.23.x
> -------------------------------------------
>
> Key: SPARK-1441
> URL: https://issues.apache.org/jira/browse/SPARK-1441
> Project: Spark
> Issue Type: Bug
> Components: Build
> Affects Versions: 1.0.0
> Reporter: witgo
> Attachments: mvn.log, sbt.log
>
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
