[
https://issues.apache.org/jira/browse/KYLIN-4895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17322547#comment-17322547
]
Gaurav Rawat commented on KYLIN-4895:
-------------------------------------
When will this be fixed previously in kylin 4.0 we used to use this setting
{code:java}
kylin.engine.spark-conf.spark.submit.deployMode=cluster
{code}
but now it no longer works in cluster mode . Even when using
kylin.engine.spark-conf.spark.master=yarn . I think this should be fixed in
kylin 4.0 soon
> change spark deploy mode of kylin4.0 engine from local to cluster
> -----------------------------------------------------------------
>
> Key: KYLIN-4895
> URL: https://issues.apache.org/jira/browse/KYLIN-4895
> Project: Kylin
> Issue Type: New Feature
> Components: Job Engine
> Affects Versions: v4.0.0-alpha
> Reporter: tianhui
> Assignee: tianhui
> Priority: Major
> Fix For: v4.0.0-GA
>
>
> In cloud native environment, the memory of pod is quite limited. But the
> spark Driver can use a huge amount of memory, in job engine pod, which is
> quite difficult to manage when there are more and more projects.
> So it's better to put spark driver on yarn, and kylin only maintain its
> status.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
