[
https://issues.apache.org/jira/browse/SPARK-34406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17281541#comment-17281541
]
Hyukjin Kwon commented on SPARK-34406:
--------------------------------------
You should use Yarn cluster mode that will evenly distribute the drivers.
> When we submit spark core tasks frequently, the submitted nodes will have a
> lot of resource pressure
> ----------------------------------------------------------------------------------------------------
>
> Key: SPARK-34406
> URL: https://issues.apache.org/jira/browse/SPARK-34406
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Affects Versions: 3.0.1
> Reporter: hao
> Priority: Major
>
> When we submit spark core tasks frequently, the submitted node will have a
> lot of resource pressure, because spark will create a process instead of a
> thread for each submitted task. In fact, there is a lot of resource
> consumption. When the QPS of the submitted task is very high, the submission
> will fail due to insufficient resources. I would like to ask how to optimize
> the amount of resources consumed by spark core submission
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
