[
https://issues.apache.org/jira/browse/YARN-10503?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17307529#comment-17307529
]
Qi Zhu edited comment on YARN-10503 at 3/24/21, 3:35 AM:
---------------------------------------------------------
Thanks [~ebadger] for review.
[~pbacsko] Your suggestion is valid, it make sense to me.
I have updated in latest patch consistent with YARN-9936 description:
* In an absolute amount of resources ( e.g. memory=4GB,vcores=20,yarn.io/gpu=4
). The amount of all resources in the queues has to be less than or equal to
all resources in the cluster.{color:#de350b}Actually, the above is not
supported, we only support memory and vcores now in absolute mode, we should
extend in {color}YARN-10503.
[~pbacsko] [~gandras] [~ebadger]
If you any other advice about this?
Thanks.
was (Author: zhuqi):
Thanks [~ebadger] for review.
[~pbacsko] Your suggestion is valid, it make sense to me.
I have updated in latest patch consistent with YARN-9936 description:
* In an absolute amount of resources ( e.g. memory=4GB,vcores=20,yarn.io/gpu=4
). The amount of all resources in the queues has to be less than or equal to
all resources in the cluster.{color:#de350b}Actually, the above is not
supported, we only support memory and vcores now in absolute mode, we should
extend in {color}YARN-10503.
[~pbacsko] [~gandras] [~ebadger]
If you any other advice about this?
> Support queue capacity in terms of absolute resources with custom
> resourceType.
> -------------------------------------------------------------------------------
>
> Key: YARN-10503
> URL: https://issues.apache.org/jira/browse/YARN-10503
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Qi Zhu
> Assignee: Qi Zhu
> Priority: Critical
> Attachments: YARN-10503.001.patch, YARN-10503.002.patch,
> YARN-10503.003.patch, YARN-10503.004.patch
>
>
> Now the absolute resources are memory and cores.
> {code:java}
> /**
> * Different resource types supported.
> */
> public enum AbsoluteResourceType {
> MEMORY, VCORES;
> }{code}
> But in our GPU production clusters, we need to support more resourceTypes.
> It's very import for cluster scaling when with different resourceType
> absolute demands.
>
> This Jira will handle GPU first.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
