[ 
https://issues.apache.org/jira/browse/YARN-9174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16735048#comment-16735048
 ] 

Arun Suresh commented on YARN-9174:
-----------------------------------

Thanks for the patch [~jhung]

To be honest, I am not very happy with the changes to the 
{{NMStateStoreService}} - the {{updateContainerResourceMapping}} method does 
not looks like it should belong in the state store service. I assume the 
requirement is that we we add the assigned resource to the container only if 
the {{storeAssignedResource}} completes successfully - this probably should 
have been done via a try catch by the caller of {{storeAssignedResources}}.

But I understand this is a backport, so we should retain the logic of the 
original patch and maybe raise a subsequent refactoring JIRA.

+1 otherwise.


> branch-3.0/branch-2 refactoring of GpuDevice class
> --------------------------------------------------
>
>                 Key: YARN-9174
>                 URL: https://issues.apache.org/jira/browse/YARN-9174
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Jonathan Hung
>            Assignee: Jonathan Hung
>            Priority: Major
>         Attachments: YARN-9174-YARN-8200.001.patch, 
> YARN-9174-YARN-8200.branch3.001.patch, YARN-9174-YARN-8200.branch3.002.patch
>
>
> YARN-7224 does two main things:
>  # refactors Gpu device numbers to a separate GpuDevice class,
>  # adds Docker support for Gpus
> This ticket is for doing *only* the GpuDevice class refactoring so we have 
> this logic in branch-3.0 and branch-2



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org

Reply via email to