[ 
https://issues.apache.org/jira/browse/YARN-4484?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15223667#comment-15223667
 ] 

Wangda Tan commented on YARN-4484:
----------------------------------

[~sunilg],

Thanks for elaborating. What I mentioned before is the case that a queue has >0 
available resources on default partition and other partition, so available 
resource to the queue is default_partition.available + 
other_partitions.available.

In your example, x_partition's available is 0. Could you update the test case 
to not use all quotas in x_partition?

> Available Resource calculation for a queue is not correct when used with 
> labels
> -------------------------------------------------------------------------------
>
>                 Key: YARN-4484
>                 URL: https://issues.apache.org/jira/browse/YARN-4484
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: capacity scheduler
>    Affects Versions: 2.7.1
>            Reporter: Sunil G
>            Assignee: Sunil G
>         Attachments: 0001-YARN-4484.patch, 0002-YARN-4484.patch, 
> 0003-YARN-4484-v2.patch, 0003-YARN-4484.patch, 0004-YARN-4484.patch
>
>
> To calculate available resource for a queue, we have to get the total 
> resource allocated for all labels in queue compare to its usage. 
> Also address the comments given in 
> [YARN-4304-comments|https://issues.apache.org/jira/browse/YARN-4304?focusedCommentId=15064874&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15064874
>  ] given by [~leftnoteasy] for same.
> ClusterMetrics related issues will also get handled once we fix this.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to