[
https://issues.apache.org/jira/browse/YARN-4003?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15064234#comment-15064234
]
Sunil G commented on YARN-4003:
-------------------------------
Hi [~curino]
Thank you for the clarification. Yes, there is no cleaner solution. But I
think, we cud calculate the AMUsed capacity of all ReservationQueue under one
parent (PlanQueue)
So could we have an api like below in {{PlanQueue}} and use along with
{{getAMResourceLimit}} to ensure that we do not cross the limit of parent
queue. I might be wrong, pls correct me if so.
{code}
public synchronized Resource sumOfChildAMUsedCapacities() {
Resource ret = Resources.createResource(0);
for (CSQueue l : childQueues) {
Resources.addTo(ret, l.getQueueResourceUsage().getAMUsed());
}
return ret;
}
{code}
> ReservationQueue inherit getAMResourceLimit() from LeafQueue, but behavior is
> not consistent
> --------------------------------------------------------------------------------------------
>
> Key: YARN-4003
> URL: https://issues.apache.org/jira/browse/YARN-4003
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Reporter: Carlo Curino
> Assignee: Carlo Curino
> Attachments: YARN-4003.patch
>
>
> The inherited behavior from LeafQueue (limit AM % based on capacity) is not a
> good fit for ReservationQueue (that have highly dynamic capacity).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
