[
https://issues.apache.org/jira/browse/YARN-5830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15795778#comment-15795778
]
Karthik Kambatla commented on YARN-5830:
----------------------------------------
Discussed with Yufei offline.
I have a couple of high-level concerns:
# When identifying a node and containers to preempt, the patch differentiates
between a set with AM containers and a set without. However, it doesn't look
into the number of AM containers.
# The method {{identifyContainersToPreempt}} is too long and might benefit from
more modularity.
One approach we discussed was to add a new method,
{{identifyContainersToPreemptOnNode}} that works on a single node and takes
maximum number of allowed AM containers.
> Avoid preempting AM containers
> ------------------------------
>
> Key: YARN-5830
> URL: https://issues.apache.org/jira/browse/YARN-5830
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: fairscheduler
> Reporter: Karthik Kambatla
> Assignee: Yufei Gu
> Attachments: YARN-5830.001.patch, YARN-5830.002.patch
>
>
> While considering containers for preemption, avoid AM containers unless
> absolutely necessary.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
