[
https://issues.apache.org/jira/browse/YARN-6706?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16081275#comment-16081275
]
Haibo Chen commented on YARN-6706:
----------------------------------
bq. Not sure if startPendingContainers() before and after en-queuing is usefull
You are referring to the else statement in
ContainerScheduler#scheduleContainer() right? The reason of doing
startPendingContainers() both before and after en-queuing is so that we always
respect the max queue limit for OPPORTUNISTIC containers.
If we were to always do enqueueing first and then startPendingContainers(), we
could end up going over the OPPR container queue length.
For GUARANTEED containers, killOpportunisticContainers is needed if the
GUARANTEED container stays in the queue after startPendingContainers(). If I
misunderstood your comment, can you elaborate a little more.
> Refactor ContainerScheduler to make oversubscription change easier
> ------------------------------------------------------------------
>
> Key: YARN-6706
> URL: https://issues.apache.org/jira/browse/YARN-6706
> Project: Hadoop YARN
> Issue Type: Sub-task
> Affects Versions: 3.0.0-alpha3
> Reporter: Haibo Chen
> Assignee: Haibo Chen
> Attachments: YARN-6706.01.patch, YARN-6706-YARN-1011.00.patch,
> YARN-6706-YARN-1011.01.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
