[
https://issues.apache.org/jira/browse/YARN-6863?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Daniel Templeton updated YARN-6863:
-----------------------------------
Description: In {{FSPreemptionThread.identifyContainersToPreemptOnNode()}},
we add every container we encounter to the preemption list until we meet the
desired target. As we head into resource types, that behavior will become more
of a problem, but it's technically an issue already because the fair scheduler
supports requests with 0 vcores or 0 memory. (was: In
{{FSPreemptionThread.identifyContainersToPreempt()}}, we add every container we
encounter to the preemption list until we meet the desired target. As we head
into resource types, that behavior will become more of a problem, but it's
technically an issue already because the fair scheduler supports requests with
0 vcores or 0 memory.)
> Fair Scheduler preemption thread should check that a container has the needed
> resources before adding it to the preemption list
> -------------------------------------------------------------------------------------------------------------------------------
>
> Key: YARN-6863
> URL: https://issues.apache.org/jira/browse/YARN-6863
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler
> Affects Versions: 3.0.0-alpha4
> Reporter: Daniel Templeton
> Priority: Minor
>
> In {{FSPreemptionThread.identifyContainersToPreemptOnNode()}}, we add every
> container we encounter to the preemption list until we meet the desired
> target. As we head into resource types, that behavior will become more of a
> problem, but it's technically an issue already because the fair scheduler
> supports requests with 0 vcores or 0 memory.
--
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
