[
https://issues.apache.org/jira/browse/YARN-6210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15877405#comment-15877405
]
Karthik Kambatla commented on YARN-6210:
----------------------------------------
Thanks for the prompt review, Daniel.
The updated patch incorporates all your suggestions but for the following:
# {{FSAppAttempt#fairShareStarvation()}}: I have reverted the meaning of
_starved_. The reason for this is to ensure we don't mark an app whose demand
is fully met, but the allocation is under its fairshare. The code also seems
simpler and less confusing this way.
# {{TestFairScheduler.testReservationWithMultiplePriorities()}}: The
reservation-at-lower-priority assert is retained. I did drop the asserts for
scheduler resources, but retained the checks around running containers. Since
the test is for verifying node reservation behavior, other asserts are
misleading.
> FS: Node reservations can interfere with preemption
> ---------------------------------------------------
>
> Key: YARN-6210
> URL: https://issues.apache.org/jira/browse/YARN-6210
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler
> Affects Versions: 2.9.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Attachments: YARN-6210.1.patch, YARN-6210.2.patch, YARN-6210.3.patch
>
>
> Today, on a saturated cluster, apps with pending demand reserve nodes. A new
> app might not be able to preempt resources because these nodes are already
> reserved. This can be reproduced by the example in YARN-6151.
> Since node reservations are to prevent starvation of apps requesting large
> containers, triggering these reservations only on starved applications would
> avoid this situation.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
