[
https://issues.apache.org/jira/browse/YARN-5783?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15649382#comment-15649382
]
Karthik Kambatla commented on YARN-5783:
----------------------------------------
bq. Overall, I think "starved" should be a scheduler-agnostic scheduling state
of an app.
[~leftnoteasy] - that is an interesting thought. In the current patch for
FairScheduler preemption, strictly speaking we are not essentially marking the
application as starved. Marking them so makes it easier to handle any changes
to the starvation much more gracefully. Given it is related, but not necessary,
I would like to follow up on that separately. We should also think carefully
about adding scheduler-only states for applications.
[~templedf] - mind taking another look?
> Verify applications are identified starved
> ------------------------------------------
>
> Key: YARN-5783
> URL: https://issues.apache.org/jira/browse/YARN-5783
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: fairscheduler
> Affects Versions: 2.8.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Labels: oct16-medium
> Attachments: yarn-5783.YARN-4752.1.patch,
> yarn-5783.YARN-4752.2.patch, yarn-5783.YARN-4752.3.patch,
> yarn-5783.YARN-4752.4.patch, yarn-5783.YARN-4752.5.patch,
> yarn-5783.YARN-4752.6.patch, yarn-5783.YARN-4752.7.patch,
> yarn-5783.YARN-4752.8.patch
>
>
> JIRA to track unit tests to verify the identification of starved
> applications. An application should be marked starved only when:
> # Cluster allocation is over the configured threshold for preemption.
> # Preemption is enabled for a queue and any of the following:
> ## The queue is under its minshare for longer than minsharePreemptionTimeout
> ## One of the queue’s applications is under its fairshare for longer than
> fairsharePreemptionTimeout.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
