[
https://issues.apache.org/jira/browse/YUNIKORN-231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17156421#comment-17156421
]
Wilfred Spiegelenburg commented on YUNIKORN-231:
------------------------------------------------
Attached a document describing the limitations of the implementation that was
added.
Based on the limitations we need to look at this again to fully implement auto
scaling support.
> Add ability to update pod condition based on accurate scheduling state
> ----------------------------------------------------------------------
>
> Key: YUNIKORN-231
> URL: https://issues.apache.org/jira/browse/YUNIKORN-231
> Project: Apache YuniKorn
> Issue Type: Improvement
> Components: core - scheduler, shim - kubernetes
> Reporter: Weiwei Yang
> Assignee: Weiwei Yang
> Priority: Major
> Labels: pull-request-available
> Fix For: 0.9
>
> Attachments: YuniKorn and cluster auto-scaler.pdf
>
>
> The cluster auto-scaler calculates the desire number of nodes based on pods
> in *unscheduable* state. Currently, yunikorn sets a pod to unscheduable state
> after hard coded 3 seconds. This would trigger auto-scaling happen
> unnecessarily sometimes.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
