[
https://issues.apache.org/jira/browse/YUNIKORN-3269?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Manikandan R resolved YUNIKORN-3269.
------------------------------------
Fix Version/s: 1.9.0
Resolution: Fixed
> Kubernetes shim should respect node cordon state
> ------------------------------------------------
>
> Key: YUNIKORN-3269
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3269
> Project: Apache YuniKorn
> Issue Type: Bug
> Components: shim - kubernetes
> Reporter: Xi Chen
> Assignee: Xi Chen
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.9.0
>
>
> YuniKorn’s Kubernetes shim does not fully preserve Kubernetes node cordon
> state when syncing nodes to scheduler-core.
> A Kubernetes node with spec.unschedulable=true should remain unavailable for
> scheduling in YuniKorn. Today, the shim registers nodes with scheduler-core
> and can restore them to schedulable state without checking
> Node.Spec.Unschedulable. The node update path also does not translate
> cordon/uncordon changes into corresponding scheduler-core drain/restore
> actions.
> This can make a Kubernetes SchedulingDisabled node appear schedulable to
> YuniKorn.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
