[
https://issues.apache.org/jira/browse/YUNIKORN-1434?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Peter Bacsko updated YUNIKORN-1434:
-----------------------------------
Component/s: core - scheduler
(was: shim - kubernetes)
> Mark allocation & ask as being preempted
> ----------------------------------------
>
> Key: YUNIKORN-1434
> URL: https://issues.apache.org/jira/browse/YUNIKORN-1434
> Project: Apache YuniKorn
> Issue Type: Bug
> Components: core - scheduler
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Major
>
> Currently, ask & allocation objects could be preempted multiple times if
> there's no room for a daemonset pod, flooding the logs:
> {noformat}
> 2022-11-08T04:50:00.395Z INFO objects/application.go:1128
> Triggering preemption process for daemon set ask {"ds allocation key":
> "0d92110b-ea80-4cd5-962c-8d604de8ebe6"}
> 2022-11-08T04:50:00.395Z INFO objects/application.go:1139 Found
> victims for daemon set ask preemption {"ds allocation key":
> "0d92110b-ea80-4cd5-962c-8d604de8ebe6"}
> 2022-11-08T04:50:00.395Z INFO cache/application.go:707 try to
> release pod from application {"appID":
> "spark-b5e4b7814eca423bb89eee938051267c", "allocationUUID":
> "d307350a-b18c-4497-a808-b5d14b17e69c", "terminationType":
> "PREEMPTED_BY_SCHEDULER"}
> {noformat}
> The above lines are printed multiple times until there's enough room.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
