[GitHub] [yunikorn-core] craigcondit commented on a diff in pull request #458: [YUNIKORN-1434] Mark allocation as being preempted

[GitBox <gi...@apache.org>]

craigcondit commented on code in PR #458:
URL: https://github.com/apache/yunikorn-core/pull/458#discussion_r1035278006


##########
pkg/scheduler/objects/application.go:
##########
@@ -1093,6 +1097,9 @@ func (sa *Application) tryReservedAllocate(headRoom *resources.Resource, nodeIte
 
 		// Do we need a specific node?
 		if ask.GetRequiredNode() != "" {
+			if ask.HasTriggeredPreemption() || time.Since(ask.GetLastPreemptionAttempt()) < sa.preemptionAttemptInterval {

Review Comment:
   I'm not sure we want to do this. Once we have triggered preemption, the pods will be scheduled for termination, but may not yet have terminated (either gracefully or not). If graceful termination takes longer than 5 seconds, we're going to attempt to trigger again, which could result in evicting even more pods. In the worst case scenario, we could kill everything else running on the node.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@yunikorn.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org