Jan Šafránek
Jan Šafránek
/retest > There are no nodes that your pod can schedule to
This behavior is controller by Pod's strategy: https://github.com/kubernetes/kubernetes/blob/4abf29c5c86349c7d2c769d267f4552ced5d0c15/pkg/registry/core/pod/strategy.go#L174-L176 Not sure what SIG is responsible for it. Starting with: /sig api-machinery /sig architecture
/remove-sig node This is not related to node in any way. Pods are deleted in the API server and as I pointed above and node / kubelet cannot do anything...
> terminationGracePeriod: 0 is the pod author making pods default to force deletion I disagree here. API says it means that pod's containers will be killed without waiting. That's IMO...
> TerminationGracePeriodSeconds defines the default value for the delete operation's GracePeriodSeconds If that's true, then it should be explicitly written in the field description, with a huge warning about zero...
@carlory, can you please split it into two PRs? The plugin + feature gate removal should be easy to merge. A second PR with the annotation removal might need some...
Thanks! /lgtm /approve
I can see it helped with the events. /lgtm /approve
/label acknowledge-critical-fixes-only It fixes a TRT component readiness issue.
/retest ???