Patrick Ohly

In this case, "behavioral changes or user-facing changes" would imply updating the KEP upfront. I can't think of anything right now, so I suppose it's fine to not track it.

/milestone clear No user-visible changes planned for 1.34, let's stop tracking this for this cycle.

/wg device-management /sig scheduling As discussed during KubeCon NA 2024, there is precedence for enabling new functionality in the kube-scheduler as beta with enabled by default. A feature gate needs...

This is only about situations where the scheduler needs to free up devices to make them available for other pods. A device that is removed from a ResourceSlice doesn't fall...

> I can think of several reasons a cluster provisioner may not want to support plugins, I'm not surprised about this 😄 > In the case of DRA, aren't drivers...

A downward API would be simpler than my proposal and sufficient for portability. Cleanly separating plugins is just the icing on the cake. But the main question remains: if the...

/hold Needs to be merged after https://github.com/kubernetes/test-infra/pull/35594.

[pull-kind-conformance-parallel-dual-stack-ipv4-ipv6](https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/kubernetes-sigs_kind/4015/pull-kind-conformance-parallel-dual-stack-ipv4-ipv6/1971517704107986944) timed out because https://github.com/kubernetes/test-infra/pull/35594 hasn't been merged yet. That PR ensures that serial tests don't run in that job, avoiding the timeout.

If we don't want to run the serial tests, then we need to tweak https://github.com/kubernetes/test-infra/pull/35594 to exclude them. Let's discuss there. Doing it via e2e-k8s.sh is the wrong place for...

/test pull-kind-conformance-parallel-ipv6 Retesting to see whether the job works here (it fails in https://github.com/kubernetes-sigs/kind/pull/4046).