`kubectl drain --disable-eviction --dry-run=server` hangs indefinitely

[kfess]

What happened:

kubectl drain --disable-eviction --dry-run=server target-node hangs indefinitely when there are pods on the target node.

What you expected to happen:

The command should complete immediately without hanging, as it does with --dry-run=client or without --disable-eviction.

How to reproduce it (as minimally and precisely as possible):

# With kind multi-node cluster
kubectl apply -f pod.yaml (any pod manifest)
kubectl drain kind-worker --disable-eviction --ignore-daemonsets --dry-run=server
# Hangs here indefinitely

Anything else we need to know?:

I think it appears that the deletePods path (used when --disable-eviction is set) enters waitForDelete after sending the dry-run delete request, waiting for pods that are never actually removed.

https://github.com/kubernetes/kubernetes/blob/fbe282098364806aed50a44fa69bcd2a0fc2ad8b/staging/src/k8s.io/kubectl/pkg/drain/drain.go#L411

Environment:

• Kubernetes version: v1.35
• OS: Linux

/sig cli
/kind bug
/area kubectl
/assign

[k8s-ci-robot]

This issue is currently awaiting triage.

SIG CLI takes a lead on issue triage for this repo, but any Kubernetes member can accept issues by applying the triage/accepted label.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[kfess]

Also when --disable-eviction is used, --dry-run=client prints "evicting pod" instead of "deleting pod".

# In my environment
k drain --disable-eviction --ignore-daemonsets --dry-run=client  kind-worker
node/kind-worker cordoned (dry run)
Warning: ignoring DaemonSet-managed Pods: kube-system/kindnet-pqw24, kube-system/kube-proxy-77bbt
evicting pod default/nginx-deployment-6f6756cdcb-j6spl (dry run) # here
evicting pod default/nginx-deployment-6f6756cdcb-x4sm4 (dry run) # here
node/kind-worker2 drained (dry run)

And --dry-run=server currently hangs (which this PR fixes), but even after the fix, no per-pod message will be printed.

I kept these untouched to keep this PR minimal.