// with later. The new M with the same ID could even try to
	// steal back this P from itself!
	//
	// The runtime is careful to make sure that any GoCreateSyscall
	// event will enter the runtime emitting events for reacquiring a P.
	//
	// Note: we might have a P here. The P might not be released
	// eagerly by the runtime, and it might get stolen back later

			return
		}
	}

	// once a pod is terminated by UID, it cannot reenter the pod worker (until the UID is purged by housekeeping)
	if status.IsFinished() {
		klog.V(4).InfoS("Pod is finished processing, no further updates", "pod", klog.KRef(ns, name), "podUID", uid, "updateType", options.UpdateType)

* gce/kube-down: Parallelize IGM deletion, batch more ([#27302](https://github.com/kubernetes/kubernetes/pull/27302), [@zmerlynn](https://github.com/zmerlynn))
* Enable dynamic allocation of heapster/eventer cpu request/limit ([#27185](https://github.com/kubernetes/kubernetes/pull/27185), [@gmarek](https://github.com/gmarek))

	if pendingDelayedClaims > 0 {
		// We could start a pod scheduling attempt to refresh the
		// potential nodes list.  But pod scheduling attempts are
		// expensive and doing them too often causes the pod to enter
		// backoff. Let's wait instead for all drivers to reply.
		if loggerV := logger.V(6); loggerV.Enabled() {