What would you like to be added?

k8s version is 1.28.1.

The problem handling process is as follows:

1. Use STS to create a pod webswingservice.
2. A CSI plug-in is customized. webswingservice depends on this plug-in.
3. The first startup process is smooth.
4. The node is powered off once after startup.
5. After the system is powered on, the customized CSI fault is not rectified. As a result, volumes fail to be mounted to webswingservice.
6. The webswingservice does not execute the probe because the volume fails to be attached. Therefore, the phase and conditions.Ready in the pod are still in the ready state.
7. In this case, the pod status is not ready, but the workload is not updated. As a result, the inconsistency occurs.
8. The following figure shows the proble

[root@master1 ~]# kubectl get po -A | grep webswingservice
webswingservice-0	0/2	Unknown	47h
webswingservice-1	0/2	Unknown	47h
webswingservice-2	0/2	Unknown	47h
[root@master1 ~]#
[root@master1 ~]#
[root@master1 ~]#
[root@master1 ~]# kubectl get sts -n test webswingservice
NAME	ReADY	AGE
webswingservice	3/3	47h
[root@master1 ~]#

Code snippet:

1. The probe is executed only when the volume is successfully attached to the SyncPod.
   

   kubernetes/pkg/kubelet/kubelet.go

   Lines 1911 to 1933 in 9d9e08b

   	// Wait for volumes to attach/mount
   	if err := kl.volumeManager.WaitForAttachAndMount(ctx, pod); err != nil {
   	if !wait.Interrupted(err) {
   	kl.recorder.Eventf(pod, v1.EventTypeWarning, events.FailedMountVolume, "Unable to attach or mount volumes: %v", err)
   	klog.ErrorS(err, "Unable to attach or mount volumes for pod; skipping pod", "pod", klog.KObj(pod))
   	}
   	return false, err
   	}
   	// Fetch the pull secrets for the pod
   	pullSecrets := kl.getPullSecretsForPod(pod)
   	// Ensure the pod is being probed
   	kl.probeManager.AddPod(pod)
   	if utilfeature.DefaultFeatureGate.Enabled(features.InPlacePodVerticalScaling) {
   	// Handle pod resize here instead of doing it in HandlePodUpdates because
   	// this conveniently retries any Deferred resize requests
   	// TODO(vinaykul,InPlacePodVerticalScaling): Investigate doing this in HandlePodUpdates + periodic SyncLoop scan
   	// See: https://github.com/kubernetes/kubernetes/pull/102884#discussion_r663160060
   	if kl.podWorkers.CouldHaveRunningContainers(pod.UID) && !kubetypes.IsStaticPod(pod) {
   	pod = kl.handlePodResourcesResize(pod)
   	}

2. The sts checks whether the pod is ready.
   

   kubernetes/pkg/controller/statefulset/stateful_set_utils.go

   Lines 461 to 464 in 9d9e08b

   	// isRunningAndReady returns true if pod is in the PodRunning Phase, if it has a condition of PodReady.
   	func isRunningAndReady(pod *v1.Pod) bool {
   	return pod.Status.Phase == v1.PodRunning && podutil.IsPodReady(pod)
   	}

Why is this needed?

Is this a problem? Is there any solution?