diff --git a/pkg/minikube/bootstrapper/bsutil/kverify/pod_ready.go b/pkg/minikube/bootstrapper/bsutil/kverify/pod_ready.go
index 55d1c5ad2b..8c5dc8c054 100644
--- a/pkg/minikube/bootstrapper/bsutil/kverify/pod_ready.go
+++ b/pkg/minikube/bootstrapper/bsutil/kverify/pod_ready.go
@@ -20,8 +20,6 @@ package kverify
 import (
 	"context"
 	"fmt"
-	"regexp"
-	"strings"
 	"time"
 
 	core "k8s.io/api/core/v1"
@@ -45,7 +43,6 @@ func WaitExtra(cs *kubernetes.Clientset, labels []string, timeout time.Duration)
 		return fmt.Errorf("error listing pods in %q namespace: %w", meta.NamespaceSystem, err)
 	}
 
-	unstartedMinion := regexp.MustCompile(`node "(.*-m[0-9]+)" has status "Ready":"Unknown"`)
 	for _, pod := range pods.Items {
 		if time.Since(start) > timeout {
 			return fmt.Errorf("timed out waiting %v for all system-critical and pods with labels %v to be %q", timeout, labels, core.NodeReady)
@@ -66,15 +63,7 @@ func WaitExtra(cs *kubernetes.Clientset, labels []string, timeout time.Duration)
 			}
 			if match || pod.Spec.PriorityClassName == "system-cluster-critical" || pod.Spec.PriorityClassName == "system-node-critical" {
 				if err := waitPodCondition(cs, pod.Name, pod.Namespace, core.PodReady, timeout); err != nil {
-					// ignore unstarted minion nodes, non-running pods and non-pending pods, eg:
-					// error: `waitPodCondition: node "docker-containerd-multinode-wait-m02" hosting pod "kube-proxy-frz7b" in "kube-system" namespace is currently not "Ready": node "docker-containerd-multinode-wait-m02" has status "Ready":"Unknown"`, or
-					// error: `waitPodCondition: error getting pod "coredns-74ff55c5b-57fhw" in "kube-system" namespace (skipping!): pods "coredns-74ff55c5b-57fhw" not found`
-					// error: `waitPodCondition: pod "coredns-74ff55c5b-57fhw" in "kube-system" namespace has status phase "Failed" (skipping!)`
-					skipPod := strings.Contains(err.Error(), "skipping!")
-					if !unstartedMinion.MatchString(err.Error()) && !skipPod {
-						klog.Errorf("WaitExtra: %v", err)
-					}
-					continue
+					klog.Errorf("WaitExtra: %v", err)
 				}
 				break
 			}
@@ -132,7 +121,7 @@ func podConditionStatus(cs *kubernetes.Clientset, name, namespace string, condit
 	// check if undelying node is Ready - in case we got stale data about the pod
 	if pod.Spec.NodeName != "" {
 		if status, reason := nodeConditionStatus(cs, pod.Spec.NodeName, core.NodeReady); status != core.ConditionTrue {
-			return core.ConditionUnknown, fmt.Sprintf("node %q hosting pod %q in %q namespace is currently not %q: %v", pod.Spec.NodeName, name, namespace, core.NodeReady, reason)
+			return core.ConditionUnknown, fmt.Sprintf("node %q hosting pod %q in %q namespace is currently not %q (skipping!): %v", pod.Spec.NodeName, name, namespace, core.NodeReady, reason)
 		}
 	}
 
diff --git a/pkg/minikube/bootstrapper/kubeadm/kubeadm.go b/pkg/minikube/bootstrapper/kubeadm/kubeadm.go
index 8067c492b9..f89001612c 100644
--- a/pkg/minikube/bootstrapper/kubeadm/kubeadm.go
+++ b/pkg/minikube/bootstrapper/kubeadm/kubeadm.go
@@ -699,7 +699,21 @@ func (k *Bootstrapper) restartControlPlane(cfg config.ClusterConfig) error {
 		return errors.Wrap(err, "system pods")
 	}
 
-	// must be called after applyCNI
+	if err := kverify.NodePressure(client); err != nil {
+		adviseNodePressure(err, cfg.Name, cfg.Driver)
+	}
+
+	// This can fail during upgrades if the old pods have not shut down yet
+	addonPhase := func() error {
+		_, err := k.c.RunCmd(exec.Command("/bin/bash", "-c", fmt.Sprintf("%s phase addon all --config %s", baseCmd, conf)))
+		return err
+	}
+	if err = retry.Expo(addonPhase, 100*time.Microsecond, 30*time.Second); err != nil {
+		klog.Warningf("addon install failed, wil retry: %v", err)
+		return errors.Wrap(err, "addons")
+	}
+
+	// must be called after applyCNI and `kubeadm phase addon all` (ie, coredns redeploy)
 	if cfg.VerifyComponents[kverify.ExtraKey] {
 		// after kubelet is restarted (with 'kubeadm init phase kubelet-start' above),
 		// it appears as to be immediately Ready as well as all kube-system pods (last observed state),
@@ -728,20 +742,6 @@ func (k *Bootstrapper) restartControlPlane(cfg config.ClusterConfig) error {
 		}
 	}
 
-	if err := kverify.NodePressure(client); err != nil {
-		adviseNodePressure(err, cfg.Name, cfg.Driver)
-	}
-
-	// This can fail during upgrades if the old pods have not shut down yet
-	addonPhase := func() error {
-		_, err := k.c.RunCmd(exec.Command("/bin/bash", "-c", fmt.Sprintf("%s phase addon all --config %s", baseCmd, conf)))
-		return err
-	}
-	if err = retry.Expo(addonPhase, 100*time.Microsecond, 30*time.Second); err != nil {
-		klog.Warningf("addon install failed, wil retry: %v", err)
-		return errors.Wrap(err, "addons")
-	}
-
 	if err := bsutil.AdjustResourceLimits(k.c); err != nil {
 		klog.Warningf("unable to adjust resource limits: %v", err)
 	}