Merge pull request #30318 from QiWang19/drainless-update

OCPNODE-3659: Not fail upgrade checks if all nodes are ready

Author: openshift-merge-bot[bot]

test/e2e/upgrade/upgrade.go

@@ -41,6 +41,7 @@ import (
 	"github.com/openshift/origin/test/e2e/upgrade/adminack"
 	"github.com/openshift/origin/test/e2e/upgrade/dns"
 	"github.com/openshift/origin/test/e2e/upgrade/manifestdelete"
+	mc "github.com/openshift/origin/test/extended/machine_config"
 	"github.com/openshift/origin/test/extended/prometheus"
 	"github.com/openshift/origin/test/extended/util/disruption"
 	"github.com/openshift/origin/test/extended/util/operator"
@@ -625,6 +626,21 @@ func clusterUpgrade(f *framework.Framework, c configv1client.Interface, dc dynam
 		func() (error, bool) {
 			framework.Logf("Waiting on pools to be upgraded")
 			if err := wait.PollImmediate(10*time.Second, 30*time.Minute, func() (bool, error) {
+
+				nodes, err := kubeClient.CoreV1().Nodes().List(context.Background(), metav1.ListOptions{})
+				if err != nil {
+					framework.Logf("error getting nodes %v", err)
+					return false, nil
+				}
+
+				allNodesReady := true
+				for _, node := range nodes.Items {
+					if !mc.IsNodeReady(node) {
+						allNodesReady = false
+						break
+					}
+				}
+
 				mcps := dc.Resource(schema.GroupVersionResource{
 					Group:    "machineconfiguration.openshift.io",
 					Version:  "v1",
@@ -641,7 +657,7 @@ func clusterUpgrade(f *framework.Framework, c configv1client.Interface, dc dynam
 					allUpdated = allUpdated && updated
 
 					// Invariant: when CVO reaches level, MCO is required to have rolled out control plane updates
-					if p.GetName() == "master" && requiresUpdate && errMasterUpdating == nil {
+					if p.GetName() == "master" && requiresUpdate && !allNodesReady && errMasterUpdating == nil {
 						errMasterUpdating = fmt.Errorf("the %q pool should be updated before the CVO reports available at the new version", p.GetName())
 						framework.Logf("Invariant violation detected: %s", errMasterUpdating)
 					}

test/extended/machine_config/helpers.go

@@ -494,7 +494,7 @@ func ValidateMCNForNodeInPool(oc *exutil.CLI, clientSet *machineconfigclient.Cli
 // `GetRandomNode` gets a random node from with a given role and checks whether the node is ready. If no
 // nodes are ready, it will wait for up to 5 minutes for a node to become available.
 func GetRandomNode(oc *exutil.CLI, role string) corev1.Node {
-	if node := getRandomNode(oc, role); isNodeReady(node) {
+	if node := getRandomNode(oc, role); IsNodeReady(node) {
 		return node
 	}
 
@@ -503,7 +503,7 @@ func GetRandomNode(oc *exutil.CLI, role string) corev1.Node {
 	framework.Logf("No ready nodes found with role '%s', waiting up to %s for a ready node to become available", role, waitPeriod)
 	var targetNode corev1.Node
 	o.Eventually(func() bool {
-		if node := getRandomNode(oc, role); isNodeReady(node) {
+		if node := getRandomNode(oc, role); IsNodeReady(node) {
 			targetNode = node
 			return true
 		}
@@ -549,7 +549,7 @@ func GetAllNodes(oc *exutil.CLI) ([]corev1.Node, error) {
 }
 
 // `isNodeReady` determines if a given node is ready
-func isNodeReady(node corev1.Node) bool {
+func IsNodeReady(node corev1.Node) bool {
 	// If the node is cordoned, it is not ready.
 	if node.Spec.Unschedulable {
 		return false
@@ -1038,7 +1038,7 @@ func GetNewReadyNodeInMachine(oc *exutil.CLI, machineName string) (corev1.Node,
 
 		// Check if node is in desiredStatus
 		framework.Logf("Checking if node '%v' is ready.", node.Name)
-		if isNodeReady(node) {
+		if IsNodeReady(node) {
 			framework.Logf("Node '%v' is ready.", node.Name)
 			desiredNode = node
 			err = nil