Merge pull request #5270 from pablintino/ocpbugs-61016

OCPBUGS-61016: Fix wait failure on MCO pods

Author: openshift-merge-bot[bot]

test/extended/mco_ocb.go

@@ -302,7 +302,10 @@ func ValidateSuccessfulMOSC(mosc *MachineOSConfig, checkers []Checker) {
 	o.Eventually(mOSBuilder.Get, "2m", "30s").WithArguments(`{.status.availableReplicas}`).Should(o.Equal("1"),
 		"The machine-os-builder deployment was created but the available number of replicas is not the expected one")
 
-	exutil.AssertAllPodsToBeReady(mosc.GetOC(), MachineConfigNamespace)
+	// Ensure all pods are running
+	// Do not consider pods created by jobs, as the build pod uses an init container
+	// to build that takes time to be ready and its configuration will be asserted later
+	exutil.AssertAllPodsToBeReadyWithSelector(mosc.GetOC(), MachineConfigNamespace, "!batch.kubernetes.io/job-name")
 	logger.Infof("OK!\n")
 
 	exutil.By("Check that the  machine-os-builder is using leader election without failing")

test/extended/util/pods.go

@@ -96,14 +96,18 @@ func GetPodName(oc *CLI, namespace, podLabel, node string) (string, error) {
 
 // AssertAllPodsToBeReadyWithPollerParams assert all pods in NS are in ready state until timeout in a given namespace
 // Pros: allow user to customize poller parameters
-func AssertAllPodsToBeReadyWithPollerParams(oc *CLI, namespace string, interval, timeout time.Duration) {
+func AssertAllPodsToBeReadyWithPollerParams(oc *CLI, namespace string, interval, timeout time.Duration, selector string) {
 	err := wait.PollUntilContextTimeout(context.Background(), interval, timeout, false, func(_ context.Context) (bool, error) {
 
 		// get the status flag for all pods
 		// except the ones which are in Complete Status.
 		// it use 'ne' operator which is only compatible with 4.10+ oc versions
 		template := "'{{- range .items -}}{{- range .status.conditions -}}{{- if ne .reason \"PodCompleted\" -}}{{- if eq .type \"Ready\" -}}{{- .status}} {{\" \"}}{{- end -}}{{- end -}}{{- end -}}{{- end -}}'"
-		stdout, err := oc.AsAdmin().Run("get").Args("pods", "-n", namespace).Template(template).Output()
+		baseArgs := []string{"pods", "-n", namespace}
+		if selector != "" {
+			baseArgs = append(baseArgs, "-l", selector)
+		}
+		stdout, err := oc.AsAdmin().Run("get").Args(baseArgs...).Template(template).Output()
 		if err != nil {
 			e2e.Logf("the err:%v, and try next round", err)
 			return false, nil
@@ -118,5 +122,11 @@ func AssertAllPodsToBeReadyWithPollerParams(oc *CLI, namespace string, interval,
 
 // AssertAllPodsToBeReady assert all pods in NS are in ready state until timeout in a given namespace
 func AssertAllPodsToBeReady(oc *CLI, namespace string) {
-	AssertAllPodsToBeReadyWithPollerParams(oc, namespace, 10*time.Second, 4*time.Minute)
+	AssertAllPodsToBeReadyWithPollerParams(oc, namespace, 10*time.Second, 4*time.Minute, "")
+}
+
+// AssertAllPodsToBeReadyWithSelector assert all pods in NS are in ready state until timeout in a given namespace
+// The selector parameter follows the regular oc/kubectl format for the --selector option.
+func AssertAllPodsToBeReadyWithSelector(oc *CLI, namespace, selector string) {
+	AssertAllPodsToBeReadyWithPollerParams(oc, namespace, 10*time.Second, 4*time.Minute, selector)
 }