Updating EndpointSlice e2e tests to be less flaky and easier to debug

This adjusts tests that were waiting for Pods to be ready to instead
just wait for them to have IPs assigned to them. This relies on the
associated publishNotReadyAddresses field on Services. Additionally this
increases the the length of time we'll wait for EndpointSlices to be garbage
collected from 12s to 30s. Finally, this adds additional logging to
ExpectNoError calls so it's easier to understand where and why a test
failed.

test/e2e/network/endpointslice.go

@@ -49,14 +49,14 @@ var _ = SIGDescribe("EndpointSlice", func() {
 		namespace := "default"
 		name := "kubernetes"
 		endpoints, err := cs.CoreV1().Endpoints(namespace).Get(context.TODO(), name, metav1.GetOptions{})
-		framework.ExpectNoError(err)
+		framework.ExpectNoError(err, "error creating Endpoints resource")
 		if len(endpoints.Subsets) != 1 {
 			framework.Failf("Expected 1 subset in endpoints, got %d: %#v", len(endpoints.Subsets), endpoints.Subsets)
 		}
 
 		endpointSubset := endpoints.Subsets[0]
 		endpointSlice, err := cs.DiscoveryV1beta1().EndpointSlices(namespace).Get(context.TODO(), name, metav1.GetOptions{})
-		framework.ExpectNoError(err)
+		framework.ExpectNoError(err, "error creating EndpointSlice resource")
 		if len(endpointSlice.Ports) != len(endpointSubset.Ports) {
 			framework.Failf("Expected EndpointSlice to have %d ports, got %d: %#v", len(endpointSubset.Ports), len(endpointSlice.Ports), endpointSlice.Ports)
 		}
@@ -85,7 +85,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 		})
 
 		// Expect Endpoints resource to be created.
-		if err := wait.PollImmediate(2*time.Second, 12*time.Second, func() (bool, error) {
+		if err := wait.PollImmediate(2*time.Second, wait.ForeverTestTimeout, func() (bool, error) {
 			_, err := cs.CoreV1().Endpoints(svc.Namespace).Get(context.TODO(), svc.Name, metav1.GetOptions{})
 			if err != nil {
 				return false, nil
@@ -97,7 +97,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 
 		// Expect EndpointSlice resource to be created.
 		var endpointSlice discoveryv1beta1.EndpointSlice
-		if err := wait.PollImmediate(2*time.Second, 12*time.Second, func() (bool, error) {
+		if err := wait.PollImmediate(2*time.Second, wait.ForeverTestTimeout, func() (bool, error) {
 			endpointSliceList, err := cs.DiscoveryV1beta1().EndpointSlices(svc.Namespace).List(context.TODO(), metav1.ListOptions{
 				LabelSelector: "kubernetes.io/service-name=" + svc.Name,
 			})
@@ -126,10 +126,10 @@ var _ = SIGDescribe("EndpointSlice", func() {
 		}
 
 		err := cs.CoreV1().Services(svc.Namespace).Delete(context.TODO(), svc.Name, metav1.DeleteOptions{})
-		framework.ExpectNoError(err)
+		framework.ExpectNoError(err, "error deleting Service")
 
 		// Expect Endpoints resource to be deleted when Service is.
-		if err := wait.PollImmediate(2*time.Second, 12*time.Second, func() (bool, error) {
+		if err := wait.PollImmediate(2*time.Second, wait.ForeverTestTimeout, func() (bool, error) {
 			_, err := cs.CoreV1().Endpoints(svc.Namespace).Get(context.TODO(), svc.Name, metav1.GetOptions{})
 			if err != nil {
 				if apierrors.IsNotFound(err) {
@@ -143,7 +143,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 		}
 
 		// Expect EndpointSlice resource to be deleted when Service is.
-		if err := wait.PollImmediate(2*time.Second, 12*time.Second, func() (bool, error) {
+		if err := wait.PollImmediate(2*time.Second, wait.ForeverTestTimeout, func() (bool, error) {
 			endpointSliceList, err := cs.DiscoveryV1beta1().EndpointSlices(svc.Namespace).List(context.TODO(), metav1.ListOptions{
 				LabelSelector: "kubernetes.io/service-name=" + svc.Name,
 			})
@@ -219,6 +219,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 			},
 			Spec: v1.ServiceSpec{
 				Selector:                 map[string]string{labelPod1: labelValue},
+				PublishNotReadyAddresses: true,
 				Ports: []v1.ServicePort{{
 					Name:       "example",
 					Port:       80,
@@ -234,6 +235,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 			},
 			Spec: v1.ServiceSpec{
 				Selector:                 map[string]string{labelShared12: labelValue},
+				PublishNotReadyAddresses: true,
 				Ports: []v1.ServicePort{{
 					Name:       "http",
 					Port:       80,
@@ -249,6 +251,7 @@ var _ = SIGDescribe("EndpointSlice", func() {
 			},
 			Spec: v1.ServiceSpec{
 				Selector:                 map[string]string{labelPod3: labelValue},
+				PublishNotReadyAddresses: true,
 				Ports: []v1.ServicePort{{
 					Name:       "example-no-match",
 					Port:       80,
@@ -259,30 +262,26 @@ var _ = SIGDescribe("EndpointSlice", func() {
 		})
 
 		err := wait.Poll(5*time.Second, 3*time.Minute, func() (bool, error) {
-			if !podClient.PodIsReady(pod1.Name) {
-				framework.Logf("Pod 1 not ready yet")
-				return false, nil
-			}
-
-			if !podClient.PodIsReady(pod2.Name) {
-				framework.Logf("Pod 2 not ready yet")
-				return false, nil
-			}
-
 			var err error
 			pod1, err = podClient.Get(context.TODO(), pod1.Name, metav1.GetOptions{})
 			if err != nil {
 				return false, err
 			}
+			if len(pod1.Status.PodIPs) == 0 {
+				return false, nil
+			}
 
 			pod2, err = podClient.Get(context.TODO(), pod2.Name, metav1.GetOptions{})
 			if err != nil {
 				return false, err
 			}
+			if len(pod2.Status.PodIPs) == 0 {
+				return false, nil
+			}
 
 			return true, nil
 		})
-		framework.ExpectNoError(err)
+		framework.ExpectNoError(err, "timed out waiting for Pods to have IPs assigned")
 
 		ginkgo.By("referencing a single matching pod")
 		expectEndpointsAndSlices(cs, f.Namespace.Name, svc1, []*v1.Pod{pod1}, 1, 1, false)
@@ -560,6 +559,6 @@ func ensurePodTargetRef(pod *v1.Pod, targetRef *v1.ObjectReference) {
 // createServiceReportErr creates a Service and reports any associated error.
 func createServiceReportErr(cs clientset.Interface, ns string, service *v1.Service) *v1.Service {
 	svc, err := cs.CoreV1().Services(ns).Create(context.TODO(), service, metav1.CreateOptions{})
-	framework.ExpectNoError(err)
+	framework.ExpectNoError(err, "error deleting Service")
 	return svc
 }