Merge pull request #40330 from janetkuo/kill-failed-daemon-pods

Automatic merge from submit-queue

DaemonSet controller actively kills failed pods (to recreate them)

Ref #36482, @erictune @yujuhong @mikedanese @kargakis @lukaszo @piosz @kubernetes/sig-apps-bugs 

This also helps with DaemonSet update

```release-note
```

pkg/controller/daemon/daemoncontroller.go

@@ -461,26 +461,42 @@ func (dsc *DaemonSetsController) manage(ds *extensions.DaemonSet) error {
 		return fmt.Errorf("couldn't get list of nodes when syncing daemon set %#v: %v", ds, err)
 	}
 	var nodesNeedingDaemonPods, podsToDelete []string
+	var failedPodsObserved int
 	for _, node := range nodeList.Items {
 		_, shouldSchedule, shouldContinueRunning, err := dsc.nodeShouldRunDaemonPod(&node, ds)
 		if err != nil {
 			continue
 		}
 
-		daemonPods, isRunning := nodeToDaemonPods[node.Name]
+		daemonPods, exists := nodeToDaemonPods[node.Name]
 
 		switch {
-		case shouldSchedule && !isRunning:
+		case shouldSchedule && !exists:
 			// If daemon pod is supposed to be running on node, but isn't, create daemon pod.
 			nodesNeedingDaemonPods = append(nodesNeedingDaemonPods, node.Name)
-		case shouldContinueRunning && len(daemonPods) > 1:
+		case shouldContinueRunning:
-			// If daemon pod is supposed to be running on node, but more than 1 daemon pod is running, delete the excess daemon pods.
+			// If a daemon pod failed, delete it
-			// Sort the daemon pods by creation time, so the the oldest is preserved.
+			// If there's no daemon pods left on this node, we will create it in the next sync loop
-			sort.Sort(podByCreationTimestamp(daemonPods))
+			var daemonPodsRunning []*v1.Pod
-			for i := 1; i < len(daemonPods); i++ {
+			for i := range daemonPods {
-				podsToDelete = append(podsToDelete, daemonPods[i].Name)
+				pod := daemonPods[i]
+				if pod.Status.Phase == v1.PodFailed {
+					glog.V(2).Infof("Found failed daemon pod %s/%s on node %s, will try to kill it", pod.Namespace, node.Name, pod.Name)
+					podsToDelete = append(podsToDelete, pod.Name)
+					failedPodsObserved++
+				} else {
+					daemonPodsRunning = append(daemonPodsRunning, pod)
+				}
 			}
-		case !shouldContinueRunning && isRunning:
+			// If daemon pod is supposed to be running on node, but more than 1 daemon pod is running, delete the excess daemon pods.
+			// Sort the daemon pods by creation time, so the oldest is preserved.
+			if len(daemonPodsRunning) > 1 {
+				sort.Sort(podByCreationTimestamp(daemonPodsRunning))
+				for i := 1; i < len(daemonPodsRunning); i++ {
+					podsToDelete = append(podsToDelete, daemonPods[i].Name)
+				}
+			}
+		case !shouldContinueRunning && exists:
 			// If daemon pod isn't supposed to run on node, but it is, delete all daemon pods on node.
 			for i := range daemonPods {
 				podsToDelete = append(podsToDelete, daemonPods[i].Name)
@@ -547,6 +563,10 @@ func (dsc *DaemonSetsController) manage(ds *extensions.DaemonSet) error {
 	for err := range errCh {
 		errors = append(errors, err)
 	}
+	// Throw an error when the daemon pods fail, to use ratelimiter to prevent kill-recreate hot loop
+	if failedPodsObserved > 0 {
+		errors = append(errors, fmt.Errorf("deleted %d failed pods of DaemonSet %s/%s", failedPodsObserved, ds.Namespace, ds.Name))
+	}
 	return utilerrors.NewAggregate(errors)
 }
 
@@ -774,7 +794,7 @@ func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *exten
 				predicates.ErrTaintsTolerationsNotMatch:
 				return false, false, false, fmt.Errorf("unexpected reason: GeneralPredicates should not return reason %s", reason.GetReason())
 			default:
-				glog.V(4).Infof("unknownd predicate failure reason: %s", reason.GetReason())
+				glog.V(4).Infof("unknown predicate failure reason: %s", reason.GetReason())
 				wantToRun, shouldSchedule, shouldContinueRunning = false, false, false
 				emitEvent = true
 			}

pkg/controller/daemon/daemoncontroller_test.go

@@ -138,6 +138,14 @@ func addPods(podStore cache.Store, nodeName string, label map[string]string, num
 	}
 }
 
+func addFailedPods(podStore cache.Store, nodeName string, label map[string]string, number int) {
+	for i := 0; i < number; i++ {
+		pod := newPod(fmt.Sprintf("%s-", nodeName), nodeName, label)
+		pod.Status = v1.PodStatus{Phase: v1.PodFailed}
+		podStore.Add(pod)
+	}
+}
+
 func newTestController(initialObjects ...runtime.Object) (*DaemonSetsController, *controller.FakePodControl, *fake.Clientset) {
 	clientset := fake.NewSimpleClientset(initialObjects...)
 	informerFactory := informers.NewSharedInformerFactory(clientset, nil, controller.NoResyncPeriodFunc())
@@ -653,6 +661,31 @@ func TestObservedGeneration(t *testing.T) {
 	}
 }
 
+// DaemonSet controller should kill all failed pods and create at most 1 pod on every node.
+func TestDaemonKillFailedPods(t *testing.T) {
+	tests := []struct {
+		numFailedPods, numNormalPods, expectedCreates, expectedDeletes int
+		test                                                           string
+	}{
+		{numFailedPods: 0, numNormalPods: 1, expectedCreates: 0, expectedDeletes: 0, test: "normal (do nothing)"},
+		{numFailedPods: 0, numNormalPods: 0, expectedCreates: 1, expectedDeletes: 0, test: "no pods (create 1)"},
+		{numFailedPods: 1, numNormalPods: 0, expectedCreates: 0, expectedDeletes: 1, test: "1 failed pod (kill 1), 0 normal pod (create 0; will create in the next sync)"},
+		{numFailedPods: 1, numNormalPods: 3, expectedCreates: 0, expectedDeletes: 3, test: "1 failed pod (kill 1), 3 normal pods (kill 2)"},
+		{numFailedPods: 2, numNormalPods: 1, expectedCreates: 0, expectedDeletes: 2, test: "2 failed pods (kill 2), 1 normal pod"},
+	}
+
+	for _, test := range tests {
+		t.Logf("test case: %s\n", test.test)
+		manager, podControl, _ := newTestController()
+		addNodes(manager.nodeStore.Store, 0, 1, nil)
+		addFailedPods(manager.podStore.Indexer, "node-0", simpleDaemonSetLabel, test.numFailedPods)
+		addPods(manager.podStore.Indexer, "node-0", simpleDaemonSetLabel, test.numNormalPods)
+		ds := newDaemonSet("foo")
+		manager.dsStore.Add(ds)
+		syncAndValidateDaemonSets(t, manager, ds, podControl, test.expectedCreates, test.expectedDeletes)
+	}
+}
+
 func TestNodeShouldRunDaemonPod(t *testing.T) {
 	cases := []struct {
 		podsOnNode                                       []*v1.Pod

test/e2e/daemon_set.go

@@ -32,6 +32,7 @@ import (
 	extensionsinternal "k8s.io/kubernetes/pkg/apis/extensions"
 	extensions "k8s.io/kubernetes/pkg/apis/extensions/v1beta1"
 	"k8s.io/kubernetes/pkg/client/clientset_generated/clientset"
+	"k8s.io/kubernetes/pkg/controller"
 	"k8s.io/kubernetes/pkg/kubectl"
 	"k8s.io/kubernetes/test/e2e/framework"
 
@@ -59,6 +60,20 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 	var f *framework.Framework
 
 	AfterEach(func() {
+		// Clean up
+		daemonsets, err := f.ClientSet.Extensions().DaemonSets(f.Namespace.Name).List(metav1.ListOptions{})
+		Expect(err).NotTo(HaveOccurred(), "unable to dump DaemonSets")
+		if daemonsets != nil && len(daemonsets.Items) > 0 {
+			for _, ds := range daemonsets.Items {
+				By(fmt.Sprintf("Deleting DaemonSet %q with reaper", ds.Name))
+				dsReaper, err := kubectl.ReaperFor(extensionsinternal.Kind("DaemonSet"), f.InternalClientset)
+				Expect(err).NotTo(HaveOccurred())
+				err = dsReaper.Stop(f.Namespace.Name, ds.Name, 0, nil)
+				Expect(err).NotTo(HaveOccurred())
+				err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnNoNodes(f, ds.Spec.Template.Labels))
+				Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to be reaped")
+			}
+		}
 		if daemonsets, err := f.ClientSet.Extensions().DaemonSets(f.Namespace.Name).List(metav1.ListOptions{}); err == nil {
 			framework.Logf("daemonset: %s", runtime.EncodeOrDie(api.Codecs.LegacyCodec(api.Registry.EnabledVersions()...), daemonsets))
 		} else {
@@ -69,7 +84,7 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 		} else {
 			framework.Logf("unable to dump pods: %v", err)
 		}
-		err := clearDaemonSetNodeLabels(f.ClientSet)
+		err = clearDaemonSetNodeLabels(f.ClientSet)
 		Expect(err).NotTo(HaveOccurred())
 	})
 
@@ -92,38 +107,9 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 	It("should run and stop simple daemon", func() {
 		label := map[string]string{daemonsetNameLabel: dsName}
 
-		framework.Logf("Creating simple daemon set %s", dsName)
+		By(fmt.Sprintf("Creating simple DaemonSet %q", dsName))
-		_, err := c.Extensions().DaemonSets(ns).Create(&extensions.DaemonSet{
+		_, err := c.Extensions().DaemonSets(ns).Create(newDaemonSet(dsName, image, label))
-			ObjectMeta: metav1.ObjectMeta{
-				Name: dsName,
-			},
-			Spec: extensions.DaemonSetSpec{
-				Template: v1.PodTemplateSpec{
-					ObjectMeta: metav1.ObjectMeta{
-						Labels: label,
-					},
-					Spec: v1.PodSpec{
-						Containers: []v1.Container{
-							{
-								Name:  dsName,
-								Image: image,
-								Ports: []v1.ContainerPort{{ContainerPort: 9376}},
-							},
-						},
-					},
-				},
-			},
-		})
 		Expect(err).NotTo(HaveOccurred())
-		defer func() {
-			framework.Logf("Check that reaper kills all daemon pods for %s", dsName)
-			dsReaper, err := kubectl.ReaperFor(extensionsinternal.Kind("DaemonSet"), f.InternalClientset)
-			Expect(err).NotTo(HaveOccurred())
-			err = dsReaper.Stop(ns, dsName, 0, nil)
-			Expect(err).NotTo(HaveOccurred())
-			err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnNoNodes(f, label))
-			Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to be reaped")
-		}()
 
 		By("Check that daemon pods launch on every node of the cluster.")
 		Expect(err).NotTo(HaveOccurred())
@@ -133,48 +119,21 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 		Expect(err).NotTo(HaveOccurred())
 
 		By("Stop a daemon pod, check that the daemon pod is revived.")
-		podClient := c.Core().Pods(ns)
+		podList := listDaemonPods(c, ns, label)
-
-		selector := labels.Set(label).AsSelector()
-		options := metav1.ListOptions{LabelSelector: selector.String()}
-		podList, err := podClient.List(options)
-		Expect(err).NotTo(HaveOccurred())
-		Expect(len(podList.Items)).To(BeNumerically(">", 0))
 		pod := podList.Items[0]
-		err = podClient.Delete(pod.Name, nil)
+		err = c.Core().Pods(ns).Delete(pod.Name, nil)
 		Expect(err).NotTo(HaveOccurred())
 		err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
 		Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to revive")
-
 	})
 
 	It("should run and stop complex daemon", func() {
 		complexLabel := map[string]string{daemonsetNameLabel: dsName}
 		nodeSelector := map[string]string{daemonsetColorLabel: "blue"}
-		framework.Logf("Creating daemon with a node selector %s", dsName)
+		framework.Logf("Creating daemon %q with a node selector", dsName)
-		_, err := c.Extensions().DaemonSets(ns).Create(&extensions.DaemonSet{
+		ds := newDaemonSet(dsName, image, complexLabel)
-			ObjectMeta: metav1.ObjectMeta{
+		ds.Spec.Template.Spec.NodeSelector = nodeSelector
-				Name: dsName,
+		_, err := c.Extensions().DaemonSets(ns).Create(ds)
-			},
-			Spec: extensions.DaemonSetSpec{
-				Selector: &metav1.LabelSelector{MatchLabels: complexLabel},
-				Template: v1.PodTemplateSpec{
-					ObjectMeta: metav1.ObjectMeta{
-						Labels: complexLabel,
-					},
-					Spec: v1.PodSpec{
-						NodeSelector: nodeSelector,
-						Containers: []v1.Container{
-							{
-								Name:  dsName,
-								Image: image,
-								Ports: []v1.ContainerPort{{ContainerPort: 9376}},
-							},
-						},
-					},
-				},
-			},
-		})
 		Expect(err).NotTo(HaveOccurred())
 
 		By("Initially, daemon pods should not be running on any nodes.")
@@ -198,17 +157,14 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 		Expect(err).NotTo(HaveOccurred(), "error removing labels on node")
 		Expect(wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnNoNodes(f, complexLabel))).
 			NotTo(HaveOccurred(), "error waiting for daemon pod to not be running on nodes")
-
-		By("We should now be able to delete the daemon set.")
-		Expect(c.Extensions().DaemonSets(ns).Delete(dsName, nil)).NotTo(HaveOccurred())
-
 	})
 
 	It("should run and stop complex daemon with node affinity", func() {
 		complexLabel := map[string]string{daemonsetNameLabel: dsName}
 		nodeSelector := map[string]string{daemonsetColorLabel: "blue"}
-		framework.Logf("Creating daemon with a node affinity %s", dsName)
+		framework.Logf("Creating daemon %q with a node affinity", dsName)
-		affinity := &v1.Affinity{
+		ds := newDaemonSet(dsName, image, complexLabel)
+		ds.Spec.Template.Spec.Affinity = &v1.Affinity{
 			NodeAffinity: &v1.NodeAffinity{
 				RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
 					NodeSelectorTerms: []v1.NodeSelectorTerm{
@@ -225,29 +181,7 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 				},
 			},
 		}
-		_, err := c.Extensions().DaemonSets(ns).Create(&extensions.DaemonSet{
+		_, err := c.Extensions().DaemonSets(ns).Create(ds)
-			ObjectMeta: metav1.ObjectMeta{
-				Name: dsName,
-			},
-			Spec: extensions.DaemonSetSpec{
-				Selector: &metav1.LabelSelector{MatchLabels: complexLabel},
-				Template: v1.PodTemplateSpec{
-					ObjectMeta: metav1.ObjectMeta{
-						Labels: complexLabel,
-					},
-					Spec: v1.PodSpec{
-						Affinity: affinity,
-						Containers: []v1.Container{
-							{
-								Name:  dsName,
-								Image: image,
-								Ports: []v1.ContainerPort{{ContainerPort: 9376}},
-							},
-						},
-					},
-				},
-			},
-		})
 		Expect(err).NotTo(HaveOccurred())
 
 		By("Initially, daemon pods should not be running on any nodes.")
@@ -271,13 +205,67 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
 		Expect(err).NotTo(HaveOccurred(), "error removing labels on node")
 		Expect(wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnNoNodes(f, complexLabel))).
 			NotTo(HaveOccurred(), "error waiting for daemon pod to not be running on nodes")
+	})
 
-		By("We should now be able to delete the daemon set.")
+	It("should retry creating failed daemon pods", func() {
-		Expect(c.Extensions().DaemonSets(ns).Delete(dsName, nil)).NotTo(HaveOccurred())
+		label := map[string]string{daemonsetNameLabel: dsName}
 
+		By(fmt.Sprintf("Creating a simple DaemonSet %q", dsName))
+		_, err := c.Extensions().DaemonSets(ns).Create(newDaemonSet(dsName, image, label))
+		Expect(err).NotTo(HaveOccurred())
+
+		By("Check that daemon pods launch on every node of the cluster.")
+		Expect(err).NotTo(HaveOccurred())
+		err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
+		Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")
+		err = checkDaemonStatus(f, dsName)
+		Expect(err).NotTo(HaveOccurred())
+
+		By("Set a daemon pod's phase to 'Failed', check that the daemon pod is revived.")
+		podList := listDaemonPods(c, ns, label)
+		pod := podList.Items[0]
+		pod.ResourceVersion = ""
+		pod.Status.Phase = v1.PodFailed
+		_, err = c.Core().Pods(ns).UpdateStatus(&pod)
+		Expect(err).NotTo(HaveOccurred(), "error failing a daemon pod")
+		err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
+		Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to revive")
 	})
 })
 
+func newDaemonSet(dsName, image string, label map[string]string) *extensions.DaemonSet {
+	return &extensions.DaemonSet{
+		ObjectMeta: metav1.ObjectMeta{
+			Name: dsName,
+		},
+		Spec: extensions.DaemonSetSpec{
+			Template: v1.PodTemplateSpec{
+				ObjectMeta: metav1.ObjectMeta{
+					Labels: label,
+				},
+				Spec: v1.PodSpec{
+					Containers: []v1.Container{
+						{
+							Name:  dsName,
+							Image: image,
+							Ports: []v1.ContainerPort{{ContainerPort: 9376}},
+						},
+					},
+				},
+			},
+		},
+	}
+}
+
+func listDaemonPods(c clientset.Interface, ns string, label map[string]string) *v1.PodList {
+	selector := labels.Set(label).AsSelector()
+	options := metav1.ListOptions{LabelSelector: selector.String()}
+	podList, err := c.Core().Pods(ns).List(options)
+	Expect(err).NotTo(HaveOccurred())
+	Expect(len(podList.Items)).To(BeNumerically(">", 0))
+	return podList
+}
+
 func separateDaemonSetNodeLabels(labels map[string]string) (map[string]string, map[string]string) {
 	daemonSetLabels := map[string]string{}
 	otherLabels := map[string]string{}
@@ -354,7 +342,9 @@ func checkDaemonPodOnNodes(f *framework.Framework, selector map[string]string, n
 
 		nodesToPodCount := make(map[string]int)
 		for _, pod := range pods {
-			nodesToPodCount[pod.Spec.NodeName] += 1
+			if controller.IsPodActive(&pod) {
+				nodesToPodCount[pod.Spec.NodeName] += 1
+			}
 		}
 		framework.Logf("nodesToPodCount: %#v", nodesToPodCount)