vendor update for CSI 0.3.0

vendor/k8s.io/kubernetes/pkg/controller/daemon/BUILD

@@ -23,7 +23,7 @@ go_library(
         "//pkg/kubelet/types:go_default_library",
         "//pkg/scheduler/algorithm:go_default_library",
         "//pkg/scheduler/algorithm/predicates:go_default_library",
-        "//pkg/scheduler/schedulercache:go_default_library",
+        "//pkg/scheduler/cache:go_default_library",
         "//pkg/util/labels:go_default_library",
         "//pkg/util/metrics:go_default_library",
         "//vendor/github.com/golang/glog:go_default_library",
@@ -33,6 +33,7 @@ go_library(
         "//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/runtime:go_default_library",
+        "//vendor/k8s.io/apimachinery/pkg/types:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/util/errors:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/util/intstr:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/util/json:go_default_library",
@@ -68,6 +69,8 @@ go_test(
         "//pkg/api/v1/pod:go_default_library",
         "//pkg/apis/core:go_default_library",
         "//pkg/controller:go_default_library",
+        "//pkg/features:go_default_library",
+        "//pkg/kubelet/apis:go_default_library",
         "//pkg/kubelet/types:go_default_library",
         "//pkg/scheduler/algorithm:go_default_library",
         "//pkg/securitycontext:go_default_library",

vendor/k8s.io/kubernetes/pkg/controller/daemon/daemon_controller.go

@@ -23,11 +23,14 @@ import (
 	"sync"
 	"time"
 
+	"github.com/golang/glog"
+
 	apps "k8s.io/api/apps/v1"
 	"k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/labels"
+	"k8s.io/apimachinery/pkg/types"
 	utilerrors "k8s.io/apimachinery/pkg/util/errors"
 	utilruntime "k8s.io/apimachinery/pkg/util/runtime"
 	"k8s.io/apimachinery/pkg/util/sets"
@@ -53,10 +56,8 @@ import (
 	kubelettypes "k8s.io/kubernetes/pkg/kubelet/types"
 	"k8s.io/kubernetes/pkg/scheduler/algorithm"
 	"k8s.io/kubernetes/pkg/scheduler/algorithm/predicates"
-	"k8s.io/kubernetes/pkg/scheduler/schedulercache"
+	schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
 	"k8s.io/kubernetes/pkg/util/metrics"
-
-	"github.com/golang/glog"
 )
 
 const (
@@ -112,6 +113,8 @@ type DaemonSetsController struct {
 	historyStoreSynced cache.InformerSynced
 	// podLister get list/get pods from the shared informers's store
 	podLister corelisters.PodLister
+	// podNodeIndex indexes pods by their nodeName
+	podNodeIndex cache.Indexer
 	// podStoreSynced returns true if the pod store has been synced at least once.
 	// Added as a member to the struct to allow injection for testing.
 	podStoreSynced cache.InformerSynced
@@ -134,8 +137,7 @@ type DaemonSetsController struct {
 func NewDaemonSetsController(daemonSetInformer appsinformers.DaemonSetInformer, historyInformer appsinformers.ControllerRevisionInformer, podInformer coreinformers.PodInformer, nodeInformer coreinformers.NodeInformer, kubeClient clientset.Interface) (*DaemonSetsController, error) {
 	eventBroadcaster := record.NewBroadcaster()
 	eventBroadcaster.StartLogging(glog.Infof)
-	// TODO: remove the wrapper when every clients have moved to use the clientset.
-	eventBroadcaster.StartRecordingToSink(&v1core.EventSinkImpl{Interface: v1core.New(kubeClient.CoreV1().RESTClient()).Events("")})
+	eventBroadcaster.StartRecordingToSink(&v1core.EventSinkImpl{Interface: kubeClient.CoreV1().Events("")})
 
 	if kubeClient != nil && kubeClient.CoreV1().RESTClient().GetRateLimiter() != nil {
 		if err := metrics.RegisterMetricAndTrackRateLimiterUsage("daemon_controller", kubeClient.CoreV1().RESTClient().GetRateLimiter()); err != nil {
@@ -191,6 +193,12 @@ func NewDaemonSetsController(daemonSetInformer appsinformers.DaemonSetInformer,
 		DeleteFunc: dsc.deletePod,
 	})
 	dsc.podLister = podInformer.Lister()
+
+	// This custom indexer will index pods based on their NodeName which will decrease the amount of pods we need to get in simulate() call.
+	podInformer.Informer().GetIndexer().AddIndexers(cache.Indexers{
+		"nodeName": indexByPodNodeName,
+	})
+	dsc.podNodeIndex = podInformer.Informer().GetIndexer()
 	dsc.podStoreSynced = podInformer.Informer().HasSynced
 
 	nodeInformer.Informer().AddEventHandler(cache.ResourceEventHandlerFuncs{
@@ -207,6 +215,18 @@ func NewDaemonSetsController(daemonSetInformer appsinformers.DaemonSetInformer,
 	return dsc, nil
 }
 
+func indexByPodNodeName(obj interface{}) ([]string, error) {
+	pod, ok := obj.(*v1.Pod)
+	if !ok {
+		return []string{}, nil
+	}
+	// We are only interested in active pods with nodeName set
+	if len(pod.Spec.NodeName) == 0 || pod.Status.Phase == v1.PodSucceeded || pod.Status.Phase == v1.PodFailed {
+		return []string{}, nil
+	}
+	return []string{pod.Spec.NodeName}, nil
+}
+
 func (dsc *DaemonSetsController) deleteDaemonset(obj interface{}) {
 	ds, ok := obj.(*apps.DaemonSet)
 	if !ok {
@@ -763,7 +783,7 @@ func (dsc *DaemonSetsController) getDaemonPods(ds *apps.DaemonSet) ([]*v1.Pod, e
 	return cm.ClaimPods(pods)
 }
 
-// getNodesToDaemonPods returns a map from nodes to daemon pods (corresponding to ds) running on the nodes.
+// getNodesToDaemonPods returns a map from nodes to daemon pods (corresponding to ds) created for the nodes.
 // This also reconciles ControllerRef by adopting/orphaning.
 // Note that returned Pods are pointers to objects in the cache.
 // If you want to modify one, you need to deep-copy it first.
@@ -775,9 +795,16 @@ func (dsc *DaemonSetsController) getNodesToDaemonPods(ds *apps.DaemonSet) (map[s
 	// Group Pods by Node name.
 	nodeToDaemonPods := make(map[string][]*v1.Pod)
 	for _, pod := range claimedPods {
-		nodeName := pod.Spec.NodeName
+		nodeName, err := util.GetTargetNodeName(pod)
+		if err != nil {
+			glog.Warningf("Failed to get target node name of Pod %v/%v in DaemonSet %v/%v",
+				pod.Namespace, pod.Name, ds.Namespace, ds.Name)
+			continue
+		}
+
 		nodeToDaemonPods[nodeName] = append(nodeToDaemonPods[nodeName], pod)
 	}
+
 	return nodeToDaemonPods, nil
 }
 
@@ -802,12 +829,76 @@ func (dsc *DaemonSetsController) resolveControllerRef(namespace string, controll
 	return ds
 }
 
+// podsShouldBeOnNode figures out the DaemonSet pods to be created and deleted on the given node:
+//   - nodesNeedingDaemonPods: the pods need to start on the node
+//   - podsToDelete: the Pods need to be deleted on the node
+//   - failedPodsObserved: the number of failed pods on node
+//   - err: unexpected error
+func (dsc *DaemonSetsController) podsShouldBeOnNode(
+	node *v1.Node,
+	nodeToDaemonPods map[string][]*v1.Pod,
+	ds *apps.DaemonSet,
+) (nodesNeedingDaemonPods, podsToDelete []string, failedPodsObserved int, err error) {
+
+	wantToRun, shouldSchedule, shouldContinueRunning, err := dsc.nodeShouldRunDaemonPod(node, ds)
+	if err != nil {
+		return
+	}
+
+	daemonPods, exists := nodeToDaemonPods[node.Name]
+	dsKey, _ := cache.MetaNamespaceKeyFunc(ds)
+	dsc.removeSuspendedDaemonPods(node.Name, dsKey)
+
+	switch {
+	case wantToRun && !shouldSchedule:
+		// If daemon pod is supposed to run, but can not be scheduled, add to suspended list.
+		dsc.addSuspendedDaemonPods(node.Name, dsKey)
+	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:
+		// If a daemon pod failed, delete it
+		// If there's non-daemon pods left on this node, we will create it in the next sync loop
+		var daemonPodsRunning []*v1.Pod
+		for _, pod := range daemonPods {
+			if pod.DeletionTimestamp != nil {
+				continue
+			}
+			if pod.Status.Phase == v1.PodFailed {
+				msg := fmt.Sprintf("Found failed daemon pod %s/%s on node %s, will try to kill it", pod.Namespace, pod.Name, node.Name)
+				glog.V(2).Infof(msg)
+				// Emit an event so that it's discoverable to users.
+				dsc.eventRecorder.Eventf(ds, v1.EventTypeWarning, FailedDaemonPodReason, msg)
+				podsToDelete = append(podsToDelete, pod.Name)
+				failedPodsObserved++
+			} else {
+				daemonPodsRunning = append(daemonPodsRunning, pod)
+			}
+		}
+		// 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(podByCreationTimestampAndPhase(daemonPodsRunning))
+			for i := 1; i < len(daemonPodsRunning); i++ {
+				podsToDelete = append(podsToDelete, daemonPodsRunning[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 _, pod := range daemonPods {
+			podsToDelete = append(podsToDelete, pod.Name)
+		}
+	}
+
+	return nodesNeedingDaemonPods, podsToDelete, failedPodsObserved, nil
+}
+
 // manage manages the scheduling and running of Pods of ds on nodes.
 // After figuring out which nodes should run a Pod of ds but not yet running one and
 // which nodes should not run a Pod of ds but currently running one, it calls function
 // syncNodes with a list of pods to remove and a list of nodes to run a Pod of ds.
 func (dsc *DaemonSetsController) manage(ds *apps.DaemonSet, hash string) error {
-	// Find out which nodes are running the daemon pods controlled by ds.
+	// Find out the pods which are created for the nodes by DaemonSet.
 	nodeToDaemonPods, err := dsc.getNodesToDaemonPods(ds)
 	if err != nil {
 		return fmt.Errorf("couldn't get node to daemon pod mapping for daemon set %q: %v", ds.Name, err)
@@ -822,55 +913,16 @@ func (dsc *DaemonSetsController) manage(ds *apps.DaemonSet, hash string) error {
 	var nodesNeedingDaemonPods, podsToDelete []string
 	var failedPodsObserved int
 	for _, node := range nodeList {
-		wantToRun, shouldSchedule, shouldContinueRunning, err := dsc.nodeShouldRunDaemonPod(node, ds)
+		nodesNeedingDaemonPodsOnNode, podsToDeleteOnNode, failedPodsObservedOnNode, err := dsc.podsShouldBeOnNode(
+			node, nodeToDaemonPods, ds)
+
 		if err != nil {
 			continue
 		}
 
-		daemonPods, exists := nodeToDaemonPods[node.Name]
-		dsKey, _ := cache.MetaNamespaceKeyFunc(ds)
-		dsc.removeSuspendedDaemonPods(node.Name, dsKey)
-
-		switch {
-		case wantToRun && !shouldSchedule:
-			// If daemon pod is supposed to run, but can not be scheduled, add to suspended list.
-			dsc.addSuspendedDaemonPods(node.Name, dsKey)
-		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:
-			// If a daemon pod failed, delete it
-			// If there's non-daemon pods left on this node, we will create it in the next sync loop
-			var daemonPodsRunning []*v1.Pod
-			for _, pod := range daemonPods {
-				if pod.DeletionTimestamp != nil {
-					continue
-				}
-				if pod.Status.Phase == v1.PodFailed {
-					msg := fmt.Sprintf("Found failed daemon pod %s/%s on node %s, will try to kill it", pod.Namespace, pod.Name, node.Name)
-					glog.V(2).Infof(msg)
-					// Emit an event so that it's discoverable to users.
-					dsc.eventRecorder.Eventf(ds, v1.EventTypeWarning, FailedDaemonPodReason, msg)
-					podsToDelete = append(podsToDelete, pod.Name)
-					failedPodsObserved++
-				} else {
-					daemonPodsRunning = append(daemonPodsRunning, pod)
-				}
-			}
-			// 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, daemonPodsRunning[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 _, pod := range daemonPods {
-				podsToDelete = append(podsToDelete, pod.Name)
-			}
-		}
+		nodesNeedingDaemonPods = append(nodesNeedingDaemonPods, nodesNeedingDaemonPodsOnNode...)
+		podsToDelete = append(podsToDelete, podsToDeleteOnNode...)
+		failedPodsObserved += failedPodsObservedOnNode
 	}
 
 	// Label new pods using the hash label value of the current history when creating them
@@ -934,7 +986,26 @@ func (dsc *DaemonSetsController) syncNodes(ds *apps.DaemonSet, podsToDelete, nod
 		for i := pos; i < pos+batchSize; i++ {
 			go func(ix int) {
 				defer createWait.Done()
-				err := dsc.podControl.CreatePodsOnNode(nodesNeedingDaemonPods[ix], ds.Namespace, &template, ds, metav1.NewControllerRef(ds, controllerKind))
+				var err error
+
+				podTemplate := &template
+
+				if utilfeature.DefaultFeatureGate.Enabled(features.ScheduleDaemonSetPods) {
+					podTemplate = template.DeepCopy()
+					// The pod's NodeAffinity will be updated to make sure the Pod is bound
+					// to the target node by default scheduler. It is safe to do so because there
+					// should be no conflicting node affinity with the target node.
+					podTemplate.Spec.Affinity = util.ReplaceDaemonSetPodNodeNameNodeAffinity(
+						podTemplate.Spec.Affinity, nodesNeedingDaemonPods[ix])
+					podTemplate.Spec.Tolerations = util.AppendNoScheduleTolerationIfNotExist(podTemplate.Spec.Tolerations)
+
+					err = dsc.podControl.CreatePodsWithControllerRef(ds.Namespace, podTemplate,
+						ds, metav1.NewControllerRef(ds, controllerKind))
+				} else {
+					err = dsc.podControl.CreatePodsOnNode(nodesNeedingDaemonPods[ix], ds.Namespace, podTemplate,
+						ds, metav1.NewControllerRef(ds, controllerKind))
+				}
+
 				if err != nil && errors.IsTimeout(err) {
 					// Pod is created but its initialization has timed out.
 					// If the initialization is successful eventually, the
@@ -1058,7 +1129,7 @@ func (dsc *DaemonSetsController) updateDaemonSetStatus(ds *apps.DaemonSet, hash
 				currentNumberScheduled++
 				// Sort the daemon pods by creation time, so that the oldest is first.
 				daemonPods, _ := nodeToDaemonPods[node.Name]
-				sort.Sort(podByCreationTimestamp(daemonPods))
+				sort.Sort(podByCreationTimestampAndPhase(daemonPods))
 				pod := daemonPods[0]
 				if podutil.IsPodReady(pod) {
 					numberReady++
@@ -1095,7 +1166,7 @@ func (dsc *DaemonSetsController) updateDaemonSetStatus(ds *apps.DaemonSet, hash
 func (dsc *DaemonSetsController) syncDaemonSet(key string) error {
 	startTime := time.Now()
 	defer func() {
-		glog.V(4).Infof("Finished syncing daemon set %q (%v)", key, time.Now().Sub(startTime))
+		glog.V(4).Infof("Finished syncing daemon set %q (%v)", key, time.Since(startTime))
 	}()
 
 	namespace, name, err := cache.SplitMetaNamespaceKey(key)
@@ -1126,6 +1197,18 @@ func (dsc *DaemonSetsController) syncDaemonSet(key string) error {
 		return fmt.Errorf("couldn't get key for object %#v: %v", ds, err)
 	}
 
+	// If the DaemonSet is being deleted (either by foreground deletion or
+	// orphan deletion), we cannot be sure if the DaemonSet history objects
+	// it owned still exist -- those history objects can either be deleted
+	// or orphaned. Garbage collector doesn't guarantee that it will delete
+	// DaemonSet pods before deleting DaemonSet history objects, because
+	// DaemonSet history doesn't own DaemonSet pods. We cannot reliably
+	// calculate the status of a DaemonSet being deleted. Therefore, return
+	// here without updating status for the DaemonSet being deleted.
+	if ds.DeletionTimestamp != nil {
+		return nil
+	}
+
 	// Construct histories of the DaemonSet, and get the hash of current history
 	cur, old, err := dsc.constructHistory(ds)
 	if err != nil {
@@ -1133,7 +1216,7 @@ func (dsc *DaemonSetsController) syncDaemonSet(key string) error {
 	}
 	hash := cur.Labels[apps.DefaultDaemonSetUniqueLabelKey]
 
-	if ds.DeletionTimestamp != nil || !dsc.expectations.SatisfiedExpectations(dsKey) {
+	if !dsc.expectations.SatisfiedExpectations(dsKey) {
 		// Only update status.
 		return dsc.updateDaemonSetStatus(ds, hash)
 	}
@@ -1209,30 +1292,27 @@ func (dsc *DaemonSetsController) simulate(newPod *v1.Pod, node *v1.Node, ds *app
 		})
 	}
 
-	pods := []*v1.Pod{}
-
-	podList, err := dsc.podLister.List(labels.Everything())
+	objects, err := dsc.podNodeIndex.ByIndex("nodeName", node.Name)
 	if err != nil {
 		return nil, nil, err
 	}
-	for _, pod := range podList {
-		if pod.Spec.NodeName != node.Name {
-			continue
-		}
-		if pod.Status.Phase == v1.PodSucceeded || pod.Status.Phase == v1.PodFailed {
-			continue
-		}
-		// ignore pods that belong to the daemonset when taking into account whether
-		// a daemonset should bind to a node.
-		if metav1.IsControlledBy(pod, ds) {
-			continue
-		}
-		pods = append(pods, pod)
-	}
 
-	nodeInfo := schedulercache.NewNodeInfo(pods...)
+	nodeInfo := schedulercache.NewNodeInfo()
 	nodeInfo.SetNode(node)
 
+	for _, obj := range objects {
+		// Ignore pods that belong to the daemonset when taking into account whether a daemonset should bind to a node.
+		// TODO: replace this with metav1.IsControlledBy() in 1.12
+		pod, ok := obj.(*v1.Pod)
+		if !ok {
+			continue
+		}
+		if isControlledByDaemonSet(pod, ds.GetUID()) {
+			continue
+		}
+		nodeInfo.AddPod(pod)
+	}
+
 	_, reasons, err := Predicates(newPod, nodeInfo)
 	return reasons, nodeInfo, err
 }
@@ -1267,6 +1347,9 @@ func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *apps.
 		return false, false, false, err
 	}
 
+	// TODO(k82cn): When 'ScheduleDaemonSetPods' upgrade to beta or GA, remove unnecessary check on failure reason,
+	//              e.g. InsufficientResourceError; and simplify "wantToRun, shouldSchedule, shouldContinueRunning"
+	//              into one result, e.g. selectedNode.
 	var insufficientResourceErr error
 	for _, r := range reasons {
 		glog.V(4).Infof("DaemonSet Predicates failed on node %s for ds '%s/%s' for reason: %v", node.Name, ds.ObjectMeta.Namespace, ds.ObjectMeta.Name, r.GetReason())
@@ -1341,11 +1424,50 @@ func NewPod(ds *apps.DaemonSet, nodeName string) *v1.Pod {
 	return newPod
 }
 
+// nodeSelectionPredicates runs a set of predicates that select candidate nodes for the DaemonSet;
+// the predicates include:
+//   - PodFitsHost: checks pod's NodeName against node
+//   - PodMatchNodeSelector: checks pod's NodeSelector and NodeAffinity against node
+func nodeSelectionPredicates(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
+	var predicateFails []algorithm.PredicateFailureReason
+	fit, reasons, err := predicates.PodFitsHost(pod, meta, nodeInfo)
+	if err != nil {
+		return false, predicateFails, err
+	}
+	if !fit {
+		predicateFails = append(predicateFails, reasons...)
+	}
+
+	fit, reasons, err = predicates.PodMatchNodeSelector(pod, meta, nodeInfo)
+	if err != nil {
+		return false, predicateFails, err
+	}
+	if !fit {
+		predicateFails = append(predicateFails, reasons...)
+	}
+	return len(predicateFails) == 0, predicateFails, nil
+}
+
 // Predicates checks if a DaemonSet's pod can be scheduled on a node using GeneralPredicates
 // and PodToleratesNodeTaints predicate
 func Predicates(pod *v1.Pod, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
 	var predicateFails []algorithm.PredicateFailureReason
-	critical := utilfeature.DefaultFeatureGate.Enabled(features.ExperimentalCriticalPodAnnotation) && kubelettypes.IsCriticalPod(pod)
+
+	// If ScheduleDaemonSetPods is enabled, only check nodeSelector and nodeAffinity.
+	if utilfeature.DefaultFeatureGate.Enabled(features.ScheduleDaemonSetPods) {
+		fit, reasons, err := nodeSelectionPredicates(pod, nil, nodeInfo)
+		if err != nil {
+			return false, predicateFails, err
+		}
+		if !fit {
+			predicateFails = append(predicateFails, reasons...)
+		}
+
+		return len(predicateFails) == 0, predicateFails, nil
+	}
+
+	critical := utilfeature.DefaultFeatureGate.Enabled(features.ExperimentalCriticalPodAnnotation) &&
+		kubelettypes.IsCriticalPod(pod)
 
 	fit, reasons, err := predicates.PodToleratesNodeTaints(pod, nil, nodeInfo)
 	if err != nil {
@@ -1384,14 +1506,32 @@ func (o byCreationTimestamp) Less(i, j int) bool {
 	return o[i].CreationTimestamp.Before(&o[j].CreationTimestamp)
 }
 
-type podByCreationTimestamp []*v1.Pod
+type podByCreationTimestampAndPhase []*v1.Pod
 
-func (o podByCreationTimestamp) Len() int      { return len(o) }
-func (o podByCreationTimestamp) Swap(i, j int) { o[i], o[j] = o[j], o[i] }
+func (o podByCreationTimestampAndPhase) Len() int      { return len(o) }
+func (o podByCreationTimestampAndPhase) Swap(i, j int) { o[i], o[j] = o[j], o[i] }
+
+func (o podByCreationTimestampAndPhase) Less(i, j int) bool {
+	// Scheduled Pod first
+	if len(o[i].Spec.NodeName) != 0 && len(o[j].Spec.NodeName) == 0 {
+		return true
+	}
+
+	if len(o[i].Spec.NodeName) == 0 && len(o[j].Spec.NodeName) != 0 {
+		return false
+	}
 
-func (o podByCreationTimestamp) Less(i, j int) bool {
 	if o[i].CreationTimestamp.Equal(&o[j].CreationTimestamp) {
 		return o[i].Name < o[j].Name
 	}
 	return o[i].CreationTimestamp.Before(&o[j].CreationTimestamp)
 }
+
+func isControlledByDaemonSet(p *v1.Pod, uuid types.UID) bool {
+	for _, ref := range p.OwnerReferences {
+		if ref.Controller != nil && *ref.Controller && ref.UID == uuid {
+			return true
+		}
+	}
+	return false
+}

vendor/k8s.io/kubernetes/pkg/controller/daemon/daemon_controller_test.go

@@ -43,6 +43,8 @@ import (
 	podutil "k8s.io/kubernetes/pkg/api/v1/pod"
 	api "k8s.io/kubernetes/pkg/apis/core"
 	"k8s.io/kubernetes/pkg/controller"
+	"k8s.io/kubernetes/pkg/features"
+	kubeletapis "k8s.io/kubernetes/pkg/kubelet/apis"
 	kubelettypes "k8s.io/kubernetes/pkg/kubelet/types"
 	"k8s.io/kubernetes/pkg/scheduler/algorithm"
 	"k8s.io/kubernetes/pkg/securitycontext"
@@ -145,7 +147,7 @@ func updateStrategies() []*apps.DaemonSetUpdateStrategy {
 
 func newNode(name string, label map[string]string) *v1.Node {
 	return &v1.Node{
-		TypeMeta: metav1.TypeMeta{APIVersion: legacyscheme.Registry.GroupOrDie(v1.GroupName).GroupVersion.String()},
+		TypeMeta: metav1.TypeMeta{APIVersion: "v1"},
 		ObjectMeta: metav1.ObjectMeta{
 			Name:      name,
 			Labels:    label,
@@ -195,7 +197,7 @@ func newPod(podName string, nodeName string, label map[string]string, ds *apps.D
 	}
 
 	pod := &v1.Pod{
-		TypeMeta: metav1.TypeMeta{APIVersion: legacyscheme.Registry.GroupOrDie(v1.GroupName).GroupVersion.String()},
+		TypeMeta: metav1.TypeMeta{APIVersion: "v1"},
 		ObjectMeta: metav1.ObjectMeta{
 			GenerateName: podName,
 			Labels:       newLabels,
@@ -267,6 +269,31 @@ func (f *fakePodControl) CreatePodsOnNode(nodeName, namespace string, template *
 	return nil
 }
 
+func (f *fakePodControl) CreatePodsWithControllerRef(namespace string, template *v1.PodTemplateSpec, object runtime.Object, controllerRef *metav1.OwnerReference) error {
+	f.Lock()
+	defer f.Unlock()
+	if err := f.FakePodControl.CreatePodsWithControllerRef(namespace, template, object, controllerRef); err != nil {
+		return fmt.Errorf("failed to create pod for DaemonSet")
+	}
+
+	pod := &v1.Pod{
+		ObjectMeta: metav1.ObjectMeta{
+			Labels:    template.Labels,
+			Namespace: namespace,
+		},
+	}
+
+	pod.Name = names.SimpleNameGenerator.GenerateName(fmt.Sprintf("%p-", pod))
+
+	if err := legacyscheme.Scheme.Convert(&template.Spec, &pod.Spec, nil); err != nil {
+		return fmt.Errorf("unable to convert pod template: %v", err)
+	}
+
+	f.podStore.Update(pod)
+	f.podIDMap[pod.Name] = pod
+	return nil
+}
+
 func (f *fakePodControl) DeletePod(namespace string, podID string, object runtime.Object) error {
 	f.Lock()
 	defer f.Unlock()
@@ -423,6 +450,97 @@ func TestSimpleDaemonSetLaunchesPods(t *testing.T) {
 	}
 }
 
+// When ScheduleDaemonSetPods is enabled, DaemonSets without node selectors should
+// launch pods on every node by NodeAffinity.
+func TestSimpleDaemonSetScheduleDaemonSetPodsLaunchesPods(t *testing.T) {
+	t.Skip("disabled for 1.10")
+
+	enabled := utilfeature.DefaultFeatureGate.Enabled(features.ScheduleDaemonSetPods)
+	// Rollback feature gate.
+	defer func() {
+		if !enabled {
+			utilfeature.DefaultFeatureGate.Set("ScheduleDaemonSetPods=false")
+		}
+	}()
+
+	utilfeature.DefaultFeatureGate.Set("ScheduleDaemonSetPods=true")
+
+	nodeNum := 5
+
+	for _, strategy := range updateStrategies() {
+		ds := newDaemonSet("foo")
+		ds.Spec.UpdateStrategy = *strategy
+		manager, podControl, _, err := newTestController(ds)
+		if err != nil {
+			t.Fatalf("error creating DaemonSets controller: %v", err)
+		}
+		addNodes(manager.nodeStore, 0, nodeNum, nil)
+		manager.dsStore.Add(ds)
+		syncAndValidateDaemonSets(t, manager, ds, podControl, nodeNum, 0, 0)
+
+		// Check for ScheduleDaemonSetPods feature
+		if len(podControl.podIDMap) != nodeNum {
+			t.Fatalf("failed to create pods for DaemonSet when enabled ScheduleDaemonSetPods.")
+		}
+
+		nodeMap := make(map[string]*v1.Node)
+		for _, node := range manager.nodeStore.List() {
+			n := node.(*v1.Node)
+			nodeMap[n.Name] = n
+		}
+
+		if len(nodeMap) != nodeNum {
+			t.Fatalf("not enough nodes in the store, expected: %v, got: %v",
+				nodeNum, len(nodeMap))
+		}
+
+		for _, pod := range podControl.podIDMap {
+			if len(pod.Spec.NodeName) != 0 {
+				t.Fatalf("the hostname of pod %v should be empty, but got %s",
+					pod.Name, pod.Spec.NodeName)
+			}
+			if pod.Spec.Affinity == nil {
+				t.Fatalf("the Affinity of pod %s is nil.", pod.Name)
+			}
+			if pod.Spec.Affinity.NodeAffinity == nil {
+				t.Fatalf("the NodeAffinity of pod %s is nil.", pod.Name)
+			}
+
+			nodeSelector := pod.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution
+			if nodeSelector == nil {
+				t.Fatalf("the node selector of pod %s is nil.", pod.Name)
+			}
+			if len(nodeSelector.NodeSelectorTerms) != 1 {
+				t.Fatalf("incorrect node selector terms number of pod %s, expected: 1, got: %d.",
+					pod.Name, len(nodeSelector.NodeSelectorTerms))
+			}
+
+			if len(nodeSelector.NodeSelectorTerms[0].MatchExpressions) != 1 {
+				t.Fatalf("incorrect expression number of pod %s node selector term, expected: 1, got: %d.",
+					pod.Name, len(nodeSelector.NodeSelectorTerms[0].MatchExpressions))
+			}
+
+			exp := nodeSelector.NodeSelectorTerms[0].MatchExpressions[0]
+			if exp.Key == kubeletapis.LabelHostname {
+				if exp.Operator != v1.NodeSelectorOpIn {
+					t.Fatalf("the operation of hostname NodeAffinity is not %v", v1.NodeSelectorOpIn)
+				}
+
+				if len(exp.Values) != 1 {
+					t.Fatalf("incorrect hostname in node affinity: expected 1, got %v", len(exp.Values))
+				}
+
+				delete(nodeMap, exp.Values[0])
+			}
+		}
+
+		if len(nodeMap) != 0 {
+			t.Fatalf("did not foud pods on nodes %+v", nodeMap)
+		}
+	}
+
+}
+
 // Simulate a cluster with 100 nodes, but simulate a limit (like a quota limit)
 // of 10 pods, and verify that the ds doesn't make 100 create calls per sync pass
 func TestSimpleDaemonSetPodCreateErrors(t *testing.T) {
@@ -1545,6 +1663,7 @@ func TestNodeShouldRunDaemonPod(t *testing.T) {
 		predicateName                                    string
 		podsOnNode                                       []*v1.Pod
 		nodeCondition                                    []v1.NodeCondition
+		nodeUnschedulable                                bool
 		ds                                               *apps.DaemonSet
 		wantToRun, shouldSchedule, shouldContinueRunning bool
 		err                                              error
@@ -1800,6 +1919,24 @@ func TestNodeShouldRunDaemonPod(t *testing.T) {
 			shouldSchedule:        true,
 			shouldContinueRunning: true,
 		},
+		{
+			predicateName: "ShouldRunDaemonPodOnUnscheduableNode",
+			ds: &apps.DaemonSet{
+				Spec: apps.DaemonSetSpec{
+					Selector: &metav1.LabelSelector{MatchLabels: simpleDaemonSetLabel},
+					Template: v1.PodTemplateSpec{
+						ObjectMeta: metav1.ObjectMeta{
+							Labels: simpleDaemonSetLabel,
+						},
+						Spec: resourcePodSpec("", "50M", "0.5"),
+					},
+				},
+			},
+			nodeUnschedulable:     true,
+			wantToRun:             true,
+			shouldSchedule:        true,
+			shouldContinueRunning: true,
+		},
 	}
 
 	for i, c := range cases {
@@ -1807,6 +1944,7 @@ func TestNodeShouldRunDaemonPod(t *testing.T) {
 			node := newNode("test-node", simpleDaemonSetLabel)
 			node.Status.Conditions = append(node.Status.Conditions, c.nodeCondition...)
 			node.Status.Allocatable = allocatableResources("100M", "1")
+			node.Spec.Unschedulable = c.nodeUnschedulable
 			manager, _, _, err := newTestController()
 			if err != nil {
 				t.Fatalf("error creating DaemonSets controller: %v", err)
@@ -1815,6 +1953,7 @@ func TestNodeShouldRunDaemonPod(t *testing.T) {
 			for _, p := range c.podsOnNode {
 				manager.podStore.Add(p)
 				p.Spec.NodeName = "test-node"
+				manager.podNodeIndex.Add(p)
 			}
 			c.ds.Spec.UpdateStrategy = *strategy
 			wantToRun, shouldSchedule, shouldContinueRunning, err := manager.nodeShouldRunDaemonPod(node, c.ds)

vendor/k8s.io/kubernetes/pkg/controller/daemon/util/BUILD

@@ -19,6 +19,7 @@ go_library(
         "//vendor/k8s.io/api/apps/v1:go_default_library",
         "//vendor/k8s.io/api/core/v1:go_default_library",
         "//vendor/k8s.io/api/extensions/v1beta1:go_default_library",
+        "//vendor/k8s.io/apimachinery/pkg/api/equality:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
         "//vendor/k8s.io/apiserver/pkg/util/feature:go_default_library",
     ],
@@ -43,8 +44,12 @@ go_test(
     embed = [":go_default_library"],
     deps = [
         "//pkg/api/testapi:go_default_library",
+        "//pkg/features:go_default_library",
+        "//pkg/kubelet/apis:go_default_library",
+        "//pkg/scheduler/algorithm:go_default_library",
         "//vendor/k8s.io/api/core/v1:go_default_library",
         "//vendor/k8s.io/api/extensions/v1beta1:go_default_library",
         "//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
+        "//vendor/k8s.io/apiserver/pkg/util/feature:go_default_library",
     ],
 )

vendor/k8s.io/kubernetes/pkg/controller/daemon/util/daemonset_util.go

@@ -23,6 +23,7 @@ import (
 	apps "k8s.io/api/apps/v1"
 	"k8s.io/api/core/v1"
 	extensions "k8s.io/api/extensions/v1beta1"
+	apiequality "k8s.io/apimachinery/pkg/api/equality"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	utilfeature "k8s.io/apiserver/pkg/util/feature"
 	podutil "k8s.io/kubernetes/pkg/api/v1/pod"
@@ -133,3 +134,117 @@ func SplitByAvailablePods(minReadySeconds int32, pods []*v1.Pod) ([]*v1.Pod, []*
 	}
 	return availablePods, unavailablePods
 }
+
+// ReplaceDaemonSetPodNodeNameNodeAffinity replaces the RequiredDuringSchedulingIgnoredDuringExecution
+// NodeAffinity of the given affinity with a new NodeAffinity that selects the given nodeName.
+// Note that this function assumes that no NodeAffinity conflicts with the selected nodeName.
+func ReplaceDaemonSetPodNodeNameNodeAffinity(affinity *v1.Affinity, nodename string) *v1.Affinity {
+	nodeSelReq := v1.NodeSelectorRequirement{
+		Key:      algorithm.NodeFieldSelectorKeyNodeName,
+		Operator: v1.NodeSelectorOpIn,
+		Values:   []string{nodename},
+	}
+
+	nodeSelector := &v1.NodeSelector{
+		NodeSelectorTerms: []v1.NodeSelectorTerm{
+			{
+				MatchFields: []v1.NodeSelectorRequirement{nodeSelReq},
+			},
+		},
+	}
+
+	if affinity == nil {
+		return &v1.Affinity{
+			NodeAffinity: &v1.NodeAffinity{
+				RequiredDuringSchedulingIgnoredDuringExecution: nodeSelector,
+			},
+		}
+	}
+
+	if affinity.NodeAffinity == nil {
+		affinity.NodeAffinity = &v1.NodeAffinity{
+			RequiredDuringSchedulingIgnoredDuringExecution: nodeSelector,
+		}
+		return affinity
+	}
+
+	nodeAffinity := affinity.NodeAffinity
+
+	if nodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution == nil {
+		nodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution = nodeSelector
+		return affinity
+	}
+
+	// Replace node selector with the new one.
+	nodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms = []v1.NodeSelectorTerm{
+		{
+			MatchFields: []v1.NodeSelectorRequirement{nodeSelReq},
+		},
+	}
+
+	return affinity
+}
+
+// AppendNoScheduleTolerationIfNotExist appends unschedulable toleration to `.spec` if not exist; otherwise,
+// no changes to `.spec.tolerations`.
+func AppendNoScheduleTolerationIfNotExist(tolerations []v1.Toleration) []v1.Toleration {
+	unschedulableToleration := v1.Toleration{
+		Key:      algorithm.TaintNodeUnschedulable,
+		Operator: v1.TolerationOpExists,
+		Effect:   v1.TaintEffectNoSchedule,
+	}
+
+	unschedulableTaintExist := false
+
+	for _, t := range tolerations {
+		if apiequality.Semantic.DeepEqual(t, unschedulableToleration) {
+			unschedulableTaintExist = true
+			break
+		}
+	}
+
+	if !unschedulableTaintExist {
+		tolerations = append(tolerations, unschedulableToleration)
+	}
+
+	return tolerations
+}
+
+// GetTargetNodeName get the target node name of DaemonSet pods. If `.spec.NodeName` is not empty (nil),
+// return `.spec.NodeName`; otherwise, retrieve node name of pending pods from NodeAffinity. Return error
+// if failed to retrieve node name from `.spec.NodeName` and NodeAffinity.
+func GetTargetNodeName(pod *v1.Pod) (string, error) {
+	if len(pod.Spec.NodeName) != 0 {
+		return pod.Spec.NodeName, nil
+	}
+
+	// If ScheduleDaemonSetPods was enabled before, retrieve node name of unscheduled pods from NodeAffinity
+	if pod.Spec.Affinity == nil ||
+		pod.Spec.Affinity.NodeAffinity == nil ||
+		pod.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution == nil {
+		return "", fmt.Errorf("no spec.affinity.nodeAffinity.requiredDuringSchedulingIgnoredDuringExecution for pod %s/%s",
+			pod.Namespace, pod.Name)
+	}
+
+	terms := pod.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms
+	if len(terms) < 1 {
+		return "", fmt.Errorf("no nodeSelectorTerms in requiredDuringSchedulingIgnoredDuringExecution of pod %s/%s",
+			pod.Namespace, pod.Name)
+	}
+
+	for _, term := range terms {
+		for _, exp := range term.MatchFields {
+			if exp.Key == algorithm.NodeFieldSelectorKeyNodeName &&
+				exp.Operator == v1.NodeSelectorOpIn {
+				if len(exp.Values) != 1 {
+					return "", fmt.Errorf("the matchFields value of '%s' is not unique for pod %s/%s",
+						algorithm.NodeFieldSelectorKeyNodeName, pod.Namespace, pod.Name)
+				}
+
+				return exp.Values[0], nil
+			}
+		}
+	}
+
+	return "", fmt.Errorf("no node name found for pod %s/%s", pod.Namespace, pod.Name)
+}

vendor/k8s.io/kubernetes/pkg/controller/daemon/util/daemonset_util_test.go

@@ -18,12 +18,17 @@ package util
 
 import (
 	"fmt"
+	"reflect"
 	"testing"
 
 	"k8s.io/api/core/v1"
 	extensions "k8s.io/api/extensions/v1beta1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	utilfeature "k8s.io/apiserver/pkg/util/feature"
 	"k8s.io/kubernetes/pkg/api/testapi"
+	"k8s.io/kubernetes/pkg/features"
+	kubeletapis "k8s.io/kubernetes/pkg/kubelet/apis"
+	"k8s.io/kubernetes/pkg/scheduler/algorithm"
 )
 
 func newPod(podName string, nodeName string, label map[string]string) *v1.Pod {
@@ -168,3 +173,427 @@ func int64Ptr(i int) *int64 {
 	li := int64(i)
 	return &li
 }
+
+func TestReplaceDaemonSetPodNodeNameNodeAffinity(t *testing.T) {
+	tests := []struct {
+		affinity *v1.Affinity
+		hostname string
+		expected *v1.Affinity
+	}{
+		{
+			affinity: nil,
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchExpressions: []v1.NodeSelectorRequirement{
+									{
+										Key:      kubeletapis.LabelHostname,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					PreferredDuringSchedulingIgnoredDuringExecution: []v1.PreferredSchedulingTerm{
+						{
+							Preference: v1.NodeSelectorTerm{
+								MatchExpressions: []v1.NodeSelectorRequirement{
+									{
+										Key:      kubeletapis.LabelHostname,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					PreferredDuringSchedulingIgnoredDuringExecution: []v1.PreferredSchedulingTerm{
+						{
+							Preference: v1.NodeSelectorTerm{
+								MatchExpressions: []v1.NodeSelectorRequirement{
+									{
+										Key:      kubeletapis.LabelHostname,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1", "host_2"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: nil,
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchExpressions: []v1.NodeSelectorRequirement{
+									{
+										Key:      "hostname",
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_2"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpNotIn,
+										Values:   []string{"host_2"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+		{
+			affinity: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										// NOTE: Only `metadata.name` is valid key in `MatchFields` in 1.11;
+										//       added this case for compatibility: the feature works as normal
+										//       when new Keys introduced.
+										Key:      "metadata.foo",
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"bar"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			hostname: "host_1",
+			expected: &v1.Affinity{
+				NodeAffinity: &v1.NodeAffinity{
+					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+						NodeSelectorTerms: []v1.NodeSelectorTerm{
+							{
+								MatchFields: []v1.NodeSelectorRequirement{
+									{
+										Key:      algorithm.NodeFieldSelectorKeyNodeName,
+										Operator: v1.NodeSelectorOpIn,
+										Values:   []string{"host_1"},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+	}
+
+	for i, test := range tests {
+		got := ReplaceDaemonSetPodNodeNameNodeAffinity(test.affinity, test.hostname)
+		if !reflect.DeepEqual(test.expected, got) {
+			t.Errorf("Failed to append NodeAffinity in case %d, got: %v, expected: %v",
+				i, got, test.expected)
+		}
+	}
+}
+
+func forEachFeatureGate(t *testing.T, tf func(t *testing.T), gates ...utilfeature.Feature) {
+	for _, fg := range gates {
+		func() {
+			enabled := utilfeature.DefaultFeatureGate.Enabled(fg)
+			defer func() {
+				utilfeature.DefaultFeatureGate.Set(fmt.Sprintf("%v=%t", fg, enabled))
+			}()
+
+			for _, f := range []bool{true, false} {
+				utilfeature.DefaultFeatureGate.Set(fmt.Sprintf("%v=%t", fg, f))
+				t.Run(fmt.Sprintf("%v (%t)", fg, f), tf)
+			}
+		}()
+	}
+}
+
+func TestGetTargetNodeName(t *testing.T) {
+	testFun := func(t *testing.T) {
+		tests := []struct {
+			pod         *v1.Pod
+			nodeName    string
+			expectedErr bool
+		}{
+			{
+				pod: &v1.Pod{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "pod1",
+						Namespace: "default",
+					},
+					Spec: v1.PodSpec{
+						NodeName: "node-1",
+					},
+				},
+				nodeName: "node-1",
+			},
+			{
+				pod: &v1.Pod{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "pod2",
+						Namespace: "default",
+					},
+					Spec: v1.PodSpec{
+						Affinity: &v1.Affinity{
+							NodeAffinity: &v1.NodeAffinity{
+								RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+									NodeSelectorTerms: []v1.NodeSelectorTerm{
+										{
+											MatchFields: []v1.NodeSelectorRequirement{
+												{
+													Key:      algorithm.NodeFieldSelectorKeyNodeName,
+													Operator: v1.NodeSelectorOpIn,
+													Values:   []string{"node-1"},
+												},
+											},
+										},
+									},
+								},
+							},
+						},
+					},
+				},
+				nodeName: "node-1",
+			},
+			{
+				pod: &v1.Pod{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "pod3",
+						Namespace: "default",
+					},
+					Spec: v1.PodSpec{
+						Affinity: &v1.Affinity{
+							NodeAffinity: &v1.NodeAffinity{
+								RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+									NodeSelectorTerms: []v1.NodeSelectorTerm{
+										{
+											MatchFields: []v1.NodeSelectorRequirement{
+												{
+													Key:      algorithm.NodeFieldSelectorKeyNodeName,
+													Operator: v1.NodeSelectorOpIn,
+													Values:   []string{"node-1", "node-2"},
+												},
+											},
+										},
+									},
+								},
+							},
+						},
+					},
+				},
+				expectedErr: true,
+			},
+			{
+				pod: &v1.Pod{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "pod4",
+						Namespace: "default",
+					},
+					Spec: v1.PodSpec{},
+				},
+				expectedErr: true,
+			},
+		}
+
+		for _, test := range tests {
+			got, err := GetTargetNodeName(test.pod)
+			if test.expectedErr != (err != nil) {
+				t.Errorf("Unexpected error, expectedErr: %v, err: %v", test.expectedErr, err)
+			} else if !test.expectedErr {
+				if test.nodeName != got {
+					t.Errorf("Failed to get target node name, got: %v, expected: %v", got, test.nodeName)
+				}
+			}
+		}
+	}
+
+	forEachFeatureGate(t, testFun, features.ScheduleDaemonSetPods)
+}