Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

In DaemonSet e2e test, don't check nodes with NoSchedule taints #42370

Merged
merged 2 commits into from
Mar 4, 2017
Merged

In DaemonSet e2e test, don't check nodes with NoSchedule taints #42370

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7bdf54a
In DaemonSet e2e test, don't check nodes with NoSchedule taints
janetkuo Mar 2, 2017
16b88e7
Add unit test for daemonset with network unavailable node
janetkuo Mar 2, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
23 changes: 14 additions & 9 deletions pkg/controller/daemon/daemoncontroller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -785,9 +785,7 @@ func (dsc *DaemonSetsController) syncDaemonSet(key string) error {
// Returns true when a daemonset should continue running on a node if a daemonset pod is already
// running on that node.
func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *extensions.DaemonSet) (wantToRun, shouldSchedule, shouldContinueRunning bool, err error) {
newPod := &v1.Pod{Spec: ds.Spec.Template.Spec, ObjectMeta: ds.Spec.Template.ObjectMeta}
newPod.Namespace = ds.Namespace
newPod.Spec.NodeName = node.Name
newPod := NewPod(ds, node.Name)
critical := utilfeature.DefaultFeatureGate.Enabled(features.ExperimentalCriticalPodAnnotation) && kubelettypes.IsCriticalPod(newPod)

// Because these bools require an && of all their required conditions, we start
Expand Down Expand Up @@ -864,14 +862,14 @@ func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *exten

nodeInfo := schedulercache.NewNodeInfo(pods...)
nodeInfo.SetNode(node)
_, reasons, err := daemonSetPredicates(newPod, nodeInfo)
_, reasons, err := Predicates(newPod, nodeInfo)
if err != nil {
glog.Warningf("daemonSetPredicates failed on ds '%s/%s' due to unexpected error: %v", ds.ObjectMeta.Namespace, ds.ObjectMeta.Name, err)
glog.Warningf("DaemonSet Predicates failed on node %s for ds '%s/%s' due to unexpected error: %v", node.Name, ds.ObjectMeta.Namespace, ds.ObjectMeta.Name, err)
return false, false, false, err
}

for _, r := range reasons {
glog.V(4).Infof("daemonSetPredicates failed on ds '%s/%s' for reason: %v", ds.ObjectMeta.Namespace, ds.ObjectMeta.Name, r.GetReason())
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())
switch reason := r.(type) {
case *predicates.InsufficientResourceError:
dsc.eventRecorder.Eventf(ds, v1.EventTypeNormal, FailedPlacementReason, "failed to place pod on %q: %s", node.ObjectMeta.Name, reason.Error())
Expand Down Expand Up @@ -908,7 +906,7 @@ func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *exten
predicates.ErrPodAffinityNotMatch,
predicates.ErrServiceAffinityViolated:
glog.Warningf("unexpected predicate failure reason: %s", reason.GetReason())
return false, false, false, fmt.Errorf("unexpected reason: daemonSetPredicates should not return reason %s", reason.GetReason())
return false, false, false, fmt.Errorf("unexpected reason: DaemonSet Predicates should not return reason %s", reason.GetReason())
default:
glog.V(4).Infof("unknown predicate failure reason: %s", reason.GetReason())
wantToRun, shouldSchedule, shouldContinueRunning = false, false, false
Expand All @@ -922,9 +920,16 @@ func (dsc *DaemonSetsController) nodeShouldRunDaemonPod(node *v1.Node, ds *exten
return
}

// daemonSetPredicates checks if a DaemonSet's pod can be scheduled on a node using GeneralPredicates
func NewPod(ds *extensions.DaemonSet, nodeName string) *v1.Pod {
newPod := &v1.Pod{Spec: ds.Spec.Template.Spec, ObjectMeta: ds.Spec.Template.ObjectMeta}
newPod.Namespace = ds.Namespace
newPod.Spec.NodeName = nodeName
return newPod
}

// Predicates checks if a DaemonSet's pod can be scheduled on a node using GeneralPredicates
// and PodToleratesNodeTaints predicate
func daemonSetPredicates(pod *v1.Pod, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
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)

Expand Down
16 changes: 16 additions & 0 deletions pkg/controller/daemon/daemoncontroller_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -448,6 +448,22 @@ func TestSufficientCapacityNodeDaemonLaunchesPod(t *testing.T) {
syncAndValidateDaemonSets(t, manager, ds, podControl, 1, 0)
}

// DaemonSet should launch a pod on a node with taint NetworkUnavailable condition.
func TestNetworkUnavailableNodeDaemonLaunchesPod(t *testing.T) {
manager, podControl, _ := newTestController()

node := newNode("network-unavailable", nil)
node.Status.Conditions = []v1.NodeCondition{
{Type: v1.NodeNetworkUnavailable, Status: v1.ConditionTrue},
}
manager.nodeStore.Add(node)

ds := newDaemonSet("simple")
manager.dsStore.Add(ds)

syncAndValidateDaemonSets(t, manager, ds, podControl, 1, 0)
}

// DaemonSets not take any actions when being deleted
func TestDontDoAnythingIfBeingDeleted(t *testing.T) {
podSpec := resourcePodSpec("not-too-much-mem", "75M", "75m")
Expand Down
2 changes: 2 additions & 0 deletions test/e2e/BUILD
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -128,6 +128,7 @@ go_library(
"//pkg/cloudprovider/providers/gce:go_default_library",
"//pkg/cloudprovider/providers/vsphere:go_default_library",
"//pkg/controller:go_default_library",
"//pkg/controller/daemon:go_default_library",
"//pkg/controller/deployment/util:go_default_library",
"//pkg/controller/endpoint:go_default_library",
"//pkg/controller/job:go_default_library",
Expand All @@ -149,6 +150,7 @@ go_library(
"//pkg/util/version:go_default_library",
"//pkg/volume/util/volumehelper:go_default_library",
"//plugin/pkg/admission/serviceaccount:go_default_library",
"//plugin/pkg/scheduler/schedulercache:go_default_library",
"//test/e2e/chaosmonkey:go_default_library",
"//test/e2e/common:go_default_library",
"//test/e2e/framework:go_default_library",
Expand Down
49 changes: 34 additions & 15 deletions test/e2e/daemon_set.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,9 @@ import (
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/controller/daemon"
"k8s.io/kubernetes/pkg/kubectl"
"k8s.io/kubernetes/plugin/pkg/scheduler/schedulercache"
"k8s.io/kubernetes/test/e2e/framework"

. "github.com/onsi/ginkgo"
Expand Down Expand Up @@ -109,12 +111,12 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
label := map[string]string{daemonsetNameLabel: dsName}

By(fmt.Sprintf("Creating simple DaemonSet %q", dsName))
_, err := c.Extensions().DaemonSets(ns).Create(newDaemonSet(dsName, image, label))
ds, 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))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")
err = checkDaemonStatus(f, dsName)
Expect(err).NotTo(HaveOccurred())
Expand All @@ -124,7 +126,7 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
pod := podList.Items[0]
err = c.Core().Pods(ns).Delete(pod.Name, nil)
Expect(err).NotTo(HaveOccurred())
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to revive")
})

Expand Down Expand Up @@ -212,12 +214,12 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
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))
ds, 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))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")
err = checkDaemonStatus(f, dsName)
Expect(err).NotTo(HaveOccurred())
Expand All @@ -229,24 +231,24 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {
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))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to revive")
})

It("Should not update pod when spec was updated and update strategy is on delete", func() {
label := map[string]string{daemonsetNameLabel: dsName}

framework.Logf("Creating simple daemon set %s", dsName)
_, err := c.Extensions().DaemonSets(ns).Create(newDaemonSet(dsName, image, label))
ds, 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))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")

By("Update daemon pods image.")
ds, err := c.Extensions().DaemonSets(ns).Get(dsName, metav1.GetOptions{})
ds, err = c.Extensions().DaemonSets(ns).Get(dsName, metav1.GetOptions{})
ds.Spec.Template.Spec.Containers[0].Image = redisImage
_, err = c.Extensions().DaemonSets(ns).Update(ds)
Expect(err).NotTo(HaveOccurred())
Expand All @@ -257,24 +259,24 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {

By("Check that daemon pods are still running on every node of the cluster.")
Expect(err).NotTo(HaveOccurred())
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")
})

It("Should update pod when spec was updated and update strategy is RollingUpdate", func() {
label := map[string]string{daemonsetNameLabel: dsName}

framework.Logf("Creating simple daemon set %s", dsName)
_, err := c.Extensions().DaemonSets(ns).Create(newDaemonSet(dsName, image, label))
ds, 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))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")

By("Update daemon pods image.")
ds, err := c.Extensions().DaemonSets(ns).Get(dsName, metav1.GetOptions{})
ds, err = c.Extensions().DaemonSets(ns).Get(dsName, metav1.GetOptions{})
ds.Spec.Template.Spec.Containers[0].Image = redisImage
ds.Spec.UpdateStrategy = extensions.DaemonSetUpdateStrategy{Type: extensions.RollingUpdateDaemonSetStrategyType}
_, err = c.Extensions().DaemonSets(ns).Update(ds)
Expand All @@ -286,7 +288,7 @@ var _ = framework.KubeDescribe("Daemon set [Serial]", func() {

By("Check that daemon pods are still running on every node of the cluster.")
Expect(err).NotTo(HaveOccurred())
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label))
err = wait.Poll(dsRetryPeriod, dsRetryTimeout, checkRunningOnAllNodes(f, label, ds))
Expect(err).NotTo(HaveOccurred(), "error waiting for daemon pod to start")
})

Expand Down Expand Up @@ -421,18 +423,35 @@ func checkDaemonPodOnNodes(f *framework.Framework, selector map[string]string, n
}
}

func checkRunningOnAllNodes(f *framework.Framework, selector map[string]string) func() (bool, error) {
func checkRunningOnAllNodes(f *framework.Framework, selector map[string]string, ds *extensions.DaemonSet) func() (bool, error) {
return func() (bool, error) {
nodeList, err := f.ClientSet.Core().Nodes().List(metav1.ListOptions{})
framework.ExpectNoError(err)
nodeNames := make([]string, 0)
for _, node := range nodeList.Items {
if !canScheduleOnNode(node, ds) {
framework.Logf("DaemonSet pods can't tolerate node %s with taints %+v, skip checking this node", node, node.Spec.Taints)
continue
}
nodeNames = append(nodeNames, node.Name)
}
return checkDaemonPodOnNodes(f, selector, nodeNames)()
}
}

// canScheduleOnNode checks if a given DaemonSet can schedule pods on the given node
func canScheduleOnNode(node v1.Node, ds *extensions.DaemonSet) bool {
newPod := daemon.NewPod(ds, node.Name)
nodeInfo := schedulercache.NewNodeInfo()
nodeInfo.SetNode(&node)
fit, _, err := daemon.Predicates(newPod, nodeInfo)
if err != nil {
framework.Failf("Can't test DaemonSet predicates for node %s: %v", node.Name, err)
return false
}
return fit
}

func checkRunningOnNoNodes(f *framework.Framework, selector map[string]string) func() (bool, error) {
return checkDaemonPodOnNodes(f, selector, make([]string, 0))
}
Expand Down