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

fix terminationGracePeriod blocked by preStop #115835

Merged
merged 2 commits into from
Jun 14, 2023
Merged

fix terminationGracePeriod blocked by preStop #115835

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
b4dfae2
fix terminationGracePeriod blocked by preStop
HirazawaUi Feb 16, 2023
f520f6f
Add Pod Termination e2e tests to test termination grace period seconds
gjkim42 Mar 24, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 7 additions & 5 deletions pkg/kubelet/kuberuntime/kuberuntime_container.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -725,19 +725,21 @@ func (m *kubeGenericRuntimeManager) killContainer(ctx context.Context, pod *v1.P
}
m.recordContainerEvent(pod, containerSpec, containerID.ID, v1.EventTypeNormal, events.KillingContainer, message)

if gracePeriodOverride != nil {
gracePeriod = *gracePeriodOverride
klog.V(3).InfoS("Killing container with a grace period override", "pod", klog.KObj(pod), "podUID", pod.UID,
"containerName", containerName, "containerID", containerID.String(), "gracePeriod", gracePeriod)
}

// Run the pre-stop lifecycle hooks if applicable and if there is enough time to run it
if containerSpec.Lifecycle != nil && containerSpec.Lifecycle.PreStop != nil && gracePeriod > 0 {
gracePeriod = gracePeriod - m.executePreStopHook(ctx, pod, containerID, containerSpec, gracePeriod)
}

// always give containers a minimal shutdown window to avoid unnecessary SIGKILLs
if gracePeriod < minimumGracePeriodInSeconds {
gracePeriod = minimumGracePeriodInSeconds
}
if gracePeriodOverride != nil {
gracePeriod = *gracePeriodOverride
klog.V(3).InfoS("Killing container with a grace period override", "pod", klog.KObj(pod), "podUID", pod.UID,
"containerName", containerName, "containerID", containerID.String(), "gracePeriod", gracePeriod)
}

klog.V(2).InfoS("Killing container with a grace period", "pod", klog.KObj(pod), "podUID", pod.UID,
"containerName", containerName, "containerID", containerID.String(), "gracePeriod", gracePeriod)
Expand Down
80 changes: 80 additions & 0 deletions test/e2e_node/container_lifecycle_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -515,4 +515,84 @@ var _ = SIGDescribe("[NodeConformance] Containers Lifecycle ", func() {
// while the regular container did
framework.ExpectNoError(results.HasRestarted(regular1))
})

ginkgo.When("a pod cannot terminate gracefully", func() {
testPod := func(name string, gracePeriod int64) *v1.Pod {
return &v1.Pod{
ObjectMeta: metav1.ObjectMeta{
Name: name,
},
Spec: v1.PodSpec{
Containers: []v1.Container{
{
Name: "busybox",
Image: imageutils.GetE2EImage(imageutils.BusyBox),
Command: []string{
"sleep",
"10000",
},
},
},
TerminationGracePeriodSeconds: &gracePeriod,
},
}
}

// To account for the time it takes to delete the pod, we add a 10 second
// buffer. The 10 second buffer is arbitrary, but it should be enough to
// account for the time it takes to delete the pod.
bufferSeconds := int64(10)

ginkgo.It("should respect termination grace period seconds [NodeConformance]", func() {
client := e2epod.NewPodClient(f)
gracePeriod := int64(30)

ginkgo.By("creating a pod with a termination grace period seconds")
pod := testPod("pod-termination-grace-period", gracePeriod)
pod = client.Create(context.TODO(), pod)

ginkgo.By("ensuring the pod is running")
err := e2epod.WaitForPodRunningInNamespace(context.TODO(), f.ClientSet, pod)
framework.ExpectNoError(err)

ginkgo.By("deleting the pod gracefully")
err = client.Delete(context.TODO(), pod.Name, metav1.DeleteOptions{})
framework.ExpectNoError(err)

ginkgo.By("ensuring the pod is terminated within the grace period seconds + buffer seconds")
err = e2epod.WaitForPodNotFoundInNamespace(context.TODO(), f.ClientSet, pod.Name, pod.Namespace, time.Duration(gracePeriod+bufferSeconds)*time.Second)
framework.ExpectNoError(err)
})

ginkgo.It("should respect termination grace period seconds with long-running preStop hook [NodeConformance]", func() {
client := e2epod.NewPodClient(f)
gracePeriod := int64(30)

ginkgo.By("creating a pod with a termination grace period seconds and long-running preStop hook")
pod := testPod("pod-termination-grace-period", gracePeriod)
pod.Spec.Containers[0].Lifecycle = &v1.Lifecycle{
PreStop: &v1.LifecycleHandler{
Exec: &v1.ExecAction{
Command: []string{
"sleep",
"10000",
},
},
},
}
pod = client.Create(context.TODO(), pod)

ginkgo.By("ensuring the pod is running")
err := e2epod.WaitForPodRunningInNamespace(context.TODO(), f.ClientSet, pod)
framework.ExpectNoError(err)

ginkgo.By("deleting the pod gracefully")
err = client.Delete(context.TODO(), pod.Name, metav1.DeleteOptions{})
framework.ExpectNoError(err)

ginkgo.By("ensuring the pod is terminated within the grace period seconds + buffer seconds")
err = e2epod.WaitForPodNotFoundInNamespace(context.TODO(), f.ClientSet, pod.Name, pod.Namespace, time.Duration(gracePeriod+bufferSeconds)*time.Second)
framework.ExpectNoError(err)
})
})
})