mirror of
https://github.com/k3s-io/kubernetes.git
synced 2025-08-06 10:43:56 +00:00
e2e: KEP-3857: Recursive Read-only (RRO) mounts
Usage: ``` make test-e2e-node \ TEST_ARGS='--service-feature-gates=RecursiveReadOnlyMounts=true --kubelet-flags="--feature-gates=RecursiveReadOnlyMounts=true"' \ FOCUS="Mount recursive read-only" SKIP="" ``` Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
This commit is contained in:
parent
c7f52b34f3
commit
d4925ce8f8
@ -91,6 +91,9 @@ var (
|
||||
// TODO: document the feature (owning SIG, when to use this feature for a test)
|
||||
PodResources = framework.WithNodeFeature(framework.ValidNodeFeatures.Add("PodResources"))
|
||||
|
||||
// RecursiveReadOnlyMounts (SIG-node, used for testing recursive read-only mounts <https://kep.k8s.io/3857>)
|
||||
RecursiveReadOnlyMounts = framework.WithNodeFeature(framework.ValidNodeFeatures.Add("RecursiveReadOnlyMounts"))
|
||||
|
||||
// TODO: document the feature (owning SIG, when to use this feature for a test)
|
||||
ResourceMetrics = framework.WithNodeFeature(framework.ValidNodeFeatures.Add("ResourceMetrics"))
|
||||
|
||||
|
428
test/e2e_node/mount_rro_linux_test.go
Normal file
428
test/e2e_node/mount_rro_linux_test.go
Normal file
@ -0,0 +1,428 @@
|
||||
/*
|
||||
Copyright 2024 The Kubernetes Authors.
|
||||
|
||||
Licensed under the Apache License, Version 2.0 (the "License");
|
||||
you may not use this file except in compliance with the License.
|
||||
You may obtain a copy of the License at
|
||||
|
||||
http://www.apache.org/licenses/LICENSE-2.0
|
||||
|
||||
Unless required by applicable law or agreed to in writing, software
|
||||
distributed under the License is distributed on an "AS IS" BASIS,
|
||||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
See the License for the specific language governing permissions and
|
||||
limitations under the License.
|
||||
*/
|
||||
|
||||
package e2enode
|
||||
|
||||
import (
|
||||
"context"
|
||||
|
||||
"github.com/onsi/ginkgo/v2"
|
||||
"github.com/onsi/gomega"
|
||||
|
||||
v1 "k8s.io/api/core/v1"
|
||||
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
|
||||
"k8s.io/apimachinery/pkg/util/uuid"
|
||||
"k8s.io/kubernetes/test/e2e/framework"
|
||||
e2epod "k8s.io/kubernetes/test/e2e/framework/pod"
|
||||
e2eskipper "k8s.io/kubernetes/test/e2e/framework/skipper"
|
||||
"k8s.io/kubernetes/test/e2e/nodefeature"
|
||||
admissionapi "k8s.io/pod-security-admission/api"
|
||||
"k8s.io/utils/ptr"
|
||||
)
|
||||
|
||||
// Usage:
|
||||
// make test-e2e-node TEST_ARGS='--service-feature-gates=RecursiveReadOnlyMounts=true --kubelet-flags="--feature-gates=RecursiveReadOnlyMounts=true"' FOCUS="Mount recursive read-only" SKIP=""
|
||||
var _ = SIGDescribe("Mount recursive read-only [LinuxOnly]", framework.WithSerial(), nodefeature.RecursiveReadOnlyMounts, func() {
|
||||
f := framework.NewDefaultFramework("mount-rro")
|
||||
f.NamespacePodSecurityLevel = admissionapi.LevelPrivileged
|
||||
ginkgo.Describe("Mount recursive read-only", func() {
|
||||
ginkgo.Context("when the runtime supports recursive read-only mounts", func() {
|
||||
f.It("should accept recursive read-only mounts", func(ctx context.Context) {
|
||||
ginkgo.By("waiting for the node to be ready", func() {
|
||||
waitForNodeReady(ctx)
|
||||
if !supportsRRO(ctx, f) {
|
||||
e2eskipper.Skipf("runtime does not support recursive read-only mounts")
|
||||
}
|
||||
}) // By
|
||||
var pod *v1.Pod
|
||||
ginkgo.By("creating a pod", func() {
|
||||
pod = e2epod.NewPodClient(f).Create(ctx,
|
||||
podForRROSupported("mount-rro-"+string(uuid.NewUUID()), f.Namespace.Name))
|
||||
framework.ExpectNoError(e2epod.WaitForPodSuccessInNamespace(ctx, f.ClientSet, pod.Name, pod.Namespace))
|
||||
var err error
|
||||
pod, err = f.ClientSet.CoreV1().Pods(pod.Namespace).Get(ctx, pod.Name, metav1.GetOptions{})
|
||||
framework.ExpectNoError(err)
|
||||
}) // By
|
||||
ginkgo.By("checking containerStatuses.volumeMounts", func() {
|
||||
gomega.Expect(pod.Status.InitContainerStatuses).To(gomega.HaveLen(3)) // "mount", "test", "unmount"
|
||||
volMountStatuses := pod.Status.InitContainerStatuses[1].VolumeMounts
|
||||
var verifiedVolMountStatuses int
|
||||
for _, f := range volMountStatuses {
|
||||
switch f.Name {
|
||||
case "mnt":
|
||||
switch f.MountPath {
|
||||
case "/mnt-rro", "/mnt-rro-if-possible":
|
||||
gomega.Expect(*f.RecursiveReadOnly).To(gomega.Equal(v1.RecursiveReadOnlyEnabled))
|
||||
verifiedVolMountStatuses++
|
||||
case "/mnt-rro-disabled", "/mnt-ro":
|
||||
gomega.Expect(*f.RecursiveReadOnly).To(gomega.Equal(v1.RecursiveReadOnlyDisabled))
|
||||
verifiedVolMountStatuses++
|
||||
case "/mnt-rw":
|
||||
gomega.Expect(f.RecursiveReadOnly).To(gomega.BeNil())
|
||||
verifiedVolMountStatuses++
|
||||
default:
|
||||
framework.Failf("unexpected mount path: %q", f.MountPath)
|
||||
}
|
||||
default: // implicit secret volumes, etc.
|
||||
// NOP
|
||||
}
|
||||
}
|
||||
gomega.Expect(verifiedVolMountStatuses).To(gomega.Equal(5))
|
||||
}) // By
|
||||
}) // It
|
||||
f.It("should reject invalid recursive read-only mounts", func(ctx context.Context) {
|
||||
ginkgo.By("waiting for the node to be ready", func() {
|
||||
waitForNodeReady(ctx)
|
||||
if !supportsRRO(ctx, f) {
|
||||
e2eskipper.Skipf("runtime does not support recursive read-only mounts")
|
||||
}
|
||||
}) // By
|
||||
ginkgo.By("specifying RRO without RO", func() {
|
||||
pod := &v1.Pod{
|
||||
ObjectMeta: metav1.ObjectMeta{
|
||||
Name: "mount-rro-invalid-" + string(uuid.NewUUID()),
|
||||
Namespace: f.Namespace.Name,
|
||||
},
|
||||
Spec: v1.PodSpec{
|
||||
RestartPolicy: v1.RestartPolicyNever,
|
||||
Containers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "busybox",
|
||||
Command: []string{"echo", "this container should fail"},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyEnabled),
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
Volumes: []v1.Volume{
|
||||
{
|
||||
Name: "mnt",
|
||||
VolumeSource: v1.VolumeSource{
|
||||
EmptyDir: &v1.EmptyDirVolumeSource{},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
pod = e2epod.NewPodClient(f).Create(ctx, pod)
|
||||
framework.ExpectNoError(e2epod.WaitForPodContainerToFail(ctx, f.ClientSet, pod.Namespace, pod.Name, 0, "CreateContainerConfigError", framework.PodStartShortTimeout))
|
||||
var err error
|
||||
pod, err = f.ClientSet.CoreV1().Pods(f.Namespace.Name).Get(ctx, pod.Name, metav1.GetOptions{})
|
||||
framework.ExpectNoError(err)
|
||||
gomega.Expect(pod.Status.ContainerStatuses[0].State.Waiting.Message).To(
|
||||
gomega.ContainSubstring("failed to resolve recursive read-only mode: volume \"mnt\" requested recursive read-only mode, but it is not read-only"))
|
||||
}) // By
|
||||
// See also the unit test [pkg/kubelet.TestResolveRecursiveReadOnly] for more invalid conditions (e.g., incompatible mount propagation)
|
||||
}) // It
|
||||
}) // Context
|
||||
ginkgo.Context("when the runtime does not support recursive read-only mounts", func() {
|
||||
f.It("should accept non-recursive read-only mounts", func(ctx context.Context) {
|
||||
ginkgo.By("waiting for the node to be ready", func() {
|
||||
waitForNodeReady(ctx)
|
||||
if supportsRRO(ctx, f) {
|
||||
e2eskipper.Skipf("runtime supports recursive read-only mounts")
|
||||
}
|
||||
}) // By
|
||||
var pod *v1.Pod
|
||||
ginkgo.By("creating a pod", func() {
|
||||
pod = e2epod.NewPodClient(f).Create(ctx,
|
||||
podForRROUnsupported("mount-ro-"+string(uuid.NewUUID()), f.Namespace.Name))
|
||||
framework.ExpectNoError(e2epod.WaitForPodSuccessInNamespace(ctx, f.ClientSet, pod.Name, pod.Namespace))
|
||||
var err error
|
||||
pod, err = f.ClientSet.CoreV1().Pods(pod.Namespace).Get(ctx, pod.Name, metav1.GetOptions{})
|
||||
framework.ExpectNoError(err)
|
||||
}) // By
|
||||
ginkgo.By("checking containerStatuses.volumeMounts", func() {
|
||||
gomega.Expect(pod.Status.InitContainerStatuses).To(gomega.HaveLen(3)) // "mount", "test", "unmount"
|
||||
volMountStatuses := pod.Status.InitContainerStatuses[1].VolumeMounts
|
||||
var verifiedVolMountStatuses int
|
||||
for _, f := range volMountStatuses {
|
||||
switch f.Name {
|
||||
case "mnt":
|
||||
switch f.MountPath {
|
||||
case "/mnt-rro-if-possible", "/mnt-rro-disabled", "/mnt-ro":
|
||||
gomega.Expect(*f.RecursiveReadOnly).To(gomega.Equal(v1.RecursiveReadOnlyDisabled))
|
||||
verifiedVolMountStatuses++
|
||||
case "/mnt-rw":
|
||||
gomega.Expect(f.RecursiveReadOnly).To(gomega.BeNil())
|
||||
verifiedVolMountStatuses++
|
||||
default:
|
||||
framework.Failf("unexpected mount path: %q", f.MountPath)
|
||||
}
|
||||
default: // implicit secret volumes, etc.
|
||||
// NOP
|
||||
}
|
||||
}
|
||||
gomega.Expect(verifiedVolMountStatuses).To(gomega.Equal(4))
|
||||
}) // By
|
||||
}) // It
|
||||
f.It("should reject recursive read-only mounts", func(ctx context.Context) {
|
||||
ginkgo.By("waiting for the node to be ready", func() {
|
||||
waitForNodeReady(ctx)
|
||||
if supportsRRO(ctx, f) {
|
||||
e2eskipper.Skipf("runtime supports recursive read-only mounts")
|
||||
}
|
||||
}) // By
|
||||
ginkgo.By("specifying RRO explicitly", func() {
|
||||
pod := &v1.Pod{
|
||||
ObjectMeta: metav1.ObjectMeta{
|
||||
Name: "mount-rro-unsupported-" + string(uuid.NewUUID()),
|
||||
Namespace: f.Namespace.Name,
|
||||
},
|
||||
Spec: v1.PodSpec{
|
||||
RestartPolicy: v1.RestartPolicyNever,
|
||||
Containers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "busybox",
|
||||
Command: []string{"echo", "this container should fail"},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
ReadOnly: true,
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyEnabled),
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
Volumes: []v1.Volume{
|
||||
{
|
||||
Name: "mnt",
|
||||
VolumeSource: v1.VolumeSource{
|
||||
EmptyDir: &v1.EmptyDirVolumeSource{},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
pod = e2epod.NewPodClient(f).Create(ctx, pod)
|
||||
framework.ExpectNoError(e2epod.WaitForPodContainerToFail(ctx, f.ClientSet, pod.Namespace, pod.Name, 0, "CreateContainerConfigError", framework.PodStartShortTimeout))
|
||||
var err error
|
||||
pod, err = f.ClientSet.CoreV1().Pods(f.Namespace.Name).Get(ctx, pod.Name, metav1.GetOptions{})
|
||||
framework.ExpectNoError(err)
|
||||
gomega.Expect(pod.Status.ContainerStatuses[0].State.Waiting.Message).To(
|
||||
gomega.ContainSubstring("failed to resolve recursive read-only mode: volume \"mnt\" requested recursive read-only mode, but it is not supported by the runtime"))
|
||||
}) // By
|
||||
}) // It
|
||||
}) // Context
|
||||
}) // Describe
|
||||
}) // SIGDescribe
|
||||
|
||||
func supportsRRO(ctx context.Context, f *framework.Framework) bool {
|
||||
nodeList, err := f.ClientSet.CoreV1().Nodes().List(ctx, metav1.ListOptions{})
|
||||
framework.ExpectNoError(err)
|
||||
// Assuming that there is only one node, because this is a node e2e test.
|
||||
gomega.Expect(nodeList.Items).To(gomega.HaveLen(1))
|
||||
node := nodeList.Items[0]
|
||||
for _, f := range node.Status.RuntimeClasses {
|
||||
if f.Name == "" && f.Features != nil && *f.Features.RecursiveReadOnlyMounts {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
func podForRROSupported(name, ns string) *v1.Pod {
|
||||
return &v1.Pod{
|
||||
ObjectMeta: metav1.ObjectMeta{
|
||||
Name: name,
|
||||
Namespace: ns,
|
||||
},
|
||||
Spec: v1.PodSpec{
|
||||
RestartPolicy: v1.RestartPolicyNever,
|
||||
InitContainers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "mount",
|
||||
Command: []string{"sh", "-euxc", "mkdir -p /mnt/tmpfs && mount -t tmpfs none /mnt/tmpfs"},
|
||||
SecurityContext: &v1.SecurityContext{
|
||||
Privileged: ptr.To(true),
|
||||
},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
MountPropagation: ptr.To(v1.MountPropagationBidirectional),
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "test",
|
||||
Command: []string{"sh", "-euxc", `
|
||||
for f in rro rro-if-possible; do touch /mnt-$f/tmpfs/foo 2>&1 | grep "Read-only"; done
|
||||
for f in rro-disabled ro rw; do touch /mnt-$f/tmpfs/foo; done
|
||||
`},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rro",
|
||||
ReadOnly: true,
|
||||
MountPropagation: ptr.To(v1.MountPropagationNone), // explicit
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyEnabled),
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rro-if-possible",
|
||||
ReadOnly: true,
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyIfPossible),
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rro-disabled",
|
||||
ReadOnly: true,
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyDisabled), // explicit
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-ro",
|
||||
ReadOnly: true,
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rw",
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "unmount",
|
||||
Command: []string{"umount", "/mnt/tmpfs"},
|
||||
SecurityContext: &v1.SecurityContext{
|
||||
Privileged: ptr.To(true),
|
||||
},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
MountPropagation: ptr.To(v1.MountPropagationBidirectional),
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
Containers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "completion",
|
||||
Command: []string{"echo", "OK"},
|
||||
},
|
||||
},
|
||||
Volumes: []v1.Volume{
|
||||
{
|
||||
Name: "mnt",
|
||||
VolumeSource: v1.VolumeSource{
|
||||
EmptyDir: &v1.EmptyDirVolumeSource{},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
}
|
||||
|
||||
func podForRROUnsupported(name, ns string) *v1.Pod {
|
||||
return &v1.Pod{
|
||||
ObjectMeta: metav1.ObjectMeta{
|
||||
Name: name,
|
||||
Namespace: ns,
|
||||
},
|
||||
Spec: v1.PodSpec{
|
||||
RestartPolicy: v1.RestartPolicyNever,
|
||||
InitContainers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "mount",
|
||||
Command: []string{"sh", "-euxc", "mkdir -p /mnt/tmpfs && mount -t tmpfs none /mnt/tmpfs"},
|
||||
SecurityContext: &v1.SecurityContext{
|
||||
Privileged: ptr.To(true),
|
||||
},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
MountPropagation: ptr.To(v1.MountPropagationBidirectional),
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "test",
|
||||
Command: []string{"sh", "-euxc", `
|
||||
for f in rro-if-possible rro-disabled ro rw; do touch /mnt-$f/tmpfs/foo; done
|
||||
`},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rro-if-possible",
|
||||
ReadOnly: true,
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyIfPossible),
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rro-disabled",
|
||||
ReadOnly: true,
|
||||
RecursiveReadOnly: ptr.To(v1.RecursiveReadOnlyDisabled), // explicit
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-ro",
|
||||
ReadOnly: true,
|
||||
},
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt-rw",
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "unmount",
|
||||
Command: []string{"umount", "/mnt/tmpfs"},
|
||||
SecurityContext: &v1.SecurityContext{
|
||||
Privileged: ptr.To(true),
|
||||
},
|
||||
VolumeMounts: []v1.VolumeMount{
|
||||
{
|
||||
Name: "mnt",
|
||||
MountPath: "/mnt",
|
||||
MountPropagation: ptr.To(v1.MountPropagationBidirectional),
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
Containers: []v1.Container{
|
||||
{
|
||||
Image: busyboxImage,
|
||||
Name: "completion",
|
||||
Command: []string{"echo", "OK"},
|
||||
},
|
||||
},
|
||||
Volumes: []v1.Volume{
|
||||
{
|
||||
Name: "mnt",
|
||||
VolumeSource: v1.VolumeSource{
|
||||
EmptyDir: &v1.EmptyDirVolumeSource{},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
}
|
Loading…
Reference in New Issue
Block a user