mirror of
https://github.com/k3s-io/kubernetes.git
synced 2025-08-08 03:33:56 +00:00
added init containers tests to simplify the sidecar KEP large PR
This commit is contained in:
parent
00eee07272
commit
fc0d2cd32f
@ -35,11 +35,13 @@ import (
|
|||||||
|
|
||||||
"k8s.io/kubernetes/test/e2e/framework"
|
"k8s.io/kubernetes/test/e2e/framework"
|
||||||
e2epod "k8s.io/kubernetes/test/e2e/framework/pod"
|
e2epod "k8s.io/kubernetes/test/e2e/framework/pod"
|
||||||
|
imageutils "k8s.io/kubernetes/test/utils/image"
|
||||||
)
|
)
|
||||||
|
|
||||||
type execCommand struct {
|
type execCommand struct {
|
||||||
ExitCode int
|
ExitCode int
|
||||||
Delay int
|
Delay int
|
||||||
|
StartDelay int
|
||||||
}
|
}
|
||||||
|
|
||||||
func ExecCommand(name string, c execCommand) []string {
|
func ExecCommand(name string, c execCommand) []string {
|
||||||
@ -55,7 +57,13 @@ func ExecCommand(name string, c execCommand) []string {
|
|||||||
fmt.Fprintf(&cmd, "touch %s; ", containerLog)
|
fmt.Fprintf(&cmd, "touch %s; ", containerLog)
|
||||||
fmt.Fprintf(&cmd, "cat %s >> /dev/termination-log; ", containerLog)
|
fmt.Fprintf(&cmd, "cat %s >> /dev/termination-log; ", containerLog)
|
||||||
|
|
||||||
fmt.Fprintf(&cmd, "echo %s '%s Starting' | tee -a %s >> /dev/termination-log; ", timeCmd, name, containerLog)
|
fmt.Fprintf(&cmd, "echo %s '%s Starting %d' | tee -a %s >> /dev/termination-log; ", timeCmd, name, c.StartDelay, containerLog)
|
||||||
|
if c.StartDelay != 0 {
|
||||||
|
fmt.Fprintf(&cmd, "sleep %d; ", c.StartDelay)
|
||||||
|
}
|
||||||
|
// You can check started file to see if the container has started
|
||||||
|
fmt.Fprintf(&cmd, "touch started; ")
|
||||||
|
fmt.Fprintf(&cmd, "echo %s '%s Started' | tee -a %s >> /dev/termination-log; ", timeCmd, name, containerLog)
|
||||||
fmt.Fprintf(&cmd, "echo %s '%s Delaying %d' | tee -a %s >> /dev/termination-log; ", timeCmd, name, c.Delay, containerLog)
|
fmt.Fprintf(&cmd, "echo %s '%s Delaying %d' | tee -a %s >> /dev/termination-log; ", timeCmd, name, c.Delay, containerLog)
|
||||||
if c.Delay != 0 {
|
if c.Delay != 0 {
|
||||||
fmt.Fprintf(&cmd, "sleep %d; ", c.Delay)
|
fmt.Fprintf(&cmd, "sleep %d; ", c.Delay)
|
||||||
@ -65,6 +73,19 @@ func ExecCommand(name string, c execCommand) []string {
|
|||||||
return []string{"sh", "-c", cmd.String()}
|
return []string{"sh", "-c", cmd.String()}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// WaitForPodInitContainerRestartCount waits for the given Pod init container
|
||||||
|
// to achieve at least a given restartCount
|
||||||
|
func WaitForPodInitContainerRestartCount(ctx context.Context, c clientset.Interface, namespace, podName string, initContainerIndex int, desiredRestartCount int32, timeout time.Duration) error {
|
||||||
|
conditionDesc := fmt.Sprintf("init container %d started", initContainerIndex)
|
||||||
|
return e2epod.WaitForPodCondition(ctx, c, namespace, podName, conditionDesc, timeout, func(pod *v1.Pod) (bool, error) {
|
||||||
|
if initContainerIndex > len(pod.Status.InitContainerStatuses)-1 {
|
||||||
|
return false, nil
|
||||||
|
}
|
||||||
|
containerStatus := pod.Status.InitContainerStatuses[initContainerIndex]
|
||||||
|
return containerStatus.RestartCount >= desiredRestartCount, nil
|
||||||
|
})
|
||||||
|
}
|
||||||
|
|
||||||
// WaitForPodContainerRestartCount waits for the given Pod container to achieve at least a given restartCount
|
// WaitForPodContainerRestartCount waits for the given Pod container to achieve at least a given restartCount
|
||||||
func WaitForPodContainerRestartCount(ctx context.Context, c clientset.Interface, namespace, podName string, containerIndex int, desiredRestartCount int32, timeout time.Duration) error {
|
func WaitForPodContainerRestartCount(ctx context.Context, c clientset.Interface, namespace, podName string, containerIndex int, desiredRestartCount int32, timeout time.Duration) error {
|
||||||
conditionDesc := fmt.Sprintf("container %d started", containerIndex)
|
conditionDesc := fmt.Sprintf("container %d started", containerIndex)
|
||||||
@ -77,6 +98,29 @@ func WaitForPodContainerRestartCount(ctx context.Context, c clientset.Interface,
|
|||||||
})
|
})
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// WaitForPodInitContainerToFail waits for the given Pod init container to fail with the given reason, specifically due to
|
||||||
|
// invalid container configuration. In this case, the container will remain in a waiting state with a specific
|
||||||
|
// reason set, which should match the given reason.
|
||||||
|
func WaitForPodInitContainerToFail(ctx context.Context, c clientset.Interface, namespace, podName string, containerIndex int, reason string, timeout time.Duration) error {
|
||||||
|
conditionDesc := fmt.Sprintf("container %d failed with reason %s", containerIndex, reason)
|
||||||
|
return e2epod.WaitForPodCondition(ctx, c, namespace, podName, conditionDesc, timeout, func(pod *v1.Pod) (bool, error) {
|
||||||
|
switch pod.Status.Phase {
|
||||||
|
case v1.PodPending:
|
||||||
|
if len(pod.Status.InitContainerStatuses) == 0 {
|
||||||
|
return false, nil
|
||||||
|
}
|
||||||
|
containerStatus := pod.Status.InitContainerStatuses[containerIndex]
|
||||||
|
if containerStatus.State.Waiting != nil && containerStatus.State.Waiting.Reason == reason {
|
||||||
|
return true, nil
|
||||||
|
}
|
||||||
|
return false, nil
|
||||||
|
case v1.PodFailed, v1.PodRunning, v1.PodSucceeded:
|
||||||
|
return false, fmt.Errorf("pod was expected to be pending, but it is in the state: %s", pod.Status.Phase)
|
||||||
|
}
|
||||||
|
return false, nil
|
||||||
|
})
|
||||||
|
}
|
||||||
|
|
||||||
const (
|
const (
|
||||||
PostStartPrefix = "PostStart"
|
PostStartPrefix = "PostStart"
|
||||||
)
|
)
|
||||||
@ -133,9 +177,21 @@ var _ = SIGDescribe("[NodeConformance] Containers Lifecycle ", func() {
|
|||||||
Name: regular1,
|
Name: regular1,
|
||||||
Image: busyboxImage,
|
Image: busyboxImage,
|
||||||
Command: ExecCommand(regular1, execCommand{
|
Command: ExecCommand(regular1, execCommand{
|
||||||
Delay: 1,
|
StartDelay: 5,
|
||||||
ExitCode: 0,
|
Delay: 1,
|
||||||
|
ExitCode: 0,
|
||||||
}),
|
}),
|
||||||
|
StartupProbe: &v1.Probe{
|
||||||
|
ProbeHandler: v1.ProbeHandler{
|
||||||
|
Exec: &v1.ExecAction{
|
||||||
|
Command: []string{
|
||||||
|
"test",
|
||||||
|
"-f",
|
||||||
|
"started",
|
||||||
|
},
|
||||||
|
},
|
||||||
|
},
|
||||||
|
},
|
||||||
},
|
},
|
||||||
},
|
},
|
||||||
},
|
},
|
||||||
@ -145,18 +201,22 @@ var _ = SIGDescribe("[NodeConformance] Containers Lifecycle ", func() {
|
|||||||
|
|
||||||
/// generates an out file output like:
|
/// generates an out file output like:
|
||||||
//
|
//
|
||||||
// 1678337827 45930.43 init-1 Starting
|
// 1682076093 4905.79 init-1 Starting 0
|
||||||
// 1678337827 45930.43 init-1 Delaying 1
|
// 1682076093 4905.80 init-1 Started
|
||||||
// 1678337828 45931.43 init-1 Exiting
|
// 1682076093 4905.80 init-1 Delaying 1
|
||||||
// 1678337829 45932.52 init-2 Starting
|
// 1682076094 4906.80 init-1 Exiting
|
||||||
// 1678337829 45932.53 init-2 Delaying 1
|
// 1682076095 4907.70 init-2 Starting 0
|
||||||
// 1678337830 45933.53 init-2 Exiting
|
// 1682076095 4907.71 init-2 Started
|
||||||
// 1678337831 45934.47 init-3 Starting
|
// 1682076095 4907.71 init-2 Delaying 1
|
||||||
// 1678337831 45934.47 init-3 Delaying 1
|
// 1682076096 4908.71 init-2 Exiting
|
||||||
// 1678337832 45935.47 init-3 Exiting
|
// 1682076097 4909.74 init-3 Starting 0
|
||||||
// 1678337833 45936.58 regular-1 Starting
|
// 1682076097 4909.74 init-3 Started
|
||||||
// 1678337833 45936.58 regular-1 Delaying 1
|
// 1682076097 4909.74 init-3 Delaying 1
|
||||||
// 1678337834 45937.58 regular-1 Exiting
|
// 1682076098 4910.75 init-3 Exiting
|
||||||
|
// 1682076099 4911.70 regular-1 Starting 5
|
||||||
|
// 1682076104 4916.71 regular-1 Started
|
||||||
|
// 1682076104 4916.71 regular-1 Delaying 1
|
||||||
|
// 1682076105 4917.72 regular-1 Exiting
|
||||||
|
|
||||||
client := e2epod.NewPodClient(f)
|
client := e2epod.NewPodClient(f)
|
||||||
podSpec = client.Create(context.TODO(), podSpec)
|
podSpec = client.Create(context.TODO(), podSpec)
|
||||||
@ -404,6 +464,66 @@ var _ = SIGDescribe("[NodeConformance] Containers Lifecycle ", func() {
|
|||||||
framework.ExpectNoError(results.ExitsBefore(prefixedName(PostStartPrefix, regular1), regular2))
|
framework.ExpectNoError(results.ExitsBefore(prefixedName(PostStartPrefix, regular1), regular2))
|
||||||
})
|
})
|
||||||
|
|
||||||
|
ginkgo.When("have init container in a Pod with restartPolicy=Never", func() {
|
||||||
|
|
||||||
|
ginkgo.When("an init container fails to start because of a bad image", ginkgo.Ordered, func() {
|
||||||
|
|
||||||
|
init1 := "init1-1"
|
||||||
|
regular1 := "regular-1"
|
||||||
|
|
||||||
|
podSpec := &v1.Pod{
|
||||||
|
ObjectMeta: metav1.ObjectMeta{
|
||||||
|
Name: "sidecar-runs-with-pod",
|
||||||
|
},
|
||||||
|
Spec: v1.PodSpec{
|
||||||
|
RestartPolicy: v1.RestartPolicyNever,
|
||||||
|
InitContainers: []v1.Container{
|
||||||
|
{
|
||||||
|
Name: init1,
|
||||||
|
Image: imageutils.GetE2EImage(imageutils.InvalidRegistryImage),
|
||||||
|
Command: ExecCommand(init1, execCommand{
|
||||||
|
Delay: 600,
|
||||||
|
ExitCode: 0,
|
||||||
|
}),
|
||||||
|
},
|
||||||
|
},
|
||||||
|
Containers: []v1.Container{
|
||||||
|
{
|
||||||
|
Name: regular1,
|
||||||
|
Image: busyboxImage,
|
||||||
|
Command: ExecCommand(regular1, execCommand{
|
||||||
|
Delay: 1,
|
||||||
|
ExitCode: 0,
|
||||||
|
}),
|
||||||
|
},
|
||||||
|
},
|
||||||
|
},
|
||||||
|
}
|
||||||
|
|
||||||
|
preparePod(podSpec)
|
||||||
|
var results containerOutputList
|
||||||
|
|
||||||
|
ginkgo.It("should mark a Pod as failed and produce log", func() {
|
||||||
|
client := e2epod.NewPodClient(f)
|
||||||
|
podSpec = client.Create(context.TODO(), podSpec)
|
||||||
|
|
||||||
|
// sidecar should be in image pull backoff
|
||||||
|
err := WaitForPodInitContainerToFail(context.TODO(), f.ClientSet, podSpec.Namespace, podSpec.Name, 0, "ImagePullBackOff", f.Timeouts.PodStart)
|
||||||
|
framework.ExpectNoError(err)
|
||||||
|
|
||||||
|
podSpec, err = client.Get(context.Background(), podSpec.Name, metav1.GetOptions{})
|
||||||
|
framework.ExpectNoError(err)
|
||||||
|
results = parseOutput(podSpec)
|
||||||
|
})
|
||||||
|
ginkgo.It("should not start an init container", func() {
|
||||||
|
framework.ExpectNoError(results.DoesntStart(init1))
|
||||||
|
})
|
||||||
|
ginkgo.It("should not start a regular container", func() {
|
||||||
|
framework.ExpectNoError(results.DoesntStart(regular1))
|
||||||
|
})
|
||||||
|
})
|
||||||
|
})
|
||||||
|
|
||||||
})
|
})
|
||||||
|
|
||||||
type containerOutput struct {
|
type containerOutput struct {
|
||||||
@ -426,8 +546,8 @@ func (o containerOutputList) String() string {
|
|||||||
|
|
||||||
// RunTogether returns an error the lhs and rhs run together
|
// RunTogether returns an error the lhs and rhs run together
|
||||||
func (o containerOutputList) RunTogether(lhs, rhs string) error {
|
func (o containerOutputList) RunTogether(lhs, rhs string) error {
|
||||||
lhsStart := o.findIndex(lhs, "Starting", 0)
|
lhsStart := o.findIndex(lhs, "Started", 0)
|
||||||
rhsStart := o.findIndex(rhs, "Starting", 0)
|
rhsStart := o.findIndex(rhs, "Started", 0)
|
||||||
|
|
||||||
lhsFinish := o.findIndex(lhs, "Finishing", 0)
|
lhsFinish := o.findIndex(lhs, "Finishing", 0)
|
||||||
rhsFinish := o.findIndex(rhs, "Finishing", 0)
|
rhsFinish := o.findIndex(rhs, "Finishing", 0)
|
||||||
@ -452,7 +572,7 @@ func (o containerOutputList) RunTogether(lhs, rhs string) error {
|
|||||||
|
|
||||||
// StartsBefore returns an error if lhs did not start before rhs
|
// StartsBefore returns an error if lhs did not start before rhs
|
||||||
func (o containerOutputList) StartsBefore(lhs, rhs string) error {
|
func (o containerOutputList) StartsBefore(lhs, rhs string) error {
|
||||||
lhsStart := o.findIndex(lhs, "Starting", 0)
|
lhsStart := o.findIndex(lhs, "Started", 0)
|
||||||
|
|
||||||
if lhsStart == -1 {
|
if lhsStart == -1 {
|
||||||
return fmt.Errorf("couldn't find that %s ever started, got %v", lhs, o)
|
return fmt.Errorf("couldn't find that %s ever started, got %v", lhs, o)
|
||||||
@ -467,6 +587,24 @@ func (o containerOutputList) StartsBefore(lhs, rhs string) error {
|
|||||||
return nil
|
return nil
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// DoesntStartAfter returns an error if lhs started after rhs
|
||||||
|
func (o containerOutputList) DoesntStartAfter(lhs, rhs string) error {
|
||||||
|
rhsStart := o.findIndex(rhs, "Starting", 0)
|
||||||
|
|
||||||
|
if rhsStart == -1 {
|
||||||
|
return fmt.Errorf("couldn't find that %s ever started, got %v", rhs, o)
|
||||||
|
}
|
||||||
|
|
||||||
|
// this works even for the same names (restart case)
|
||||||
|
lhsStart := o.findIndex(lhs, "Started", rhsStart+1)
|
||||||
|
|
||||||
|
if lhsStart != -1 {
|
||||||
|
return fmt.Errorf("expected %s to not start after %s, got %v", lhs, rhs, o)
|
||||||
|
}
|
||||||
|
|
||||||
|
return nil
|
||||||
|
}
|
||||||
|
|
||||||
// ExitsBefore returns an error if lhs did not end before rhs
|
// ExitsBefore returns an error if lhs did not end before rhs
|
||||||
func (o containerOutputList) ExitsBefore(lhs, rhs string) error {
|
func (o containerOutputList) ExitsBefore(lhs, rhs string) error {
|
||||||
lhsExit := o.findIndex(lhs, "Exiting", 0)
|
lhsExit := o.findIndex(lhs, "Exiting", 0)
|
||||||
@ -486,7 +624,7 @@ func (o containerOutputList) ExitsBefore(lhs, rhs string) error {
|
|||||||
|
|
||||||
// Starts returns an error if the container was not found to have started
|
// Starts returns an error if the container was not found to have started
|
||||||
func (o containerOutputList) Starts(name string) error {
|
func (o containerOutputList) Starts(name string) error {
|
||||||
if idx := o.findIndex(name, "Starting", 0); idx == -1 {
|
if idx := o.findIndex(name, "Started", 0); idx == -1 {
|
||||||
return fmt.Errorf("couldn't find that %s ever started, got %v", name, o)
|
return fmt.Errorf("couldn't find that %s ever started, got %v", name, o)
|
||||||
}
|
}
|
||||||
return nil
|
return nil
|
||||||
@ -494,7 +632,7 @@ func (o containerOutputList) Starts(name string) error {
|
|||||||
|
|
||||||
// DoesntStart returns an error if the container was found to have started
|
// DoesntStart returns an error if the container was found to have started
|
||||||
func (o containerOutputList) DoesntStart(name string) error {
|
func (o containerOutputList) DoesntStart(name string) error {
|
||||||
if idx := o.findIndex(name, "Starting", 0); idx != -1 {
|
if idx := o.findIndex(name, "Started", 0); idx != -1 {
|
||||||
return fmt.Errorf("find %s started, but didn't expect to, got %v", name, o)
|
return fmt.Errorf("find %s started, but didn't expect to, got %v", name, o)
|
||||||
}
|
}
|
||||||
return nil
|
return nil
|
||||||
@ -508,6 +646,22 @@ func (o containerOutputList) Exits(name string) error {
|
|||||||
return nil
|
return nil
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// HasRestarted returns an error if the container was not found to have restarted
|
||||||
|
func (o containerOutputList) HasRestarted(name string) error {
|
||||||
|
idx := o.findIndex(name, "Starting", 0)
|
||||||
|
if idx == -1 {
|
||||||
|
return fmt.Errorf("couldn't find that %s ever started, got %v", name, o)
|
||||||
|
}
|
||||||
|
|
||||||
|
idx = o.findIndex(name, "Starting", idx+1)
|
||||||
|
|
||||||
|
if idx == -1 {
|
||||||
|
return fmt.Errorf("couldn't find that %s ever restarted, got %v", name, o)
|
||||||
|
}
|
||||||
|
|
||||||
|
return nil
|
||||||
|
}
|
||||||
|
|
||||||
func (o containerOutputList) findIndex(name string, command string, startIdx int) int {
|
func (o containerOutputList) findIndex(name string, command string, startIdx int) int {
|
||||||
for i, v := range o {
|
for i, v := range o {
|
||||||
if i < startIdx {
|
if i < startIdx {
|
||||||
@ -529,7 +683,10 @@ func parseOutput(pod *v1.Pod) containerOutputList {
|
|||||||
statuses = append(statuses, pod.Status.ContainerStatuses...)
|
statuses = append(statuses, pod.Status.ContainerStatuses...)
|
||||||
var buf bytes.Buffer
|
var buf bytes.Buffer
|
||||||
for _, cs := range statuses {
|
for _, cs := range statuses {
|
||||||
if cs.State.Terminated != nil {
|
// If the container is terminated but the reason is ContainerStatusUnknown,
|
||||||
|
// it means that the kubelet has overwritten the termination message. Read
|
||||||
|
// the LastTerminationState instead.
|
||||||
|
if cs.State.Terminated != nil && cs.State.Terminated.Reason != "ContainerStatusUnknown" {
|
||||||
buf.WriteString(cs.State.Terminated.Message)
|
buf.WriteString(cs.State.Terminated.Message)
|
||||||
} else if cs.LastTerminationState.Terminated != nil {
|
} else if cs.LastTerminationState.Terminated != nil {
|
||||||
buf.WriteString(cs.LastTerminationState.Terminated.Message)
|
buf.WriteString(cs.LastTerminationState.Terminated.Message)
|
||||||
|
Loading…
Reference in New Issue
Block a user