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

Generate AppWrapper name in tests to provide unique workloads #591

Merged
merged 1 commit into from
Jul 17, 2024
Merged

Generate AppWrapper name in tests to provide unique workloads #591

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
34 changes: 21 additions & 13 deletions test/e2e/mnist_pytorch_appwrapper_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,14 +75,20 @@ func runMnistPyTorchAppWrapper(t *testing.T, accelerator string) {
Kind: "Job",
},
ObjectMeta: metav1.ObjectMeta{
Name: "mnist",
Namespace: namespace.Name,
GenerateName: "mnist",
Namespace: namespace.Name,
},
Spec: batchv1.JobSpec{
Completions: Ptr(int32(1)),
Parallelism: Ptr(int32(1)),
Template: corev1.PodTemplateSpec{
Spec: corev1.PodSpec{
Tolerations: []corev1.Toleration{
{
Key: "nvidia.com/gpu",
Operator: corev1.TolerationOpExists,
},
},
Containers: []corev1.Container{
{
Name: "job",
Expand Down Expand Up @@ -139,9 +145,9 @@ func runMnistPyTorchAppWrapper(t *testing.T, accelerator string) {
Kind: "AppWrapper",
},
ObjectMeta: metav1.ObjectMeta{
Name: "mnist",
Namespace: namespace.Name,
Labels: map[string]string{"kueue.x-k8s.io/queue-name": localQueue.Name},
GenerateName: "mnist",
Namespace: namespace.Name,
Labels: map[string]string{"kueue.x-k8s.io/queue-name": localQueue.Name},
},
Spec: mcadv1beta2.AppWrapperSpec{
Components: []mcadv1beta2.AppWrapperComponent{
Expand All @@ -158,16 +164,18 @@ func runMnistPyTorchAppWrapper(t *testing.T, accelerator string) {
unstruct := unstructured.Unstructured{Object: awMap}
_, err = test.Client().Dynamic().Resource(appWrapperResource).Namespace(namespace.Name).Create(test.Ctx(), &unstruct, metav1.CreateOptions{})
test.Expect(err).NotTo(HaveOccurred())
test.T().Logf("Created AppWrapper %s/%s successfully", aw.Namespace, aw.Name)
test.T().Logf("Created AppWrapper %s/%s successfully", aw.Namespace, aw.GenerateName)

test.T().Logf("Waiting for AppWrapper %s/%s to be running", aw.Namespace, aw.Name)
test.Eventually(AppWrapper(test, namespace, aw.Name), TestTimeoutMedium).
Should(WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperRunning)))
test.T().Logf("Waiting for AppWrapper %s/%s to be running", aw.Namespace, aw.GenerateName)
test.Eventually(AppWrappers(test, namespace), TestTimeoutMedium).
Should(ContainElement(WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperRunning))))

test.T().Logf("Waiting for AppWrapper %s/%s to complete", job.Namespace, job.Name)
test.Eventually(AppWrapper(test, namespace, aw.Name), TestTimeoutLong).Should(
Or(
WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperSucceeded)),
WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperFailed)),
test.Eventually(AppWrappers(test, namespace), TestTimeoutLong).Should(
ContainElement(
Or(
WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperSucceeded)),
WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperFailed)),
),
))
}
20 changes: 13 additions & 7 deletions test/e2e/mnist_rayjob_raycluster_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,9 +127,9 @@ func runMnistRayJobRayClusterAppWrapper(t *testing.T, accelerator string, number
Kind: "AppWrapper",
},
ObjectMeta: metav1.ObjectMeta{
Name: rayCluster.Name,
Namespace: namespace.Name,
Labels: map[string]string{"kueue.x-k8s.io/queue-name": localQueue.Name},
GenerateName: rayCluster.Name,
Namespace: namespace.Name,
Labels: map[string]string{"kueue.x-k8s.io/queue-name": localQueue.Name},
},
Spec: mcadv1beta2.AppWrapperSpec{
Components: []mcadv1beta2.AppWrapperComponent{
Expand All @@ -145,11 +145,11 @@ func runMnistRayJobRayClusterAppWrapper(t *testing.T, accelerator string, number
unstruct := unstructured.Unstructured{Object: awMap}
_, err = test.Client().Dynamic().Resource(appWrapperResource).Namespace(namespace.Name).Create(test.Ctx(), &unstruct, metav1.CreateOptions{})
test.Expect(err).NotTo(HaveOccurred())
test.T().Logf("Created AppWrapper %s/%s successfully", aw.Namespace, aw.Name)
test.T().Logf("Created AppWrapper %s/%s successfully", aw.Namespace, aw.GenerateName)

test.T().Logf("Waiting for AppWrapper %s/%s to be running", aw.Namespace, aw.Name)
test.Eventually(AppWrapper(test, namespace, aw.Name), TestTimeoutMedium).
Should(WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperRunning)))
test.T().Logf("Waiting for AppWrapper %s/%s to be running", aw.Namespace, aw.GenerateName)
test.Eventually(AppWrappers(test, namespace), TestTimeoutMedium).
Should(ContainElement(WithTransform(AppWrapperPhase, Equal(mcadv1beta2.AppWrapperRunning))))

test.T().Logf("Waiting for RayCluster %s/%s to be running", rayCluster.Namespace, rayCluster.Name)
test.Eventually(RayCluster(test, namespace.Name, rayCluster.Name), TestTimeoutMedium).
Expand Down Expand Up @@ -266,6 +266,12 @@ func constructRayCluster(_ Test, namespace *corev1.Namespace, mnist *corev1.Conf
RayStartParams: map[string]string{},
Template: corev1.PodTemplateSpec{
Spec: corev1.PodSpec{
Tolerations: []corev1.Toleration{
{
Key: "nvidia.com/gpu",
Operator: corev1.TolerationOpExists,
},
},
Containers: []corev1.Container{
{
Name: "ray-worker",
Expand Down