undo changes

e2e/aks_model.go

@@ -44,7 +44,7 @@ func getBaseClusterModel(clusterName string) *armcontainerservice.ManagedCluster
 				{
 					Name:         to.Ptr("nodepool1"),
 					Count:        to.Ptr[int32](1),
-					VMSize:       to.Ptr("Standard_D2s_v3"),
+					VMSize:       to.Ptr("Standard_D2ds_v5"),
 					MaxPods:      to.Ptr[int32](110),
 					OSType:       to.Ptr(armcontainerservice.OSTypeLinux),
 					Type:         to.Ptr(armcontainerservice.AgentPoolTypeVirtualMachineScaleSets),

e2e/kube.go

@@ -14,7 +14,6 @@ import (
 	"github.com/Azure/azure-sdk-for-go/sdk/azcore/to"
 	"github.com/stretchr/testify/require"
 	appsv1 "k8s.io/api/apps/v1"
-	batchv1 "k8s.io/api/batch/v1"
 	corev1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
@@ -302,6 +301,11 @@ func (k *Kubeclient) EnsureDebugDaemonsets(ctx context.Context, t *testing.T, is
 		return err
 	}
 
+	err = k.CreateDaemonset(ctx, t, nvidiaDevicePluginDaemonSet())
+	if err != nil {
+		return err
+	}
+
 	return nil
 }
 
@@ -513,32 +517,6 @@ func podRunNvidiaWorkload(s *Scenario) *corev1.Pod {
 	}
 }
 
-func podAMDGPUWorkload(s *Scenario) *corev1.Pod {
-	return &corev1.Pod{
-		ObjectMeta: metav1.ObjectMeta{
-			Name:      fmt.Sprintf("%s-gpu-validation-pod", s.Runtime.KubeNodeName),
-			Namespace: defaultNamespace,
-		},
-		Spec: corev1.PodSpec{
-			Containers: []corev1.Container{
-				{
-					Name:  "gpu-validation-container",
-					Image: "mcr.microsoft.com/azuredocs/samples-tf-mnist-demo:gpu",
-					Args: []string{
-						"--max-steps", "1",
-					},
-					Resources: corev1.ResourceRequirements{
-						Limits: corev1.ResourceList{
-							"amd.com/gpu": resource.MustParse("1"),
-						},
-					},
-				},
-			},
-			RestartPolicy: corev1.RestartPolicyNever,
-		},
-	}
-}
-
 func nvidiaDevicePluginDaemonSet() *appsv1.DaemonSet {
 	return &appsv1.DaemonSet{
 		ObjectMeta: metav1.ObjectMeta{
@@ -655,34 +633,3 @@ func podEnableAMDGPUResource(s *Scenario) *corev1.Pod {
 		},
 	}
 }
-
-func jobAMDGPUWorkload(s *Scenario) *batchv1.Job {
-	return &batchv1.Job{
-		ObjectMeta: metav1.ObjectMeta{
-			Name:      fmt.Sprintf("%s-gpu-validation-job", s.Runtime.KubeNodeName),
-			Namespace: defaultNamespace,
-		},
-		Spec: batchv1.JobSpec{
-			Template: corev1.PodTemplateSpec{
-				Spec: corev1.PodSpec{
-					Containers: []corev1.Container{
-						{
-							Name:  "gpu-validation-container",
-							Image: "mcr.microsoft.com/azuredocs/samples-tf-mnist-demo:gpu",
-							Args: []string{
-								"--max-steps", "1",
-							},
-							Resources: corev1.ResourceRequirements{
-								Limits: corev1.ResourceList{
-									"amd.com/gpu": resource.MustParse("1"),
-								},
-							},
-						},
-					},
-					RestartPolicy: corev1.RestartPolicyNever,
-				},
-			},
-			BackoffLimit: to.Ptr(int32(0)), // No retries, fail immediately if something goes wrong
-		},
-	}
-}