Addressing review comments

Signed-off-by: Vishesh Tanksale <vtanksale@nvidia.com>

Author: visheshtanksale

internal/controller/nimservice_controller.go

@@ -134,7 +134,10 @@ func (r *NIMServiceReconciler) Reconcile(ctx context.Context, req ctrl.Request)
 
 	logger.Info("Reconciling", "NIMService", nimService.Name)
 
-	if nimService.Spec.MultiNode != nil && nimService.Annotations != nil {
+	if nimService.Spec.MultiNode != nil {
+		if nimService.Annotations == nil {
+			nimService.Annotations = map[string]string{}
+		}
 		if _, ok := nimService.Annotations[utils.GPUCountPerPodAnnotationKey]; !ok {
 			gpuCountPerPod, err := shared.GetGPUCountPerPod(ctx, r.GetClient(), nimService)
 			if err != nil {
@@ -272,6 +275,16 @@ func (r *NIMServiceReconciler) SetupWithManager(mgr ctrl.Manager) error {
 		return err
 	}
 
+	exists, err := k8sutil.CRDExists(r.discoveryClient, resourcev1beta2.SchemeGroupVersion.WithResource("deviceclasses"))
+	if err != nil {
+		return err
+	}
+	if exists {
+		if _, err := mgr.GetCache().GetInformer(context.Background(), &resourcev1beta2.DeviceClass{}); err != nil {
+			return err
+		}
+	}
+
 	nimServiceBuilder := ctrl.NewControllerManagedBy(mgr).
 		For(&appsv1alpha1.NIMService{}).
 		Owns(&appsv1.Deployment{}).

internal/controller/platform/kserve/nimservice.go

@@ -578,7 +578,7 @@ func (r *NIMServiceReconciler) addGPUResources(ctx context.Context, nimService *
 	// Get tensorParallelism from the profile
 	tensorParallelism, err := utils.GetTensorParallelismByProfileTags(profile.Config)
 	if err != nil {
-		logger.Error(err, "Failed to retrieve tensorParallelism")
+		logger.Error(err, "Missing nvidia.com/gpu resource request/limit and unable to retrieve tensorParallelism for NIM profile")
 		return nil, err
 	}
 	if tensorParallelism != "" {

internal/shared/resourceclaims.go

@@ -331,8 +331,7 @@ func getGPUCountFromDeviceRequests(ctx context.Context, client client.Client, re
 
 		isGPU, err := isNVIDIAGPU(ctx, client, req.Exactly.DeviceClassName)
 		if err != nil {
-			// This allows partial success scenarios
-			continue
+			return 0, fmt.Errorf("failed to check if device class %s is a GPU: %w", req.Exactly.DeviceClassName, err)
 		}
 
 		if isGPU {
@@ -353,8 +352,7 @@ func getGPUCountFromDeviceSpecs(ctx context.Context, client client.Client, devic
 
 		isGPU, err := isNVIDIAGPU(ctx, client, dev.DeviceClassName)
 		if err != nil {
-			// This allows partial success scenarios
-			continue
+			return 0, fmt.Errorf("failed to check if device class %s is a GPU: %w", dev.DeviceClassName, err)
 		}
 
 		if isGPU {