Create new controller for all node labeling operations

cmd/gpu-operator/main.go

@@ -210,6 +210,16 @@ func main() {
 		setupLog.Error(err, "unable to create controller", "controller", "NVIDIADriver")
 		os.Exit(1)
 	}
+
+	if err = (&controllers.NodeLabelingReconciler{
+		Namespace: operatorNamespace,
+		Client:    mgr.GetClient(),
+		Scheme:    mgr.GetScheme(),
+		Log:       ctrl.Log.WithName("controllers").WithName("NodeLabeling"),
+	}).SetupWithManager(ctx, mgr); err != nil {
+		setupLog.Error(err, "unable to create controller", "controller", "NodeLabeling")
+		os.Exit(1)
+	}
 	// +kubebuilder:scaffold:builder
 	if err := mgr.AddHealthzCheck("health", healthz.Ping); err != nil {
 		setupLog.Error(err, "unable to set up health check")

config/rbac/role.yaml

@@ -11,7 +11,6 @@ rules:
   - endpoints
   - events
   - namespaces
-  - nodes
   - persistentvolumeclaims
   - pods
   - pods/eviction
@@ -27,6 +26,16 @@ rules:
   - patch
   - update
   - watch
+- apiGroups:
+  - ""
+  resources:
+  - nodes
+  verbs:
+  - get
+  - list
+  - patch
+  - update
+  - watch
 - apiGroups:
   - apiextensions.k8s.io
   resources:

controllers/clusterpolicy_controller.go

@@ -70,7 +70,8 @@ type ClusterPolicyReconciler struct {
 // +kubebuilder:rbac:groups=security.openshift.io,resources=securitycontextconstraints,verbs=use,resourceNames=privileged
 // +kubebuilder:rbac:groups=rbac.authorization.k8s.io,resources=clusterroles;clusterrolebindings;roles;rolebindings,verbs=*
 // +kubebuilder:rbac:groups="",resources=namespaces;serviceaccounts;pods;pods/eviction;services;services/finalizers;endpoints,verbs=get;list;watch;create;update;patch;delete
-// +kubebuilder:rbac:groups="",resources=persistentvolumeclaims;events;configmaps;secrets;nodes,verbs=get;list;watch;create;update;patch;delete
+// +kubebuilder:rbac:groups="",resources=persistentvolumeclaims;events;configmaps;secrets,verbs=get;list;watch;create;update;patch;delete
+// +kubebuilder:rbac:groups="",resources=nodes,verbs=get;list;watch
 // +kubebuilder:rbac:groups=apps,resources=deployments;daemonsets;replicasets;statefulsets,verbs=get;list;watch;create;update;patch;delete
 // +kubebuilder:rbac:groups=apps,resources=controllerrevisions,verbs=get;list;watch
 // +kubebuilder:rbac:groups=monitoring.coreos.com,resources=servicemonitors;prometheusrules,verbs=get;list;watch;create;update;patch;delete
@@ -281,9 +282,9 @@ func addWatchNewGPUNode(r *ClusterPolicyReconciler, c controller.Controller, mgr
 			oldLabels := e.ObjectOld.GetLabels()
 			nodeName := e.ObjectNew.GetName()
 
-			gpuCommonLabelMissing := hasGPULabels(newLabels) && !hasCommonGPULabel(newLabels)
+			// Trigger when NodeLabelingReconciler sets gpu.present=true on a new GPU node.
+			gpuCommonLabelAdded := !hasCommonGPULabel(oldLabels) && hasCommonGPULabel(newLabels)
 			gpuCommonLabelOutdated := !hasGPULabels(newLabels) && hasCommonGPULabel(newLabels)
-			migManagerLabelMissing := hasMIGCapableGPU(newLabels) && !hasMIGManagerLabel(newLabels)
 			commonOperandsLabelChanged := hasOperandsDisabled(oldLabels) != hasOperandsDisabled(newLabels)
 
 			oldGPUWorkloadConfig, _ := getWorkloadConfig(oldLabels, true)
@@ -294,19 +295,17 @@ func addWatchNewGPUNode(r *ClusterPolicyReconciler, c controller.Controller, mgr
 			newOSTreeLabel := newLabels[nfdOSTreeVersionLabelKey]
 			osTreeLabelChanged := oldOSTreeLabel != newOSTreeLabel
 
-			needsUpdate := gpuCommonLabelMissing ||
+			needsUpdate := gpuCommonLabelAdded ||
 				gpuCommonLabelOutdated ||
-				migManagerLabelMissing ||
 				commonOperandsLabelChanged ||
 				gpuWorkloadConfigLabelChanged ||
 				osTreeLabelChanged
 
 			if needsUpdate {
 				r.Log.Info("Node needs an update",
 					"name", nodeName,
-					"gpuCommonLabelMissing", gpuCommonLabelMissing,
+					"gpuCommonLabelAdded", gpuCommonLabelAdded,
 					"gpuCommonLabelOutdated", gpuCommonLabelOutdated,
-					"migManagerLabelMissing", migManagerLabelMissing,
 					"commonOperandsLabelChanged", commonOperandsLabelChanged,
 					"gpuWorkloadConfigLabelChanged", gpuWorkloadConfigLabelChanged,
 					"osTreeLabelChanged", osTreeLabelChanged,