add flag to skip deployments for fake nodes

deploy/fake-gpu-operator/templates/status-updater/clusterrole.yaml

@@ -8,6 +8,7 @@ rules:
     resources:
       - pods
       - nodes
+      - nodes/status
     verbs:
       - get
       - list
@@ -28,4 +29,4 @@ rules:
     resources:
       - podgroups
     verbs:
-      - get
+      - get

internal/common/topology/types.go

@@ -48,7 +48,8 @@ type Range struct {
 }
 
 type Config struct {
-	NodeAutofill NodeAutofillSettings `yaml:"node-autofill"`
+	NodeAutofill     NodeAutofillSettings `yaml:"node-autofill"`
+	FakeNodeHandling bool                 `yaml:"fake-node-handling"`
 }
 
 type NodeAutofillSettings struct {

internal/status-updater/controllers/node/controller.go

@@ -53,6 +53,10 @@ func NewNodeController(kubeClient kubernetes.Interface, wg *sync.WaitGroup) *Nod
 				node := obj.(*v1.Node)
 				util.LogErrorIfExist(c.handler.HandleAdd(node), "Failed to handle node addition")
 			},
+			UpdateFunc: func(oldObj, newObj interface{}) {
+				newNode := newObj.(*v1.Node)
+				util.LogErrorIfExist(c.handler.HandleUpdate(newNode), "Failed to handle node addition")
+			},
 			DeleteFunc: func(obj interface{}) {
 				node := obj.(*v1.Node)
 				util.LogErrorIfExist(c.handler.HandleDelete(node), "Failed to handle node deletion")

internal/status-updater/handlers/node/fake_node_deployments.go

@@ -11,9 +11,27 @@ import (
 	"k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/utils/ptr"
 )
 
+func (p *NodeHandler) applyFakeDevicePlugin(gpuCount int, node *v1.Node) error {
+	if !isFakeNode(node) {
+		return nil
+	}
+
+	patch := fmt.Sprintf(
+		`{"status": {"capacity": {"%s": "%d"}, "allocatable": {"%s": "%d"}}}`,
+		constants.GpuResourceName, gpuCount, constants.GpuResourceName, gpuCount,
+	)
+	_, err := p.kubeClient.CoreV1().Nodes().Patch(context.TODO(), node.Name, types.MergePatchType, []byte(patch), metav1.PatchOptions{}, "status")
+	if err != nil {
+		return fmt.Errorf("failed to update node capacity and allocatable: %v", err)
+	}
+
+	return nil
+}
+
 func (p *NodeHandler) applyFakeNodeDeployments(node *v1.Node) error {
 	if !isFakeNode(node) {
 		return nil

internal/status-updater/handlers/node/handler.go

@@ -13,6 +13,7 @@ import (
 type Interface interface {
 	HandleAdd(node *v1.Node) error
 	HandleDelete(node *v1.Node) error
+	HandleUpdate(node *v1.Node) error
 }
 
 type NodeHandler struct {
@@ -30,14 +31,26 @@ func NewNodeHandler(kubeClient kubernetes.Interface) *NodeHandler {
 func (p *NodeHandler) HandleAdd(node *v1.Node) error {
 	log.Printf("Handling node addition: %s\n", node.Name)
 
-	err := p.createNodeTopologyCM(node)
+	baseTopology, err := topology.GetBaseTopologyFromCM(p.kubeClient)
 	if err != nil {
-		return fmt.Errorf("failed to create node topology ConfigMap: %w", err)
+		return fmt.Errorf("failed to get base topology: %w", err)
 	}
 
-	err = p.applyFakeNodeDeployments(node)
+	err = p.createNodeTopologyCM(node, baseTopology)
 	if err != nil {
-		return fmt.Errorf("failed to apply fake node deployments: %w", err)
+		return fmt.Errorf("failed to create node topology ConfigMap: %w", err)
+	}
+
+	if baseTopology.Config.FakeNodeHandling {
+		err = p.applyFakeDevicePlugin(baseTopology.Config.NodeAutofill.GpuCount, node)
+		if err != nil {
+			return fmt.Errorf("failed to apply fake node deployments: %w", err)
+		}
+	} else {
+		err = p.applyFakeNodeDeployments(node)
+		if err != nil {
+			return fmt.Errorf("failed to apply fake node deployments: %w", err)
+		}
 	}
 
 	return nil
@@ -58,3 +71,21 @@ func (p *NodeHandler) HandleDelete(node *v1.Node) error {
 
 	return nil
 }
+
+func (p *NodeHandler) HandleUpdate(node *v1.Node) error {
+	baseTopology, err := topology.GetBaseTopologyFromCM(p.kubeClient)
+	if err != nil {
+		return fmt.Errorf("failed to get base topology: %w", err)
+	}
+
+	if !baseTopology.Config.FakeNodeHandling {
+		return nil
+	}
+
+	gpuCount := baseTopology.Config.NodeAutofill.GpuCount
+	err = p.applyFakeDevicePlugin(gpuCount, node)
+	if err != nil {
+		return fmt.Errorf("failed to apply fake node deployments: %w", err)
+	}
+	return nil
+}

internal/status-updater/handlers/node/topology_cm.go

@@ -8,17 +8,14 @@ import (
 	v1 "k8s.io/api/core/v1"
 )
 
-func (p *NodeHandler) createNodeTopologyCM(node *v1.Node) error {
+func (p *NodeHandler) createNodeTopologyCM(
+	node *v1.Node, baseTopology *topology.BaseTopology,
+) error {
 	nodeTopology, _ := topology.GetNodeTopologyFromCM(p.kubeClient, node.Name)
 	if nodeTopology != nil {
 		return nil
 	}
 
-	baseTopology, err := topology.GetBaseTopologyFromCM(p.kubeClient)
-	if err != nil {
-		return fmt.Errorf("failed to get base topology: %w", err)
-	}
-
 	nodeAutofillSettings := baseTopology.Config.NodeAutofill
 
 	nodeTopology = &topology.NodeTopology{
@@ -28,7 +25,7 @@ func (p *NodeHandler) createNodeTopologyCM(node *v1.Node) error {
 		MigStrategy: nodeAutofillSettings.MigStrategy,
 	}
 
-	err = topology.CreateNodeTopologyCM(p.kubeClient, nodeTopology, node.Name)
+	err := topology.CreateNodeTopologyCM(p.kubeClient, nodeTopology, node.Name)
 	if err != nil {
 		return fmt.Errorf("failed to create node topology: %w", err)
 	}