fix: fix auto cluster creation failure (#229)

Author: hyunsun

VERSION

@@ -1 +1 @@
-1.3.1
+1.3.2

deployment/charts/intel-infra-provider-crds/Chart.yaml

@@ -5,8 +5,8 @@ apiVersion: v2
 name: intel-infra-provider-crds
 description: Custom Resource Definitions for Intel Cluster API Provider
 type: application
-version: 1.3.1
+version: 1.3.2
 annotations:
-  revision: 45825ac8d246455282c07fef4b1943d0828615a5
-  created: "2025-11-25T04:50:31Z"
-appVersion: 1.3.1
+  revision: 34058c65b59cd6f16dc4f45836105c007b786946
+  created: "2025-11-25T23:18:25Z"
+appVersion: 1.3.2

deployment/charts/intel-infra-provider/Chart.yaml

@@ -5,8 +5,8 @@ apiVersion: v2
 name: intel-infra-provider
 description: The Intel Cluster API Provider
 type: application
-version: 1.3.1
+version: 1.3.2
 annotations:
-  revision: 45825ac8d246455282c07fef4b1943d0828615a5
-  created: "2025-11-25T04:50:31Z"
-appVersion: 1.3.1
+  revision: 34058c65b59cd6f16dc4f45836105c007b786946
+  created: "2025-11-25T23:18:25Z"
+appVersion: 1.3.2

internal/southboundhandler/handler.go

@@ -259,9 +259,11 @@ func (h *Handler) UpdateStatus(ctx context.Context, nodeGUID string, status pb.U
 	// IntelMachine for the node doesn't exist yet
 	if intelmachine == nil {
 		// unpause the cluster if paused
-		cluster, err := getCluster(ctx, h.client, projectId, nodeGUID)
+		cluster, err := h.getCluster(ctx, projectId, nodeGUID)
 		if cluster != nil && cluster.Spec.Paused {
+			log.Debug().Msgf("Unpausing cluster %s/%s", cluster.Namespace, cluster.Name)
 			err = unpauseCluster(ctx, h.client, cluster)
+			log.Debug().Msgf("Cluster unpause result: %v", err)
 		}
 		return pb.UpdateClusterStatusResponse_NONE, err
 	}
@@ -393,16 +395,31 @@ func (h *Handler) getIntelMachine(ctx context.Context, client ctrlclient.Client,
 	return intelMachine, nil
 }
 
-func getCluster(ctx context.Context, client ctrlclient.Client, projectId string, nodeID string) (*clusterv1.Cluster, error) {
+func (h *Handler) getCluster(ctx context.Context, projectId string, nodeID string) (*clusterv1.Cluster, error) {
 	// see if there is machine binding for this node
 	var machineBindingList infrastructurev1alpha1.IntelMachineBindingList
 
-	if err := client.List(ctx, &machineBindingList,
-		ctrlclient.InNamespace(projectId),
-		ctrlclient.MatchingFields{nodeGUIDKey: nodeID}); err != nil {
+	// TODO: always convert nodeID(GUID format) to hostID before search the cluster once host
+	// unification from cluster-manager is done. For 3.2 release, try to get
+	// hostID by UUID if no machine binding is found.
+	if err := h.client.List(ctx, &machineBindingList, ctrlclient.InNamespace(projectId), ctrlclient.MatchingFields{nodeGUIDKey: nodeID}); err != nil {
 		return nil, fmt.Errorf("failed to get intel machine binding list: %w", err)
 	}
 
+	if len(machineBindingList.Items) == 0 && h.inventoryClient != nil {
+		log.Debug().Msgf("No IntelMachineBinding found for node %s in project %s, attempt to get hostID by UUID", nodeID, projectId)
+		host, err := h.inventoryClient.Client.GetHostByUUID(ctx, projectId, nodeID)
+		if err != nil {
+			return nil, fmt.Errorf("failed to get host resource by UUID: %w", err)
+		}
+
+		if err := h.client.List(ctx, &machineBindingList,
+			ctrlclient.InNamespace(projectId),
+			ctrlclient.MatchingFields{nodeGUIDKey: host.GetResourceId()}); err != nil {
+			return nil, fmt.Errorf("failed to get intel machine binding list: %w", err)
+		}
+	}
+
 	if len(machineBindingList.Items) == 0 {
 		// no cluster is available for this node
 		return nil, nil
@@ -412,9 +429,10 @@ func getCluster(ctx context.Context, client ctrlclient.Client, projectId string,
 	}
 
 	// one cluster is found for the node
+	log.Debug().Msgf("Found cluster for node %s %s/%s", nodeID, projectId, machineBindingList.Items[0].Spec.ClusterName)
 	cluster := &clusterv1.Cluster{}
 	key := types.NamespacedName{Namespace: projectId, Name: machineBindingList.Items[0].Spec.ClusterName}
-	if err := client.Get(ctx, key, cluster); err != nil {
+	if err := h.client.Get(ctx, key, cluster); err != nil {
 		return nil, err
 	}