[WIP]EKS V6 fix

api/v1alpha6/conditions_consts.go

@@ -51,3 +51,13 @@ const (
 	// FloatingIPErrorReason used when the floating ip could not be created or attached.
 	FloatingIPErrorReason = "FloatingIPError"
 )
+
+const (
+	ClusterReadyCondition clusterv1.ConditionType = "ClusterReady"
+	// ClusterNotReadyReason used when create the cluster failed.
+	ClusterNotReadyReason = "ClusterNotReady"
+	// ClusterReadyReason reports on current status of the OpenStack cluster. Ready indicates the cluster is in a OK state.
+	ClusterReadyReason = "ClusterReady"
+	// LoadBalancerReconcileErrorReason used when the openstack create loadbalancer error.
+	LoadBalancerReconcileErrorReason = "LoadBalancerReconcileError"
+)

api/v1alpha6/openstackcluster_types.go

@@ -64,7 +64,7 @@ type OpenStackClusterSpec struct {
 	APIServerLoadBalancer APIServerLoadBalancer `json:"apiServerLoadBalancer,omitempty"`
 
 	// DisableFloatingIP determines whether or not to attempt to attach a floating
-	// IP to the Instance. 
+	// IP to the Instance.
 	DisableFloatingIP bool `json:"disableFloatingIP"`
 
 	// DisableAPIServerFloatingIP determines whether or not to attempt to attach a floating
@@ -214,6 +214,8 @@ type OpenStackClusterStatus struct {
 	// and/or logged in the controller's output.
 	// +optional
 	FailureMessage *string `json:"failureMessage,omitempty"`
+
+	Conditions clusterv1.Conditions `json:"conditions,omitempty"`
 }
 
 // +kubebuilder:object:root=true
@@ -249,3 +251,13 @@ type OpenStackClusterList struct {
 func init() {
 	SchemeBuilder.Register(&OpenStackCluster{}, &OpenStackClusterList{})
 }
+
+// GetConditions returns the observations of the operational state of the OpenStackMachine resource.
+func (r *OpenStackCluster) GetConditions() clusterv1.Conditions {
+	return r.Status.Conditions
+}
+
+// SetConditions sets the underlying service state of the OpenStackMachine to the predescribed clusterv1.Conditions.
+func (r *OpenStackCluster) SetConditions(conditions clusterv1.Conditions) {
+	r.Status.Conditions = conditions
+}

api/v1alpha6/openstackmachine_types.go

@@ -90,7 +90,7 @@ type OpenStackMachineSpec struct {
 
 	// The volume metadata to boot from
 	RootVolume *RootVolume `json:"rootVolume,omitempty"`
-	
+
 	// The custome metadata to boot from
 	CustomeVolumes []*RootVolume `json:"customeVolumes,omitempty"`
 

api/v1alpha6/openstackmachine_webhook.go

@@ -88,15 +88,14 @@ func (r *OpenStackMachine) ValidateUpdate(old runtime.Object) error {
 	newOpenStackMachineSpec := newOpenStackMachine["spec"].(map[string]interface{})
 	oldOpenStackMachineSpec := oldOpenStackMachine["spec"].(map[string]interface{})
 
-
 	// allow changes to providerID always
-	if oldOpenStackMachineSpec["providerID"] != nil  || newOpenStackMachineSpec["providerID"] != nil {
+	if oldOpenStackMachineSpec["providerID"] != nil || newOpenStackMachineSpec["providerID"] != nil {
 		delete(oldOpenStackMachineSpec, "providerID")
 		delete(newOpenStackMachineSpec, "providerID")
 	}
 
 	// allow changes to instanceID always
-	if oldOpenStackMachineSpec["instanceID"] != nil  || newOpenStackMachineSpec["instanceID"] != nil {
+	if oldOpenStackMachineSpec["instanceID"] != nil || newOpenStackMachineSpec["instanceID"] != nil {
 		delete(oldOpenStackMachineSpec, "instanceID")
 		delete(newOpenStackMachineSpec, "instanceID")
 	}

config/crd/bases/infrastructure.cluster.x-k8s.io_openstackclusters.yaml

@@ -5308,6 +5308,52 @@ spec:
                 - name
                 - rules
                 type: object
+              conditions:
+                description: Conditions provide observations of the operational state
+                  of a Cluster API resource.
+                items:
+                  description: Condition defines an observation of a Cluster API resource
+                    operational state.
+                  properties:
+                    lastTransitionTime:
+                      description: Last time the condition transitioned from one status
+                        to another. This should be when the underlying condition changed.
+                        If that is not known, then using the time when the API field
+                        changed is acceptable.
+                      format: date-time
+                      type: string
+                    message:
+                      description: A human readable message indicating details about
+                        the transition. This field may be empty.
+                      type: string
+                    reason:
+                      description: The reason for the condition's last transition
+                        in CamelCase. The specific API may choose whether or not this
+                        field is considered a guaranteed API. This field may not be
+                        empty.
+                      type: string
+                    severity:
+                      description: Severity provides an explicit classification of
+                        Reason code, so the users or machines can immediately understand
+                        the current situation and act accordingly. The Severity field
+                        MUST be set only when Status=False.
+                      type: string
+                    status:
+                      description: Status of the condition, one of True, False, Unknown.
+                      type: string
+                    type:
+                      description: Type of condition in CamelCase or in foo.example.com/CamelCase.
+                        Many .condition.type values are consistent across resources
+                        like Available, but because arbitrary conditions can be useful
+                        (see .node.status.conditions), the ability to deconflict is
+                        important.
+                      type: string
+                  required:
+                  - lastTransitionTime
+                  - status
+                  - type
+                  type: object
+                type: array
               controlPlaneSecurityGroup:
                 description: 'ControlPlaneSecurityGroups contains all the information
                   about the OpenStack Security Group that needs to be applied to control

controllers/openstackcluster_controller.go

@@ -20,6 +20,11 @@ import (
 	"context"
 	"fmt"
 	"reflect"
+	"strings"
+
+	kerrors "k8s.io/apimachinery/pkg/util/errors"
+
+	"sigs.k8s.io/cluster-api/util/conditions"
 
 	"github.com/pkg/errors"
 	corev1 "k8s.io/api/core/v1"
@@ -104,10 +109,8 @@ func (r *OpenStackClusterReconciler) Reconcile(ctx context.Context, req ctrl.Req
 
 	// Always patch the openStackCluster when exiting this function so we can persist any OpenStackCluster changes.
 	defer func() {
-		if err := patchHelper.Patch(ctx, openStackCluster); err != nil {
-			if reterr == nil {
-				reterr = errors.Wrapf(err, "error patching OpenStackCluster %s/%s", openStackCluster.Namespace, openStackCluster.Name)
-			}
+		if err := patchCluster(ctx, patchHelper, openStackCluster); err != nil {
+			reterr = kerrors.NewAggregate([]error{reterr, err})
 		}
 	}()
 
@@ -132,6 +135,14 @@ func (r *OpenStackClusterReconciler) Reconcile(ctx context.Context, req ctrl.Req
 	return reconcileNormal(ctx, scope, patchHelper, cluster, openStackCluster)
 }
 
+func patchCluster(ctx context.Context, patchHelper *patch.Helper, openStackCluster *infrav1.OpenStackCluster, options ...patch.Option) error {
+	err := patchHelper.Patch(ctx, openStackCluster, options...)
+	if err != nil {
+		return err
+	}
+	return nil
+}
+
 func reconcileDelete(ctx context.Context, scope *scope.Scope, patchHelper *patch.Helper, cluster *clusterv1.Cluster, openStackCluster *infrav1.OpenStackCluster) (ctrl.Result, error) {
 	scope.Logger.Info("Reconciling Cluster delete")
 
@@ -146,15 +157,25 @@ func reconcileDelete(ctx context.Context, scope *scope.Scope, patchHelper *patch
 
 	clusterName := fmt.Sprintf("%s-%s", cluster.Namespace, cluster.Name)
 
+	var skipLBDeleting bool
+
 	if openStackCluster.Spec.APIServerLoadBalancer.Enabled {
 		loadBalancerService, err := loadbalancer.NewService(scope)
 		if err != nil {
-			return reconcile.Result{}, err
+			if strings.EqualFold(err.Error(), loadbalancer.ErrLoadBalancerNoPoint) {
+				handleUpdateOSCError(openStackCluster, fmt.Errorf("failed to init delete load balancer client: %w", err))
+				skipLBDeleting = true
+			}
+			if !skipLBDeleting {
+				return reconcile.Result{}, err
+			}
 		}
 
-		if err = loadBalancerService.DeleteLoadBalancer(openStackCluster, clusterName); err != nil {
-			handleUpdateOSCError(openStackCluster, errors.Errorf("failed to delete load balancer: %v", err))
-			return reconcile.Result{}, errors.Errorf("failed to delete load balancer: %v", err)
+		if !skipLBDeleting {
+			if err = loadBalancerService.DeleteLoadBalancer(openStackCluster, clusterName); err != nil {
+				handleUpdateOSCError(openStackCluster, errors.Errorf("failed to delete load balancer: %v", err))
+				return reconcile.Result{}, errors.Errorf("failed to delete load balancer: %v", err)
+			}
 		}
 	}
 
@@ -249,6 +270,10 @@ func deleteBastion(scope *scope.Scope, cluster *clusterv1.Cluster, openStackClus
 func reconcileNormal(ctx context.Context, scope *scope.Scope, patchHelper *patch.Helper, cluster *clusterv1.Cluster, openStackCluster *infrav1.OpenStackCluster) (ctrl.Result, error) {
 	scope.Logger.Info("Reconciling Cluster")
 
+	if openStackCluster.Status.FailureReason != nil || openStackCluster.Status.FailureMessage != nil {
+		scope.Logger.Info("Not reconciling cluster in failed state. See openStackCluster.status.failureReason, openStackCluster.status.failureMessage, or previously logged error for details")
+		return ctrl.Result{}, nil
+	}
 	// If the OpenStackCluster doesn't have our finalizer, add it.
 	controllerutil.AddFinalizer(openStackCluster, infrav1.ClusterFinalizer)
 	// Register the finalizer immediately to avoid orphaning OpenStack resources on delete
@@ -382,14 +407,14 @@ func reconcileBastion(scope *scope.Scope, cluster *clusterv1.Cluster, openStackC
 func bastionToInstanceSpec(openStackCluster *infrav1.OpenStackCluster, clusterName string, userData string) *compute.InstanceSpec {
 	name := fmt.Sprintf("%s-bastion", clusterName)
 	instanceSpec := &compute.InstanceSpec{
-		Name:          name,
-		Flavor:        openStackCluster.Spec.Bastion.Instance.Flavor,
-		SSHKeyName:    openStackCluster.Spec.Bastion.Instance.SSHKeyName,
-		Image:         openStackCluster.Spec.Bastion.Instance.Image,
-		UserData:      userData,
-		ImageUUID:     openStackCluster.Spec.Bastion.Instance.ImageUUID,
-		FailureDomain: openStackCluster.Spec.Bastion.AvailabilityZone,
-		RootVolume:    openStackCluster.Spec.Bastion.Instance.RootVolume,
+		Name:                      name,
+		Flavor:                    openStackCluster.Spec.Bastion.Instance.Flavor,
+		SSHKeyName:                openStackCluster.Spec.Bastion.Instance.SSHKeyName,
+		Image:                     openStackCluster.Spec.Bastion.Instance.Image,
+		UserData:                  userData,
+		ImageUUID:                 openStackCluster.Spec.Bastion.Instance.ImageUUID,
+		FailureDomain:             openStackCluster.Spec.Bastion.AvailabilityZone,
+		RootVolume:                openStackCluster.Spec.Bastion.Instance.RootVolume,
 		DeleteVolumeOnTermination: openStackCluster.Spec.Bastion.Instance.DeleteVolumeOnTermination,
 	}
 
@@ -512,6 +537,10 @@ func reconcileNetworkComponents(scope *scope.Scope, cluster *clusterv1.Cluster,
 	if openStackCluster.Spec.APIServerLoadBalancer.Enabled {
 		loadBalancerService, err := loadbalancer.NewService(scope)
 		if err != nil {
+			if strings.EqualFold(err.Error(), loadbalancer.ErrLoadBalancerNoPoint) {
+				handleUpdateOSCError(openStackCluster, fmt.Errorf("failed to init load balancer client: %w", err))
+				conditions.MarkFalse(openStackCluster, infrav1.ClusterReadyReason, infrav1.LoadBalancerReconcileErrorReason, clusterv1.ConditionSeverityError, err.Error())
+			}
 			return err
 		}
 
@@ -616,4 +645,5 @@ func handleUpdateOSCError(openstackCluster *infrav1.OpenStackCluster, message er
 	err := capierrors.UpdateClusterError
 	openstackCluster.Status.FailureReason = &err
 	openstackCluster.Status.FailureMessage = pointer.StringPtr(message.Error())
+	openstackCluster.Status.Ready = false
 }