Add health to status field of duros resource. (#19)

Makefile

@@ -87,7 +87,7 @@ ifeq (, $(shell which controller-gen))
 	CONTROLLER_GEN_TMP_DIR=$$(mktemp -d) ;\
 	cd $$CONTROLLER_GEN_TMP_DIR ;\
 	go mod init tmp ;\
-	go get sigs.k8s.io/controller-tools/cmd/controller-gen@v0.5.0 ;\
+	go install sigs.k8s.io/controller-tools/cmd/controller-gen@v0.6.2 ;\
 	rm -rf $$CONTROLLER_GEN_TMP_DIR ;\
 	}
 CONTROLLER_GEN=$(GOBIN)/controller-gen

api/v1/duros_types.go

@@ -27,6 +27,7 @@ import (
 // +kubebuilder:object:root=true
 // +kubebuilder:printcolumn:name="ProjectID",type=string,JSONPath=`.spec.metalProjectID`
 // +kubebuilder:printcolumn:name="StorageClasses",type=string,JSONPath=`.spec.storageClasses`
+// +kubebuilder:subresource:status
 type Duros struct {
 	metav1.TypeMeta   `json:",inline"`
 	metav1.ObjectMeta `json:"metadata,omitempty"`
@@ -57,8 +58,34 @@ type DurosStatus struct {
 	// SecretRef to the create JWT Token
 	// TODO, this can be used to detect required key rotation
 	SecretRef string `json:"secret,omitempty" description:"Reference to JWT Token generated on the duros storage side for this project"`
+
+	// ManagedResourceStatuses contains a list of statuses of resources managed by this controller
+	ManagedResourceStatuses []ManagedResourceStatus `json:"managedResourceStatuses" description:"A list of managed resource statuses"`
+}
+
+type ManagedResourceStatus struct {
+	// Name is the name of the resource described by this status
+	Name string `json:"name" description:"The name of the resource"`
+	// Group is the api group kind of the resource described by this status
+	Group string `json:"group" description:"The group kind of the resource"`
+	// State is the actual state of the managed resource
+	State HealthState `json:"state" description:"The state of this resource"`
+	// Description further describes the state of the managed resource
+	Description string `json:"description" description:"The description of the state of this component"`
+	// LastUpdateTime is the last time the status was updated
+	LastUpdateTime metav1.Time `json:"lastUpdateTime" description:"The time when this status was last updated"`
 }
 
+// HealthState describes the state of a managed resource
+type HealthState string
+
+const (
+	// HealthStateRunning indicates that the resource is running
+	HealthStateRunning HealthState = "Running"
+	// HealthStateNotRunning indicates that the resource is not running
+	HealthStateNotRunning HealthState = "Not Running"
+)
+
 // StorageClass defines the storageClass parameters
 type StorageClass struct {
 	Name         string `json:"name"`

config/crd/bases/storage.metal-stack.io_duros.yaml

@@ -4,7 +4,7 @@ apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
   annotations:
-    controller-gen.kubebuilder.io/version: v0.4.0
+    controller-gen.kubebuilder.io/version: v0.6.2
   creationTimestamp: null
   name: duros.storage.metal-stack.io
 spec:
@@ -68,15 +68,51 @@ spec:
           status:
             description: DurosStatus defines the observed state of Duros
             properties:
+              managedResourceStatuses:
+                description: ManagedResourceStatuses contains a list of statuses of
+                  resources managed by this controller
+                items:
+                  properties:
+                    description:
+                      description: Description further describes the state of the
+                        managed resource
+                      type: string
+                    group:
+                      description: Group is the api group kind of the resource described
+                        by this status
+                      type: string
+                    lastUpdateTime:
+                      description: LastUpdateTime is the last time the status was
+                        updated
+                      format: date-time
+                      type: string
+                    name:
+                      description: Name is the name of the resource described by this
+                        status
+                      type: string
+                    state:
+                      description: State is the actual state of the managed resource
+                      type: string
+                  required:
+                  - description
+                  - group
+                  - lastUpdateTime
+                  - name
+                  - state
+                  type: object
+                type: array
               secret:
                 description: SecretRef to the create JWT Token TODO, this can be used
                   to detect required key rotation
                 type: string
+            required:
+            - managedResourceStatuses
             type: object
         type: object
     served: true
     storage: true
-    subresources: {}
+    subresources:
+      status: {}
 status:
   acceptedNames:
     kind: ""

controllers/duros_controller.go

@@ -23,23 +23,26 @@ import (
 
 	"github.com/go-logr/logr"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
-	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/types"
 	ctrl "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/client"
+	"sigs.k8s.io/controller-runtime/pkg/predicate"
 
 	"github.com/metal-stack/duros-go"
 	durosv2 "github.com/metal-stack/duros-go/api/duros/v2"
 
 	storagev1 "github.com/metal-stack/duros-controller/api/v1"
 	v1 "github.com/metal-stack/duros-controller/api/v1"
+
+	appsv1 "k8s.io/api/apps/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
 // DurosReconciler reconciles a Duros object
 type DurosReconciler struct {
 	client.Client
 	Shoot       client.Client
 	Log         logr.Logger
-	Scheme      *runtime.Scheme
 	Namespace   string
 	DurosClient durosv2.DurosAPIClient
 	Endpoints   duros.EPs
@@ -66,12 +69,13 @@ func (r *DurosReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl
 		return ctrl.Result{}, nil
 	}
 	// first get the metal-api projectID
-	var duros storagev1.Duros
-	if err := r.Get(ctx, req.NamespacedName, &duros); err != nil {
+	duros := &storagev1.Duros{}
+	if err := r.Get(ctx, req.NamespacedName, duros); err != nil {
 		if apierrors.IsNotFound(err) {
 			log.Info("no duros storage defined")
-			return requeue, err
+			return ctrl.Result{}, nil
 		}
+		return requeue, err
 	}
 	err := validateDuros(duros)
 	if err != nil {
@@ -102,17 +106,90 @@ func (r *DurosReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl
 	if err != nil {
 		return requeue, err
 	}
-	return ctrl.Result{}, nil
+
+	err = r.reconcileStatus(ctx, duros)
+	if err != nil {
+		return requeue, err
+	}
+
+	return ctrl.Result{
+		// we requeue in a small interval to ensure resources are recreated quickly
+		// and status is updated regularly
+		RequeueAfter: 30 * time.Second,
+	}, nil
+}
+
+func (r *DurosReconciler) reconcileStatus(ctx context.Context, duros *storagev1.Duros) error {
+	var (
+		updateTime = metav1.NewTime(time.Now())
+		ds         = &appsv1.DaemonSet{}
+		sts        = &appsv1.StatefulSet{}
+	)
+
+	duros.Status.SecretRef = "" // TODO?
+
+	err := r.Shoot.Get(ctx, types.NamespacedName{Name: lbCSINodeName, Namespace: namespace}, ds)
+	if err != nil {
+		return fmt.Errorf("error getting daemon set: %w", err)
+	}
+
+	dsStatus := v1.ManagedResourceStatus{
+		Name:           ds.Name,
+		Group:          "DaemonSet", // ds.GetObjectKind().GroupVersionKind().String() --> this does not work :(
+		State:          v1.HealthStateRunning,
+		Description:    "All replicas are ready",
+		LastUpdateTime: updateTime,
+	}
+
+	if ds.Status.DesiredNumberScheduled != ds.Status.NumberReady {
+		dsStatus.State = v1.HealthStateNotRunning
+		dsStatus.Description = fmt.Sprintf("%d/%d replicas are ready", ds.Status.NumberReady, ds.Status.DesiredNumberScheduled)
+	}
+
+	err = r.Shoot.Get(ctx, types.NamespacedName{Name: lbCSIControllerName, Namespace: namespace}, sts)
+	if err != nil {
+		return fmt.Errorf("error getting statefulset: %w", err)
+	}
+
+	stsStatus := v1.ManagedResourceStatus{
+		Name:           sts.Name,
+		Group:          "StatefulSet", // sts.GetObjectKind().GroupVersionKind().String() --> this does not work :(
+		State:          v1.HealthStateRunning,
+		Description:    "All replicas are ready",
+		LastUpdateTime: updateTime,
+	}
+
+	replicas := int32(1)
+	if sts.Spec.Replicas != nil {
+		replicas = *sts.Spec.Replicas
+	}
+
+	if replicas != sts.Status.ReadyReplicas {
+		stsStatus.State = v1.HealthStateNotRunning
+		stsStatus.Description = fmt.Sprintf("%d/%d replicas are ready", sts.Status.ReadyReplicas, replicas)
+	}
+
+	duros.Status.ManagedResourceStatuses = []v1.ManagedResourceStatus{dsStatus, stsStatus}
+	err = r.Status().Update(ctx, duros)
+	if err != nil {
+		return fmt.Errorf("error updating status: %w", err)
+	}
+
+	r.Log.Info("status updated", "name", duros.Name)
+
+	return nil
 }
 
 // SetupWithManager boilerplate to setup the Reconciler
 func (r *DurosReconciler) SetupWithManager(mgr ctrl.Manager) error {
+	pred := predicate.GenerationChangedPredicate{} // prevents reconcile on status sub resource update
 	return ctrl.NewControllerManagedBy(mgr).
 		For(&storagev1.Duros{}).
+		WithEventFilter(pred).
 		Complete(r)
 }
 
-func validateDuros(duros v1.Duros) error {
+func validateDuros(duros *v1.Duros) error {
 	if len(duros.Spec.MetalProjectID) == 0 {
 		return fmt.Errorf("metalProjectID is empty")
 	}

controllers/storageclass.go

@@ -28,6 +28,9 @@ const (
 	provisioner = "csi.lightbitslabs.com"
 
 	storageClassCredentialsRef = "lb-csi-creds"
+
+	lbCSIControllerName = "lb-csi-controller"
+	lbCSINodeName       = "lb-csi-node"
 )
 
 var (
@@ -677,9 +680,9 @@ var (
 	}
 
 	// Node DaemonSet
-	nodeRoleLabels   = map[string]string{"app": "lb-csi-node", "role": "node"}
+	nodeRoleLabels   = map[string]string{"app": lbCSINodeName, "role": "node"}
 	csiNodeDaemonSet = apps.DaemonSet{
-		ObjectMeta: metav1.ObjectMeta{Name: "lb-csi-node", Namespace: namespace},
+		ObjectMeta: metav1.ObjectMeta{Name: lbCSINodeName, Namespace: namespace},
 		Spec: apps.DaemonSetSpec{
 			Selector: &metav1.LabelSelector{MatchLabels: nodeRoleLabels},
 			UpdateStrategy: apps.DaemonSetUpdateStrategy{
@@ -846,7 +849,7 @@ func (r *DurosReconciler) deployStorageClass(ctx context.Context, projectID stri
 		log.Info("clusterrolebindinding", "name", crb.Name, "operation", op)
 	}
 
-	sts := &apps.StatefulSet{ObjectMeta: metav1.ObjectMeta{Name: "lb-csi-controller", Namespace: namespace}}
+	sts := &apps.StatefulSet{ObjectMeta: metav1.ObjectMeta{Name: lbCSIControllerName, Namespace: namespace}}
 	op, err := controllerutil.CreateOrUpdate(ctx, r.Shoot, sts, func() error {
 
 		controllerRoleLabels := map[string]string{"app": "lb-csi-plugin", "role": "controller"}

main.go

@@ -26,6 +26,7 @@ import (
 	v2 "github.com/metal-stack/duros-go/api/duros/v2"
 	"github.com/metal-stack/v"
 	"k8s.io/apimachinery/pkg/runtime"
+	utilruntime "k8s.io/apimachinery/pkg/util/runtime"
 	clientgoscheme "k8s.io/client-go/kubernetes/scheme"
 	_ "k8s.io/client-go/plugin/pkg/client/auth/gcp"
 	_ "k8s.io/client-go/plugin/pkg/client/auth/oidc"
@@ -45,9 +46,8 @@ var (
 )
 
 func init() {
-	_ = clientgoscheme.AddToScheme(scheme)
-
-	_ = storagev1.AddToScheme(scheme)
+	utilruntime.Must(clientgoscheme.AddToScheme(scheme))
+	utilruntime.Must(storagev1.AddToScheme(scheme))
 	// +kubebuilder:scaffold:scheme
 }
 
@@ -110,7 +110,7 @@ func main() {
 			setupLog.Error(err, "unable to create shoot restconfig")
 			os.Exit(1)
 		}
-		shootClient, err = client.New(shootRestConfig, client.Options{})
+		shootClient, err = client.New(shootRestConfig, client.Options{Scheme: scheme})
 		if err != nil {
 			setupLog.Error(err, "unable to create shoot client")
 			os.Exit(1)
@@ -193,7 +193,6 @@ func main() {
 		Client:      mgr.GetClient(),
 		Shoot:       shootClient,
 		Log:         ctrl.Log.WithName("controllers").WithName("LightBits"),
-		Scheme:      mgr.GetScheme(),
 		Namespace:   namespace,
 		DurosClient: durosClient,
 		Endpoints:   durosEPs,