feat(runtime): support dataset auto-create policy

Add a new DatasetAutoCreatePolicy feature that allows runtimes to
automatically create a Dataset when one does not already exist.
Use controllerutil.SetControllerReference to set proper owner reference
on the auto-created Dataset so it is garbage-collected with the runtime.

Signed-off-by: Monika Jakhar <jakharmonika364@gmail.com>

Author: jakharmonika364

pkg/common/constants.go

@@ -225,3 +225,8 @@ const (
 	SkipPrecheckAnnotationKey             = "sidecar.fluid.io/skip-precheck"
 	HostMountPathModeOnDefaultPlatformKey = "default.fuse-sidecar.fluid.io/host-mount-path-mode"
 )
+
+const (
+	// DatasetPolicyAutoCreate indicates that a Dataset should be auto-created for the Runtime.
+	DatasetPolicyAutoCreate = "auto-create"
+)

pkg/common/label.go

@@ -82,6 +82,10 @@ const (
 	// i.e. fluid.io/check-mount-script-sha256
 	AnnotationCheckMountScriptSHA256 = LabelAnnotationPrefix + "check-mount-script-sha256"
 
+	// AnnotationDatasetPolicy is a runtime annotation that controls how Dataset is handled.
+	// i.e. fluid.io/dataset-policy
+	AnnotationDatasetPolicy = LabelAnnotationPrefix + "dataset-policy"
+
 	// AnnotationDisableRuntimeHelmValueConfig is a runtime label indicates the configmap contains helm value will not be created in setup.
 	AnnotationDisableRuntimeHelmValueConfig = "runtime." + LabelAnnotationPrefix + "disable-helm-value-config"
 

pkg/controllers/runtime_controller.go

@@ -23,6 +23,7 @@ import (
 	"strings"
 	"time"
 
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/apimachinery/pkg/util/validation"
 	"k8s.io/apimachinery/pkg/util/validation/field"
@@ -34,8 +35,7 @@ import (
 	"k8s.io/client-go/util/retry"
 	ctrl "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/client"
-
-	// "sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
+	"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
 
 	"github.com/fluid-cloudnative/fluid/pkg/dump"
 	fluiderrs "github.com/fluid-cloudnative/fluid/pkg/errors"
@@ -110,14 +110,27 @@ func (r *RuntimeReconciler) ReconcileInternal(ctx cruntime.ReconcileRequestConte
 		return utils.RequeueIfError(err)
 	}
 
+	var datasetPolicy string
+	if annotations := objectMeta.GetAnnotations(); annotations != nil {
+		datasetPolicy = annotations[common.AnnotationDatasetPolicy]
+	}
+
 	// 5.Get the dataset
 	dataset, err := r.GetDataset(ctx)
 	if err != nil {
-		// r.Recorder.Eventf(ctx.Dataset, corev1.EventTypeWarning, common.ErrorProcessRuntimeReason, "Process Runtime error %v", err)
 		if utils.IgnoreNotFound(err) == nil {
-			ctx.Log.Info("The dataset is not found", "dataset", ctx.NamespacedName)
-			dataset = nil
-			// return ctrl.Result{}, nil
+			if datasetPolicy == common.DatasetPolicyAutoCreate {
+				ctx.Log.Info("The dataset is not found, auto-creating according to policy", "dataset", ctx.NamespacedName)
+				dataset, err = r.ensureDatasetForRuntime(ctx, objectMeta)
+				if err != nil {
+					ctx.Log.Error(err, "Failed to auto-create the dataset")
+					r.Recorder.Eventf(runtime, corev1.EventTypeWarning, common.ErrorCreateDataset, "Failed to auto-create dataset: %v", err)
+					return utils.RequeueAfterInterval(5 * time.Second)
+				}
+			} else {
+				ctx.Log.Info("The dataset is not found", "dataset", ctx.NamespacedName)
+				dataset = nil
+			}
 		} else {
 			ctx.Log.Error(err, "Failed to get the ddc dataset")
 			return utils.RequeueIfError(errors.Wrap(err, "Unable to get dataset"))
@@ -172,7 +185,11 @@ func (r *RuntimeReconciler) ReconcileInternal(ctx cruntime.ReconcileRequestConte
 		}
 	} else {
 		// If dataset is nil, need to wait because the user may have not created dataset
-		ctx.Log.Info("No dataset can be bound to the runtime, waiting.")
+		if datasetPolicy == common.DatasetPolicyAutoCreate {
+			ctx.Log.Info("No dataset is available for the runtime after auto-create, waiting.", "dataset", ctx.NamespacedName)
+		} else {
+			ctx.Log.Info("No dataset can be bound to the runtime, waiting.")
+		}
 		r.Recorder.Event(runtime, corev1.EventTypeWarning, common.ErrorProcessRuntimeReason, "No dataset can be bound to the runtime, waiting.")
 		return utils.RequeueAfterInterval(time.Duration(5 * time.Second))
 	}
@@ -385,6 +402,51 @@ func (r *RuntimeReconciler) GetDataset(ctx cruntime.ReconcileRequestContext) (*d
 	return &dataset, nil
 }
 
+func (r *RuntimeReconciler) ensureDatasetForRuntime(ctx cruntime.ReconcileRequestContext, objectMeta metav1.Object) (*datav1alpha1.Dataset, error) {
+	runtime := ctx.Runtime
+	if runtime == nil {
+		return nil, fmt.Errorf("runtime is nil")
+	}
+
+	annotations := make(map[string]string)
+	for k, v := range objectMeta.GetAnnotations() {
+		annotations[k] = v
+	}
+	annotations[common.AnnotationDatasetPolicy] = common.DatasetPolicyAutoCreate
+
+	dataset := &datav1alpha1.Dataset{
+		TypeMeta: metav1.TypeMeta{
+			Kind:       datav1alpha1.Datasetkind,
+			APIVersion: datav1alpha1.GroupVersion.Group + "/" + datav1alpha1.GroupVersion.Version,
+		},
+		ObjectMeta: metav1.ObjectMeta{
+			Name:        objectMeta.GetName(),
+			Namespace:   objectMeta.GetNamespace(),
+			Annotations: annotations,
+		},
+	}
+
+	// SetControllerReference looks up the GVK from the scheme, avoiding the
+	// empty-TypeMeta problem that arises when runtime.GetObjectKind() is used
+	// on objects retrieved via controller-runtime's client.Get().
+	if err := controllerutil.SetControllerReference(runtime, dataset, r.Client.Scheme()); err != nil {
+		return nil, fmt.Errorf("failed to set controller reference on auto-created dataset: %w", err)
+	}
+
+	err := r.Create(ctx, dataset)
+	if err != nil && !apierrors.IsAlreadyExists(err) {
+		return nil, err
+	}
+
+	if err := r.Get(ctx, ctx.NamespacedName, dataset); err != nil {
+		return nil, err
+	}
+
+	r.Recorder.Eventf(runtime, corev1.EventTypeNormal, common.Succeed, "Auto-created Dataset %s for Runtime", dataset.Name)
+
+	return dataset, nil
+}
+
 func (r *RuntimeReconciler) CheckIfReferenceDatasetIsSupported(ctx cruntime.ReconcileRequestContext) (bool, string) {
 	mounted := base.GetPhysicalDatasetFromMounts(ctx.Dataset.Spec.Mounts)