Change GitRepo deletion.

charts/fleet-crd/templates/crds.yaml

@@ -6388,6 +6388,14 @@ spec:
             sha256sum:
               description: SHA256Sum of the Content field
               type: string
+            status:
+              description: ContentStatus defines the observed state of Content
+              properties:
+                referenceCount:
+                  description: ReferenceCount is the number of BundleDeployments that
+                    currently reference this Content resource.
+                  type: integer
+              type: object
           type: object
       served: true
       storage: true

integrationtests/gitjob/controller/suite_test.go

@@ -17,6 +17,7 @@ import (
 	appsv1 "k8s.io/api/apps/v1"
 	corev1 "k8s.io/api/core/v1"
 
+	"github.com/rancher/fleet/internal/cmd/controller/gitops"
 	"github.com/rancher/fleet/internal/cmd/controller/gitops/reconciler"
 	ctrlreconciler "github.com/rancher/fleet/internal/cmd/controller/reconciler"
 	"github.com/rancher/fleet/internal/cmd/controller/target"
@@ -89,6 +90,9 @@ var _ = BeforeSuite(func() {
 	})
 	Expect(err).ToNot(HaveOccurred())
 
+	Expect(gitops.AddRepoNameLabelIndexer(ctx, mgr)).ToNot(HaveOccurred())
+	Expect(gitops.AddImageScanGitRepoIndexer(ctx, mgr)).ToNot(HaveOccurred())
+
 	ctlr := gomock.NewController(GinkgoT())
 
 	// redirect logs to a buffer that we can read in the tests

internal/cmd/controller/finalize/finalize.go

@@ -6,15 +6,11 @@ import (
 	"strings"
 
 	"github.com/rancher/fleet/pkg/apis/fleet.cattle.io/v1alpha1"
-	"github.com/rancher/wrangler/v3/pkg/kv"
-
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/types"
-	"k8s.io/client-go/util/retry"
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
-	"sigs.k8s.io/controller-runtime/pkg/log"
 )
 
 const (
@@ -66,64 +62,6 @@ func PurgeBundles(ctx context.Context, c client.Client, gitrepo types.Namespaced
 	return nil
 }
 
-// PurgeContent tries to delete the content resource related with the given bundle deployment.
-func PurgeContent(ctx context.Context, c client.Client, name, deplID string) error {
-	contentID, _ := kv.Split(deplID, ":")
-	content := &v1alpha1.Content{}
-	if err := c.Get(ctx, types.NamespacedName{Name: contentID}, content); err != nil {
-		return client.IgnoreNotFound(err)
-	}
-
-	logger := log.FromContext(ctx).WithName("purge-content").WithValues("contentID", contentID, "finalizerName", name)
-
-	nn := types.NamespacedName{Name: content.Name}
-	if controllerutil.ContainsFinalizer(content, name) {
-		err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
-			if err := c.Get(ctx, nn, content); err != nil {
-				return client.IgnoreNotFound(err)
-			}
-
-			controllerutil.RemoveFinalizer(content, name)
-
-			return c.Update(ctx, content)
-		})
-		if err != nil {
-			return err
-		}
-
-		logger.V(1).Info("Removed finalizer from content resource")
-	}
-
-	if len(content.Finalizers) == 0 {
-		if err := c.Delete(ctx, content); err != nil {
-			return err
-		}
-		logger.V(1).Info("Deleted content resource")
-	}
-
-	return nil
-}
-
-// PurgeImageScans deletes all ImageScan resources related with the given GitRepo namespaces name.
-func PurgeImageScans(ctx context.Context, c client.Client, gitrepo types.NamespacedName) error {
-	images := &v1alpha1.ImageScanList{}
-	err := c.List(ctx, images, client.InNamespace(gitrepo.Namespace))
-	if err != nil {
-		return err
-	}
-
-	for _, image := range images.Items {
-		if image.Spec.GitRepoName == gitrepo.Name {
-			err := c.Delete(ctx, &image)
-			if err != nil {
-				return err
-			}
-		}
-
-	}
-	return nil
-}
-
 // PurgeNamespace deletes the given namespace if deleteNamespace is set to true.
 // It ignores the following namespaces, that are considered as default by fleet or kubernetes:
 // fleet-local, cattle-fleet-system, fleet-default, cattle-fleet-clusters-system, default
@@ -168,3 +106,14 @@ func EnsureFinalizer(ctx context.Context, c client.Client, obj client.Object, fi
 	controllerutil.AddFinalizer(obj, finalizer)
 	return c.Update(ctx, obj)
 }
+
+func PurgeTargetNamespaceIfNeeded(ctx context.Context, c client.Client, gitrepo *v1alpha1.GitRepo) error {
+	deleteNamespace := gitrepo.Spec.DeleteNamespace
+	namespace := gitrepo.Spec.TargetNamespace
+
+	if gitrepo.Spec.KeepResources {
+		deleteNamespace = false
+	}
+
+	return PurgeNamespace(ctx, c, deleteNamespace, namespace)
+}

internal/cmd/controller/gitops/operator.go

@@ -20,11 +20,13 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	clog "sigs.k8s.io/controller-runtime/pkg/log"
 	"sigs.k8s.io/controller-runtime/pkg/log/zap"
+	"sigs.k8s.io/controller-runtime/pkg/manager"
 	metricsserver "sigs.k8s.io/controller-runtime/pkg/metrics/server"
 
 	command "github.com/rancher/fleet/internal/cmd"
 	"github.com/rancher/fleet/internal/cmd/controller/gitops/reconciler"
 	fcreconciler "github.com/rancher/fleet/internal/cmd/controller/reconciler"
+	"github.com/rancher/fleet/internal/config"
 	"github.com/rancher/fleet/internal/metrics"
 	"github.com/rancher/fleet/internal/ssh"
 	fleet "github.com/rancher/fleet/pkg/apis/fleet.cattle.io/v1alpha1"
@@ -140,6 +142,18 @@ func (g *GitOperator) Run(cmd *cobra.Command, args []string) error {
 
 	kh := ssh.KnownHosts{EnforceHostKeyChecks: !g.SkipHostKeyChecks}
 
+	// Add an indexer for the Gitrepo name label as that will make accesses in the cache
+	// faster
+	if err := AddRepoNameLabelIndexer(ctx, mgr); err != nil {
+		return err
+	}
+
+	// Add an indexer for the GitRepo name field in ImageScans as that will make accesses in the cache
+	// faster
+	if err := AddImageScanGitRepoIndexer(ctx, mgr); err != nil {
+		return err
+	}
+
 	gitJobReconciler := &reconciler.GitJobReconciler{
 		Client:          mgr.GetClient(),
 		Scheme:          mgr.GetScheme(),
@@ -245,3 +259,40 @@ func startWebhook(ctx context.Context, namespace string, addr string, client cli
 
 	return nil
 }
+
+func AddRepoNameLabelIndexer(ctx context.Context, mgr manager.Manager) error {
+	return mgr.GetFieldIndexer().IndexField(
+		ctx,
+		&fleet.Bundle{},
+		config.RepoNameIndex,
+		func(obj client.Object) []string {
+			content, ok := obj.(*fleet.Bundle)
+			if !ok {
+				return nil
+			}
+			if name, exists := content.Labels[fleet.RepoLabel]; exists {
+				return []string{name}
+			}
+
+			return nil
+		},
+	)
+}
+
+func AddImageScanGitRepoIndexer(ctx context.Context, mgr manager.Manager) error {
+	return mgr.GetFieldIndexer().IndexField(
+		ctx,
+		&fleet.ImageScan{},
+		config.ImageScanGitRepoIndex,
+		func(obj client.Object) []string {
+			content, ok := obj.(*fleet.ImageScan)
+			if !ok {
+				return nil
+			}
+			if content.Spec.GitRepoName == "" {
+				return nil
+			}
+			return []string{content.Spec.GitRepoName}
+		},
+	)
+}

internal/cmd/controller/gitops/reconciler/gitjob_controller.go

@@ -16,6 +16,7 @@ import (
 	"github.com/rancher/fleet/internal/cmd/controller/imagescan"
 	ctrlquartz "github.com/rancher/fleet/internal/cmd/controller/quartz"
 	"github.com/rancher/fleet/internal/cmd/controller/reconciler"
+	"github.com/rancher/fleet/internal/config"
 	"github.com/rancher/fleet/internal/metrics"
 	v1alpha1 "github.com/rancher/fleet/pkg/apis/fleet.cattle.io/v1alpha1"
 	"github.com/rancher/fleet/pkg/durations"
@@ -29,6 +30,7 @@ import (
 	batchv1 "k8s.io/api/batch/v1"
 	corev1 "k8s.io/api/core/v1"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
+	"k8s.io/apimachinery/pkg/api/meta"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/apis/meta/v1/unstructured"
 	"k8s.io/apimachinery/pkg/labels"
@@ -56,6 +58,10 @@ const (
 	// make sure Prometheus scrapes them.
 	ShortLivedMetricsTTL       = 120 * time.Second
 	gitJobPollingJitterPercent = 10
+
+	// period after which the GitRepo reconciler is re-scheduled,
+	// in order to wait for the dependent resources cleanup to finish
+	requeueAfterResourceCleanup = 2 * time.Second
 )
 
 var (
@@ -179,9 +185,7 @@ func (r *GitJobReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctr
 
 	if !gitrepo.DeletionTimestamp.IsZero() {
 		if controllerutil.ContainsFinalizer(gitrepo, finalize.GitRepoFinalizer) {
-			if err := r.cleanupGitRepo(ctx, logger, gitrepo); err != nil {
-				return ctrl.Result{}, err
-			}
+			return r.handleDelete(ctx, logger, gitrepo)
 		}
 
 		return ctrl.Result{}, nil
@@ -357,25 +361,50 @@ func (r *GitJobReconciler) deletePreviousJob(ctx context.Context, logger logr.Lo
 	return r.Delete(ctx, &job)
 }
 
-func (r *GitJobReconciler) cleanupGitRepo(ctx context.Context, logger logr.Logger, gitrepo *v1alpha1.GitRepo) error {
+func (r *GitJobReconciler) handleDelete(ctx context.Context, logger logr.Logger, gitrepo *v1alpha1.GitRepo) (ctrl.Result, error) {
 	logger.Info("Gitrepo deleted, deleting bundle, image scans")
 
-	metrics.GitRepoCollector.Delete(gitrepo.Name, gitrepo.Namespace)
 	_ = r.deletePollingJob(*gitrepo)
 
-	nsName := types.NamespacedName{Name: gitrepo.Name, Namespace: gitrepo.Namespace}
-	if err := finalize.PurgeBundles(ctx, r.Client, nsName, v1alpha1.RepoLabel); err != nil {
-		return err
+	if !controllerutil.ContainsFinalizer(gitrepo, finalize.GitRepoFinalizer) {
+		return ctrl.Result{}, nil
+	}
+
+	bundles, err := r.listBundlesForGitrepo(ctx, gitrepo)
+	if err != nil {
+		return ctrl.Result{}, err
+	}
+
+	// Bundle deletion happens asynchronously: mark them for deletion and requeue
+	// This ensures the Gitrepo is kept around until all its Bundles are completely deleted.
+	if len(bundles.Items) > 0 {
+		logger.V(1).Info("GitRepo deleted, purging bundles")
+		return ctrl.Result{RequeueAfter: requeueAfterResourceCleanup}, batchDeleteDependentResources(ctx, r.Client, bundles)
 	}
 
 	// remove the job scheduled by imagescan, if any
 	_ = r.Scheduler.DeleteJob(imagescan.GitCommitKey(gitrepo.Namespace, gitrepo.Name))
 
-	if err := finalize.PurgeImageScans(ctx, r.Client, nsName); err != nil {
-		return err
+	images, err := r.listImageScansForGitrepo(ctx, gitrepo)
+	if err != nil {
+		return ctrl.Result{}, err
+	}
+
+	if len(images.Items) > 0 {
+		logger.V(1).Info("GitRepo deleted, purging imagescans")
+		return ctrl.Result{RequeueAfter: requeueAfterResourceCleanup}, batchDeleteDependentResources(ctx, r.Client, images)
+	}
+
+	// Delete the target namespace if DeleteNamespace is true
+	if err := finalize.PurgeTargetNamespaceIfNeeded(ctx, r.Client, gitrepo); err != nil {
+		return ctrl.Result{}, err
 	}
 
-	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
+	metrics.GitRepoCollector.Delete(gitrepo.Name, gitrepo.Namespace)
+
+	// we don't have pending Bundles nor ImageScans, we can remove the finalizer
+	nsName := types.NamespacedName{Name: gitrepo.Name, Namespace: gitrepo.Namespace}
+	err = retry.RetryOnConflict(retry.DefaultRetry, func() error {
 		if err := r.Get(ctx, nsName, gitrepo); err != nil {
 			return err
 		}
@@ -386,10 +415,10 @@ func (r *GitJobReconciler) cleanupGitRepo(ctx context.Context, logger logr.Logge
 	})
 
 	if client.IgnoreNotFound(err) != nil {
-		return err
+		return ctrl.Result{}, err
 	}
 
-	return nil
+	return ctrl.Result{}, nil
 }
 
 // shouldCreateJob checks if the conditions to create a new job are met.
@@ -600,6 +629,41 @@ func (r *GitJobReconciler) managePollingJob(logger logr.Logger, gitrepo v1alpha1
 	return jobUpdatedOrCreated, nil
 }
 
+func (r *GitJobReconciler) listBundlesForGitrepo(ctx context.Context, gitrepo *v1alpha1.GitRepo) (*v1alpha1.BundleList, error) {
+	list := &v1alpha1.BundleList{}
+	// if err := r.List(ctx, list,
+	// 	client.MatchingFields{
+	// 		"metadata.labels.gitrepo-name+namespace": GetNamespaceAndNameID(gitrepo.Namespace, gitrepo.Name),
+	// 	},
+	// ); err != nil {
+	// 	return nil, err
+	// }
+
+	err := r.List(ctx, list, client.MatchingLabels{v1alpha1.RepoLabel: gitrepo.Name}, client.InNamespace(gitrepo.Namespace))
+	if err != nil {
+		return nil, err
+	}
+	return list, nil
+}
+
+func (r *GitJobReconciler) listImageScansForGitrepo(ctx context.Context, gitrepo *v1alpha1.GitRepo) (*v1alpha1.ImageScanList, error) {
+	list := &v1alpha1.ImageScanList{}
+
+	if err := r.List(ctx, list,
+		client.InNamespace(gitrepo.Namespace),
+		client.MatchingFields{
+			config.ImageScanGitRepoIndex: gitrepo.Name,
+		},
+	); err != nil {
+		return nil, err
+	}
+	return list, nil
+}
+
+func GetNamespaceAndNameID(namespace, name string) string {
+	return fmt.Sprintf("%s/%s", namespace, name)
+}
+
 func generationChanged(r *v1alpha1.GitRepo) bool {
 	// checks if generation changed.
 	// it ignores the case when Status.ObservedGeneration=0 because that's
@@ -837,3 +901,27 @@ func getNextCommit(status v1alpha1.GitRepoStatus) string {
 
 	return commit
 }
+
+func batchDeleteDependentResources(ctx context.Context, c client.Client, list client.ObjectList) error {
+	var errs []error
+
+	_ = meta.EachListItem(list, func(obj runtime.Object) error {
+		o, ok := obj.(client.Object)
+		if !ok {
+			errs = append(errs, fmt.Errorf("item does not implement client.Object: %T", obj))
+			return nil // continue iterating
+		}
+		if o.GetDeletionTimestamp() != nil {
+			// already being deleted
+			return nil
+		}
+
+		if err := c.Delete(ctx, o); err != nil {
+			errs = append(errs, err)
+		}
+
+		return nil // continue iterating no matter what
+	})
+
+	return errors.Join(errs...)
+}