Fix discard job

internal/controller/mantlebackup_controller.go

@@ -1391,6 +1391,11 @@ func (r *MantleBackupReconciler) startImport(
 	backup *mantlev1.MantleBackup,
 	target *snapshotTarget,
 ) (ctrl.Result, error) {
+	if !r.doesMantleBackupHaveSyncModeAnnot(backup) {
+		// SetSynchronizingg is not called yet or the cache is stale.
+		return ctrl.Result{Requeue: true}, nil
+	}
+
 	if result, err := r.isExportDataAlreadyUploaded(ctx, backup); err != nil || !result.IsZero() {
 		return result, err
 	}
@@ -1415,6 +1420,12 @@ func (r *MantleBackupReconciler) startImport(
 	return ctrl.Result{}, nil
 }
 
+func (r *MantleBackupReconciler) doesMantleBackupHaveSyncModeAnnot(backup *mantlev1.MantleBackup) bool {
+	annots := backup.GetAnnotations()
+	syncMode, ok := annots[annotSyncMode]
+	return ok && (syncMode == syncModeFull || syncMode == syncModeIncremental)
+}
+
 func (r *MantleBackupReconciler) prepareObjectStorageClient(ctx context.Context) error {
 	if r.objectStorageClient != nil {
 		return nil
@@ -1678,6 +1689,7 @@ blkdiscard /dev/discard-rbd
 						DevicePath: "/dev/discard-rbd",
 					},
 				},
+				ImagePullPolicy: corev1.PullIfNotPresent,
 			},
 		}
 

internal/controller/mantlerestore_controller_test.go

@@ -349,9 +349,11 @@ func (test *mantleRestoreControllerUnitTest) testDeleteRestoringPV() {
 		err = k8sClient.Update(ctx, &pv)
 		Expect(err).NotTo(HaveOccurred())
 
-		err = k8sClient.Get(ctx, client.ObjectKey{Name: test.reconciler.restoringPVName(restore)}, &pv)
-		Expect(err).To(HaveOccurred())
-		Expect(errors.IsNotFound(err)).To(BeTrue())
+		Eventually(ctx, func(g Gomega) {
+			err = k8sClient.Get(ctx, client.ObjectKey{Name: test.reconciler.restoringPVName(restore)}, &pv)
+			g.Expect(err).To(HaveOccurred())
+			g.Expect(errors.IsNotFound(err)).To(BeTrue())
+		}).Should(Succeed())
 	})
 
 	It("should skip deleting the PV if it does not exist", func(ctx SpecContext) {

internal/controller/persistentvolume_controller.go

@@ -64,6 +64,11 @@ func (r *PersistentVolumeReconciler) Reconcile(ctx context.Context, req ctrl.Req
 		return ctrl.Result{}, fmt.Errorf("failed to get PersistentVolume: %w", err)
 	}
 
+	// Make sure the PV has the finalizer.
+	if !controllerutil.ContainsFinalizer(&pv, RestoringPVFinalizerName) {
+		return ctrl.Result{}, nil
+	}
+
 	// Check if the PV is managed by the target Ceph cluster.
 	clusterID, err := getCephClusterIDFromSCName(ctx, r.client, pv.Spec.StorageClassName)
 	if err != nil {
@@ -77,11 +82,6 @@ func (r *PersistentVolumeReconciler) Reconcile(ctx context.Context, req ctrl.Req
 		return ctrl.Result{}, nil
 	}
 
-	// Make sure the PV has the finalizer.
-	if !controllerutil.ContainsFinalizer(&pv, RestoringPVFinalizerName) {
-		return ctrl.Result{}, nil
-	}
-
 	// Make sure the PV has a deletionTimestamp.
 	if pv.GetDeletionTimestamp().IsZero() {
 		return ctrl.Result{}, nil

internal/controller/replication.go

@@ -183,12 +183,11 @@ func (s *SecondaryServer) CreateOrUpdateMantleBackup(
 		return nil, fmt.Errorf("CreateOrUpdate failed: %w", err)
 	}
 
-	// Update the status here because ctrl.CreateOrUpdate doesn't change the status.
-	if err := updateStatus(ctx, s.client, &backup, func() error {
-		backup.Status.CreatedAt = backupReceived.Status.CreatedAt
-		return nil
-	}); err != nil {
-		return nil, fmt.Errorf("updateStatus failed: %w", err)
+	// Use Patch here because updateStatus is likely to fail due to "the object has been modified" error.
+	newBackup := backup.DeepCopy()
+	newBackup.Status.CreatedAt = backupReceived.Status.CreatedAt
+	if err := s.client.Status().Patch(ctx, newBackup, client.MergeFrom(&backup)); err != nil {
+		return nil, fmt.Errorf("status patch failed: %w", err)
 	}
 
 	return &proto.CreateOrUpdateMantleBackupResponse{}, nil

test/e2e/multik8s/suite_test.go

@@ -89,7 +89,7 @@ func replicationTestSuite() {
 					return errors.New("status of SyncedToRemote condition is not True")
 				}
 				return nil
-			}, "5m", "1s").Should(Succeed())
+			}, "10m", "1s").Should(Succeed())
 
 			By("checking PVC is replicated")
 			Eventually(func() error {