fix(hibernation): clear hibernate annotation before wake fast-path (#10)

* fix(hibernation): clear hibernate annotation before wake fast-path

reconcileWake's cloned-and-running fast-path returned Active without
clearing vm.cocoonstack.io/hibernate. A pod already awake but carrying
hibernate=true residue kept the annotation set; a subsequent
Desire=Hibernate then no-ops PatchHibernateState (annotation already
matches), so the CR could flip to Hibernated against a stale tag without
vk-cocoon taking a fresh snapshot — a later wake clones stale/divergent
state.

Move the clear above the fast-path so it runs unconditionally on any
wake reconcile. PatchHibernateState already no-ops when the annotation
matches, so the common (already-false) case adds no extra write.

Fixes #2.

* chore(hibernation): drop comments from wake fix

Author: CMGS

hibernation/reconciler_test.go

@@ -463,6 +463,58 @@ func TestReconcileWakeRecoversFromFailed(t *testing.T) {
 	}
 }
 
+func TestReconcileWakeClearsHibernateResidueOnFastPath(t *testing.T) {
+	hib := &cocoonv1.CocoonHibernation{
+		ObjectMeta: metav1.ObjectMeta{Name: "hib", Namespace: "ns", Finalizers: []string{finalizerName}},
+		Spec: cocoonv1.CocoonHibernationSpec{
+			Desire: cocoonv1.HibernationDesireWake,
+			PodRef: cocoonv1.HibernationPodRef{Name: "demo-0"},
+		},
+		Status: cocoonv1.CocoonHibernationStatus{Phase: cocoonv1.CocoonHibernationPhaseWaking},
+	}
+	pod := &corev1.Pod{
+		ObjectMeta: metav1.ObjectMeta{Name: "demo-0", Namespace: "ns"},
+		Status: corev1.PodStatus{
+			ContainerStatuses: []corev1.ContainerStatus{{
+				State: corev1.ContainerState{Running: &corev1.ContainerStateRunning{}},
+			}},
+		},
+	}
+	(&meta.VMSpec{VMName: "vk-ns-demo-0", Managed: true}).Apply(pod)
+	(&meta.VMRuntime{VMID: "vmid-live"}).Apply(pod)
+	meta.HibernateState(true).Apply(pod)
+
+	scheme := testScheme(t)
+	cli := ctrlfake.NewClientBuilder().
+		WithScheme(scheme).
+		WithObjects(hib, pod).
+		WithStatusSubresource(&cocoonv1.CocoonHibernation{}).
+		Build()
+	r := &Reconciler{Client: cli, Scheme: scheme, Epoch: &fakeRegistry{}}
+
+	if _, err := r.Reconcile(t.Context(), ctrl.Request{
+		NamespacedName: types.NamespacedName{Namespace: "ns", Name: "hib"},
+	}); err != nil {
+		t.Fatalf("Reconcile: %v", err)
+	}
+
+	var outHib cocoonv1.CocoonHibernation
+	if err := cli.Get(t.Context(), types.NamespacedName{Namespace: "ns", Name: "hib"}, &outHib); err != nil {
+		t.Fatalf("get hib: %v", err)
+	}
+	if outHib.Status.Phase != cocoonv1.CocoonHibernationPhaseActive {
+		t.Errorf("phase = %q, want Active (fast-path)", outHib.Status.Phase)
+	}
+
+	var outPod corev1.Pod
+	if err := cli.Get(t.Context(), types.NamespacedName{Namespace: "ns", Name: "demo-0"}, &outPod); err != nil {
+		t.Fatalf("get pod: %v", err)
+	}
+	if meta.ReadHibernateState(&outPod) {
+		t.Error("hibernate annotation must be cleared on the wake fast-path; still true")
+	}
+}
+
 func TestHibernateDeadlineExceeded(t *testing.T) {
 	staleReady := metav1.Condition{
 		Type:               commonk8s.ConditionTypeReady,

hibernation/wake.go

@@ -17,6 +17,12 @@ func (r *Reconciler) reconcileWake(ctx context.Context, hib *cocoonv1.CocoonHibe
 	logger := log.WithFunc("hibernation.Reconciler.reconcileWake")
 	r.announceRetryFromFailed(hib, cocoonv1.HibernationDesireWake)
 
+	if meta.ReadHibernateState(pod) {
+		if err := commonk8s.PatchHibernateState(ctx, r.Client, pod, false); err != nil {
+			return ctrl.Result{}, fmt.Errorf("clear hibernate annotation: %w", err)
+		}
+	}
+
 	if vmClonedAndRunning(pod) {
 		// Drop snapshot tag (non-fatal; stale tag gets overwritten on next hibernate).
 		if err := r.Epoch.DeleteManifest(ctx, vmName, meta.HibernateSnapshotTag); err != nil {
@@ -29,12 +35,6 @@ func (r *Reconciler) reconcileWake(ctx context.Context, hib *cocoonv1.CocoonHibe
 		return ctrl.Result{}, r.setPhase(ctx, hib, cocoonv1.CocoonHibernationPhaseActive, vmName)
 	}
 
-	if meta.ReadHibernateState(pod) {
-		if err := commonk8s.PatchHibernateState(ctx, r.Client, pod, false); err != nil {
-			return ctrl.Result{}, fmt.Errorf("clear hibernate annotation: %w", err)
-		}
-	}
-
 	if phaseDeadlineExceeded(hib, cocoonv1.CocoonHibernationPhaseWaking, wakeTimeout) {
 		if r.firstTransitionAt(hib) {
 			observePhaseExit(hib, "timeout")