(fix) Helm to Boxcutter migration during OLM upgrade

When upgrading OLM from standard (Helm runtime) to experimental
(Boxcutter runtime), the BoxcutterStorageMigrator creates a
ClusterExtensionRevision from the existing Helm release. However,
the migrated revision was created without status conditions, causing
a race condition where it wasn't recognized as "Installed".

This fix sets an initial Succeeded status on migrated revisions,
ensuring they're immediately recognized and allowing version upgrades
to proceed correctly after OLM upgrades.

Fixes test-upgrade-st2ex-e2e failures.

Author: camilamacedo86

api/v1/clusterextensionrevision_types.go

@@ -32,6 +32,7 @@ const (
 	// Condition Reasons
 	ClusterExtensionRevisionReasonArchived        = "Archived"
 	ClusterExtensionRevisionReasonBlocked         = "Blocked"
+	ClusterExtensionRevisionReasonMigrated        = "Migrated"
 	ClusterExtensionRevisionReasonProbeFailure    = "ProbeFailure"
 	ClusterExtensionRevisionReasonProbesSucceeded = "ProbesSucceeded"
 	ClusterExtensionRevisionReasonReconciling     = "Reconciling"

internal/operator-controller/applier/boxcutter.go

@@ -266,6 +266,33 @@ func (m *BoxcutterStorageMigrator) Migrate(ctx context.Context, ext *ocv1.Cluste
 		return fmt.Errorf("getting created revision: %w", err)
 	}
 
+	// Set initial status on the migrated revision to mark it as succeeded.
+	//
+	// The revision must have a Succeeded=True status condition immediately after creation.
+	//
+	// A revision is only considered "Installed" (vs "RollingOut") when it has this condition.
+	// Without it, the system cannot determine what version is currently installed, which breaks:
+	//   - Version resolution (can't compute upgrade paths from unknown starting point)
+	//   - Status reporting (installed bundle appears as nil)
+	//   - Subsequent upgrades (resolution fails without knowing current version)
+	//
+	// While the ClusterExtensionRevision controller would eventually reconcile and set this status,
+	// that creates a timing gap where the ClusterExtension reconciliation happens before the status
+	// is set, causing failures during the OLM upgrade window.
+	//
+	// Since we're creating this revision from a successfully deployed Helm release, we know it
+	// represents a working installation and can safely mark it as succeeded immediately.
+	meta.SetStatusCondition(&rev.Status.Conditions, metav1.Condition{
+		Type:               ocv1.ClusterExtensionRevisionTypeSucceeded,
+		Status:             metav1.ConditionTrue,
+		Reason:             ocv1.ClusterExtensionRevisionReasonMigrated,
+		Message:            "Migrated from Helm release",
+		ObservedGeneration: rev.GetGeneration(),
+	})
+	if err := m.Client.Status().Update(ctx, rev); err != nil {
+		return fmt.Errorf("updating migrated revision status: %w", err)
+	}
+
 	return nil
 }