Globular cluster ages revisited: the case of Ruprecht 106