Is there a way to reset the cluster.on('death') handler? I've had to resort to using a global flag for whether to fork workers or not...