fixed a problem where we would write the header to clusters files twice added extra logging in monitor leader