h1fw0 goes into period of instability, restarted every two hours or so. We are investigating the problem. Provided h1fw1 continues to be stable, no data is lost.