For the past couple of days, the clock on a satellite has been one hour ahead of the correct time (CEDT) on the other three nodes in the cluster. If I change it after booting, things are OK until the next reboot, then it is ahead by exactly an hour again. Any ideas?