We have seen a similar issue in our machines, the only issue is that in an AD environment you need the time to sync before login. To combat this, we added (or edited, if one already exists) the rc.local file in /etc with the following bash script that kills ntpd, tries up to 5 times to see if a timeserver can be contacted and then restarts network time to resync the time. This seems to work ok. It probably could be cleaner but it works for us.

#!/bin/bash

killall ntpd

/usr/sbin/ntpdate
RETVAL=$?
TRIES=1
sleep 1
until [[ "$RETVAL" == "0" || "$TRIES" == "5" ]]; do
/usr/sbin/ntpdate
RETVAL=$?
sleep 1
TRIES=`expr $TRIES + 1`
done;

. /System/Library/StartupItems/NetworkTime/NetworkTime start


[ Reply to This | # ]