It's pretty "state of art" process. Try to find all footprints available:
* May you estimate what time did it happened? Was it day time or night time?
* Check your firewall log files, if it have any records for that time. Enable more advancedlogging, may be log all transactions for a few days to look for suspicious activity.
* Try to setup sniffer on your network to watch traffic to hacked server. To avoid overflow, use filters to cut off "unwanted traffic".
* If possible, buy andinstall IDS (Intrusion Detection System) to watch your network. It can send administrative alerts on most haching attempts.
* If possible, (I mean, if Operation System allows to do that), forward log messages or SNMP traps to another server.
*Perform system security scanning against server, which was hacked (and against potentially hacked servers). Install required patches, disable unused services, etc.
* Watch your users :-)
Good luck.