Issue with Zen imaging server serving 2 networks:

Our student lab network is separate and firewalled from our staff network.
Imaging through the firewall would be reduced from gig to 100meg network
speed and strain the firewall-router. The networks are physically
intermixed in the same building so we want one ZFD server for both.

We purchased and set up a server and connected it to our student lab and
staff networks on separate interfaces. (Netware 6.5 sp5, ZFD 6.5 sp2, Dell
server, Intel dual NIC.) So far all functions we want (WOL, DHCP, PXE, auto
and maintenance imaging, Remote Control, NAL) are all working EXCEPT auto
imaging on the staff network.

Manual imaging via maintenance mode works on both networks. Automatic
imaging works on the student network. Automatice imaging on the staff
network produces a 'Could not connect to image server' error and reboot
loop. The same error is displayed if I manually try to image from either
network to the other. My assumption is that a setting somewhere points the
auto imaging request to the server's address on the student network.

TFTP works fine loading PXE envronments, and documentation indicates if no
IP is configured tftp.ini, it uses the address on the same network as the
request. Don't think TFTP is the culprit.

PDHCP.INI seems to have some related settings. We are serving DHCP on the
server and have the PXEClient option set.

One immediate fix would be to turn on routing on the server and set a
static route to between it's two interfaces. We'd like to avoid this partly
for security, mostly for imaging speed and overhead on the server.