SnapDeploy PXE and DHCP on different subnets

Hey,

I am faced the following problem.
I want my bare metal systems to boot via PXE into SD 2.0.2141 but get different errors while trying various configurations.
System configuration
SnapDeploy Server:
W2K3 SP2; all SnapDeploy components installed
1 Broadcom NIC (192.168.X.X /24 -->"Deployment"-LAN)
1 Intel NIC (10.12.X.X /24 -->"Inhouse"-LAN)
Microsoft DHCP-Relay-Agent installed to route DHCP-packets from Deployment-LAN to Inhouse-LAN where the DHCP-Server is configured.

DHCP-Server (10.12.X.X)
Scope 192.168.X.X with options 66 (hostname startserver = 192.168.X.X) and 67 (name of startfile = bootwiz.sys) configured.

PXE-boot-results
Bare metal systems get an correct IP
Message: Acronis Loader startet (or similar)
Message: Download bootwiz.cfg... (twice)
Abort with Message: No configuration file found

I did a network trace to analyse the problem.
Results: everything seems fine...
1. DHCP discover, offer, request, ack... ok!
TFTP Filetransfer bootwiz.sys... ok!

!!! TFTP: Read Request; file:bootwiz.cfg... destination not reachable
reason: now the bare metal system tries to open a tftp-connection to the dhcp-server (10.12.X.X) and not to the pxe-server (192.168.X.X) as it did while downloading the bootwiz.sys!

I wonder why there is a change in destination IP?!

I also tried to install a DHCP-Server on the SnapDeploy-Server but it didn't work. Obviously this behaviour is by design as I read in another thread.

Any suggestions, experiences?

Kind Regards,
Andy