@foguser I’ll answer your questions in different order as it makes more sense to me doing it this way.
Is it even possible to get this to work?
From what I have read so far it looks like a bit of a non-standard setup that needs a bit of knowledge and setting the right configurations. See my later answers.
Did i do anything wrong in the Installation of fog?
Looks fine for me. You don’t want the FOG server to handle DHCP in your network which is perfectly fine. All the other things are clear and correct.
Have i forgotten/overlooked a configuration to get this to work?
You probably have and I am fairly sure we can point you the right way. Maybe not spot on but close enough for you to figure it out.
so fog would recognize a dhcp-request for it, but without luck.
Just to get that right in the first place. In your setup FOG does not care about DHCP at all! You choose to use your own DHCP server and this one is in charge of serving the right information to the clients. BUT the FOG server is taking a share in the whole PXE process and I guess this is where things fail for you.
If you have Citrix Fat Client PXE stuff on by default your clients will boot into that and don’t reach out to the FOG server.
I can manually configure Option 66 and 67 for the Clients to use fog.
I replicated the scope to all other DCs and checked it, but the Client did not care at all.
He kept booting and finding the citrix server-ip and file.
So are you planning to manually switch between Citrix and FOG PXE booting the clients? There is only one route you can go or you need to use chainloading to PXE boot from the FOG server first and then hand over to the Citrix if there is no task scheduled for this client (just an idea how you can combine all you have there). But let’s go back to your issue. What did you set Option 66 and 67 to? 66 should be the FOG server IP address (make sure the DHCP hands out gateway information as well as your clients need to talk across the subnets) and 67 should be
undionly.kpxe for legacy BIOS and
ipxe.efi for UEFI machines (wiki article on this topic).
My guess is that DC replication wasn’t properly working and caused the clients to still boot to the Citrix server.