I have the engine running on separate HW with eth0 being the management interface
I have 2 compute nodes with eth0 being the management interface and eth1 having a vlan trunk with all of the VM networks
If I install a guest from the CD image it all works fine and picks up an IP from the DHCP server. However if I switch the boot order to PXE first the gPXE DHCP request times out. Using tcpdump I can see the DHCP discovery packets get as far out as the physical trunk interfact eth1 (i.e. it gets past all of the virtual interfaces) but any other machines in that same VLAN don't see the DHCP request and neither does the DHCP server.
My network settings are: Name: TestCluster External Provider: NO Network label: TC Enable VLAN tagging: 306 VM network: YES MTU: Default 1500
Under 'Setup Host Networks' Boot Protocol: DHCP
Any ideas?
Is it really gPXE (which is not maintained anymore) or iPXE? If the former try to download iPXE, see ipxe.org. What about iptables? j.