iPXE discussion forum

After a recent upgrade of qemu/seabios/libvirt (but oddly enough not ipxe) I have a VM which no longer PXE boots.

I get these log messages in my dhcp server:
Jun 24 19:42:20 rich dhcpd: DHCPDISCOVER from 52:54:00:0c:4d:b9 via brkvm
Jun 24 19:42:21 rich dhcpd: DHCPOFFER on 192.168.0.199 to 52:54:00:0c:4d:b9 via brkvm
Jun 24 19:42:23 rich dhcpd: DHCPREQUEST for 192.168.1.2 (192.168.0.101) from 52:54:00:0c:4d:b9 via brkvm: wrong network.
Jun 24 19:42:23 rich dhcpd: DHCPNAK on 192.168.1.2 to 52:54:00:0c:4d:b9 via brkvm

On the console it identifies as ipxe 1.0.0+ (09c5).

It then displays net 192.168.1.2/255.255.255.0 gw 192.168.0.101
next server 192.168.0.101

So, it is sticking to an IP not assigned by DHCP, but it apparently is getting the correct gateway, but not the next server to use. Of course, that gateway will never work with that subnet/mask.

Any ideas what is going on?

Is it possible you're getting a response from qemu's internal DHCP server instead of the your main DHCP server?

I seem to recall this be discussed on the IRC channel. Was any conclusion found there?