Hello all, this is a networking/virtualization issue, and I'm tossing it here in the networking subforum but if it needs moved, please move it. This is in regards to CentOS 6

I have a machine with several network bridges that are connected to bonded ports. An example:

eth1 and eth2 are a part of bond0, which is bridged to br0
eth3 and eth4 are a part of bond1, which is bridged to br1
eth5 and eth6 are a part of bond2, but is a trunked connection that services VLANs 50 60 and 70, so there are three bonds (bond2.50, bond2.60 and bond2.70), which is bridged to br2.50, br2.60 and br2.70 respectively

This network works just fine on the physical host. Our management network is on the 50 VLAN, and is serviced through bond2.50/br2.50.

When I create a new KVM virtual guest, I set this as its bridged interface to install over the network (using kickstart) Once it boots into the kernel, it attempts to get the kickstart file, which fails every time. While this is going on, I am pinging the IP address I have set in the -x arguments of the virt-install command, and the pings are coming through fine. The interesting thing is when I tell the installer to try to get the kickstart file again, the pings suddenly stop, as if the interface goes down, the kickstart retrieval fails, and the pings begin again.

Is there some sort of delay or something I am not aware of, or is the trunked interface interfering with connections to the new virtual machine?

Here's the virt-install command: