Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have an Ubuntu box with a KVM guest VM running. They use bridging so the guest VM attaches to the physical network of its host. The guest VM has 2 NICs in 2 different bridges.

First NIC of the VM is tap5 and is in bridge br0

br0     8000.46720f5c572e   no          eth0.500
                                        tap5

Second NIC of the VM is tap2 and is in bridge br100

br100       8000.76ad2fc96661   no      eth0.100
                                        eth0.101
                                        eth0.103
                                        eth0.104
                                        eth0.105
                                        tap2

On the host, br0 has an IP and br100 does not

21: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP 
link/ether 46:72:0f:5c:57:2e brd ff:ff:ff:ff:ff:ff
inet 192.168.100.4/24 brd 192.168.10.255 scope global br0
inet6 fe80::d6ae:52ff:febe:777/64 scope link 
valid_lft forever preferred_lft forever

On the guest, its eth0 and eth1 interfaces both have IP addresses

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:3e:61:fb:7a:da brd ff:ff:ff:ff:ff:ff
inet 192.168.100.6/24 brd 192.168.100.255 scope global eth0
inet6 fe80::23e:61ff:fefb:7ada/64 scope link 
   valid_lft forever preferred_lft forever
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:3e:61:fb:7a:ea brd ff:ff:ff:ff:ff:ff
inet 172.16.50.129/25 brd 172.16.50.255 scope global eth1
inet6 fe80::23e:61ff:fefb:7aea/64 scope link 
   valid_lft forever preferred_lft forever

On the guest VM, a tcpdump of its eth1 interface (tap2) shows traffic from its eth0 interface (tap5), as if the 2 bridges are themselves bridged. This means any interface on br100 is now bridged across to br0 - which is completely broken.

root@chillispot:~# tcpdump -c 1 -n -v -i eth1 net 192.168.100.0/24
tcpdump: listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes
16:31:24.175583 IP (tos 0x0, ttl 64, id 48054, offset 0, flags [DF], proto TCP (6), length 148)
192.168.100.6.22 > 192.168.100.4.59505: Flags [P.], cksum 0x6c2b (correct), seq 1056321648:1056321744, ack 398642983, win 1700, options [nop,nop,TS val 197473436 ecr 200655363], length 96

What could be bridging the 2 bridges, except the guest VM (which is a stock Ubuntu install)? I am at a complete loss! Thanks.

share|improve this question
add comment

2 Answers 2

It's not entirely clear what you're trying to do with your configuration - what's happening with the upstream switch as far as any interconnection amongst the six VLAN's coming into eth0?

The other thing that pops out is that you've both got an IP address on eth0 as well as VLAN interfaces. Put the IP on the bridge interface if you're trying to present the same subnet to the various VLAN's (...not that this configuration entirely makes sense). Again, some additional background on what you're trying to achieve would be helpful.

For what it's worth, the usual configuration is a single VLAN in a bridge with multiple VM interfaces (i.e. multiple VM's on a given subnet) rather than the inverse.

share|improve this answer
add comment

Are those VLAN interfaces? I've seen issues with combinations of bridges and VLAN, which is why I then switched to using VDE for anything fancy like that.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.