QEMU/KVM virbridge not working

Ask Fedora
,
1

I have a bridge network connected to eth0 that works

brctl show

bridge name bridge id STP enabled interfaces
br0 8000.387c761cd98f no eth0
virbr0 8000.5254002f35df no

(not sure whaat virbr0 is, I never created it)

In the client the network is setup as

<interface type="bridge">
  <mac address="52:54:00:bb:ff:27"/>
  <source bridge="br0"/>
  <target dev="vnet5"/>
  <model type="virtio"/>
  <link state="up"/>
  <alias name="net0"/>
  <address type="pci" domain="0x0000" bus="0x01" slot="0x00" function="0x0"/>
</interface>

Yet I get no DHCPin the (windows 11) client.

Windows-IP-Konfiguration


Ethernet-Adapter Ethernet:

   Verbindungsspezifisches DNS-Suffix:
   IPv4-Adresse (Auto. Konfiguration): 169.254.93.18
   Subnetzmaske  . . . . . . . . . . : 255.255.0.0
   Standardgateway . . . . . . . . . :

How do I enable dhcp traffic from client to host network?

Renewing (ipconfig /renew) times out

do I need this:

sudo sysctl -w "net.ipv4.ip_forward=1"
sudo iptables -P FORWARD ACCEPT

Or is that some bad advice?

2

The bridge works at a lower level then IP layer of the network. setting forwarding and iptables is both bad advice.

What does ip addr show?

Someone that using a bridge with NetworkManage will need to ask for info from you.
(I use systemd-networkd on my desktop to create the bridge I use).

3

I try to use the KDE Plasma NetworkManager to create the bridge. Your question ip addr reverse to OSI 3 which is indeed at a higher level than datalink (MACs and stuff)

4

virbr0 is created by default and can be used to connect VM’s to a separate network segment using NAT to connect to the internet.

If eth0 is connected to your router, the VM should pick-up an IP address from the router and becomes part of your LAN, no NAT needed.

On my system, this leads by virt-manager to network XML:

<interface type="bridge">
  <mac address="52:54:00:76:fe:fd"/>
  <source bridge="bridge0"/>
  <model type="virtio"/>
  <address type="pci" domain="0x0000" bus="0x01" slot="0x00" function="0x0"/>
</interface>

without target dev, it’s choosen sequentially.

Be sure that both eth0 and vnet5 belong to the bridge, I get:

$ bridge link show master bridge0
2: enp2s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 master bridge0 state forwarding priority 32 cost 100 
5: vnet0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 master bridge0 state forwarding priority 32 cost 2

and the bridge should get the IP address from the network, not the eth0. in NetworkManager, eth0 should be defined as slave of br0. Then it should work.

5

I ran this:

sudo sysctl -w net.bridge.bridge-nf-call-iptables=0
sudo sysctl -w net.bridge.bridge-nf-call-ip6tables=0
6

I’m aware of that, the ip addr reveals mistakes in config.

7

I believe all the qemukvm bridge blues can be retraced to the fact that I have docker installed as well. This is not specifically covered in the many resources I’ve found online and deepseek. I had Claude signal me that docker bridging could potential thward bridging for others and was able to solve the problem (for the time being.)

8

Usually you would use podman not docker on Fedora.

9

I’m glad I got docker working. But you are envited to get all 45+ containers working on podman .