Hi,
I've been trying to migrate some of my CentOS 7 KVM hypervisors to
CentOS 8, and I have encountered the following issue while trying to
load my network config:
virsh:
error: Failed to start network test1
error: internal error: Network is already in use by interface virbr2
journalctl:
error : networkCheckRouteCollision:123 : internal error: Network is
already in use by interface virbr2
I use the following network definitions, which are a bit non-standard,
however they work perfectly on CentOS 7:
<network>
<name>test1</name>
<forward mode='open'/>
<bridge name='virbr1' stp='off' delay='0'/>
<mac address='52:54:00:c1:a7:7c'/>
<dns enable='no'/>
<ip address='10.0.0.1' netmask='255.255.255.255'></ip>
<route address='192.168.1.22' prefix='32'
gateway='10.0.0.1'/>
</network>
<network>
<name>test2</name>
<forward mode='open'/>
<bridge name='virbr2' stp='off' delay='0'/>
<mac address='52:54:00:54:09:3c'/>
<dns enable='no'/>
<ip address='10.0.0.1' netmask='255.255.255.255'></ip>
<route address='192.168.1.33' prefix='32'
gateway='10.0.0.1'/>
</network>
My question is, why is the behaviour different across CentOS releases,
despite the libvirt versions apparently matching? Output from both
systems is the same:
# virsh version
Compiled against library: libvirt 4.5.0
Using library: libvirt 4.5.0
Using API: QEMU 4.5.0
Running hypervisor: QEMU 2.12.0
What would be the best approach on patching out this check, so I can
continue using my network config?
Just to summarize the expected output:
- virbr1 and virbr2 get created, each one with an address 10.0.0.1/32.
- Two routes get created:
192.168.1.22 via 10.0.0.1 dev virbr1
192.168.1.33 via 10.0.0.1 dev virbr2
Thanks,
Filip Hruska