My Debian Bookworm VMs are taking a very long time to start because they are trying to acquire an IP address on ‘eth0’, before giving up and then getting an address on ‘enp5s0’ instead. Is it safe to just remove ‘eth0’ from the VM altogether? There are many messages here indicating that there may be all sorts of problems if I would.
You can remove any network device from an instance.
The instances are configured (by default) to receive network configuration for any network device that is configured in that instance. You may generate an image with distrobuilder that behaves differently.
If you have information against the removal, please provide pointers.
Ok, this is my bad. It turns out that there is no configuration on the system that pertains to eth0, but during boot. I was only confused by other stances of the network configuration in the distrobuilder’s YAML file, and I still don’t know how distrobuilder selects between them. Back to the problem, the system tries quite some time to activate eth0, regardless, before switching to enp5s0, which the system does have. I would still like to get rid of the ~30-45s delay waiting for eth0 to come up. The only network configuration appears to be in
/etc/systemd/network/enp5s0.network
In the attached screenshot, the IPv4 address isn’t there when the machine has been down for a while. But until the interface has switched, the machine is unresponsive.
