Timeout on VM's that have network interfaces without an ip assigned.

[xonvanetta]

Describe the bug Timeout on VM's that have network interfaces without an ip assigned. This can be caused when using something like k3s, podman/docker and other containerized applications.

To Reproduce Steps to reproduce the behavior:

1. Have a VM with an Network Interface thats missing IP
2. Having qemu guest agent enabled on this host
3. Trying to run terrafrom apply, will try until timeout is hit which is default 15m

Expected behavior Should be able to handle those interfaces and not timeout.

Screenshots

Additional context I tried to find a way to configure qemu guest agent to ignore certain networks but didn't find a lot, could have missed something to configure it to ignore those interfaces.

I can create a PR for this but the real question is how this should be handled, should 1 ip be enough or any other ideas?

Workaround Setting the agent to a lower timeout

agent {
    enabled = true
    timeout = "2s"
}

[bpg]

Thanks @xonvanetta, let me look into this. The issue looks quite similar to #100, and prob. #133. I think it would be really helpful to have a reproducible example of such network configuration, as it seems the problem is fairly common.

[xonvanetta]

Yea, i made a test in the http client that has this problem right now here

Don't know if this is how you wanted it or a real vm with the config or if the json response from proxmox is enough? Also sorry my linter/format tool went a bit wild

[bpg]

👍🏻 thanks @xonvanetta, I'll try to debug it further.