Completely unable to configure Talos in a Proxmox VM

[OxygenCobalt]

Followed the instructions here to create a VM for Talos inside a fresh Proxmox installation, and for some reason simply nothing about the installation is working.

1. The installation is constantly repeating these lines about failing to resolve Cloudflare's time server, despite the overall Proxmox installation having internet access, and the VM being connected to the correct wireless bridge. The VM even indicates that it's "connected" and is still accessible over the network from my main machine.

My network configuration:

(I configure my server on a private network)

2. When I try to generate and apply a config using this, the command "succeeds" only for the Talos VM to enter an "Installing" state indefinitely. It never reboots or returns to Maintenance. All further attempts to use talosctl are refused. A second network resolution error also appears:

Strangely, the cloudflare error also changes

3. If I try to bootstrap the talos VM (something I've seen in similar issues), it also doesn't work. I can't do it after the apply-config due to it rejecting the connections as said prior, and I can't do it before either due to it seemingly rejecting the authentication with:

error executing bootstrap: rpc error: code = Unavailable desc = connection error: desc = "transport: authentication handshake failed: tls: failed to verify certificate: x509: certificate signed by unknown authority"

4. Trying to manually configure the IP assuming no DHCP server (from here) doesn't work either. It just causes the VM to explicitly state it is disconnected while still repeating the same connection errors. It's also no longer accessible externally.

5. I have disabled all firewalls in my Proxmox instance to see if it was destroying any packets, still nothing changed.

I'm frankly just stumped at this point. Can anyone help?

[smira]

network is unreachable is one way or another problem with your default gateway or routing.

instead of screenshots, please grab console (serial) logs from Proxmox, and look around the networking setup (or post the logs here)

[OxygenCobalt]

Where do I find that? Sorry, I'm new to Proxmox

[smira]

I have never used Proxmox at all, so I can't help with it.

[sakers]

I'm struggling with this too following the same instructions. Networking is fine and other VMs on the same network bridge can access the internet perfectly well.

[smira]

It's very hard to guess what is wrong without any data.

If you can talk to Talos API, use talosctl support to create a bundle and upload it here.

[sakers]

No luck with the talosctl support command I'm afraid.

[smira]

Please see https://www.talos.dev/v1.7/introduction/getting-started/#default-talosconfig-configuration-file

But you also you haven't bootstrapped your cluster: https://www.talos.dev/v1.7/introduction/getting-started/#kubernetes-bootstrap

Probably reading that getting started document would help.

[sakers]

Cheers.

That's helped. Finally got the first node up and running.

[kastl-ars]

I found that triggering a "real" reboot (instead of using kexec) makes the actual reboot more easily recognizable.

I also found that any manual IP configuration (set via ip=... when booting off the ISO or image) was not respected with a kexec reboot. But when set in the machine configuration it was respected after a "full" reboot.

[smira]

That is totally expected, as after install the kernel command line is rebuilt completely. The ip= arg is only used for the initial network configuration. Talos supports nocloud when on Proxmox

[kastl-ars]

I know, I just wanted to point this out to the OP (as it is not super self-explanatory that there is a kexec-based reboot that is not being shown on any "physical" terminal (at least I could never see it on my machines).

[rkerno]

Guys, I had very similar issues getting my first Talos cluster up and running. I was attempting to configure a 3 node control plane. I found that setting the IP address through GRUB or editing using the F3 console key both were flakey, with similar symptoms to what you have described. Using GRUB or F3 I would end up with weird situations where some of the Talos installation used the static I was trying to assign, but other parts seemed to have used a random DHCP address. It consistently ended with nothing listening on port 6443, and unfortunately, all of the linked troubleshooting relies on port 6443 being listened to. In the end I configured DHCP with my desired static IP addresses to support the installation, with the cluster address configured in DNS with each of the three IPs configured within DHCP. The install guides would benefit by including reference to the troubleshooting capabilities of talosctl such as dmesg, cat and ls.