PXE booting nodes often get stuck pulling the initramfs

[glitchcrab]

I've been creating and destroying clusters a lot whilst I work on my CAPI+gitops setup and I often see new nodes which get stuck pulling the initramfs from Sidero - it just stops on a random percentage.

For example, I've just created 3 nodes which are sitting in the pre-bootstrap boot loop and all of them have gotten stuck in this way. If I reboot the machines then they will start working again until they get stuck on another boot at some point in the future. Sometimes they get stuck on the first boot and other times they can boot 50+ times without issues.

The VMs being booted are in the same VLAN as the node running Sidero, and I can't work out whether the problem lies with Proxmox itself, Sidero, or possibly something specific with my setup. There is no firewalling inside the VLAN which could be causing problems.

If anyone has any ideas then I'd be interested in hearing them 🙏