[bug] additional_post_k3s_commands

[privatecoder]

I was just debugging a strange issue with additional_post_k3s_commands when I realised that both, additional_pre_k3s_commands and additional_post_k3s_commands end up in runcmd as part of cloud‑init (here)

As cloud‑init needs to finish before the k3s install script runs this can never work – or am I wrong?

[privatecoder]

I now moved it to templates/master_install_script.sh and templates/worker_install_script.sh but I am not too happy with the inconsistency, i.e. additional_pre_k3s_commands being part of cloud-init while this is part of an install_script.

[privatecoder]

and I can safely confirm additional_post_k3s_commands is 100% non-functional in the current version

[tomlnd]

Can also confirm additional_post_k3s_commands is not working in the latest version.
Tried adding a persistent routing configuration in additional_pre_k3s_commands like so:

additional_pre_k3s_commands:
  - printf "[Match]\nName=enp7s0\n\n[Network]\nDHCP=yes\n\n[Route]\nDestination=x.x.x.x/16\nGateway=x.x.x.x\n\n[Route]\nDestination=x.x.x.x/24\nGateway=x.x.x.x\n" > /etc/systemd/network/10-enp7s0.network

and restart systemd-networkd in the post-commands:

additional_post_k3s_commands:
  - systemctl restart systemd-networkd

but that's not restarting the service at all.