tomp736 / kubernetes_at_hetzner Goto Github PK

On each subsequent run, terraform determines that the hcloud_server_network resource needs to be recreated causing issues with networking.

resource "hcloud_server_network" "networks" {
  for_each = local.node_networks

  server_id = module.nodes[each.value.node].id
  subnet_id = values(module.networks[each.value.network].hetzner_subnets)[0].id
}

Majority of time spent running tests is due to ansbile-playbook runs. Really unacceptable.

Terraform Apply ~5 Minutes
Ansible Apply ~ 30 Minutes+

Introduce fixed mac addresses to enable controlling interface naming using udev rules.
Interface name will be used when applying ansible roles.

The workflow is failing when attempting to executing the docker based action.

• tomp736/action-ansible
• tomp736/ansible-terraform