Add a new nodepool did not complete

[cro]

Greetings,

I was playing with Grafana Cloud which has a neat feature that tells you if your k8s cluster is over or under utilized. Currently mine is severely underutilized so I thought I would add a new nodepool with smaller machines & save some money. I added the agent-medium section to my kube.tf, intending to later change the count for agent-large to zero:

{
      name        = "agent-large",
      server_type = "cpx31",
      location    = "ash",
      labels      = [],
      taints      = [],
      count       = 3
    },

    {
      name        = "agent-medium",
      server_type = "cpx21",
      location    = "ash",
      labels      = [],
      taints      = [],
      count       = 3
    }

But the terraform apply seems to be stuck at waiting for microOS to become available. In the Hetzner UI I see the new machines but they do not have private IPs allocated. If I connect to their consoles I am presented with a login prompt, so clearly the machines are up and running. Am I missing something?

[mysticaltech]

@cro Make sure to upgrade your project to the latest and greatest version, and that you have a snapshot present. See the readme!

[cro]

I was on the latest, and I did have a snapshot. However, I had never done hcloud context create <project name>. So I renamed my Default project in the console to hetzner-k3s, did hcloud context create hetzner-k3s; hcloud context use hetzner-k3s and re-ran the terraform apply. All is well.

This is probably an artifact of how long ago I started with the project.

[mysticaltech]

@cro terraform does not use the hcloud-cli, so probably you just needed to re-run terraform, somehow! It happens.