Merge pull request 'fix: enable cloud-init networking in NixOS template' (#127) from stage into master

Reviewed-on: #127

.gitea/workflows/kubeadm-bootstrap.yml

@@ -27,7 +27,7 @@ jobs:
           fi
 
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: https://gitea.com/actions/checkout@v4
 
       - name: Create SSH key
         run: |

.gitea/workflows/kubeadm-reset.yml

@@ -27,7 +27,7 @@ jobs:
           fi
 
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: https://gitea.com/actions/checkout@v4
 
       - name: Create SSH key
         run: |

.gitea/workflows/terraform-apply.yml

@@ -16,7 +16,7 @@ jobs:
 
     steps:
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: https://gitea.com/actions/checkout@v4
 
       - name: Create secrets.tfvars
         working-directory: terraform

.gitea/workflows/terraform-destroy.yml

@@ -36,7 +36,7 @@ jobs:
           fi
 
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: https://gitea.com/actions/checkout@v4
 
       - name: Create Terraform secret files
         working-directory: terraform

.gitea/workflows/terraform-plan.yml

@@ -17,7 +17,7 @@ jobs:
 
     steps:
       - name: Checkout repository
-        uses: actions/checkout@v4
+        uses: https://gitea.com/actions/checkout@v4
 
       - name: Create secrets.tfvars
         working-directory: terraform

nixos/kubeadm/bootstrap/controller.py

@@ -110,9 +110,7 @@ class Controller:
             "-o",
             "IdentitiesOnly=yes",
             "-o",
-            "StrictHostKeyChecking=no",
-            "-o",
-            "UserKnownHostsFile=/dev/null",
+            "StrictHostKeyChecking=accept-new",
             "-i",
             self.ssh_key,
         ]
@@ -147,13 +145,7 @@ class Controller:
                     f"({attempt}/{self.ssh_ready_retries})"
                 )
                 time.sleep(self.ssh_ready_delay)
-        raise RuntimeError(
-            "Unable to authenticate to "
-            f"{ip} with users: {', '.join(self.ssh_candidates)}. "
-            "If this is a freshly cloned VM, the Proxmox source template likely does not yet include the "
-            "current cloud-init-capable NixOS template configuration from nixos/template-base. "
-            "Terraform can only clone what exists in Proxmox; it cannot retrofit cloud-init support into an old template."
-        )
+        raise RuntimeError(f"Unable to authenticate to {ip} with users: {', '.join(self.ssh_candidates)}")
 
     def remote(self, ip, cmd, check=True):
         ordered = [self.active_ssh_user] + [u for u in self.ssh_candidates if u != self.active_ssh_user]
@@ -174,7 +166,14 @@ class Controller:
         return last
 
     def prepare_known_hosts(self):
-        pass
+        ssh_dir = Path.home() / ".ssh"
+        ssh_dir.mkdir(parents=True, exist_ok=True)
+        (ssh_dir / "known_hosts").touch()
+        run_local(["chmod", "700", str(ssh_dir)])
+        run_local(["chmod", "600", str(ssh_dir / "known_hosts")])
+        for ip in self.node_ips.values():
+            run_local(["ssh-keygen", "-R", ip], check=False)
+            run_local(f"ssh-keyscan -H {shlex.quote(ip)} >> {shlex.quote(str(ssh_dir / 'known_hosts'))}", check=False)
 
     def prepare_remote_nix(self, ip):
         self.remote(ip, "sudo mkdir -p /etc/nix")

nixos/template-base/configuration.nix

@@ -10,7 +10,7 @@ in
       ./hardware-configuration.nix;
 
   networking.hostName = "k8s-base-template";
-  networking.useDHCP = lib.mkDefault true;
+  networking.useDHCP = false;
   networking.useNetworkd = true;
   networking.nameservers = [ "1.1.1.1" "8.8.8.8" ];