You can add bare metal worker nodes from Hetzner Robot to your cluster. The module handles enabling rescue mode, installing Talos, attaching servers to a vSwitch, and applying the Talos worker configuration.
Each server requires a unique private_ipv4 from the vSwitch subnet. For a default 10.0.0.0/16 network, you can calculate subnets using tofu console with cidrsubnet("10.0.0.0/16", 9, 176) to get 10.0.88.0/25.
Important:
- If enabling bare metal for the first time, you may need to restart Cilium components to pick up routing changes:
kubectl -n kube-system rollout restart ds/cilium ds/cilium-envoy
kubectl -n kube-system rollout restart deploy/cilium-operator
- Install Disk Warning: If you specify
install_disk, it must be a disk ID from /dev/disk/by-id. Installing Talos is destructive; the selected disk and all other eligible install disks will be wiped.
hcloud_robot_user = "<robot-user>"
hcloud_robot_password = "<robot-password>"
bare_metal_nodepools = [
{
name = "bare-metal"
servers = [
{ number = 1111111, private_ipv4 = "10.0.88.2" },
{ number = 2222222, private_ipv4 = "10.0.88.3" }
]
}
]
# Example with explicit install disk
hcloud_vswitch_id = 12345
bare_metal_nodepools = [
{
name = "bare-metal"
architecture = "amd64"
servers = [
{ number = 1234567, private_ipv4 = "10.0.88.2", install_disk = "ata-Samsung_SSD_870_ABC123" }
]
}
]
