Claude Code Plugins

Community-maintained marketplace

Feedback
1
0

Add a new node to the Kubernetes cluster. Use when connecting new hardware, expanding cluster capacity, or setting up worker nodes.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name add-node
description Add a new node to the Kubernetes cluster. Use when connecting new hardware, expanding cluster capacity, or setting up worker nodes.

Add New Node to Cluster

Complete workflow to add a new node to the Kubernetes cluster via Tailscale.

Steps Overview

  1. Discover the node on Tailscale
  2. Add to Ansible inventory
  3. Run bootstrap playbook
  4. Run preflight checks
  5. Add to cluster

Instructions

Step 1: Discover the Node

Check if the node is visible on Tailscale:

tailscale status | grep -i <node_name>

If not found, the user needs to install Tailscale on the target:

# On target node:
curl -fsSL https://tailscale.com/install.sh | sh
sudo tailscale up

Step 2: Get Node Details

Required information:

  • Node name: Hostname (should match Tailscale name)
  • Role: worker (default) or control-plane
  • Reserved CPU: CPU cores for local use (default: "2")
  • Reserved Memory: Memory to reserve (default: "4Gi")
  • GPU: Has GPU? (default: false)
  • Labels: Custom labels (optional)

Step 3: Update Inventory

Add to ansible/inventory/hosts.yml:

workers:
  hosts:
    <node_name>:
      ansible_host: <tailscale_ip>
      tailscale_ip: <tailscale_ip>
      reserved_cpu: "2"
      reserved_memory: "4Gi"
      node_labels:
        node-role: worker
        workstation: "true"

bootstrap:
  hosts:
    <node_name>: {}

Step 4: Setup SSH Access

Test connectivity:

ssh -o BatchMode=yes -o ConnectTimeout=5 <tailscale_ip> echo "SSH OK" 2>/dev/null

If fails:

ssh-copy-id <user>@<tailscale_ip>

Step 5: Run Bootstrap Playbook

cd /home/al/git/kubani
ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/bootstrap_node.yml --limit <node_name>

Step 6: Run Preflight Checks

ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/preflight_checks.yml --limit <node_name>

Step 7: Add Node to Cluster

Include control plane node to fetch join token:

ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/add_node.yml --limit "<node_name>,sparky"

Step 8: Verify

KUBECONFIG=/home/al/.kube/config kubectl get nodes

Troubleshooting

  • SSH failed: Run ssh-copy-id <user>@<ip> manually
  • Bootstrap failed: Usually package or network issue
  • Preflight failed: Missing Tailscale auth or network issues
  • Add node failed: Check control plane health