Merge pull request #170 from JamesTurland/k3s-deploy-overhaul

Overhaul K3S-Deploy (k3sup + kube-vip + MetalLB)
2026-09-26 08:26:38 +00:00 · 2026-08-22 15:16:59 +02:00
parent 8bea30ce55 081c11fb40
commit 088d8c2c05
6 changed files with 508 additions and 194 deletions
@@ -0,0 +1,151 @@
 # K3S HA Deploy (k3sup + kube-vip + MetalLB)
 Deploys a highly-available k3s cluster over SSH with `k3sup`, a floating
 control-plane VIP via **kube-vip**, and `type: LoadBalancer` support via
 **MetalLB**.
 > **Following the YouTube tutorial?** This script has been modernized since
 > the video — see [What changed](#what-changed-from-the-video) below.
 ## What it builds
 - A 3-server (control-plane) + 2-agent (worker) k3s cluster with embedded etcd.
 - **kube-vip** advertises a single virtual IP for the Kubernetes API across
  all control-plane nodes (survives a node failure).
 - **MetalLB** hands out real IPs to `type: LoadBalancer` services.
 ### Why both kube-vip and MetalLB?
 They solve different problems. **kube-vip** provides the *control-plane* VIP
 (one address for the API server, HA across masters). **MetalLB** provides
 *service* load balancing (`type: LoadBalancer` for your apps). The older
 kube-vip cloud-provider that overlapped MetalLB has been removed to avoid
 duplicate LoadBalancer controllers.
 ## Prerequisites
 - **Ubuntu/Debian** nodes (the script installs prerequisites with `apt`; on a
  non-apt distro it exits with a clear message).
 - **Passwordless sudo** for the SSH user on every node. Example cloud-init:
  ```yaml
  #cloud-config
  users:
    - name: <your-user>
      sudo: ["ALL=(ALL) NOPASSWD:ALL"]
      groups: [sudo]
  ```
 - An **SSH key pair** you can use to reach the nodes (the script distributes
  the public key and never clobbers your `~/.ssh/config`).
 - **5 nodes by default** (3 servers + 2 agents). The minimum for HA is 3
  servers (etcd needs a quorum); workers are optional. Adjust the node list in
  the config block if you have fewer machines.
 - Run the script from a **separate Ubuntu/Debian admin machine** (your
  laptop/workstation) that can SSH to every node — not on one of the nodes
  themselves. It installs `k3sup` and `kubectl` locally if missing.
 ## Important caveats
 - **The VIP and the MetalLB range must be outside your DHCP pool**, and must
  not overlap each other or any node IP. If your router hands out `vip` or an
  address in `lbrange` via DHCP, you'll get intermittent, hard-to-debug
  failures — reserve them on your router/DHCP server first.
 - **Same subnet / L2 only.** kube-vip (ARP mode) and MetalLB (L2) both
  advertise via ARP, so the VIP and LoadBalancer IPs must be on the **same
  subnet/VLAN** as the nodes and the clients reaching them — they are not
  routed across subnets.
 - **`interface` must match your nodes' real NIC.** It's `eth0` on many
  systems but often `ens18`/`enp0s3` on Proxmox and cloud images. Check with
  `ip -o -4 route show default`.
 - **Host keys are trusted on first use** (`accept-new`). If you re-image a
  node, clear its stale entry first: `ssh-keygen -R <node-ip>`.
 - This is a **homelab tutorial, not production-hardened** as-is (aggressive
  leader election, no etcd backups, single L2 domain).
 ## Usage
 1. Snapshot your VMs (so you can roll back — the script modifies every node).
 2. Place your SSH **private** key at `~/<certName>` or `~/.ssh/<certName>`,
   with the matching `.pub` beside it. `certName` is the key's *filename only*
   — no path, no `.pub` suffix. Modern OpenSSH defaults to `id_ed25519`, so set
   `certName` to match the key you actually have.
 3. Edit the **"YOU SHOULD ONLY NEED TO EDIT THIS SECTION"** block in
   `k3s.sh` — node IPs, `user`, `interface`, `vip`, `lbrange`, `certName`.
   The example IPs (`192.168.3.x`, VIP `.50`, range `.60-.80`) and
   `interface=eth0` are placeholders — change them for your network.
 4. `chmod +x k3s.sh && ./k3s.sh`
 5. Review the pre-flight summary and confirm. Grab a coffee.
 ### After it finishes
 The kubeconfig is **merged** into `~/.kube/config` as context `k3s-ha` (set by
 `context` in the config block), pointed at the VIP. Select it and check the
 cluster:
 ```bash
 kubectl config use-context k3s-ha
 kubectl get nodes -o wide
 curl http://<load-balancer-ip>   # the script prints the assigned IP
 ```
 Worker nodes are labeled `worker=true` and `longhorn=true` (the latter for the
 companion Longhorn storage tutorial — harmless if you don't use it).
 If a node fails to join, reset just that node and re-run the script (it's
 idempotent): `k3s-uninstall.sh` on a server, `k3s-agent-uninstall.sh` on a
 worker.
 ### Options (environment variables)
 | Variable         | Purpose                                               |
 | ---------------- | ----------------------------------------------------- |
 | `ASSUME_YES=1`   | Skip the pre-flight `[y/N]` prompt (unattended runs). |
 | `NO_COLOR=1`     | Disable colored output.                               |
 | `RAW_BASE=<url>` | Where to fetch the sibling manifests (kube-vip, ipAddressPool, l2Advertisement); default: upstream `main`. Override for a fork/branch/local copy. |
 ### Tracking the latest k3s instead of a pinned version
 In the config block set `k3sChannel="stable"` and leave `k3sVersion=""`.
 The script then always installs the current stable k3s release.
 ### Upgrading kube-vip
 The `kube-vip` manifest is generated from the pinned image (its env schema
 changes between releases). To bump:
 ```bash
 docker run --rm ghcr.io/kube-vip/kube-vip:<version> manifest daemonset \
  --interface eth0 --address 10.0.0.254 \
  --controlplane --arp --leaderElection --taint --inCluster > kube-vip
 sed -i 's/value: eth0/value: REPLACE_INTERFACE/' kube-vip
 sed -i 's/value: 10.0.0.254/value: REPLACE_VIP/' kube-vip
 ```
 `--inCluster` is required on k3s: it makes kube-vip use the `kube-vip`
 ServiceAccount (created by the RBAC the script applies) instead of the
 kubeadm `/etc/kubernetes/admin.conf` kubeconfig, which does not exist on k3s.
 Then update `KVVERSION` in `k3s.sh` to match.
 ## What changed from the video
 - **Versions:** k3s `v1.35.6+k3s1`, kube-vip `v1.2.1`, MetalLB `v0.16.0`.
 - **kube-vip** now runs on **all** control-plane nodes, and your kubeconfig
  points at the **VIP** (not master1) — fixes `localhost:8080`/API errors. A
  readiness check confirms the VIP is answering before the script continues.
 - Removed the redundant kube-vip **cloud-provider**; MetalLB alone handles
  `type: LoadBalancer`.
 - **MetalLB** installs from a single native manifest (which creates its own
  `metallb-system` namespace) — the old separate namespace apply and its
  mismatched MetalLB versions are gone.
 - **k3sup:** the join token is fetched once and reused across all nodes, and
  `k3sup ready` waits for the cluster instead of a hand-rolled poll loop.
 - **Safer SSH:** host keys are added via `ssh-keyscan`; the script no longer
  overwrites `~/.ssh/config`.
 - **Hardening:** `set -euo pipefail`, per-node time sync, `apt` prerequisite
  install with a clear message on unsupported distros, arch-aware `kubectl`,
  and it is safe to re-run.
 - **Cleaner output:** non-blinking step-by-step logging (only the banner
  still blinks — for old times' sake), a pre-flight config summary with a
  `[y/N]` confirmation (`ASSUME_YES=1` to skip), and a final summary listing
  the API VIP and the assigned LoadBalancer IP.
@@ -5,4 +5,4 @@ metadata:
  namespace: metallb-system
 spec:
  addresses:
-  - $lbrange
+  - REPLACE_LBRANGE
@@ -1,232 +1,404 @@
-#!/bin/bash
+#!/usr/bin/env bash
 #
 # JimsGarage — K3S HA deploy (k3sup + kube-vip + MetalLB)
 # Tutorial: https://youtube.com/@jims-garage
 # See ./README.md for prerequisites, configuration, and what changed
 # from the original video.
 set -euo pipefail
 # ── Banner (kept from the original tutorial; blink preserved) ──────────
 echo -e " \033[33;5m    __  _          _        ___                            \033[0m"
 echo -e " \033[33;5m    \ \(_)_ __ ___( )__    / _ \__ _ _ __ __ _  __ _  ___  \033[0m"
 echo -e " \033[33;5m     \ \ | '_ \` _ \/ __|  / /_\/ _\` | '__/ _\` |/ _\` |/ _ \ \033[0m"
 echo -e " \033[33;5m  /\_/ / | | | | | \__ \ / /_\\  (_| | | | (_| | (_| |  __/ \033[0m"
 echo -e " \033[33;5m  \___/|_|_| |_| |_|___/ \____/\__,_|_|  \__,_|\__, |\___| \033[0m"
 echo -e " \033[33;5m                                               |___/       \033[0m"
 echo -e " \033[36;5m         _  _________   ___         _        _ _           \033[0m"
 echo -e " \033[36;5m        | |/ |__ / __| |_ _|_ _  __| |_ __ _| | |          \033[0m"
 echo -e " \033[36;5m        | ' < |_ \__ \  | || ' \(_-|  _/ _\` | | |          \033[0m"
 echo -e " \033[36;5m        |_|\_|___|___/ |___|_||_/__/\__\__,_|_|_|          \033[0m"
 echo -e " \033[36;5m                                                           \033[0m"
 echo -e " \033[32;5m             https://youtube.com/@jims-garage              \033[0m"
 echo -e " \033[32;5m                                                           \033[0m"
 # ── Output helpers (non-blinking; auto-disable when not a TTY) ──────────
 if [[ -t 1 && -z "${NO_COLOR:-}" ]]; then
  c_reset=$'\033[0m'; c_step=$'\033[1;36m'; c_info=$'\033[36m'
  c_ok=$'\033[32m';   c_warn=$'\033[33m';   c_err=$'\033[31m'
 else
  c_reset=; c_step=; c_info=; c_ok=; c_warn=; c_err=
 fi
 step() { printf '\n%s━━━ %s ━━━%s\n' "$c_step" "$*" "$c_reset"; }
 info() { printf ' %s•%s %s\n'  "$c_info" "$c_reset" "$*"; }
 ok()   { printf ' %s✓%s %s\n'  "$c_ok"   "$c_reset" "$*"; }
 warn() { printf ' %s!%s %s\n'  "$c_warn" "$c_reset" "$*" >&2; }
 err()  { printf ' %s✗%s %s\n'  "$c_err"  "$c_reset" "$*" >&2; }
 die()  { err "$*"; exit 1; }
 #############################################
 # YOU SHOULD ONLY NEED TO EDIT THIS SECTION #
 #############################################
-# Version of Kube-VIP to deploy
+# k3s version to install. To always track the latest stable release
-KVVERSION="v0.6.3"
+# instead, set k3sChannel="stable" and leave k3sVersion empty (see README).
 k3sVersion="v1.35.6+k3s1"
 k3sChannel=""
-# K3S Version
+# kube-vip: must match the image tag baked into the ./kube-vip manifest.
-k3sVersion="v1.26.10+k3s2"
+# To bump, regenerate ./kube-vip (see README "Upgrading kube-vip").
 KVVERSION="v1.2.1"
-# Set the IP addresses of the master and work nodes
+# MetalLB version (used to build the manifest URL).
 METALLB_VERSION="v0.16.0"
 # Node IP addresses
 master1=192.168.3.21
 master2=192.168.3.22
 master3=192.168.3.23
 worker1=192.168.3.24
 worker2=192.168.3.25
-# User of remote machines
+# SSH user on the remote nodes
 user=ubuntu
-# Interface used on remotes
+# Network interface used on the remote nodes
 interface=eth0
-# Set the virtual IP address (VIP)
+# Virtual IP (VIP) for the HA control plane
 vip=192.168.3.50
-# Array of master nodes
+# MetalLB LoadBalancer address range
 masters=($master2 $master3)
 # Array of worker nodes
 workers=($worker1 $worker2)
 # Array of all
 all=($master1 $master2 $master3 $worker1 $worker2)
 # Array of all minus master
 allnomaster1=($master2 $master3 $worker1 $worker2)
 #Loadbalancer IP range
 lbrange=192.168.3.60-192.168.3.80
-#ssh certificate name variable
+# SSH private key name (in ~/.ssh) used to reach the nodes
 certName=id_rsa
-#ssh config file
+# kubeconfig context name to create locally
-config_file=~/.ssh/config
+context=k3s-ha
 #############################################
 #            DO NOT EDIT BELOW              #
 #############################################
 # For testing purposes - in case time is wrong due to VM snapshots
 sudo timedatectl set-ntp off
 sudo timedatectl set-ntp on
-# Move SSH certs to ~/.ssh and change permissions
+# Additional control-plane servers joined after master1
-cp /home/$user/{$certName,$certName.pub} /home/$user/.ssh
+masters=("$master2" "$master3")
-chmod 600 /home/$user/.ssh/$certName 
+# Agent (worker) nodes
-chmod 644 /home/$user/.ssh/$certName.pub
+workers=("$worker1" "$worker2")
 # All control-plane nodes (kube-vip manifest is placed on each)
 masters_all=("$master1" "${masters[@]}")
 # Every node (for SSH prep loops)
 all=("$master1" "${masters[@]}" "${workers[@]}")
-# Install k3sup to local machine if not already present
+# Base URL for the sibling manifests (kube-vip, ipAddressPool,
-if ! command -v k3sup version &> /dev/null
+# l2Advertisement). Override to test from a branch/fork/local copy, e.g.
-then
+#   RAW_BASE="file://$HOME/JimsGarage"
-    echo -e " \033[31;5mk3sup not found, installing\033[0m"
+RAW_BASE="${RAW_BASE:-https://raw.githubusercontent.com/JamesTurland/JimsGarage/main}"
-    curl -sLS https://get.k3sup.dev | sh
+manifest_base="$RAW_BASE/Kubernetes/K3S-Deploy"
-    sudo install k3sup /usr/local/bin/
+
 ssh_key="$HOME/.ssh/$certName"
 ssh_opts=(-i "$ssh_key" -o StrictHostKeyChecking=accept-new)
 # k3s version selector: prefer an explicit version, else a channel.
 if [[ -n "$k3sVersion" ]]; then
  k3s_selector=(--k3s-version "$k3sVersion"); k3s_display="$k3sVersion"
 else
-    echo -e " \033[32;5mk3sup already installed\033[0m"
+  k3s_selector=(--k3s-channel "${k3sChannel:-stable}"); k3s_display="channel:${k3sChannel:-stable}"
 fi
-# Install Kubectl if not already present
+# Control-plane k3s args: disable bundled traefik + servicelb (we use
-if ! command -v kubectl version &> /dev/null
+# MetalLB), add the VIP to the API serving cert SAN, pin the flannel interface
-then
+# + node IP, taint as control-plane. Applied to EVERY server so master2/3
-    echo -e " \033[31;5mKubectl not found, installing\033[0m"
+# declare the VIP SAN too (not just master1).
-    curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
+# Explicit --disable is kept (not k3sup --no-extras) for tutorial clarity.
-    sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl
+server_extra_args() {  # $1 = node ip
  printf '%s' "--disable traefik --disable servicelb --tls-san=$vip --flannel-iface=$interface --node-ip=$1 --node-taint node-role.kubernetes.io/control-plane=true:NoSchedule"
 }
 # k3sup prints a promo tip ("Create clusters on Mac…/slicervm.com") on every
 # install/join. Filter it out. Use sed, not grep -v: sed always exits 0, so
 # pipefail still surfaces a real k3sup failure (grep would exit 1 on no match
 # and spuriously abort under set -e). Only wraps the noisy display commands —
 # never node-token (whose stdout is captured for the token).
 k3sup_quiet() { command k3sup "$@" 2>&1 | sed '/Create clusters on Mac/d'; }
 # Apply a manifest, retrying on transient failure. MetalLB installs a
 # validating webhook served by its controller; there is a brief window where
 # the controller Deployment reports Available but the webhook endpoints/TLS
 # are not serving yet, so the first apply can fail with "no endpoints
 # available for service metallb-webhook-service". Retry to ride that out.
 kubectl_apply_retry() {  # $1 = manifest path or URL
  local i
  for i in $(seq 1 6); do
    kubectl apply -f "$1" && return 0
    warn "kubectl apply $1 failed (attempt $i/6) — retrying in 5s"
    sleep 5
  done
  die "kubectl apply $1 failed after 6 attempts"
 }
 # ── Pre-flight: show config, confirm before touching any node ──────────
 step "Pre-flight · review configuration"
 printf '  %-16s %s\n' \
  "k3s:"           "$k3s_display" \
  "kube-vip:"      "$KVVERSION" \
  "metallb:"       "$METALLB_VERSION" \
  "control plane:" "${masters_all[*]}" \
  "workers:"       "${workers[*]}" \
  "api VIP:"       "$vip (interface $interface)" \
  "lb range:"      "$lbrange" \
  "ssh user/key:"  "$user / $ssh_key" \
  "kube context:"  "$context"
 if [[ "${ASSUME_YES:-}" == "1" ]]; then
  info "ASSUME_YES=1 — proceeding without prompt"
 else
-    echo -e " \033[32;5mKubectl already installed\033[0m"
+  read -rp "$(printf '\n Proceed? [y/N] ')" reply || reply=""
  [[ "$reply" =~ ^[Yy]$ ]] || die "Aborted."
 fi
-# Check for SSH config file, create if needed, add/change Strict Host Key Checking (don't use in production!)
+# ── Step 1/9 · Local tools (k3sup, kubectl) ────────────────────────────
-
+step "Step 1/9 · Local tools (k3sup, kubectl)"
-if [ ! -f "$config_file" ]; then
+if ! command -v k3sup &>/dev/null; then
-  # Create the file and add the line
+  info "Installing k3sup"
-  echo "StrictHostKeyChecking no" > "$config_file"
+  curl -sLS https://get.k3sup.dev | sh
-  # Set permissions to read and write only for the owner
+  # The installer leaves the binary in the CWD — named "k3sup", or
-  chmod 600 "$config_file"
+  # "k3sup-<arch>" when run unprivileged (it can't self-install to
-  echo "File created and line added."
+  # /usr/local/bin). Install whichever it produced.
  k3sup_bin=""
  for f in k3sup k3sup-*; do
    [[ -f "$f" ]] && { k3sup_bin="$f"; break; }
  done
  [[ -n "$k3sup_bin" ]] || die "k3sup installer produced no binary"
  sudo install "$k3sup_bin" /usr/local/bin/k3sup
  rm -f "$k3sup_bin"
 else
-  # Check if the line exists
+  ok "k3sup present"
-  if grep -q "^StrictHostKeyChecking" "$config_file"; then
+fi
-    # Check if the value is not "no"
+if ! command -v kubectl &>/dev/null; then
-    if ! grep -q "^StrictHostKeyChecking no" "$config_file"; then
+  info "Installing kubectl"
-      # Replace the existing line
+  arch="$(dpkg --print-architecture 2>/dev/null || uname -m)"
-      sed -i 's/^StrictHostKeyChecking.*/StrictHostKeyChecking no/' "$config_file"
+  case "$arch" in
-      echo "Line updated."
+    amd64|x86_64) arch=amd64 ;;
-    else
+    arm64|aarch64) arch=arm64 ;;
-      echo "Line already set to 'no'."
+    *) die "Unsupported architecture for kubectl: $arch" ;;
-    fi
+  esac
-  else
+  kver="$(curl -L -s https://dl.k8s.io/release/stable.txt)"
-    # Add the line to the end of the file
+  curl -LO "https://dl.k8s.io/release/${kver}/bin/linux/${arch}/kubectl"
-    echo "StrictHostKeyChecking no" >> "$config_file"
+  sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl
-    echo "Line added."
+else
-  fi
+  ok "kubectl present"
 fi
-#add ssh keys for all nodes
+# ── Step 2/9 · SSH keys and known_hosts ────────────────────────────────
 step "Step 2/9 · SSH keys and known_hosts"
 mkdir -p "$HOME/.ssh"; chmod 700 "$HOME/.ssh"
 # If the key was dropped in $HOME (per the README), move it into ~/.ssh.
 if [[ -f "$HOME/$certName" && ! -f "$ssh_key" ]]; then
  info "Moving $certName into ~/.ssh"
  cp "$HOME/$certName" "$ssh_key"
  [[ -f "$HOME/$certName.pub" ]] && cp "$HOME/$certName.pub" "$ssh_key.pub"
 fi
 [[ -f "$ssh_key" ]] || die "SSH key not found: $ssh_key (see README prerequisites)"
 chmod 600 "$ssh_key"; [[ -f "$ssh_key.pub" ]] && chmod 644 "$ssh_key.pub"
 # Trust host keys without clobbering ~/.ssh/config (issue #62).
 for node in "${all[@]}"; do
-  ssh-copy-id $user@$node
+  ssh-keyscan "$node" 2>/dev/null >> "$HOME/.ssh/known_hosts" || \
    warn "ssh-keyscan failed for $node (will accept-new on first connect)"
 done
 [[ -f "$HOME/.ssh/known_hosts" ]] && sort -u "$HOME/.ssh/known_hosts" -o "$HOME/.ssh/known_hosts"
 # Distribute the public key to each node.
 for node in "${all[@]}"; do
  info "Copying SSH key to $user@$node"
  ssh-copy-id -i "$ssh_key.pub" -o StrictHostKeyChecking=accept-new "$user@$node" >/dev/null 2>&1 || \
    warn "ssh-copy-id to $node failed (key may already be present)"
 done
 ok "SSH ready for ${#all[@]} nodes"
 # ── Step 3/9 · Prepare nodes (time sync + prerequisites) ───────────────
 step "Step 3/9 · Prepare nodes (time sync + prerequisites)"
 for node in "${all[@]}"; do
  info "Preparing $node"
  rc=0
  ssh "${ssh_opts[@]}" "$user@$node" 'sudo bash -s' <<'REMOTE' || rc=$?
 set -e
 # Resync time — VM snapshots drift, which breaks k3s/k3sup installs (issue #68).
 timedatectl set-ntp off || true
 timedatectl set-ntp on  || true
 if command -v apt-get >/dev/null 2>&1; then
  export DEBIAN_FRONTEND=noninteractive NEEDRESTART_MODE=a
  ok=0
  for i in $(seq 1 10); do
    if apt-get update && apt-get install -y iptables sudo policycoreutils; then ok=1; break; fi
    echo "apt busy, retry $i/10..."; sleep 3
  done
  [ "$ok" = 1 ] || exit 91
 else
  exit 90
 fi
 REMOTE
  case "$rc" in
    0)  ok "$node prepared" ;;
    90) die "Node $node is not apt-based. This script targets Ubuntu/Debian; install iptables, sudo, and policycoreutils manually (see README)." ;;
    91) die "Node $node: apt failed after retries." ;;
    *)  die "Node $node: preparation failed (exit $rc)." ;;
  esac
 done
-# Install policycoreutils for each node
+# ── Step 4/9 · Bootstrap first control-plane node ──────────────────────
-for newnode in "${all[@]}"; do
+step "Step 4/9 · Bootstrap first control-plane node ($master1)"
-  ssh $user@$newnode -i ~/.ssh/$certName sudo su <<EOF
+mkdir -p "$HOME/.kube"
-  NEEDRESTART_MODE=a apt-get install policycoreutils -y
+k3sup_quiet install \
-  exit
+  --ip "$master1" \
-EOF
+  --user "$user" \
  echo -e " \033[32;5mPolicyCoreUtils installed!\033[0m"
 done
 # Step 1: Bootstrap First k3s Node
 mkdir ~/.kube
 k3sup install \
  --ip $master1 \
  --user $user \
  --tls-san $vip \
  --cluster \
  --k3s-version $k3sVersion \
  --k3s-extra-args "--disable traefik --disable servicelb --flannel-iface=$interface --node-ip=$master1 --node-taint node-role.kubernetes.io/master=true:NoSchedule" \
  --merge \
  --sudo \
-  --local-path $HOME/.kube/config \
+  --cluster \
-  --ssh-key $HOME/.ssh/$certName \
+  "${k3s_selector[@]}" \
-  --context k3s-ha
+  --k3s-extra-args "$(server_extra_args "$master1")" \
-echo -e " \033[32;5mFirst Node bootstrapped successfully!\033[0m"
+  --merge \
  --local-path "$HOME/.kube/config" \
  --ssh-key "$ssh_key" \
  --context "$context"
 ok "First node bootstrapped"
-# Step 2: Install Kube-VIP for HA
+# ── Step 5/9 · Install kube-vip (control-plane VIP) ────────────────────
 step "Step 5/9 · Install kube-vip (control-plane VIP)"
 # Ensure kubectl targets the cluster we just created (the merge above may
 # have left a different current-context in a pre-existing ~/.kube/config).
 kubectl config use-context "$context" >/dev/null
 info "Applying kube-vip RBAC"
 kubectl apply -f https://kube-vip.io/manifests/rbac.yaml
 info "Rendering kube-vip manifest (interface=$interface, vip=$vip)"
 curl -sfL "$manifest_base/kube-vip" -o "$HOME/kube-vip.src"
 sed "s/REPLACE_INTERFACE/$interface/g; s/REPLACE_VIP/$vip/g" \
  "$HOME/kube-vip.src" > "$HOME/kube-vip.yaml"
 for node in "${masters_all[@]}"; do
  info "Placing kube-vip manifest on $node"
  # scp to /tmp (absolute) — the mv below runs as root, where ~ is /root,
  # not the ssh user's home where scp would otherwise land the file.
  scp "${ssh_opts[@]}" "$HOME/kube-vip.yaml" "$user@$node:/tmp/kube-vip.yaml" >/dev/null
  ssh "${ssh_opts[@]}" "$user@$node" 'sudo bash -s' <<'REMOTE'
 set -e
 mkdir -p /var/lib/rancher/k3s/server/manifests
 mv /tmp/kube-vip.yaml /var/lib/rancher/k3s/server/manifests/kube-vip.yaml
 chown root:root /var/lib/rancher/k3s/server/manifests/kube-vip.yaml
 chmod 600 /var/lib/rancher/k3s/server/manifests/kube-vip.yaml
 REMOTE
 done
 ok "kube-vip deployed to ${#masters_all[@]} control-plane nodes"
 info "Pointing local kubeconfig at the VIP ($vip)"
 # Rewrite the server URL from master1 to the VIP. The dots in the IP are
 # unescaped regex metacharacters, but a false match would need the literal
 # "https://<master1-ish>:6443" string elsewhere in the kubeconfig, which does
 # not occur (cert data is base64 — no scheme or port), so this is safe here.
 sed -i "s#https://$master1:6443#https://$vip:6443#" "$HOME/.kube/config"
-# Step 3: Download kube-vip
+# Confirm the VIP is actually answering before any VIP-routed kubectl below.
-curl -sO https://raw.githubusercontent.com/JamesTurland/JimsGarage/main/Kubernetes/K3S-Deploy/kube-vip
+# This is an early tripwire for a broken kube-vip (wrong interface, RBAC, etc.)
-cat kube-vip | sed 's/$interface/'$interface'/g; s/$vip/'$vip'/g' > $HOME/kube-vip.yaml
+# instead of a confusing failure further down.
 info "Waiting for the control-plane VIP ($vip) to answer"
 vip_ready=""
 for _ in $(seq 1 30); do
  if kubectl --request-timeout=5s get --raw='/readyz' &>/dev/null; then
    vip_ready=1; break
  fi
  sleep 2
 done
 [[ -n "$vip_ready" ]] || die "Control-plane VIP $vip is not answering — check kube-vip on the masters: kubectl -n kube-system logs -l app.kubernetes.io/name=kube-vip-ds"
 ok "Control-plane VIP is answering"
-# Step 4: Copy kube-vip.yaml to master1
+# ── Step 6/9 · Fetch the cluster join token ────────────────────────────
-scp -i ~/.ssh/$certName $HOME/kube-vip.yaml $user@$master1:~/kube-vip.yaml
+step "Step 6/9 · Fetch the cluster join token"
 node_token="$(k3sup node-token --ip "$master1" --user "$user" --ssh-key "$ssh_key")" \
  || die "Failed to fetch node token from $master1"
 [[ -n "$node_token" ]] || die "Empty node token from $master1"
 ok "Join token fetched"
-
+# ── Step 7/9 · Join control-plane and worker nodes ─────────────────────
-# Step 5: Connect to Master1 and move kube-vip.yaml
+step "Step 7/9 · Join control-plane and worker nodes"
-ssh $user@$master1 -i ~/.ssh/$certName <<- EOF
+# The prefetched --node-token is what keeps these joins from ever SSHing to
-  sudo mkdir -p /var/lib/rancher/k3s/server/manifests
+# the server: without it, k3sup would try --server-user (default: root) against
-  sudo mv kube-vip.yaml /var/lib/rancher/k3s/server/manifests/kube-vip.yaml
+# --server-ip (the VIP, not an SSH target) and fail confusingly. Keep the token.
-EOF
+for node in "${masters[@]}"; do
-
+  info "Joining control-plane node $node"
-# Step 6: Add new master nodes (servers) & workers
+  k3sup_quiet join \
-for newnode in "${masters[@]}"; do
+    --ip "$node" \
-  k3sup join \
+    --user "$user" \
    --ip $newnode \
    --user $user \
    --sudo \
    --k3s-version $k3sVersion \
    --server \
-    --server-ip $master1 \
+    --server-ip "$vip" \
-    --ssh-key $HOME/.ssh/$certName \
+    --node-token "$node_token" \
-    --k3s-extra-args "--disable traefik --disable servicelb --flannel-iface=$interface --node-ip=$newnode --node-taint node-role.kubernetes.io/master=true:NoSchedule" \
+    "${k3s_selector[@]}" \
-    --server-user $user
+    --k3s-extra-args "$(server_extra_args "$node")" \
-  echo -e " \033[32;5mMaster node joined successfully!\033[0m"
+    --ssh-key "$ssh_key"
  ok "Control-plane node $node joined"
 done
-
+for node in "${workers[@]}"; do
-# add workers
+  info "Joining worker node $node"
-for newagent in "${workers[@]}"; do
+  k3sup_quiet join \
-  k3sup join \
+    --ip "$node" \
-    --ip $newagent \
+    --user "$user" \
    --user $user \
    --sudo \
-    --k3s-version $k3sVersion \
+    --server-ip "$vip" \
-    --server-ip $master1 \
+    --node-token "$node_token" \
-    --ssh-key $HOME/.ssh/$certName \
+    "${k3s_selector[@]}" \
-    --k3s-extra-args "--node-label \"longhorn=true\" --node-label \"worker=true\""
+    --k3s-extra-args "--flannel-iface=$interface --node-ip=$node --node-label longhorn=true --node-label worker=true" \
-  echo -e " \033[32;5mAgent node joined successfully!\033[0m"
+    --ssh-key "$ssh_key"
  ok "Worker node $node joined"
 done
-# Step 7: Install kube-vip as network LoadBalancer - Install the kube-vip Cloud Provider
+# Assert the full HA control plane is up before continuing. The VIP gate in
-kubectl apply -f https://raw.githubusercontent.com/kube-vip/kube-vip-cloud-provider/main/manifest/kube-vip-cloud-controller.yaml
+# Step 5 only proved master1 (a single-member etcd); a control-plane join that
 # half-succeeds would otherwise sail through as success on a degraded cluster.
 info "Waiting for all ${#masters_all[@]} control-plane nodes to be Ready"
 kubectl wait --for=condition=Ready node \
  -l node-role.kubernetes.io/control-plane --timeout=180s || \
  die "Control-plane nodes did not all become Ready — check etcd/kube-vip on the masters: kubectl get nodes; kubectl -n kube-system logs -l app.kubernetes.io/name=kube-vip-ds"
 cp_ready="$(kubectl get nodes -l node-role.kubernetes.io/control-plane --no-headers 2>/dev/null | awk '$2=="Ready"{c++} END{print c+0}')"
 [[ "$cp_ready" -eq "${#masters_all[@]}" ]] || \
  die "Expected ${#masters_all[@]} control-plane nodes Ready, found $cp_ready — etcd quorum may be degraded"
 ok "All ${#masters_all[@]} control-plane nodes Ready"
-# Step 8: Install Metallb
+# ── Step 8/9 · Install MetalLB (service LoadBalancer) ──────────────────
-kubectl apply -f https://raw.githubusercontent.com/metallb/metallb/v0.12.1/manifests/namespace.yaml
+step "Step 8/9 · Install MetalLB (service LoadBalancer)"
-kubectl apply -f https://raw.githubusercontent.com/metallb/metallb/v0.13.12/config/manifests/metallb-native.yaml
+kubectl apply -f "https://raw.githubusercontent.com/metallb/metallb/$METALLB_VERSION/config/manifests/metallb-native.yaml"
-# Download ipAddressPool and configure using lbrange above
+info "Waiting for the MetalLB controller"
-curl -sO https://raw.githubusercontent.com/JamesTurland/JimsGarage/main/Kubernetes/K3S-Deploy/ipAddressPool
+# rollout status waits on the Deployment (created synchronously by apply) and
-cat ipAddressPool | sed 's/$lbrange/'$lbrange'/g' > $HOME/ipAddressPool.yaml
+# avoids the "no matching resources found" race that `wait --for=condition=ready
-kubectl apply -f $HOME/ipAddressPool.yaml
+# pod` hits when the controller pod does not exist yet.
 kubectl -n metallb-system rollout status deploy/controller --timeout=120s
 info "Configuring address pool ($lbrange)"
 curl -sfL "$manifest_base/ipAddressPool" -o "$HOME/ipAddressPool.src"
 sed "s|REPLACE_LBRANGE|$lbrange|g" "$HOME/ipAddressPool.src" > "$HOME/ipAddressPool.yaml"
 kubectl_apply_retry "$HOME/ipAddressPool.yaml"
 curl -sfL "$manifest_base/l2Advertisement.yaml" -o "$HOME/l2Advertisement.yaml"
 kubectl_apply_retry "$HOME/l2Advertisement.yaml"
 ok "MetalLB configured"
-# Step 9: Test with Nginx
+# ── Step 9/9 · Verify cluster and LoadBalancer ─────────────────────────
-kubectl apply -f https://raw.githubusercontent.com/inlets/inlets-operator/master/contrib/nginx-sample-deployment.yaml -n default
+step "Step 9/9 · Verify cluster and LoadBalancer"
-kubectl expose deployment nginx-1 --port=80 --type=LoadBalancer -n default
+info "Waiting for the cluster to be ready"
-
+k3sup_quiet ready --context "$context" --kubeconfig "$HOME/.kube/config"
-echo -e " \033[32;5mWaiting for K3S to sync and LoadBalancer to come online\033[0m"
+info "Deploying nginx sample"
-
+kubectl apply -n default -f https://raw.githubusercontent.com/inlets/inlets-operator/master/contrib/nginx-sample-deployment.yaml
-while [[ $(kubectl get pods -l app=nginx -o 'jsonpath={..status.conditions[?(@.type=="Ready")].status}') != "True" ]]; do
+# `|| true` — expose is imperative and errors with AlreadyExists on a re-run,
-   sleep 1
+# which would abort the script under set -e.
 kubectl expose deployment nginx-1 --port=80 --type=LoadBalancer -n default 2>/dev/null || true
 info "Waiting for the nginx deployment"
 kubectl -n default rollout status deploy/nginx-1 --timeout=120s
 info "Waiting for the LoadBalancer IP"
 lb_ip=""
 for _ in $(seq 1 30); do
  lb_ip="$(kubectl get svc nginx-1 -n default -o jsonpath='{.status.loadBalancer.ingress[0].ip}' 2>/dev/null || true)"
  [[ -n "$lb_ip" ]] && break
  sleep 2
 done
-# Step 10: Deploy IP Pools and l2Advertisement
+kubectl get nodes -o wide || true
-kubectl wait --namespace metallb-system \
+step "Done · cluster ready"
-                --for=condition=ready pod \
+ready_nodes="$(kubectl get nodes --no-headers 2>/dev/null | awk '$2=="Ready"{c++} END{print c+0}')" || ready_nodes=0
-                --selector=component=controller \
+ok "$ready_nodes nodes Ready · context '$context'"
-                --timeout=120s
+ok "API server: https://$vip:6443"
-kubectl apply -f ipAddressPool.yaml
+if [[ -n "$lb_ip" ]]; then
-kubectl apply -f https://raw.githubusercontent.com/JamesTurland/JimsGarage/main/Kubernetes/K3S-Deploy/l2Advertisement.yaml
+  ok "nginx LoadBalancer IP: $lb_ip  (curl http://$lb_ip)"
-
+else
-kubectl get nodes
+  warn "nginx LoadBalancer IP not assigned yet — check: kubectl get svc -A"
-kubectl get svc
+fi
-kubectl get pods --all-namespaces -o wide
+info "Next: kubectl get pods -A"
 echo -e " \033[32;5mHappy Kubing! Access Nginx at EXTERNAL-IP above\033[0m"
@@ -1,10 +1,9 @@
 apiVersion: apps/v1
 kind: DaemonSet
 metadata:
  creationTimestamp: null
  labels:
    app.kubernetes.io/name: kube-vip-ds
-    app.kubernetes.io/version: v0.8.2
+    app.kubernetes.io/version: v1.2.1
  name: kube-vip-ds
  namespace: kube-system
 spec:
@@ -13,10 +12,9 @@ spec:
      app.kubernetes.io/name: kube-vip-ds
  template:
    metadata:
      creationTimestamp: null
      labels:
        app.kubernetes.io/name: kube-vip-ds
-        app.kubernetes.io/version: v0.8.2
+        app.kubernetes.io/version: v1.2.1
    spec:
      affinity:
        nodeAffinity:
@@ -36,36 +34,38 @@ spec:
          value: "true"
        - name: port
          value: "6443"
        - name: vip_nodename
          valueFrom:
            fieldRef:
              fieldPath: spec.nodeName
        - name: vip_interface
-          value: $interface
+          value: REPLACE_INTERFACE
-        - name: vip_cidr
+        - name: vip_subnet
          value: "32"
        - name: dns_mode
          value: first
        - name: dhcp_mode
          value: ipv4
        - name: cp_enable
          value: "true"
        - name: cp_namespace
          value: kube-system
        - name: vip_ddns
          value: "false"
        - name: svc_enable
          value: "false"
        - name: svc_leasename
          value: plndr-svcs-lock
        - name: vip_leaderelection
          value: "true"
        - name: vip_leasename
          value: plndr-cp-lock
        - name: vip_leaseduration
-          value: "5"
+          value: "15"
        - name: vip_renewdeadline
-          value: "3"
+          value: "10"
        - name: vip_retryperiod
-          value: "1"
+          value: "2"
        - name: address
-          value: $vip
+          value: REPLACE_VIP
        - name: prometheus_server
          value: :2112
-        image: ghcr.io/kube-vip/kube-vip:v0.8.2
+        image: ghcr.io/kube-vip/kube-vip:v1.2.1
-        imagePullPolicy: Always
+        imagePullPolicy: IfNotPresent
        name: kube-vip
        resources: {}
        securityContext:
@@ -73,6 +73,8 @@ spec:
            add:
            - NET_ADMIN
            - NET_RAW
            drop:
            - ALL
      hostNetwork: true
      serviceAccountName: kube-vip
      tolerations:
@@ -81,9 +83,3 @@ spec:
      - effect: NoExecute
        operator: Exists
  updateStrategy: {}
 status:
  currentNumberScheduled: 0
  desiredNumberScheduled: 0
  numberMisscheduled: 0
  numberReady: 0
@@ -1,8 +1,8 @@
 apiVersion: metallb.io/v1beta1
 kind: L2Advertisement
 metadata:
-  name: example
+  name: l2-first-pool
  namespace: metallb-system
 spec:
  ipAddressPools:
-  - first-pool
+  - first-pool
@@ -1,5 +0,0 @@
 1. Snapshot your VMs!
 1. Add certificates to your home directory
 1. copy script to home directory and make executable
 1. Run the script, grab a coffee and enjoy :) (hopefully!)