Skip to main content

Getting Started

Welcome to the Kubernerdes Enclave documentation — a self-sustaining, air-gap-ready Kubernetes cluster running Rancher Government Solutions (RGS) Carbide on a fleet of Intel NUCs.

What You'll Build

A fully operational, on-premises Kubernetes platform consisting of:

  • 4× Intel NUC nodes — one admin/bootstrap host plus a 3-node Harvester hypervisor cluster
  • Harvester HCI — open-source hyperconverged infrastructure for VMs and Kubernetes workloads
  • Rancher Manager — multi-cluster management UI deployed on a 3-node K3s cluster inside Harvester
  • Infrastructure services — ISC DHCP, BIND DNS, HAProxy load balancer, Keepalived VIP failover
  • RGS Carbide — hardened, FIPS-capable distribution layer for supply-chain-secure container images

The enclave is designed to boot from bare metal with PXE, operate without internet access after initial provisioning, and recover from single-node failures.

Architecture Overview

┌─────────────────────────────────────────────────────────────┐
│ CIDR: 10.10.12.0/22 │
│ │
│ nuc-00 (admin) nuc-01/02/03 (Harvester cluster) │
│ ┌──────────────────┐ ┌──────────────────────────────┐ │
│ │ KVM hypervisor │ │ Harvester HCI │ │
│ │ │ │ VIP: 10.10.12.100 │ │
│ │ ┌──────────────┐ │ │ ┌──────────────────────┐ │ │
│ │ │ nuc-00-01 │ │ │ │ rancher-01/02/03 │ │ │
│ │ │ DHCP + DNS │ │ │ │ K3s HA cluster │ │ │
│ │ └──────────────┘ │ │ │ VIP: 10.10.12.210 │ │ │
│ │ ┌──────────────┐ │ │ └──────────────────────┘ │ │
│ │ │ nuc-00-02 │ │ │ │ │
│ │ │ DNS secondary│ │ │ 10.10.12.101-103 │ │
│ │ └──────────────┘ │ └──────────────────────────────┘ │
│ │ ┌──────────────┐ │ │
│ │ │ nuc-00-03 │ │ │
│ │ │ HAProxy + │ │ │
│ │ │ Keepalived │ │ │
│ │ └──────────────┘ │ │
│ │ │ │
│ │ Apache + TFTP │ │
│ │ (PXE server) │ │
│ └──────────────────┘ │
└─────────────────────────────────────────────────────────────┘

Node Roles

HostIPRole
nuc-0010.10.12.10Admin host: KVM, PXE (Apache/TFTP), infra VMs
nuc-00-0110.10.12.8DHCP (ISC), DNS primary (BIND)
nuc-00-0210.10.12.9DNS secondary (BIND)
nuc-00-0310.10.12.93HAProxy, Keepalived
harvester-edge (VIP)10.10.12.100Harvester API/UI virtual IP
nuc-0110.10.12.101Harvester node 1
nuc-0210.10.12.102Harvester node 2
nuc-0310.10.12.103Harvester node 3
rancher (VIP)10.10.12.210Keepalived VIP for Rancher Manager
rancher-0110.10.12.211Rancher Manager K3s node 1 (Harvester VM)
rancher-0210.10.12.212Rancher Manager K3s node 2 (Harvester VM)
rancher-0310.10.12.213Rancher Manager K3s node 3 (Harvester VM)

Day 0/1/2 Framework

This documentation is organized around the standard operational lifecycle:

PhaseFocusWhere to Start
Day 0Design & planningDay 0 Overview
Day 1Initial deploymentDay 1 Overview
Day 2Ongoing operationsDay 2 Overview

Prerequisites

Before diving in, you should be comfortable with:

  • Linux command line (SSH, systemctl, journalctl)
  • Basic networking concepts (subnets, VLANs, DNS, DHCP)
  • YAML — for Kubernetes manifests and Harvester config
  • Kubernetes basics (pods, deployments, services)
  • KVM/libvirt virtualization concepts

Source Repository

The automation and configuration source lives at: 👉 enclave.kubernerdes.com

The repo contains Ansible playbooks, shell scripts, network configs, and Helm values files that implement everything described in this documentation.