Author: admin

  • Building an Amazon EKS Anywhere Homelab on Hyper-V (RHEL 10) – Part 1: Planning the Environment

    Kubernetes has become the de facto standard for running containerized applications, but learning it effectively often requires more than a single-node cluster or a managed cloud service.

    I wanted a platform that would let me experiment with production-like deployments, GitOps, networking, storage, upgrades, and disaster recovery—all from my own hardware.

    This series documents the complete process of building an Amazon EKS Anywhere cluster on Microsoft Hyper-V, using Red Hat Enterprise Linux 10 virtual machines and a dedicated management host.

    By the end of the series, we’ll have a fully functional Kubernetes platform capable of hosting real workloads while following many of the same operational practices used in production environments.


    Why Amazon EKS Anywhere?

    Amazon EKS Anywhere extends the Amazon EKS experience beyond AWS, allowing you to deploy Kubernetes clusters on your own infrastructure while benefiting from Amazon’s tooling and lifecycle management.

    For a homelab, it offers several advantages:

    • Kubernetes conformant distribution
    • Enterprise-grade lifecycle management
    • Consistent deployment experience
    • Excellent documentation
    • GitOps-friendly architecture
    • Ideal platform for learning production operations

    Rather than creating a Kubernetes cluster manually with kubeadm, EKS Anywhere automates much of the provisioning and ongoing cluster management, making it an excellent learning platform.


    The Goal

    The goal isn’t simply to install Kubernetes.

    The objective is to build an environment that resembles a small production platform.

    Throughout this series we’ll deploy:

    • Amazon EKS Anywhere
    • Harbor private container registry
    • Argo CD
    • Traefik
    • MetalLB
    • cert-manager
    • External Secrets Operator
    • WordPress
    • Nextcloud
    • Monitoring and observability tools

    Everything will be managed declaratively and stored in Git whenever possible.


    Hardware

    The entire environment runs on a single Hyper-V host with the following specifications:

    • CPU: AMD Ryzen 9 9900X (12 cores / 24 threads)
    • Memory: 96 GB DDR5 RAM
    • Storage: NVMe SSD
    • Hypervisor: Microsoft Hyper-V

    Although everything runs on one physical server, each Kubernetes node is deployed as an independent virtual machine, allowing the environment to behave much like a small on-premises cluster.


    Virtual Machines

    The initial lab consists of three virtual machines.

    RoleOperating SystemvCPUMemoryStorage
    Admin MachineRed Hat Enterprise Linux 10416 GB200 GB
    Control PlaneRed Hat Enterprise Linux 10416 GB100 GB
    Worker NodeRed Hat Enterprise Linux 1048 GB100 GB

    Admin Machine

    The Admin Machine is the central management host for the entire environment.

    It is responsible for provisioning and operating the Kubernetes cluster and hosts the following software:

    • Amazon EKS Anywhere CLI
    • kubectl
    • Helm
    • Docker
    • Git
    • Harbor Container Registry

    Hosting Harbor outside the Kubernetes cluster keeps the container registry available even while rebuilding or upgrading the cluster, which is especially useful when experimenting with cluster lifecycle operations or working in disconnected environments.

    Since it is responsible for managing the cluster, it is allocated additional memory to ensure stable operation as more services are added.


    Networking

    Networking is one of the biggest differences between a homelab and a managed cloud Kubernetes service.

    Because Hyper-V doesn’t provide cloud load balancers, we’ll build that functionality ourselves.

    The networking stack will include:

    • Hyper-V Virtual Switch
    • Static IP addresses
    • MetalLB for LoadBalancer Services
    • Traefik Ingress Controller
    • cert-manager with Let’s Encrypt

    This provides functionality similar to the networking experience available in cloud environments.


    Storage

    Persistent storage is essential for stateful workloads.

    Initially, local storage will be used for simplicity. Later in the series we’ll discuss additional storage options depending on workload requirements.

    Persistent volumes will be used for applications such as:

    • Harbor
    • WordPress
    • Nextcloud

    GitOps from Day One

    One of the primary goals of this homelab is to avoid manual configuration wherever possible.

    Instead of running endless kubectl apply commands, every deployment will eventually be managed through Git using Argo CD.

    This approach makes the environment reproducible, easier to maintain, and much closer to how modern Kubernetes platforms are operated in production.


    What’s Next?

    In Part 2 we’ll prepare the Hyper-V virtual machines by:

    • Installing Red Hat Enterprise Linux 10 on Admin Machine
    • Configuring networking
    • Setting hostnames
    • Installing required packages
    • Configuring SSH
    • Preparing the systems for Amazon EKS Anywhere

    Once the operating systems are ready, we’ll install the EKS Anywhere tooling and create our first Kubernetes cluster.

    If you’ve been looking for a practical guide to building an enterprise-style Kubernetes homelab without relying on a public cloud provider, I hope you’ll find this series useful.

    Let’s get started.

  • Hello world!

    Welcome to WordPress. This is your first post. Edit or delete it, then start writing!