Jeff Camacho 
Back to projects
•
7 min read

Home MicroK8s Cluster Implementation

Projected End Date: TBD
To Do
Table of Contents

Home MicroK8s Cluster Implementation: Scaling My Kubernetes Lab

Setting up a MicroK8s cluster at home is another “can-I-make-this-work” project that I’m undertaking to deepen my understanding of Kubernetes and container orchestration. The goal is to design, deploy, and manage a scalable Kubernetes cluster that builds upon my existing homelab environment. This project will allow me to experiment with cluster configurations, containerized applications, and high-availability setups in a practical, hands-on way.

Currently, my homelab consists of a single worker node. The plan is to scale this setup into a multi-node cluster using repurposed hardware, specifically e-waste systems acquired at steep discounts from Goodwill San Antonio and Goodwill Online Auctions. All compute systems used in this project will be sustainable choices, keeping costs low while reducing environmental impact.

I want this to be my “BIG” personal project for the year that will allow me to integrate much of what I’ve outlined in the requirements section.

Here. We. Go.

Objective and Scope

The project’s primary objectives are to:

  • Build a scalable MicroK8s cluster that extends my existing single-node setup.
  • Use repurposed hardware to reduce e-waste and keep costs low.
  • Automate deployment pipelines for sample applications.
  • Test high availability and efficient resource utilization across the cluster.

Hardware

The cluster will utilize the following hardware:

  • HP EliteDesk 800 G4 DM: Worker node 1
    • Intel Core i7-8700 CPU
    • 64GB RAM
    • Intel UHD Graphics 630
    • Realtek and Sandisk NVMe storage
    • Intel I219-LM Ethernet
  • HP EliteDesk 800 G4 DM: Worker node 2 (to be added)
    • Intel Core i7-8700 CPU
    • 64GB RAM
    • Intel UHD Graphics 630
    • Realtek and Sandisk NVMe storage
    • Intel I219-LM Ethernet
  • Dell Optiplex 3040: Master node
    • Intel Core i5-6500 CPU @ 3.2GHz
    • 1TB HDD
      • I plan to move to an SSD drive for the boot drive.
    • 16GB RAM

As mentioned, all hardware was sourced from Goodwill under deep discounts or low price bidding, demonstrating that high-performance computing projects can be achieved with minimal environmental impact by repurposing e-waste.

Requirements to Consider

In addition to the core infrastructure setup, the applications and services deployed within the MicroK8s pods will play a key role in defining the cluster’s functionality. Since this cluster will serve as a general home server, here are some potential applications and recommendations to include in the requirements:

Development Tools

  • VS Code Server (or similar): A web-based code editor for development accessible from any device.
  • Gitea: A lightweight, self-hosted Git service for managing repositories and collaborative projects.
  • Jenkins or GitLab CI/CD: For experimenting with continuous integration and deployment pipelines.

Media Sharing and Management

  • Plex or Jellyfin: For streaming media content (movies, TV shows, music) across devices.
  • Nextcloud: For file storage and sharing, similar to a private Google Drive.
    • Note: This may be an optional requirement. Nextcloud has been achingly hard for me to work with. User error, I guess. I’ve since gotten a Google Workspace account instead.
  • PhotoPrism: For managing and browsing personal photo collections.
    • Note: Same as previously mentioned.

DevOps and Infrastructure Tools

  • Portainer: For managing containers and Kubernetes clusters via a web UI.
  • Prometheus and Grafana: For monitoring cluster health and metrics, such as resource utilization and application performance.
  • Traefik or NGINX Ingress: For routing traffic to applications and providing HTTPS support.

Network Services

  • Pi-hole: A DNS-level ad and tracker blocker.
  • Unbound DNS: To run a local recursive DNS resolver for enhanced privacy.

Utility Applications

  • Home Assistant: For smart home automation and device management.
  • Bitwarden: A self-hosted password manager for secure credential storage.

These applications can be deployed as Kubernetes pods, offering a highly modular, easily scalable setup. The selection depends on specific use cases and resource availability.

Project Phases

To ensure focus and manageable progress, the project is divided into these key phases:

Planning Phase

  • Evaluate the existing homelab environment and its scalability potential.
  • Research and acquire additional hardware to expand the cluster:
  • Utilize the existing HP EliteDesk 800 G4 DM as the first worker node.
  • Add a second HP EliteDesk 800 G4 DM as a second worker node.
  • Use the Dell Optiplex 3040 as the master node.
  • Define cluster requirements, including the applications mentioned above.
  • Develop a project timeline, budget, and checklist of required tools and configurations.

Implementation Phase

  • Add and configure additional nodes to the cluster.
  • Install MicroK8s on each node and establish communication between them.
  • Set up networking, persistent storage, and cluster monitoring.
  • Deploy a sample application to validate the expanded cluster’s functionality.

Testing and Troubleshooting Phase

  • Test cluster performance under various scenarios, including failover and load balancing.
  • Monitor metrics like CPU, memory, and storage usage to identify bottlenecks.
  • Troubleshoot and resolve any issues with node communication or application performance.

Documentation and Review Phase

  • Document the scaling process, including architecture diagrams and configuration steps.
  • Conduct a retrospective to review what worked and identify areas for improvement.
  • Generate a performance report for resource allocation and cluster stability.

Deliverables

The following deliverables will provide tangible outputs from the project:

  • A cluster architecture diagram showcasing the multi-node setup.
  • Step-by-step documentation for scaling from a single node to a multi-node cluster.
  • Screenshots or recordings of a sample application running in the cluster.
  • A report detailing resource usage, performance, and lessons learned.

Key Outcomes

By expanding my current homelab into a multi-node Kubernetes cluster, I aim to:

  • Gain a practical understanding of scaling Kubernetes environments.
  • Build expertise in managing physical and software-defined infrastructure.
  • Demonstrate project management skills through structured planning and execution.
  • Showcase the potential for repurposing e-waste in high-performance IT projects.

Closing Statements

To be clear, this is about… building. Recuding. Recycling. Reusing.

There are plenty of projects that I’ve seen kick off where the top-of-the-line equipment is used right out of the gate, no matter what the actual requirement is. This pertains to manufacturing, construction, cloud computing, hell… even cooking. If we’re savvy, and know what we can do with what we have around us, we can create a fantastic product with minimal impact on our environment, our pocket, and our patience.

This can all be summed up in one project vision statement:

By scaling up from a single worker node with repurposed hardware, I’ll develop skills that are directly transferable to real-world IT infrastructure scenarios while reducing environmental impact.

Wish me luck. 🙃

All the best.

  • Jeff