Welcome to my repo where I maintain everything related to my homelab which adheres to Infrastructure as Code (IaC) and GitOps practices where possible. This allows me to have a single source of written truth for my homelab, declaring how and where I want it setup. I have a Kubernetes cluster that runs most of the services in my homelab but I also have a few services running as Docker containers on my NAS.

This allows me to:

• Version control my changes, allowing easy rollback of breaking patches/tinkering/etc
• Allow for easy reinstall/disaster recovery of a cluster, as everything except persistent data is defined here.
• Version control and declare hardware provisioning, ensuring repeatable and robust hardware configuration.
• This can be achieved with tools such as Terraform and Ansible for those wanting to use a more standard OS & deployment.
• With Talos and a few scripts that I have written, I can define and provision a cluster easily with as few manual steps as possible.
• Renovate and a few scripts makes it easy to handle Talos and Kubernetes updates. Renovate will create pull requests when new updates are available. When the pull requests are merged to the main branch I only need to run a commands to upgrade the cluster.

My motivation for having a homelab is that it is a great way to learn and educate myself and pick up new skills that I might have use for in at my work.

Besides that I'm also running some services that are used daily by myself, family & friends.

These are what I consider the main features of my homelab. You can also see this list as an index to the documentation. If you want to set up your own cluster and use my repo as a guide I suggest that you read through the documentation in this order:

• Automated bare metal provisioning of Kubernetes nodes
• Automated Kubernetes installation and management
• Installing and managing applications using GitOps
• Semi-automatic rolling upgrades for OS and Kubernetes
• Automatically update apps (with approval)
• Modular architecture, easy to add or remove features/components
• Automated certificate management
• Automatically update DNS records for exposed services
• Distributed storage
• Monitoring and alerting
• Automated offsite backups

Why do things manually when you can automate it? I try to automate as many aspects of my homelab as possible.

• Talos is my OS of choice on all Kubernetes nodes.
• Renovate keeps track of dependencies and creates a pull request when there is something to update.
• I then merge those pull requests in to the main branch
• Flux will then update the main cluster.
• Tell Talos to automatically update and reboot Kubernetes nodes if needed by running one script.
• Tell Talos to automatically update Kubernetes by running one script.
• Services will automatically be updated (and restarted if needed).

Updates to the Docker containers running on my NAS are handled in a similar way.

• There is a clone of the repository on the NAS
• A cronjob pulls the updates and does a docker-compose up -d --build to update the Docker containers.

I use Ansible to provision and configure other hardware in my homelab like my NAS, backup server and pikvm.

As a last resort I use Taskfile and write bash scripts to run repetitive tasks.

I have services that I've choosen to host outside of the Kubernetes cluster. I host a Plex server that I wan't to run on the more powerfull NAS so there is no point in running Plex in the cluster and pin it to a node.

There is also a NFS and a Samba server running in Docker on the NAS for easier access to file shares, backups and media.

Every now and then I run in to problems and I usually do one of two things when I fix them:

• I either create a task using Taskfile so it's easy to do the same thing over and over again
• Or I write down the solution. This is usually done when it's to complex to create a task for it.