All workloads are in the deployments folder and sorted into folders by namespace.
This repository isn't really a tutorial on how to set up a Kubernetes cluster, checkout my k3s-gitops-arm repo for more of a A-Z guide on how to setup a cluster on some Raspberry Pis.
k3s or k8s
k3s was my choice in deployment because of how easy and quick it is to get going with k3sup. I am also using the Docker CRI under k3s instead of the default containerd. This is helpful to me if I ever want to set up Continuous Integration in the future.
All my Kubernetes worker and master nodes below are running bare metal Ubuntu 18.04.3. Using a Hypervisor seemed like a bit overkill, all the devices would be running 1 VM anyways.
- 1x OdroidH2 w/ 256GB NVMe and 16GB RAM for the Kubernetes master node
- 3x NUC8i5BEH w/ 1TB NVMe and 32GB RAM for the rook-ceph/storage nodes
- 2x NUC8i7BEH w/ 500GB SSD and 64GB RAM for the worker nodes
- 5x Sonnet Thunderbolt to 10Gb SFP+ for the Intel NUC Kubernetes worker and storage nodes
- 1x Qnap 8 bay NAS w/ 12TB drives for media and some deployment volumes
Load Balancer IPs
MetalLB IP Address Range: