My own devops config using Raspberry Pi based k8s cluster including Owncloud, Website hosting, DLNA server, Torrents seeder and possibly anything you can imagine.
You need at least a Raspberry Pi with Kubernetes installed. I personnally followed this tutorial.
Create TLS certificate that we'll use for UI:
openssl req -x509 -nodes -days 3650 -newkey rsa:2048 -keyout ./tls.key -out ./tls.crt -subj "/CN=*.example.com"
Replace example.com by anything you want
First of all, you need to setup an Ingress Controller in order to handle applications load balancing. More info here. I chose to implement Traefik.
Paste these values in secret that is described in traefik-controller/tls_secret.yaml:
cat tls.crt | base64 -w0
cat tls.key | base64 -w0
Install the traefik controller:
$ kubectl apply -f ./traefik-controller/
From now on, every services deployed can follow these config patterns : 3 files -> deployment.yaml (pod specs), service.yaml (exposed service), ingress.yaml (routing and naming).
We'll use Helm in order to variabalize our config and handle services as packages. See Helm Quickstart Guide
You need to create your own config file with values specific to your needs. Every possible values are here.
helm install --name nextcloud -f /path/to/custom/values.yaml ./nextcloud
You need to create your own config file with values specific to your needs. Every possible values are here.
helm install --name plex -f /path/to/custom/values.yaml ./plex
To claim server for the first time, you need to be on the same subnet as explained here. This is why I enabled NodePort for the plex service by default in order to point to http://MASTER_NODE_IP:32400/web for the first install. Then you can point to the configurated hostname.
You need to create your own config file with values specific to your needs. Every possible values are here.
helm install --name plex -f /path/to/custom/values.yaml ./plex