Lab Cluster

My home lab cluster.

Building your own

VM setup

• Setup n VMs / bare metal machines with Ubuntu 18.04. The K8s control plan + etcd nodes should have 1.5 GB memory. I've given the worker nodes 8 GB memory.
• Install docker on each VM.

wget -O - https://raw.githubusercontent.com/just1689/kubernetes-setup-rke/master/setup-node.sh | bash

• Create a user on each VM.
• Allow your new user to run docker x commands.

sudo usermod -aG docker $USER

• Generate ssh keys with ssh-key-gen
• Copy your key over to each vm with ssh-copy-id user@address
• Turn swap off on each VM by running sudo swapoff -a
• To allow for cleaning up as root, setup cert base ssh access to root to each VM. You may not need this. It's only for recovering from unsupported upgrade paths etc.

Installing Kubernetes

Modify the cluster.yml by changing the entries under nodes. Pay particular attention to the IP addresses, the user field and the roles.

Once the cluster.yml file reflects your desired cluster configuration run:

rke up

• Once it is up you will find a new file called kube_config_cluster.yml
• Overwrite your home kube config with the following command cp kube_config_cluster.yml ~/.kube/config
• You will then be able to run kubectl get nodes and see the nodes.

Installing Apps

Longhorn

kubectl apply -f apps/01-longhorn/ 

Prometheus

kubectl apply -f apps/02-prometheus/ 

Grafana

kubectl apply -f apps/03-grafana/ 

Loki

kubectl apply -f apps/04-loki/ 

Lens

Lens can show you the cluster health of the nodes.

Lens can also show the health of your workloads. You can watch this space to see the various lab apps come up.

Issues

SCSI

sudo apt install open-iscsi -y