Tiernan's Comms Closet k3s

Day 2 of #100daysofhomelab

January 4, 2023 tiernano

Day 2 of #100daysofhomelab and more messing with Kubernetes… So far, I have built, torn down, rebuilt and torn down a second time… and now building for a third time! Techno Tims Ansible scripts for the Win! A couple of notes for today:

the script uses K3s version 1.24.8-k3s1. at some stage yesterday I tried changing this to 1.26.0-k3s1, the latest version from the K3s GitHub page… This was a bad idea. Rancher does not like this, and, well, I don’t know what I am doing, so I want to see what Rancher does…
ideally, you would have multiple master nodes, but, me being the lazy git that I am, only set up 1… but it does look like it could be changeable later on…
I have a total of 6 VMs running my K3S cluster: 3 are 4 Cores with 8GB RAM running on GodBoxV2, which is now running Proxmox. The other 3 are each running on my HP Micro Server, the Quad 2.5Gb Celeron Box and an 8th Gen Intel NUC… each is given 2 cores and 4GB RAM. That gives me a total of (roughly) 18 cores and 36GB RAM. Each VM has around 50 GB of storage and using Longhorn, I have around 250GB of space (master does not seem to contribute space). Replicas are set to 3, so not quite a full 250GB.
Why Kubernetes? Well, I have 2 VMs currently running my fleet of docker containers. I have lost count of how many i actually have. So, my plan is to use Kubernetes to move all them from those single docker boxes, and have them more distributed and more HA. This will allow me to move stuff around easier, or at least i think it will… At the very least, i get to play with new tech! 🙂

More work on the cluster is required. This blog is hosted in-house on one of the docker instances… Hopefully, at some stage, it will be moved to the K3s cluster! That would be the first major move!

Day 1 of #100daysofhomelab

January 3, 2023 tiernano

0 Comments

I have decided to start my #100daysofhomelab journey again, so today is day 1. I have been working on a K3s cluster in the house, and so far, I have to start again… going to rebuild it again tomorrow at some stage…

Lots of Links

some notes for myself:

Service Account for Dashboard

to create the Service account, create a file, ca.yml, and enter the following:

apiVersion: v1
kind: ServiceAccount
metadata:
  name: <username>
  namespace: kube-system

next, create a file called cluster-role-binding.yml with the following:

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: <username>
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: <username>
  namespace: kube-system

make sure username matches!

run the following commands:

kubectl apply -f sa.yml

kubectl apply -f cluster-role-binding.yml

kubectl -n kube-system create token <username>

Installing OpenSCSI and NFS (required for Longhorn) with Ansible

Ansible Script

---
- hosts: k3s
  become: true
  
  tasks:
  - name: Update and upgrade apt packages
    become: true
    apt:
      upgrade: yes
      update_cache: yes
      cache_valid_time: 600 
  - name: install packages
    become: true
    apt: 
      pkg:
      - nfs-common
      - open-iscsi

  - name: Make sure open-iscsi is enabled and running
    ansible.builtin.systemd:
      enabled: true
      state: started
      name: open-iscsi