Sensei-CHO

Setting up Cilium LoadBalancer and Nginx ingress on a k3s cluster

In a recent post I talked about setting up a k3s cluster with IPv6 enabled and High Availability. The next upgrade is making the applications we are going to deploy in our cluster accessible from the outside world !

In this write-up we'll see how to setup 2 important things in Kubernetes clusters:

• LoadBalancer: Uses an IP external to the cluster to redirect traffic to an application internally.
• Ingress-Controller: Serves as a reverse-proxy to redirect traffic to application based on rules like DNS entries.

Combining a LoadBalancer and a Ingress-Controller allows us to redirect traffic coming from outside of our cluster to internal applications using specific rules.

Sensei-CHO

Mise en place d'un système de documentation assisté par IA

Toutes les commandes sont à exécuter en root

Installation de docker

Docker permet de lancer des applications de façon simplifié, multi-plateforme et multi-version.

apt install curl

Sensei-CHO

Setting up a K3S Highly Available cluster with IPv6 in 3 minutes !

In this write-up we'll see how to deploy a k3s cluster with high availability and dual-stack IPv4/IPv6 networking.

Most use cases of Kubernetes requires HA in their environment. In my case (which is learning networking and Kubernetes with low-cost hardware) I'll need low consumption and IPv6 too !

In this write-up we will deploy a 6 nodes k3s cluster with 3 master nodes and 3 workers.

So how are we going to get around this ?

Sensei-CHO

# Définir le nombre de messages à récupérer
$nombreDeMessages = 10

# Récupérer les 10 derniers messages de journal pour le niveau critique
$messagesCritiques = Get-WinEvent -LogName "System" -FilterXPath "*[System[(Level=1)]]" -MaxEvents $nombreDeMessages

# Récupérer les 10 derniers messages de journal pour le niveau avertissement
$messagesAvertissement = Get-WinEvent -LogName "System" -FilterXPath "*[System[(Level=2)]]" -MaxEvents $nombreDeMessages

# Récupérer les 10 derniers messages de journal pour le niveau erreur

Sensei-CHO

Proxmox cluster with dstributed storage for HA with single disk nodes

Preparing nodes

Every steps listed here have to be done on each proxmoxnode

Verify installation

• verify networking interfaces and addresses
• add no-subscription repos

Sensei-CHO

# This script is a proof of concept to try MicroCloud, it is not production ready or even working properly...
echo "[+]Refreshing LXD"
echo
snap refresh lxd --channel=5.19/stable

echo "Initialising LXD"
echo
lxd init --auto

echo "[+]Creating SDN Network"

Sensei-CHO

global
    log /dev/log    local0
    log /dev/log    local1 notice
    chroot /var/lib/haproxy
    stats socket /run/haproxy_admin.sock mode 660 level admin expose-fd listeners
    stats timeout 30s
    user haproxy
    group haproxy
    daemon
    ca-base /etc/ssl/certs

Sensei-CHO

apiVersion: v1
kind: Namespace
metadata:
  name: proxmox-expose
---
kind: Service
metadata:
  name: pve-svc
  namespace: proxmox-expose
spec: