Current section: ~[..........]000%uptime4y 002d 15:14:38Time in enterprise networking since October 2022Marbella--:--:--Local time in Marbella ES EN
HOMELAB(1)Kernel process manualHOMELAB(1)
man homelab
Homelab K3s
01Namecomm
homelab — Infrastructure · DevOps · Self-hosting
02Synopsiscmdline
$ homelab --status=running --year=2026
A high-availability cluster at home.
PID
04
Status
Running
Year
2026
Role
Design, build & operations
no public link
fig.04k3s.topology
x000 y000
Topology: Proxmox › load balancer › 3 K3s control-plane nodes › 4 workers › MetalLB, NFS NAS and services.
03Descriptionreadme
My lab: Proxmox as hypervisor, a K3s cluster with three control-plane nodes and four workers, MetalLB load balancing, NFS storage on a NAS and self-hosted services with hardware transcoding.
04Problemissue
At work I run networks that can’t go down. At home I wanted to understand everything above them: virtualisation, orchestration, shared storage and how a service behaves when a node disappears.
05How it workstask
01
HA K3s control plane: 3 masters behind a load balancer.
02
4 workers with a MetalLB IP pool to expose services.
03
An OpenMediaVault NAS exporting NFS volumes to the cluster.
04
Self-hosted Vaultwarden and Jellyfin; Intel Quick Sync / VA-API transcoding on an N100.
05
Virtual networks, bridges and segmented subnets on Proxmox.
06Architecturemaps
Architecture · Homelab K3s
No.
Layer
Technology
L0
Hypervisor
Proxmox VE · bridge vmbr0
L1
Control plane
K3s · 3 masters · LB
L2
Compute
K3s · 4 workers
L3
Network
MetalLB · segmented subnets
L4
Storage
OpenMediaVault · NFS
L5
Services
Vaultwarden · Jellyfin
07Stackenviron
Proxmox
K3s
Kubernetes
MetalLB
Docker
NFS
Linux
08Statusstatus
Running
Running and constantly evolving. It’s where I test before anything goes to production.