Homelab

radar23129th January 2023 at 8:58am

  • Contents
    • Introduction
    • Kubernetes Cluster
    • LXD Hosts
    • Virtualization Servers
      • Starbug
      • Reddwarf
      • Skutter01
      • Skutter02
      • Hollister
      • Holly
    • NAS
    • Homelab Services
    • Monitoring
      • Prometheus & Grafana
      • Monitorix
      • LXDWare
      • Nagios Core
      • Uptime Kuma
      • Lens

Introduction

Server Rack

I've a relatively simple homelab, set up to primarily run virtualized workloads.

I have four Raspberry Pi4-8GB's set up as the core of a Kubernetes cluster. The cluster also includes three LXD container nodes, running on the virtualization servers. This gives me a seven node multi-architecture cluster, maximizing the flexibility for running workloads of either arm64 or amd64 architecture.

There are also six "servers" and a Synology NAS. The servers are used as virtualization servers, while the NAS is used as a file server.

Kubernetes Cluster

  • Kubernetes cluster (22 cores, 44GB RAM)
    • Multi-architecture
    • Raspberry Pi4-8GB (x4)
      • 4 cores (16 cores arm64 node total)
      • 8GB RAM (32GB RAM arm64 node total)
    • amd64 LXD containers (x3) (running on the LXD cluster)
      • 2 cores (6 cores amd64 node total)
      • 4GB RAM (12GB RAM amd64 node total)

The Kubernetes cluster is the primary location for running workloads on the homelab. The core of the cluster are four Raspberry Pi4-8GB's, and this is where most of the cluster workloads are intended to be run. However, as it is multi-architecture I can also run amd64 workloads on the LXD container nodes.

Raspberry Pi4 Cluster

Homelab K8s Cluster

LXD Hosts

A number of applications are running on LXD containers and virtual-machines, usually under nested docker instances. These are applications that might not run correctly under Kubernetes, applications under test or development, or simply standalone hosts.

LXD Hosts

Virtualization Servers

There are six virtualization servers, set up to run workloads on;

  • LXD containers or VM's, or
  • Docker containers (nested within an LXD container or a VM).

Starbug

  • Dell T610
    • Dual Xeon CPU (24 cores)
    • 64GB RAM
    • Primary virtualization server

Reddwarf

  • Retired gaming PC
    • i7 CPU (8 cores)
    • 16GB RAM

Servers

There are also two mini PCs and two old laptops being used as additional virtualization servers.

Skutter01

  • Beelink GK35
    • J4105 CPU (4 cores)
    • 8GB RAM

Skutter02

  • Beelink GK35
    • J4105 CPU (4 cores)
    • 8GB RAM

Hollister

  • HP Pavilion dv6 Notebook
    • Intel i5 CPU (4 cores)
    • 6GB RAM

Holly

  • Acer Aspire 7551 Notebook
    • AMD Athlon CPU (2 cores)
    • 4GB RAM

NAS

  • Synology DS420J
  • Seagate ST3000DM NAS HDDs (x4)

The NAS houses four 3TB HDD's, which are set up in a Synology SHR configuration, providing approximately 9TB of usable space. The NAS provides file shares and media storage for the network. There is also an NFS share, which provides persistent storage to the applications running on the Kubernetes cluster.

There is a 2TB external USB HDD attached to the NAS that is used as a destination for both the NAS backups, as well as server duplicity rsync backups from across the LAN.

NAS

Homelab Services

This diagram shows the applications and services currently running on the homelab, and how they are structured.

Services

Monitoring

I use a number of tools for monitoring the applications and services running on the homelab, including Prometheus and Grafana, Monitorix, LXDWare, Nagios Core, Uptime Kuma, as well as a number of command line management utilities for Kubernetes, LXD, KVM and Docker.

Prometheus & Grafana

Prometheus and Grafana

Monitorix

Monitorix

LXDWare

LXDWare

Nagios Core

Nagios

Uptime Kuma

I'm currently in the process of getting Uptime Kuma set up to provide additional monitoring as well as limited notifications.

Uptime Kuma

Lens

I am also using Lens to provide monitoring of the health and state of the Kubernetes cluster. This is only temporary and will soon be moving Kubernetes monitoring to Prometheus and Grafana.

Lens

....................
(created: 2021-06-10, last modified: 2023-01-29 at 08:58:09)