home_setup/README.md

Table of Contents

• Older patch notes/version
• TLDR Changelog
• Devices
• List of current devices:
• Standalone
• Kluster
• Which services I'm hosting
• Home Network
• Discord Bots
• Public DNS
• CRDs
• Observability
• CI/CD
• Git servers
• Media related
• Downsides of my current setup

Older patch notes/version

Select different tags.

TLDR Changelog

• Replaced the old standalone Docker/NFS server for a Proxmox/NFS instance.

• Added 2 VMs as worker nodes to the cluster, they will be used/are intended for x64 bit images.

• One of the new added worker VMs receives a GPU through Proxmox PCI pass through.

• Some services might have been removed or added.

Devices

List of current devices:

Standalone

Gateway: 192.168.1.1
Proxmox/NFS: somwhere.

Kluster

Kubernetes Cluster

• Pi 4 with 4GB running as a Master. (Masterk/Pi4)

• A pair of Orange PI 5, so far all of them are the 8GB of RAM version. (Slave01-2)

• Proxmox VMs, both with 3 CPU cores and 8GB of RAM (Slave03-4)

• Slave04 contains a GPU through Proxmox CPU pass through.

Masterk: 192.168.1.9
Slave01: 192.168.1.10
Slave02: 192.168.1.11
Slave03: 192.168.1.12
Slave04: 192.168.1.13

Node                 Available(GPUs)  Used(GPUs)
pi4.filter.home      0                0
slave01.filter.home  0                0
slave02.filter.home  0                0
slave03.filter.home  0                0
slave04.filter.home  1                0

Which services I'm hosting

Home Network

• CoreDNS
• DHCPd

Discord Bots

• https://gitea.filterhome.xyz/ofilter/Steam_Invite_Discord (both Master and Dev branches)
• Shlink + ShlinkUI (deployed as it has functionality with the Steam Discord Bot from above)

Public DNS

• Godaddy
• Duckdns

CRDs

• Cert Manager
• Istio
• Nvidia Gpu Operator
• NFS Volume Provisioner
• MetalLB

Observability

• Grafana
• Prometheus
• Kiali
• Jaeger

CI/CD

• Jenkins master + dynamic agent(s)
• Docker Registry
• Skaffold (Client/User side, not running on the Kubernetes cluster, yet relies on it to create multiarch docker images)

Git servers

• Gitea

Media related

• Tube
• Firebrowser
• Filebrowser
• Jellyfin
• qBitTorrent

Downsides of my current setup

• Only 1 Kubernetes master node, therefore no full High Availability
• Only 1 NFS server / no HA NFS server, therefore if the NFS server is down most of the services on the Kubernetes cluster will also be down as they depend on such NFS