MS-S1 MAX Server Setup

Caution

100% AI-Generated - Not Verified or Tested

This entire documentation (290+ pages) was generated by Claude (Anthropic). The content has not been manually verified or tested. Commands, configurations, and procedures may contain errors or may not work as described. Always review commands before executing, especially those requiring root/sudo. Use at your own risk.

View Live Documentation

---

Comprehensive setup guide for the Minisforum MS-S1 MAX as a home server running Ubuntu, ZFS, KVM, Docker, and local LLMs.

Hardware Overview

Component	Specification
System	Minisforum MS-S1 MAX
CPU	AMD Ryzen AI Max+ 395 (16 cores / 32 threads, Zen 5)
GPU	AMD Radeon Graphics (RDNA 3.5, 40 CUs, integrated)
RAM	128GB DDR5-5600 (unified, shared between CPU/GPU)
Storage	2TB NVMe (system) + 4TB NVMe (ZFS pool)
TDP	55-120W configurable

The APU architecture enables running 70B+ parameter LLMs that wouldn't fit in discrete GPU VRAM (typically 24GB max).

Documentation Sections

Section	Description	Link
Getting Started	Hardware overview, BIOS setup, prerequisites	View
Ubuntu Server	Installation, security hardening, systemd, logging	View
ZFS Storage	Pool creation, datasets, snapshots, Docker integration	View
Bcachefs	Next-gen COW filesystem, DKMS setup, encryption, caching	View
Virtualization	KVM setup, GPU passthrough, Windows 11 VM	View
Docker	Compose, development stacks, Ollama, monitoring, Nextcloud, Plex	View
Netplan	Network configuration, bridges, bonds, VLANs	View
Firewall & Security	UFW, iptables, nftables, Docker/KVM networking	View
SSH	Client/server config, tunneling, file transfer	View
Tailscale	Mesh VPN, MagicDNS, exit nodes, Docker integration	View
Bash	Shell fundamentals, scripting, tools, chezmoi, direnv, Starship	View
AI & Local LLMs	ROCm, llama.cpp, Ollama, AI coding tools (Claude Code, Aider, Cline)	View
Operations	Monitoring, backup, recovery, secrets management	View

Key Topics

• Ubuntu Server - Installation, CIS hardening, AppArmor, Fail2ban, auditd
• ZFS - Pool creation, datasets, snapshots, Docker integration
• Bcachefs - Next-gen filesystem, encryption, compression, SSD caching
• KVM/QEMU - Virtualization, GPU passthrough, Windows 11 VM
• Docker - Compose, development stacks, Ollama stack, monitoring, container services
• Networking - Netplan, UFW, iptables/nftables, namespaces
• SSH - Keys, agent forwarding, tunneling, jump hosts
• Tailscale - WireGuard mesh VPN, ACLs, subnet routing
• Bash - Scripting, chezmoi, direnv, 1Password CLI, modern CLI tools, Starship
• Local LLMs - ROCm setup, llama.cpp, Ollama, AI coding tools (Claude Code, Aider, Cline, Continue)

Local Development

# Clone repository
git clone https://github.com/mortenoh/msai-setup.git
cd msai-setup

# Install dependencies (requires uv)
uv sync

# Start local server
uv run mkdocs serve

Then open http://localhost:8000 in your browser.

Building Static Site

uv run mkdocs build

Output goes to site/ directory.

Project Structure

docs/
  getting-started/    # Hardware, BIOS, prerequisites
  ubuntu/             # Installation, security, systemd, logging
  zfs/                # Pool creation, datasets, snapshots
  bcachefs/           # Next-gen COW filesystem documentation
  virtualization/     # KVM, GPU passthrough, Windows VM
  docker/             # Container setup, Nextcloud, Plex
  netplan/            # Network configuration
  networking/         # Firewall, UFW, Docker/KVM networking
  ssh/                # Client, server, tunneling, file transfer
  tailscale/          # Mesh VPN, features, integration
  bash/               # Fundamentals, scripting, tools, Starship
  ai/                 # GPU setup, inference engines, models
  operations/         # Monitoring, backup, recovery

License

This documentation is provided as-is for educational purposes.

Acknowledgments

• Generated entirely by Claude (Anthropic)
• Built with MkDocs Material
• Deployed via GitHub Pages