Home - ajgillis04/GillisDockerDepot GitHub Wiki

Docker Home Wiki

Introduction

Welcome to the GillisDockerDepot Wiki. The repository is a modular, multi-server framework for running Docker services across several machines: a NAS, Windows desktops running Docker in WSL 2, a Linux workstation and a Raspberry Pi.

This wiki covers the environment setup, how the repository is laid out, and a setup guide for each service. Guides for services that are no longer in the main stack are kept, because they may still help someone.

About the examples: addresses, domains and names in these guides are placeholders. 192.168.1.x is a home LAN, example.com is your domain, and server1, server2 and so on are your machines. Swap in your own values.


Navigation


Repository Architecture

GillisDockerDepot uses a modular template model. Each service has its own compose file, and each server's master compose file lists (include:) the services that run on it.

GillisDockerDepot/
├── appdata/                         # Application data (git-ignored)
├── compose/
│   ├── server1/ … server8/          # The services each server runs
│   ├── templates/                   # Reference templates for every service
│   ├── php/                         # Shared PHP settings mounted into PHP containers
│   ├── nextcloud/                   # Extra Nextcloud config (HSTS header)
│   └── chowdown-redirect/           # Small nginx config that redirects an old address
├── scripts/                         # Helper scripts and example app configs
├── secrets/                         # Passwords and keys as files (git-ignored)
├── wiki/                            # Screenshots used by this wiki
├── .env                             # Ports, paths and per-server values (git-ignored)
├── .env.example                     # Every variable the compose files use, without values
└── docker-compose-server1.yaml      # Master compose file for server 1 (one per server)

How to Deploy a New Service

  1. Find the template in compose/templates/ (for example compose/templates/chrome.yaml).
  2. Copy it into the server's folder:
    cp compose/templates/chrome.yaml compose/server1/chrome.yaml
  3. Adjust ports, paths and environment values in that file and in .env.
  4. Add it to the server's master file, under include::
    include:
      - ${COMPOSE_DIR}/chrome.yaml
  5. Start it:
    docker compose -p mediaserver -f docker-compose-server1.yaml up -d chrome

To stop using a service, comment out its include: line rather than deleting the file, and remove the container with docker rm -f <name>.


How the Servers Fit Together

Server Role in this setup Main services
server1 Main NAS: media, files, home automation, websites Plex, *arr apps, Nextcloud, Home Assistant, Frigate, Vaultwarden, Mealie, Pi-hole (primary), Cloudflare tunnel
server2 Windows desktop with an NVIDIA GPU (Docker in WSL 2) Ollama, Open WebUI, ComfyUI, SearXNG, Kometa, Tdarr node; also the home WireGuard VPN server
server3 Raspberry Pi Pi-hole (secondary DNS), Nebula-Sync
server4, server7 Family members' servers Media stacks similar to server1
server5 Remote cabin PC Its own VPN, Guacamole, Homepage, Cloudflare tunnel
server6 Laptop Tdarr node
server8 Linux workstation Second Ollama node, Speaches (voice), Tdarr node, Vaultwarden standby

Remote access goes through a Cloudflared Tunnel Setup protected by Cloudflare Access: Zero Trust, with WireGuard Setup Guide as the private way in.

Back to Top


Docker Environment Setup

Linux Setup (Recommended)

  1. Install Docker:
    sudo apt update
    sudo apt install -y docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
  2. Check it works, and let your user run Docker without sudo:
    docker --version
    sudo usermod -aG docker $USER   # log out and back in afterwards
    docker run hello-world

QNAP Setup

  1. Install Container Station from the App Center.
  2. Copy the repository to a share, for example /share/Docker/GillisDockerDepot, with your .env.
  3. Run compose over SSH. QNAP doesn't put Docker on the default path, so add it first:
    export PATH=$PATH:/share/CACHEDEV1_DATA/.qpkg/container-station/bin
    docker compose -p mediaserver -f docker-compose-server1.yaml up -d

Windows Setup (WSL 2)

  1. Install WSL 2 and Ubuntu (PowerShell as Administrator):
    wsl --install
    wsl --set-default-version 2
    wsl --install -d Ubuntu
    Docker Setup Windows WSL
  2. Install Docker Desktop, enable Use the WSL 2 based engine, and turn on WSL integration for Ubuntu. Alternatively, install Docker Engine directly inside Ubuntu. WSL Integration
  3. Keep heavy data inside WSL. Folders on Windows drives (/mnt/c, /mnt/d) are slow from WSL. Put model files and databases in Docker volumes, which live on the WSL disk.
  4. Ports from other machines: WSL 2 has its own internal address, so other machines on your LAN can't reach its ports directly. A Windows port proxy (netsh interface portproxy) forwards them; the WSL address can change after a reboot, so update the rule when it does.

Back to Top


Service Configuration Guides

Media, Libraries & Automation

Files, Home & Everyday Apps

Local AI

Networking, Access & Remote Tools

Infrastructure & Maintenance

Other Guides (not in the current stack, kept for reference)

Back to Top


Keeping Things Updated

  • Images: Watchtower Setup Guide updates most containers once a month (06:00 on the 1st). Each container opts in or out with the label com.centurylinklabs.watchtower.enable.
  • Compose changes: Repo Updater Setup Guide runs on every server that receives changes. On the 1st of the month (09:00) it pulls the repository, works out which containers the new commits change, and emails a list with an Upgrade button. Its page also has Pull from GitHub now, so you don't have to wait after pushing a change. Only the changed containers are recreated, and the ones that could lock you out (the tunnel, Guacamole) are left for you to do by hand.
  • Some containers are deliberately pinned to a version and updated by hand: databases (MySQL, MariaDB), Gluetun and the apps sharing its network, Nextcloud's major versions, and anything whose upgrade needs a migration step. Their compose files say why.
  • Make changes on your desktop, push them, and let the servers pull. Don't edit tracked files on a server; the next pull would refuse to run.

Back to Top


Essential Management Commands

  • Check the compose files are valid:
    docker compose -p mediaserver -f docker-compose-server1.yaml config -q
  • Start or update one service (and nothing else):
    docker compose -p mediaserver -f docker-compose-server1.yaml up -d --no-deps <service_name>
  • Force recreate a container:
    docker compose -p mediaserver -f docker-compose-server1.yaml up -d --force-recreate --no-deps <service_name>
  • Follow a container's logs:
    docker logs -f <container_name>
  • See what would change before an up: compare a running container's com.docker.compose.config-hash label with docker compose ... config --hash="*". Treat it as a hint only: containers that Watchtower has updated can show as different, and up may then recreate them. Pass up only the services you mean to change.
  • Remove unused images (careful: --volumes also deletes unused volumes):
    docker image prune -a

Back to Top

⚠️ **GitHub.com Fallback** ⚠️