Ollama 🤖🧠

Welcome to Ollama, your friendly neighborhood local LLM service! Say goodbye to cloud dependency and hello to AI-driven tasks right at home. This repository is your trusty guide for deploying Ollama and its WebUI on Synology NAS and macOS using Docker Compose. Let's dive in!

---

Overview 📝

This repository is inspired by the genius of Lixandru Marius Bogdan. We've crafted a one-stop-shop for deploying Ollama, making your setup as easy as pie (or should we say, as easy as a well-trained AI?). Whether you're on Synology NAS (via DSM Container Manager or Portainer) or macOS, we've got you covered!

Performance Notice ⚠️

Heads up! While Ollama plays nice in Docker on Synology NAS and macOS, performance can be a bit sluggish. Synology NAS might feel like it's running on a treadmill, and macOS might give your CPU a workout without tapping into that sweet GPU power. For the best experience on macOS, install Ollama natively and use this Docker Compose setup just for the WebUI. Check out Tips for Running Ollama Natively on macOS for the scoop on setting up a native Ollama service. And if you want to connect the WebUI to your native Ollama, see Connecting WebUI to Native Ollama.

Features

• 🧠 Local AI Model Serving - Run LLM models locally, no cloud strings attached.
• 🌐 Web UI for Easy Interaction - Chat with your models through the Ollama WebUI.
• 🏠 Synology & macOS Compatible - Seamless operation across platforms.
• ⚡ Lightweight & Efficient - Minimal resource hogging with optimized performance.

For setup instructions, hop on over to:

• 📖 Ollama Setup Guide

---

Configuring IPAM and Network Firewall 🌍

Let's talk Docker IPAM (IP Address Management). It's the superhero of container networking! To keep your containers chatting smoothly, you might need to tweak your firewall settings.

IPAM Configuration

Time to get your hands dirty! Configure the following in your .env file:

# Define the subnet range for the network
COMPOSE_NETWORK_SUBNET="${COMPOSE_NETWORK_SUBNET:-172.24.0.0/16}"

# Define the IP range for containers
COMPOSE_NETWORK_IP_RANGE="${COMPOSE_NETWORK_IP_RANGE:-172.24.5.0/24}"

# Define the network gateway
COMPOSE_NETWORK_GATEWAY="${COMPOSE_NETWORK_GATEWAY:-172.24.5.254}"

Updating Firewall Settings on Synology NAS 🔥

Ready to unleash the power of communication for your Docker network? Update your Synology Firewall settings:

1. Open Control Panel → Security (under Connectivity).
2. Navigate to the Firewall tab → Click Edit Rules.
3. Click Create to add a new rule:
   • Ports: Select All
   • Source IP: Select Specific IP
   • Click Select → Choose Subnet
   • Enter 172.24.0.0 for IP Address and 255.255.0.0 for Subnet mask/Prefix length
   • Action: Select Allow
4. Click OK to apply the changes.

Now your containers can chat freely! 🎉

For more details, check out the Docker Compose IPAM documentation.

---

Deployment 🚀

Let's get Ollama up and running! This guide is your roadmap to deploying Ollama with Docker Compose. You can do it on:

• 🏠 Synology NAS (DSM Container Manager or Portainer)
• 💻 macOS (Docker Desktop)

Using the Makefile 🛠️

We've included a Makefile to make your life easier. Here's what you can do:

Command	Description
make up	Fire up the Ollama service stack.
make down	Stop and remove containers and volumes.
make logs	Peek at real-time logs of the running containers.
make open	Launch the Ollama WebUI in your favorite web browser.
make run	Start the stack, open the WebUI, and show logs.
make clean	Just like make down, but with flair.
make help	Get a rundown of available commands.

1. Copy and Edit the Environment File 📜

We've got a sample example.env file for you! Copy it and tweak it to your liking:

cp example.env .env
vim .env  # Edit as needed

2. Deploy Using Makefile

Ready, set, deploy! Run this command to bring up the service stack:

make up

3. Access the Web Interface 🌍

Once deployed, open the Ollama WebUI with:

make open

Or go the manual route:

• 🌐 Ollama WebUI: http://localhost:8271 (on macOS)
• 📡 Ollama WebUI: http://<synology-ip>:8271 (on Synology NAS)

To see real-time logs, just run:

make logs

To stop and remove the stack, do this:

make down

---

Troubleshooting 🛠️

Ollama WebUI not connecting to Ollama? 🤖

• Double-check the OLLAMA_BASE_URL in your .env file.

• Is the Ollama container running? Check with:

  docker ps | grep ollama

• Need a restart? Just run:

  docker restart ollama

Can't access the web interface? 🔥

• Verify the NAS IP and port (8271) are correct.

• Check running containers:

  docker ps

• Restart the stack like a pro:

  make down
  make up

---

Docker Compose Configuration 📄

The docker-compose.yml file is where the magic happens. Tweak it as needed!

• 📄 View docker-compose.yml

---

Tips for Running Ollama Natively on macOS 💡

We've included a plist file to manage the native Ollama service as a macOS Launch Agent:

• 📄 View com.ollama.serve.plist

How it works

Launch Agents are macOS-specific services managed by launchd. The provided plist file configures macOS to run Ollama as a background service on user login or on demand.

Installation

Copy the plist file to your LaunchAgents directory:

cp config/plist/com.ollama.serve.plist ~/Library/LaunchAgents/

Commands to start/stop the service

Time to control your Ollama service:

# Start the ollama service
launchctl load ~/Library/LaunchAgents/com.ollama.serve.plist

# Stop the ollama service
launchctl unload ~/Library/LaunchAgents/com.ollama.serve.plist

Optional Aliases

Want to make your life even easier? Add these to your .bashrc, .zshrc, or preferred aliases file:

alias ollamaStart="launchctl load ~/Library/LaunchAgents/com.ollama.serve.plist"
alias ollamaStop="launchctl unload ~/Library/LaunchAgents/com.ollama.serve.plist"

Monitoring Logs

Keep an eye on the action with:

tail -F /tmp/ollama.log /tmp/ollama.err

---

Connecting WebUI to macOS Native Ollama 🔗

Want to connect the Ollama WebUI to a natively running Ollama service on macOS? Here's how:

1. Open docker-compose.yml.

2. Comment out the entire ollama service block.

3. Comment out the depends_on section under the ollama-webui service block.

4. In your .env file, set:

   OLLAMA_HOSTNAME="${OLLAMA_HOSTNAME:-localhost}"
   LLAMA_WEBUI_BASE_URL="${OLLAMA_WEBUI_BASE_URL:-http://ollama:11434}"

Now your WebUI is ready to connect to your native Ollama service!

Running the WebUI from the command line

After configuring your .env file and updating docker-compose.yml, run:

docker-compose up -d ollama-webui

To check logs for the WebUI service:

docker-compose logs -f ollama-webui

To stop the WebUI service:

docker-compose stop ollama-webui

Conclusion 🎉

You've successfully set up Ollama for local AI model deployment! 🚀 Now you're ready to unleash the power of local AI without the cloud's shackles.

For advanced configurations, don't forget to visit the Ollama documentation. Happy AI-ing!