Installation Guide

This guide provides detailed instructions for installing Ollama on various Linux distributions, NixOS, and using Docker containers.

System Requirements

Before installing Ollama, ensure your system meets these minimum requirements:

• CPU: 64-bit Intel/AMD (x86_64) or ARM64 processor

• RAM: 8GB minimum (16GB+ recommended for larger models)

• Storage: 10GB+ free space (varies by model size)

• Operating System: Linux (kernel 4.15+), macOS 12.0+, or Windows 10/11

• GPU (optional but recommended):

  • NVIDIA GPU with CUDA 11.4+ support

  • AMD GPU with ROCm 5.4.3+ support

  • Intel Arc GPU with OneAPI support

Linux Installation (Direct Method)

Using the Install Script (Recommended)

For most Linux distributions, the simplest installation method is using the official install script:

curl -fsSL https://ollama.com/install.sh | sh

This script automatically detects your Linux distribution and installs the appropriate package.

Manual Installation (Debian/Ubuntu)

For Debian-based distributions (Ubuntu, Debian, Linux Mint, etc.):

# Download the latest .deb package
wget https://github.com/ollama/ollama/releases/latest/download/ollama-linux-amd64.deb

# Install the package
sudo dpkg -i ollama-linux-amd64.deb

# Install any missing dependencies
sudo apt-get install -f

Manual Installation (Red Hat/Fedora)

For Red Hat-based distributions (RHEL, Fedora, CentOS, etc.):

# Download the latest .rpm package
wget https://github.com/ollama/ollama/releases/latest/download/ollama-linux-x86_64.rpm

# Install the package
sudo rpm -i ollama-linux-x86_64.rpm

Manual Installation (Binary Installation)

If packages are not available for your distribution:

# Download the latest binary
wget https://github.com/ollama/ollama/releases/latest/download/ollama-linux-amd64

# Make it executable
chmod +x ollama-linux-amd64

# Move to a directory in PATH
sudo mv ollama-linux-amd64 /usr/local/bin/ollama

NixOS Installation

Ollama is available in the Nixpkgs collection, making it easy to install on NixOS.

Using Nix Package Manager

nix-env -iA nixos.ollama

NixOS Configuration (Configuration.nix)

For a system-wide installation, add Ollama to your configuration.nix:

{ config, pkgs, ... }:

{
  # Enable Ollama service
  services.ollama = {
    enable = true;
    acceleration = "cuda"; # Options: none, cuda, rocm, or oneapi
    package = pkgs.ollama;
  };
  
  # Add ollama package to system packages
  environment.systemPackages = with pkgs; [
    ollama
  ];
}

After updating your configuration, apply the changes:

sudo nixos-rebuild switch

Using Home Manager

If you're using Home Manager:

{ config, pkgs, ... }:

{
  home.packages = with pkgs; [
    ollama
  ];
}

Docker Installation

Running Ollama in Docker provides a consistent environment across different systems.

Basic Docker Setup

Pull and run the official Ollama Docker image:

# Pull the latest Ollama image
docker pull ollama/ollama:latest

# Run Ollama container
docker run -d \
  --name ollama \
  -p 11434:11434 \
  -v ollama:/root/.ollama \
  ollama/ollama

Docker Compose Setup

Create a docker-compose.yml file:

version: '3'

services:
  ollama:
    image: ollama/ollama:latest
    container_name: ollama
    volumes:
      - ollama_data:/root/.ollama
    ports:
      - "11434:11434"
    restart: unless-stopped

volumes:
  ollama_data:

Launch with Docker Compose:

docker-compose up -d

Docker with GPU Support (NVIDIA)

To enable NVIDIA GPU support:

# Install NVIDIA Container Toolkit
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/libnvidia-container/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit

# Run Ollama with GPU support
docker run -d \
  --name ollama \
  --gpus all \
  -p 11434:11434 \
  -v ollama:/root/.ollama \
  ollama/ollama

Post-Installation Setup

After installing Ollama, perform these steps to complete the setup:

1. Start the Ollama service:

   ollama serve

2. Test the installation by running a model:

   ollama pull mistral
   ollama run mistral

3. Verify API access:

   curl http://localhost:11434/api/generate -d '{
     "model": "mistral",
     "prompt": "Hello, how are you?"
   }'

Troubleshooting

Common Issues

1. Permission Denied Errors:

   sudo chown -R $USER:$USER ~/.ollama

2. Network Connectivity Issues:

   # Verify Ollama service is running
   ps aux | grep ollama
   
   # Check if port 11434 is open
   sudo lsof -i:11434

3. GPU Not Detected:

   # Verify CUDA installation
   nvidia-smi
   
   # Check Ollama logs
   journalctl -u ollama

Next Steps

Now that you have Ollama installed, proceed to:

1. Configure Ollama for optimal performance
2. Explore available models
3. Set up GPU acceleration for faster inference

For DevOps engineers, check out DevOps Usage Examples to see how Ollama can be integrated into your workflows.