# Ubuntu / Debian
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER
newgrp docker

# Verify
docker --version

docker run -d \
  --name ollama \
  -v ollama-data:/root/.ollama \
  -p 11434:11434 \
  --restart unless-stopped \
  ollama/ollama:latest

# Check container status
docker ps | grep ollama

# Test API responsiveness
curl http://localhost:11434/api/version

# One-line installer
curl -fsSL https://ollama.com/install.sh | sh

# Start the service
systemctl start ollama
systemctl enable ollama  # Auto-start on boot

# Verify
ollama --version

# Meta Llama 3.2 (most popular open-source model)
ollama pull llama3.2

# Mistral (excellent European open-source model)
ollama pull mistral

# Qwen 2.5 (Alibaba, strong Chinese language support)
ollama pull qwen2.5

# DeepSeek R1 (Chinese reasoning model, free and open)
ollama pull deepseek-r1:7b

# Codestral (Mistral's code-specialized model)
ollama pull codestral

# Gemma 2 (Google open-source model)
ollama pull gemma2

# Larger parameter versions (need more RAM)
ollama pull llama3.2:latest    # 3B
ollama pull llama3.1:8b        # 8B
ollama pull qwen2.5:14b        # 14B

# Smaller quantized versions (save RAM)
ollama pull llama3.2:1b        # 1B (ultra-light)
ollama pull llama3.2:0.5b      # 0.5B (mobile-grade)

# Interactive chat
ollama run llama3.2

# One-shot query
ollama run qwen2.5 "Write a quicksort algorithm in Python"

# List available models
curl http://localhost:11434/api/tags

# Streaming chat
curl http://localhost:11434/api/chat -d '{
  "model": "llama3.2",
  "messages": [{"role": "user", "content": "Hello, introduce yourself"}],
  "stream": true
}'

# Non-streaming
curl http://localhost:11434/api/generate -d '{
  "model": "qwen2.5",
  "prompt": "How to implement multithreading in Python?",
  "stream": false
}'

from openai import OpenAI

# Point to local Ollama
client = OpenAI(
    base_url="http://your-vps-ip:11434/v1/",
    api_key="not-needed"  # Ollama doesn't require an API key
)

response = client.chat.completions.create(
    model="llama3.2",
    messages=[
        {"role": "system", "content": "You are a professional programmer assistant"},
        {"role": "user", "content": "Write a REST API with FastAPI"}
    ],
    stream=True
)

for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

# Check GPU availability
nvidia-smi

# GPU-enabled Ollama (NVIDIA CUDA)
docker run -d \
  --name ollama \
  --gpus all \
  -v ollama-data:/root/.ollama \
  -p 11434:11434 \
  --restart unless-stopped \
  ollama/ollama:latest-gpu

# Check if Ollama is using GPU
ollama ps

# Expected output:
# NAME            ID              SIZE      PROCESSOR           UNTIL
# llama3.2:latest  abc123...      2.0 GB    CUDA:0 (NVIDIA ...)  Now

ollama list
# or
curl http://localhost:11434/api/tags

ollama rm llama3.1        # Remove specific model
ollama rm llama3.2:3b     # Remove specific version

# Build custom model
ollama create my-llama -f ./Modelfile

# Run custom model
ollama run my-llama "Hello"

docker run -d \
  --name open-webui \
  -v open-webui:/app/backend/data \
  -p 8080:8080 \
  --restart unless-stopped \
  ghcr.io/open-webui/open-webui:main

docker run -d \
  --name textgen \
  -p 7860:7860 \
  -v textgen-data:/data \
  --restart unless-stopped \
  ghcr.io/oobabooga/text-generation-webui:latest

server {
    listen 443 ssl;
    server_name ollama.yourdomain.com;

    ssl_certificate     /etc/nginx/ssl/cert.pem;
    ssl_certificate_key /etc/nginx/ssl/key.pem;

    location / {
        proxy_pass http://localhost:11434;
        proxy_http_version 1.1;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;

        # WebSocket support (required for Open WebUI)
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
        proxy_read_timeout 86400;
    }
}

# Only open necessary ports
sudo ufw allow 11434/tcp   # Ollama API
sudo ufw allow 8080/tcp    # Open WebUI (optional)
sudo ufw allow 443/tcp     # HTTPS
sudo ufw enable

# Real-time monitoring
htop

# Check Ollama processes
ollama ps

# Docker resource usage
docker stats ollama

# Memory and GPU
free -h
nvidia-smi  # GPU users

# Backup all models
tar czf ollama-models-backup-$(date +%Y%m%d).tar.gz \
  -C /root/.ollama .

# Or use rsync for incremental backup
rsync -avz /root/.ollama/ backup-server:/ollama-backup/ollama/

# Check disk I/O
iotop

# Ensure models are stored on SSD/NVMe
# Docker volumes default to /var/lib/docker — verify it's on SSD
df -h /var/lib/docker

# Check available memory
free -h

# Use smaller models
ollama rm llama3.1:8b
ollama pull llama3.2:3b    # Smaller model

# Or limit container resources
docker run -d --memory="8g" --cpus="2" ...

# Check Ollama service status
systemctl status ollama
# or
docker logs ollama

# Check port listening
ss -tlnp | grep 11434

# Firewall
sudo ufw status

# Verify NVIDIA drivers
nvidia-smi

# Verify Docker GPU support
docker run --rm --gpus all nvidia/cuda:12.0.0-base-ubuntu22.04 nvidia-smi

# Restart Ollama container
docker restart ollama