Skip to content

LINUX-FULL-SETUP.md

Latest commit

 

History

History
203 lines (143 loc) · 4.41 KB

LINUX-FULL-SETUP.md

File metadata and controls

203 lines (143 loc) · 4.41 KB

Full Linux Setup — SSH Checklist

Scenario: You've SSH'd into the fresh Ubuntu Server from your Mac. Do these steps in order.

1. NVIDIA Drivers

sudo apt update && sudo apt install -y nvidia-driver-570
sudo reboot

Wait 30 seconds, SSH back in, then verify:

nvidia-smi

Should show RTX 5090, 32GB.

2. Docker + NVIDIA Container Toolkit

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

# NVIDIA Container Toolkit
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | \
  sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg
curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
  sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
  sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
sudo apt update && sudo apt install -y nvidia-container-toolkit
sudo nvidia-ctk runtime configure --runtime=docker
sudo systemctl restart docker

# Verify GPU in Docker
docker run --rm --gpus all nvidia/cuda:12.6.0-base-ubuntu24.04 nvidia-smi

3. Python + Git

sudo apt install -y python3.12 python3-pip python3.12-venv git curl

4. Claude Code (CLI)

# Install Node.js (required for Claude Code)
curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash -
sudo apt install -y nodejs

# Install Claude Code
npm install -g @anthropic-ai/claude-code

# Login (follow the browser auth flow — it will give you a URL to open on your Mac)
claude

5. Clone the Project

# Set up SSH key for GitHub (if not already)
ssh-keygen -t ed25519 -C "your@email.com"
cat ~/.ssh/id_ed25519.pub
# Copy the output and add it at: https://github.com/settings/ssh/new

# Clone
git clone git@github.com:JoshuaPubNub/vibercoded.git ~/vibercoded
cd ~/vibercoded

# Install Python dependencies
pip install -e ".[dev]"

6. Start vLLM

cd ~/vibercoded/gpu-machine
docker compose up -d

Wait for model to load (~2-3 minutes):

until curl -s http://localhost:8000/health > /dev/null 2>&1; do sleep 5; echo "waiting..."; done
echo "vLLM ready!"

Test it:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"Kbenkhaled/Qwen3.5-27B-NVFP4","messages":[{"role":"user","content":"Hello"}],"max_tokens":10}'

7. Start AgentCreator

cd ~/vibercoded
nohup python3 -c "from agent_creator.main import main; main()" > data/server.log 2>&1 &

Verify:

curl http://localhost:7000/api/llm/health

8. Tailscale (for team access)

curl -fsSL https://tailscale.com/install.sh | sh
sudo tailscale up

Follow the auth URL. After that, your Tailscale IP is:

tailscale ip -4

Team accesses UI at: http://YOUR_TAILSCALE_IP:7000

9. Build All Agents

cd ~/vibercoded
python3 scripts/build_all_agents.py --output data/build_report.md

10. Run Tests

cd ~/vibercoded
python3 tests/integration/test_runner.py --output data/test_report.md

Quick Reference — Daily Commands

# SSH in
ssh joshua@100.x.x.x

# Check everything is running
nvidia-smi                                    # GPU status
docker ps                                     # Running containers
curl http://localhost:8000/health              # vLLM health
curl http://localhost:7000/api/llm/health      # AgentCreator health

# Start services after reboot
cd ~/vibercoded/gpu-machine && docker compose up -d    # vLLM
cd ~/vibercoded && nohup python3 -c "from agent_creator.main import main; main()" > data/server.log 2>&1 &  # AgentCreator

# View logs
tail -f ~/vibercoded/data/server.log          # AgentCreator logs
docker logs gpu-machine-vllm-1 --tail 20      # vLLM logs

# Stop everything
pkill -f agent_creator                        # Stop AgentCreator
docker stop $(docker ps -q)                   # Stop all containers

# Run Claude Code
cd ~/vibercoded && claude

What's Different from Windows

Windows (old) Linux (new)
Inference speed 3-5 tok/s 60-80 tok/s
GPU clock lock Needed every reboot Not needed
pin_memory Disabled (WSL) Enabled (native)
Docker GPU Via WSL2 layer Direct
Access Desktop + JumpDesktop SSH + Tailscale
Agent build time (197) 4-6 hours ~45 minutes
⚡