Skip to main content

System Requirements

This page is a quick pre-flight checklist and FAQ for VirtueRed deployments. Use it to validate environment readiness and answer common setup questions before you start containers.

Requirements Checklist​

Core Requirements (All Deployments)​

ItemRequirementHow to Check
OSLinux host recommendeduname -a
Docker Engine25.0+docker version
Docker Compose2.24+docker compose version
Disk200+ GB freedf -h
Ports3000, 4401, 8000, 5432 availabless -lntp

GPU Requirements (Evaluator Only)​

ItemRequirementHow to Check
NVIDIA Driver535+ (CUDA 12.x compatible)nvidia-smi
NVIDIA Container Toolkit1.15+nvidia-container-toolkit --version
Docker GPU Access--gpus all worksdocker run --rm --gpus all nvidia/cuda:12.2.0-base-ubuntu22.04 nvidia-smi

Environment & Tooling Check (Pre-Flight)​

  1. Docker daemon access

    • Run: docker ps
    • Expected: command returns a list of containers (even if empty) without permission errors.

  2. Docker Compose

    • Run: docker compose version
    • Expected: version 2.24+.

  3. GPU access (Evaluator only)

    • Run: nvidia-smi
    • Expected: GPU list and driver version appear.
    • Run: docker run --rm --gpus all nvidia/cuda:12.2.0-base-ubuntu22.04 nvidia-smi
    • Expected: the container sees the GPU.

  4. Disk space

    • Run: df -h
    • Expected: 200+ GB free on the partition where Docker stores images and volumes.

  5. Required ports

    • Run: ss -lntp | grep -E ':(3000|4401|8000|5432)\\b'
    • Expected: no output (ports are free), or change ports in your compose file.

Common Questions (FAQ)​

Q: How do I check Docker daemon access?​

A: Run docker ps. If you see a permission error, add your user to the docker group and re-login:

sudo usermod -aG docker $USER
newgrp docker

Q: Docker says "Cannot connect to the Docker daemon". What should I do?​

A: Ensure the Docker service is running:

sudo systemctl status docker
sudo systemctl start docker

Q: How do I verify NVIDIA GPU support for Docker?​

A: You need both the NVIDIA driver and the NVIDIA Container Toolkit:

nvidia-smi
nvidia-container-toolkit --version
docker run --rm --gpus all nvidia/cuda:12.2.0-base-ubuntu22.04 nvidia-smi

If nvidia-smi works on the host but fails in the container, the Container Toolkit is missing or Docker is not configured with the NVIDIA runtime.

Q: Do I need GPU drivers if I am not using the Evaluator?​

A: No. The Evaluator service is the only component that requires GPU support.

Q: What if required ports are already in use?​

A: Update the port mappings in your docker-compose.yaml and any related environment variables (for example, BACKEND_PORT, FRONTEND_PORT, EVALUATOR_PORT, POSTGRES_PORT).

Q: Where should I look if containers fail to start?​

A: Check container logs and service health:

docker compose ps
docker compose logs backend
docker compose logs evaluator

Validation Before Go-Live​

  • All services show healthy in docker compose ps.
  • Backend /health endpoint returns HTTP 200.
  • Frontend loads and can reach the backend.
  • Evaluator (if enabled) responds on /health and can run a sample evaluation.