ComfyUI Docker Deployment

This project provides a robust Docker and Docker Compose configuration for quickly deploying [ComfyUI]. It is pre-configured for full data persistence and NVIDIA GPU acceleration, creating a stable and productive environment.

---

Prerequisites

• Docker & Docker Compose
• An NVIDIA GPU
• The appropriate NVIDIA drivers installed for your Linux distribution
• The NVIDIA Container Toolkit

---

Project Structure

• Dockerfile: The recipe for building the ComfyUI Docker image.
• docker-compose.yml: The orchestration file that defines how to run the container, which ports to expose, and how to link directories for data persistence.
• .env: An environment file for variables used by Docker Compose.
• run.sh: A convenience script for launching the application.
• ComfyUI_data/: The most important directory. This is the folder on your host machine that serves as the permanent storage for all your ComfyUI data.
• filebrowser_data/: Stores persistent data and configuration for the Filebrowser service (database and settings).

---

Detailed Breakdown of the ComfyUI_data Contents

• models/: This is the heart of your ComfyUI setup. You will place all your model files here, such as:
  • Checkpoints: models/checkpoints/your_model.safetensors
  • LoRAs: models/loras/your_lora.safetensors
  • VAEs: models/vae/your_vae.pt
  • ControlNet Models: models/controlnet/your_controlnet.pth
  • And any other model types.
• input/: Place any images you want to use as inputs for your workflows (e.g., for img2img, ControlNet, or loading reference images) in this directory.
• output/: ComfyUI will save all your generated images, videos, and other outputs here. Your results will be safely stored on your host machine.
• custom_nodes/: A critical folder for extensibility. When you install third-party nodes (e.g., via git clone), they live here. Persisting this folder saves you from having to reinstall all your extensions every time you restart or update the container.
• comfyui_config.json: This small but important file stores your personal UI settings, such as enabled options, the last used workflow, and other configuration tweaks. Persisting it makes your user experience seamless.

---

Detailed Breakdown of the filebrowser_data Contents

• database/: Contains the Filebrowser SQLite database file (filebrowser.db) for user and file metadata.
• config/: Stores Filebrowser configuration files (such as settings.json).

---

Installation and Setup

Step 1: Install the NVIDIA Container Toolkit

(This section contains the standard procedure for enabling GPU access in Docker).

Add the NVIDIA repository (for Debian/Ubuntu):

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

Install the packages:

sudo apt-get update
export NVIDIA_CONTAINER_TOOLKIT_VERSION=1.17.8-1
  sudo apt-get install -y \
      nvidia-container-toolkit=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      nvidia-container-toolkit-base=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      libnvidia-container-tools=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      libnvidia-container1=${NVIDIA_CONTAINER_TOOLKIT_VERSION}

Configure and restart Docker:

sudo nvidia-ctk runtime configure --runtime=docker

Or

sudo tee /etc/docker/daemon.json > /dev/null <<EOF
{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}
EOF

Then restart docker

sudo systemctl restart docker

For Rootless docker container runtime mode check

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html

---

Step 2: Clone and Prepare the Project

Clone this repository:

git clone https://github.com/EugeneTasks/docker-comfyui.git
cd docker-comfyui

Set Permissions (Crucial Step!):

sudo chmod -R 777 ComfyUI_data/

sudo chmod -R 777 filebrowser_data/

Create the required host directories:

Create further structure (such as ComfyUI_data/models/checkpoints, ComfyUI_data/models/loras, etc.) as needed following the instructions of ComfyUI. In the example below you need to create a vae, text_encoders and diffusion_models folders in ComfyUI_data/models dir

---

Step 3: Launch ComfyUI

After completing the setup, launch the application using Docker Compose.

docker-compose up -d --build

• --build: This flag rebuilds the image if you've made changes to the Dockerfile.
• -d: This flag runs the container in detached mode (in the background).