Skip to main content

Package and deploy a worker image

Learn how to package a handler function into a Docker image and deploy it as a Serverless worker on RunPod.

Requirements​

To deploy a worker image, you need:

Project organization​

Organize your project files in a directory structure like this:

project_directory
├── Dockerfile              # Instructions for building the Docker image
├── src
│   └── handler.py          # Your handler function
└── builder
    └── requirements.txt    # Dependencies required by your handler

Your requirements.txt file should list all Python packages your handler needs:

# Example requirements.txt
runpod==1.7.0
torch==2.0.1
pillow==9.5.0
transformers==4.30.2

Creating a Dockerfile​

The Dockerfile tells Docker how to build your worker image. Create a file named Dockerfile (no extension) in your project's root directory:

FROM python:3.11.1-slim

WORKDIR /

# Copy and install requirements
COPY builder/requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy your handler code
COPY src/handler.py .

# Command to run when the container starts
CMD ["python", "-u", "/handler.py"]

This Dockerfile starts with a Python base image, installs your dependencies, copies your handler code, and specifies the command to run when the container starts.

Including models and external files​

If your handler uses machine learning models or other external files, include them in the Docker image:

FROM python:3.11.1-slim

WORKDIR /

# Copy and install requirements
COPY builder/requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy your code and model files
COPY src/handler.py .
COPY models/ /models/

# Set environment variables if needed
ENV MODEL_PATH=/models/my_model.pt

# Command to run when the container starts
CMD ["python", "-u", "/handler.py"]

Always include your model files directly in the Docker image rather than downloading them at runtime to ensure faster startup times and more reliable execution.

Building the Docker image​

From your terminal, navigate to your project directory and build the Docker image:

docker build --platform linux/amd64 -t [DOCKER_USERNAME]/[WORKER_NAME]:v1.0.0 .

Replace [DOCKER_USERNAME] with your Docker Hub username, [WORKER_NAME] with a descriptive name for your worker, and v1.0.0 with an appropriate version tag.

note

The --platform linux/amd64 flag is required to ensure compatibility with RunPod's infrastructure.

Testing your image locally​

Before pushing it to the registry, you should test your Docker image locally:

docker run -it [DOCKER_USERNAME]/[WORKER_NAME]:v1.0.0

If your handler is properly configured with a test input, you should see it process the test input and provide output.

Pushing the image to Docker Hub​

Make your image available to RunPod by pushing it to Docker Hub:

# Log in to Docker Hub
docker login

# Push the image
docker push [DOCKER_USERNAME]/[WORKER_NAME]:v1.0.0

Once your image is in the Docker container registry, you can create a Serverless endpoint through the RunPod console.

Image versioning​

For production workloads, use SHA tags for absolute reproducibility:

# Get the SHA after pushing
docker inspect --format='{{index .RepoDigests 0}}' [DOCKER_USERNAME]/[WORKER_NAME]:v1.0.0

# Use the SHA when deploying
# [DOCKER_USERNAME]/[WORKER_NAME]:v1.0.0@sha256:4d3d4b3c5a5c2b3a5a5c3b2a5a4d2b3a2b3c5a3b2a5d2b3a3b4c3d3b5c3d4a3

Versioning best practices:

  • Never rely on the :latest tag for production.
  • Use semantic versioning AND SHA tags for clarity and reproducibility.
  • Document the specific image SHA in your deployment documentation.
  • Keep images as small as possible for faster startup times.

Troubleshooting deployment issues​

If your worker fails to start or process requests:

  1. Check the logs in the RunPod console for error messages.
  2. Verify your handler function works correctly in local testing.
  3. Ensure all dependencies are properly installed in the Docker image.
  4. Check that your Docker image is compatible with the selected GPU type.
  5. Verify your input format matches what your handler expects.

Next steps​

After successfully deploying your worker, you can: