VisualStudio Code server images based on https://github.com/cdr/code-server

• Hosted on GitHub Container Registry (ghcr.io) to avoid DockerHub pull limitations, and easily deploy on clusters (such as Kubernetes).
• Additionally installed on the CPU image: Python3, NodeJS (npm, yarn), Java JDK 17s, PHP, Fortran

Automatically updated

The image on ghcr.io is automatically updated every week (Monday at 3:00 GMT+1) by a GitHub Actions workflow to match the latest tag of codercom/code-server

This image extends the Dockerfile defined at https://github.com/cdr/code-server

Code server on CPU

Run

docker run --rm -it -p 8080:8080 -e PASSWORD=password -v $(pwd):/home/coder/project ghcr.io/maastrichtu-ids/code-server:latest

In the container:

• User, with sudo privileges: coder
• Workspace path: /home/coder

You can also provide the URL of a git repository to be cloned at start, if a requirements.txt, yarn.lock or package-lock.json are present, they will be automatically installed

docker run --rm -it -p 8080:8080 -e PASSWORD=password -e GIT_URL=https://github.com/MaastrichtU-IDS/play-fair ghcr.io/maastrichtu-ids/code-server:latest

Build

To build the image:

docker build -t ghcr.io/maastrichtu-ids/code-server:latest .

Push

docker push ghcr.io/maastrichtu-ids/code-server:latest

Code server on Nvidia GPU

Images hosted on the GitHub Container Registry: https://github.com/orgs/MaastrichtU-IDS/packages/container/package/code-server-gpu

Based on Docker images provided by Nvidia:

• Tensorflow: https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow
• PyTorch: https://ngc.nvidia.com/catalog/containers/nvidia:pytorch

The best way to update the images is to update the version of the environment variables TENSORFLOW_IMAGE and PYTORCH_IMAGE in the publish-docker-gpu.yml workflow, and push the changes to the main branch, the new images version will be built and published by GitHub Actions

You can also build the images locally.

Build Tensorflow:

docker build --build-arg NVIDIA_IMAGE=nvcr.io/nvidia/tensorflow:21.05-tf2-py3 -t ghcr.io/maastrichtu-ids/code-server-gpu:tensorflow-21.05-tf2-py3 -f Dockerfile.gpu .

Build PyTorch:

docker build --build-arg NVIDIA_IMAGE=nvcr.io/nvidia/pytorch:21.05-py3 -t ghcr.io/maastrichtu-ids/code-server-gpu:pytorch-21.05-py3 -f Dockerfile.gpu .

Test to run it locally:

docker run -it --rm -p 8081:8081 -e PASSWORD=password ghcr.io/maastrichtu-ids/code-server-gpu:tensorflow-21.05-tf2-py3

Push:

docker push ghcr.io/maastrichtu-ids/code-server-gpu:tensorflow-21.05-tf2-py3