Run notebooks via a variety of Spawners
Tag a new version to trigger GitHub Actions to build & push a new image to GitHub Packages
- new users via nativeauthenticator
- dockerspawn
- cull idle notebooks
docker build . -t ghcr.io/drejom/jupyterhub:latest
docker-compose up -d
docker-compose downOr
docker volume create workbench
docker network create workbench
docker run -it --rm -p 8000:8000 \
-e DOCKER_NETWORK_NAME=workbench \
-e DOCKER_NOTEBOOK_IMAGE=jupyter/minimal-notebook \
-e DOCKER_NOTEBOOK_DIR=/home/jovyan \
-e DOCKER_SPAWN_CMD=start-singleuser.sh \
-e JUPYTERHUB_ADMIN=${USER} \
-v workbench:/data \
-v ./jupyterhub_config_testing.py:/srv/jupyterhub/jupyterhub_config_testing.py \
-v /var/run/docker.sock:/var/run/docker.sock \
--name jupyterhub \
--network workbench \
ghcr.io/drejom/jupyterhub:latest
# Remove when done
docker volume rm workbench
docker network rm workbench- fix native authenticator
- traefik proxy
- ssh spawner
- Slurm spawner
- dashboards
- launch other services via hub (eg llm apps)
- jupyter-server-proxy
- H2O LLM Studio
- Memgraph (https://medium.com/memgraph/memgraph-lab-2-5-0-is-out-232228bb6187)
- code https://github.com/seblum/jupyterhub-server-image/tree/main
- code, pluto, docserver, bookserver https://github.com/Rahuketu86/RemoteConnect
- OpenVINO for Intel GPU
- vLLM https://vllm.readthedocs.io/
- bia-bob
- jupyter-server-proxy
Medium-scale JupyterHub deployments (with Traefik)