I have been playing around with a Pangeo deployment on GCS. I am super impressed with how easy it was to setup and get using with the instructions provided on Pangeo.io. I have found the cluster easy to use and very helpful.
I’m going to share access with my colleagues but I want to include some example notebooks for them to run. Is it possible to include example notebooks in a deployment? And what is the best way to do it?
I have tried building my own image following the instructions here https://github.com/pangeo-data/pangeo-stacks#customize-images-with-the--onbuild-variants
I built the image locally using
docker build in a repo with the notebooks I wanted to include, pushed it to our google container registry and then updated the docker image in the helm charts. JupyterHub loads with the correct image, but the notebooks aren’t in there.
Good point Joe!
We used to provide pre-loaded examples on some jupyterhubs but have since stopped doing so. Our recommended practice is for users to use git to bring in example content, for example
git clone https://github.com/pangeo-data/pangeo-ocean-examples.git
bring in one of our main demo repos.
In the pangeo-verse, we like to assume everyone is comfortable using git from the command line.
But we can make this even easier by using nbgitpuller to create a button or link that automatically does this without touching the command line. Like this one I use for my Earth and Environmental Data Science class.
This is what we commonly do for classes, workshops, and tutorials.
Nice! Thanks @rabernat - great solution that save me from messing around with persistent volumes in Kubernetes.
To be clear, if you did want to add content to the home directory, it’s not difficult either. You would just edit the repo2docker-based image used to build the user’s jupyter server environment. An example configuration is here:
But in general we try to keep these as simple and small as possible.