Request for how to support the requirement of a user having more than one Jupyterlab pod running at the same time

We are creating multiple custom Docker images for different use cases. We have a requirement that a user can have more than one Jupyter Lab pod running at the same time.

Ideally, I would like to append a nonce or a timestamp on to the end of the generated pod name.

Would the a user to share the PVC across all of his/her running pods.

What is the best way to accomplish this? Does having multiple pods running for the same user at the same time present any type of issue?

1 Like

JupyterHub supports “Named servers”, it sounds like this will fulfil your requirements:
https://jupyterhub.readthedocs.io/en/stable/reference/config-user-env.html#named-servers

The default name for PVC includes the servername:

So each pod should have it’s own PVC. In theory if you wanted all pods to share the same PVC you could do that by changing the template to omit the servername, but your Kubernetes cluster will need to support storage volumes that allow multiple readwrite mounts.

1 Like

Thank you very much.