How can I run Jupyter Notebooks on my institution's HPC cluster?



I want to use Jupyter notebook to run an analysis which needs access to data stored on the cluster. How can I launch a notebook server on the cluster?

CURATOR: Raminder Singh


I recent document this for Harvard users but its very generic. This can help to run Jupyter server remotely on compute nodes of the cluster and connect to your computational notebook using a local browser of your workstation/laptop.


On the Shared Computing Cluster (SCC) at Boston University, the steps to run Jupyter Notebooks are outlined on the following webpage:


The University of Arizona uses Ohio Supercomputing’s OnDemand software to provide Jupyter notebooks and remote desktops. It’s free, and easily installed and configured by an experienced admin. Run’s on CentOS/RHEL 7 and perhaps other Linux variants.

Intro is at

Install guide and other info is at


A lot of clusters have web interfaces these days, but sometimes you just need basic ssh forwarding and wrappers to help users do it! To help with this, we have the forward tool at Stanford:

And several “tiny tutorials” in the README to help you get started. The ssh configurations are managed in the “ssh” folder and the batch submission jobs in the “batches” folder, so if you have a new cluster or job that you’ like to add, just open an issue and I’d be happy to do this for you (and write another mini tutorial!).


At Iowa State University, we have Jupyter installed as a module (via spack), and instructions to login via port-forwarding over an ssh tunnel are here:

Note, a user on nearly any system could use to install jupyter in their own directory and follow these same basic steps, wthether it’s been provided by the institution or not