JupyterHub at CCS is moving to MCC Open OnDemand

CCS has transitioned from the legacy JupyterHub service. JupyterLab sessions are now provided through Open OnDemand (OOD) on the Morgan Compute Cluster (MCC) as interactive SLURM jobs.

This page explains how to start a Jupyter session and manage custom Python environments.

Start a JupyterLab Session

1. Open the MCC Open OnDemand portal:
   https://mcc-ood.ccs.uky.edu

2. Log in using CI Logon credentials.

3. From the main dashboard, select:
   Interactive Apps → JupyterLab (Interactive)

4. Complete the job submission form:

   • Select your SLURM account (example: coa_PIlinkblue_uksr)

   • Adjust resource settings if needed

5. Click Launch.

6. When the job starts, select Connect to Jupyter.

Default Resource Settings

• 1 CPU core

• 4 GB memory

• Up to 180 days wall time

Create a Custom Conda Environment and Kernel

You can create a Python environment within your Jupyter session.

Step 1 — Open a Terminal in JupyterLab

Select:

File → New → Terminal

Step 2 — Create a Conda Environment

Example:

Step 3 — Activate the Environment

Step 4 — Install Additional Packages

Example:

Step 5 — Register the Environment as a Jupyter Kernel

Step 6 — Select the Kernel in JupyterLab

Use:

Kernel → Change Kernel

Select Python (CustomEnv1).

Storage Recommendations

Custom environments are persistent when stored in:

• $HOME (default location)

• Project storage space

Avoid creating environments in scratch directories, as these may be cleared.

For more information, see the File System Basics guide.

Remove a Custom Kernel

List Installed Kernels

Remove a Kernel (Example)

Remove a Conda Environment

List Environments

or

Remove Environment (Example)

Get Help

If you encounter issues, submit a support request through the CCS Service Desk.