Open OnDemand: Web-Based Access to the Illinois Campus Cluster

Open OnDemand is a powerful, browser-based interface that gives researchers, students, and instructors intuitive and seamless access to the Illinois Campus Cluster. Built with flexibility in mind, Open OnDemand enables users to manage files, run jobs, and launch interactive applications—all from a standard web browser, with no need to install or configure local client software.

Whether you’re a seasoned HPC user or just getting started, Open OnDemand simplifies your experience and shortens your time-to-science.

What You Can Do with Open OnDemand

  • Jupyter Notebooks

    Run interactive Jupyter Notebooks on ICCP compute nodes with full access to the Python kernel. The notebooks execute on cluster resources but are accessed directly from your web browser—ideal for data analysis, machine learning, or instructional environments.

  • VS Code Server

    Launch a full Visual Studio Code Server session in your browser. Code, edit, and debug projects on the ICCP, with built-in terminal access and support for extensions, Git, and remote file management.

  • File Management

    Easily upload, download, organize, and edit files through a drag-and-drop file manager. Preview CSV, text, image, and notebook files before moving them to storage or jobs.

  • Terminal Access

    Use a fully functional shell interface from within your browser. This gives you direct command-line access to cluster login nodes without the need for SSH software.

Accessing Open OnDemand

URL:

https://cc-ood.campuscluster.illinois.edu

Login:

  • Use your Illinois NetID and password.

  • Two-factor authentication (Duo) is required for security.

How It Works

Open OnDemand sits between your web browser and the ICCP backend infrastructure. When you launch applications or submit jobs through the portal:

  • Compute work is executed on actual ICCP compute nodes.

  • File actions are reflected in your home, project, or scratch storage on ICCP systems. All the heavy lifting is done by the HPC systems—you just interact with them through a secure, graphical front end.

Information Required to Launch Interactive Sessions

When launching an interactive app like Jupyter Lab or VS Code, you will be prompted to fill in several fields. These map directly to Slurm job submission parameters and are required to allocate compute resources properly.

Here’s what you need to know:

Field

What It Means

Name of account

Your ICCP project or allocation name (e.g., chem001). You can find your available accounts by running /sw/cc.users/tools/my.accounts on the login node.

Partition

The compute queue (e.g., cpu, gpu, short, long) where your job will run. Choose one appropriate to your project’s resources.

Duration of job

How long the job should run, in Slurm format: DD-HH:MM:SS (e.g., 00-02:00:00 for two hours).

Name of reservation

Only fill this out if you’ve been given a reservation for a specific event or timeframe. Otherwise, leave it blank.

Number of CPUs

Number of CPU cores to allocate for the session. A typical Jupyter session might use 2–4.

Amount of RAM

Use Slurm format, like 4096M or 10G. If left blank, 1200 MB per core is allocated by default.

Number of GPUs

Only enter a number if your job requires GPU acceleration. Leave blank for CPU-only sessions.

Email notification

Optional. Check this to receive an email when the session starts. Useful for long queue times.

Working directory

Set the default path where your session starts (e.g., $HOME, $PROJECT). If unsure, leave as is.

These values help ensure your job runs successfully and that the correct allocation is charged for compute time. |

Example Use Cases

Use Case

Tools in Open OnDemand

Data science, ML, AI Simulation or modeling

Jupyter Notebooks, VS Code, terminal

Bioinformatics workflows

Jupyter, RStudio, file manager

Course support & teaching

Browser-only access, VS Code for labs,

File conversion, plotting

command-line tools

Tips for Getting Started

  • Log in and explore: Visit the Open OnDemand portal and explore the interface. Check your available storage under the File menu.

  • Launch a Jupyter session: Try running a Jupyter notebook with Python on a single compute node using an interactive job.

  • Submit a job: Use the Job Composer to build and submit a basic batch job. You can also submit traditional Slurm scripts.

  • Open VS Code: Start a VS Code Server session to work on your code from anywhere.

  • Use the terminal: Need full control? Launch a terminal session right in your browser.

Security & Best Practices

  • Always log out when you’re done using the portal, especially on shared or public computers.

  • Don’t run long or compute-intensive tasks on the login node—use the Job Composer or Interactive Sessions instead.

  • Save your work regularly when using graphical apps or notebooks.

Support and Documentation

If you need help with Open OnDemand on the Illinois Campus Cluster, submit a help ticket: help@campuscluster.illinois.edu

Stay Updated

New features and applications are added regularly. Watch for:

  • MATLAB support (coming soon)

  • More containerized tools

  • GPU-accelerated interactive sessions

We recommend checking the Open OnDemand announcements on the ICCP Status page for updates and downtime notices.