Open OnDemand

Open OnDemand (OOD) is an open-source, web-based portal developed by the Ohio Supercomputer Center that provides users with a graphical interface to access and manage High-Performance Computing (HPC) resources from any device with a web browser. It simplifies interaction with complex HPC systems by offering features like:

A file management system
A job management system (including a job composer and tools for monitoring submitted jobs)
Interactive command-line shell access
Interactive applications such as Jupyter Notebook and RStudio with graphical desktop environments

OOD provides convenient web access to ReD resources via your Okta login for Mac, Windows, and Linux users, all without requiring command-line expertise or software installation.*

OpenOnDemand access

After connecting to the URC sso page, select the Open OnDemand Okta tile.

In a few moments, the Open OnDemand Dashboard page for ReD's AWS Parallel Cluster will appear.
Note: In some cases, it might take 30 to 60 seconds for the Dashboard page to appear. This window delay is a known issue; work is in progress to resolve this problem.

From the Dashboard, one can initiate a number of actions to perform work:

Performing Work

Running GUI Applications

Submit an RStudio jobs to the cluster:

a. Select InteractiveApps -> URC Rstudio Server

On the submission form you, select the appropriate options for your work. Some options may carry over from past runs of the same application. Select Launch when done.

Note: - Although you might see other partitions in the drop-down menu, you will not be able to submit to a different group's partitions unless you belong in that group. - GPU options only appply to partitions with GPU resources - By detault, interactive jobs dispatch as one task and only on one node (--ntasks=1 and --nodes=1)

b. This will bring you to the “My interactive sessions” page which will show a stanza for each of the job you submitted that are either in the queue, running or completed.

c. Once the job changes state to “Running”, you can click on the button to connect to the Rstudio session, and that will open another tab connecting to your running rstudio session.

Note: the compute servers are provisioned on demand when a new job gets submitted, so it might take about 10 min for the job to start.

d. Upon clicking “Connect to Rstudio Server” it will open the Rstudio app

Run Command-Line Applications

Open an interactive terminal to the login node:

a. From the main dashboard click on “Clusters -> _urcds-red-pc-prod Shell Access” . that will open a new tab with a terminal access to the login node of the cluster

b. From there you can in theory submit batch jobs or start interactive slurm jobs.

Viewing Running Jobs

Open OnDemand has the My Interactive Sessions item in the Dashboard's navigation bar. Use this screen to see what interactive jobs are currently running, and interactive jobs finished or exited in the last six days.

Browsing Files

Open OnDemand has an embedded Files application that enables you to browse filesystems that you have access to. By default, your home directory is included in the menu; other filesystems may be included in the menu as well.

Note: Certain features have been disable for security reasons or to support compliance by preventing direct import/export of data from the environment:

Uploads and downloads have been disabled.
You will not be able to view or edit files directly via the web view.
Lab folders are not accessible at this time (the design team has minimized risk by not mounting lab shared on the OOD host).

Some operations are still available by the OOD shell – this application places you directly on the cluster head node.

Work with Shared/Network Storage

Info about Network Storage:

In addition to local storage on the compute nodes, the following storage is network storage and available across the cluster.

a. The personal home folders for the users are in /shared/home/username; so for your users those would be:
```
/shared/home/christopher.souther2
/shared/home/christopher.souther3
/shared/home/christopher.souther4
```
b. Users christopher.souther2 and christopher.souther3 belong to group urc-<LabName1>-lab while user christopher.souther4 belongs to urc-<LabName2>-lab

The designated Lab folders – the place where usually user research labs host their sensitive research data - for those 2 labs are respectively
```
/data/labs/<LabName1>_lab
/data/labs/<LabName2>_lab
```

Software

Please see our separate Software documentation for specific details on installed research and supporting software, and how one can leverage use them in your interactive and batch jobs.

* Text adapted from Google's AI Overveiw to the query "What is Open OnDemand?"

Last Updated: Oct 10, 2025