Skip to content

Using the Web-Based Shell

Open OnDemand provides a browser-based terminal connected to a VHPC login node. The shell uses the same VHPC account, storage, permissions, and software environment available through a traditional SSH connection.

Use the shell for command-line tasks such as managing files, finding software, preparing job scripts, submitting Slurm jobs, and reviewing results.

Open the Shell

From the Open OnDemand navigation bar:

  1. Open Clusters.
  2. Select VHPC Shell Access.

The terminal opens in a new browser tab and displays a command prompt on a VHPC login node.

Appropriate Login-Node Use

VHPC login nodes are shared systems intended for tasks such as:

  • Managing files
  • Editing scripts
  • Loading software modules
  • Compiling software
  • Submitting jobs
  • Monitoring jobs
  • Performing lightweight data inspection

Do not run long-running, CPU-intensive, memory-intensive, or GPU workloads directly in the shell.

Submit computational work through Slurm or use an interactive application.

Warning

Processes that consume excessive resources on a login node may be limited or terminated to protect the service for other users.

Work with Files

The shell and the Open OnDemand file manager access the same underlying VHPC storage.

Common commands include:

pwd
ls
cd directory-name
mkdir directory-name

Files created or modified in the shell are visible in the Open OnDemand file manager after refreshing the directory listing.

The file manager’s Home Directory, Work, and Staging shortcuts correspond to your primary VHPC storage locations.

Find Available Software

Open Clusters → Module Browser to search for software modules available on VHPC.

To use the Module Browser:

  1. Enter part of a software or module name in Search Modules.
  2. Select VHPC from the cluster filter when needed.
  3. Expand a matching result.
  4. Select the required version.
  5. Copy the displayed module load command.

The Module Browser may also show dependencies required by the selected module.

Note

The Module Browser helps you find software, but it does not load the module into your environment. Run the displayed command in the shell or include it in a Slurm job script.

The Last updated value shows when the searchable module catalog was most recently generated.

Load a Software Module

Run the command provided by the Module Browser, for example:

module load hdf5/1.12.1

You can also search from the command line:

module spider application-name

To view currently loaded modules:

module list

To remove a loaded module:

module unload module-name

Use the exact module name and version appropriate for your work.

Submit a Slurm Job

Submit a Slurm job script with:

sbatch myjob.slurm

Slurm returns a numeric job ID when it accepts the job.

For example:

Submitted batch job 123456

Keep the job ID when troubleshooting or reviewing the job later.

Monitor Jobs

To display your active jobs from the shell:

squeue -u "$USER"

You can also use Jobs → Active Jobs for a browser-based view.

To cancel a job:

scancel job-id

Replace job-id with the numeric Slurm job ID.

For more information, see Monitoring and Managing Jobs.

Review a Completed Job

The Active Jobs page focuses on queued and running jobs. Use Slurm accounting when you need information about an older or completed job:

sacct -j job-id

Job output and error files can be viewed from the shell or through the file manager.

Check VHPC System Status

Open Clusters → System Status for a high-level summary of current VHPC activity.

The page displays information such as:

  • Available and in-use nodes
  • Available and in-use processors
  • Available and in-use GPUs
  • Running jobs
  • Queued jobs

This is a general capacity snapshot. It does not predict when a particular job will start.

A job’s wait time also depends on its requested resources, partition, account, scheduling priority, limits, and other Slurm requirements.

Use the Shell and File Manager Together

Keeping the shell and file manager open in separate tabs can be useful when you want to:

  • Upload a script and submit it
  • Make a light edit in the browser and test it
  • Review job output
  • Check file permissions
  • Download completed results

Changes made in either interface affect the same files.

Close the Shell

When finished, exit the shell with:

exit

You can then close the browser tab.

Closing the tab without exiting disconnects the visible terminal. Commands running directly in the shell may stop when the connection closes or expires.

Do not rely on the browser shell to keep unattended work running. Submit that work through Slurm.

Troubleshooting

The shell does not open

Confirm that:

  • You are connected to KSU VPN-Groups
  • You are signed in to Open OnDemand
  • Your VHPC account is active
  • Your browser allows Open OnDemand to open a new tab

Return to the dashboard and try Clusters → VHPC Shell Access again.

A command is not available

The software may require an environment module.

Search for it with Clusters → Module Browser or:

module spider application-name

I cannot access a directory

The shell uses your normal VHPC permissions.

Check the path and permissions with:

ls -ld path

If access was recently granted, sign out of Open OnDemand and sign in again.

My command stopped after I closed the tab

Commands started directly in the shell may stop when the terminal disconnects.

Submit long-running or unattended work through Slurm instead.

The Module Browser does not show expected software

Try searching with only part of the application or module name.

Also check the Last updated time. If recently installed software remains missing after the catalog has been refreshed, contact the Service Desk.

Next Steps

Getting Help

If you cannot open the shell, cannot find expected software, or receive an unexpected VHPC error, contact the Kennesaw State University Service Desk.

Method Students Faculty/Staff
Web ServiceNow ServiceNow
E-Mail studenthelpdesk@kennesaw.edu service@kennesaw.edu
Phone 470-578-3555 470-578-6999

When reporting a problem, include:

  • Your NetID
  • The approximate date and time the problem occurred
  • The command or task you were attempting
  • The complete error message
  • The relevant file path or Slurm job ID, when applicable