Code Server on SeaWulf

Code Server

Code Server is a web-based version of Visual Studio Code (VS Code) that lets you write, edit, and run code remotely on SeaWulf using your browser. It’s ideal for users who want a lightweight, interactive development environment without installing VS Code locally. You can access your home, scratch, and project directories, and use the integrated terminal for compiling and running code.

Getting Started

  1. Open the SeaWulf Open OnDemand portal: https://sn-ood.seawulf.stonybrook.edu/
  2. Log in using your NetID and DUO two-factor authentication.
  3. Launch Code Server:
    • Click the Code Server icon on the main dashboard
    • OR navigate to Interactive Apps → Code Server

Session Configuration

  • Queue: Choose a queue suitable for your workload (e.g., short-40core-shared).
  • Cores: 1 core is usually sufficient for light editing; increase if compiling/testing.
  • Time Limit: Set based on how long you plan to work.
  • Nodes: Keep at 1 unless using parallel builds.
  • Email Notification: Optional; will notify your Stony Brook email when the session starts.
Tip: Request only the resources you need to reduce queue wait times.

Click Launch to begin. Your session status will show Queued → Starting → Running. Once running, a new page will open with a VS Code interface and session controls.

Using Code Server

Once your session starts, you have access to a full VS Code environment:

pinned apps
 

File Management

  • Open and edit files in your home, scratch, or project directories.
  • Use the integrated file explorer to move, copy, and manage files.

Terminal & Compilation

  • Use the integrated terminal for compiling code, running scripts, or executing commands.
  • Commands behave as if you are directly logged in via SSH.

Extensions & Environment

  • Some VS Code extensions are supported; install only what you need.
  • Configure your shell profile, Python virtual environments, or other language-specific tools for development.
Important: Save your work frequently in home or project directories to avoid data loss.

Recommendations

  • Close idle sessions to free up cluster resources for other users.
  • Be mindful of CPU and memory usage to avoid long queue times.
  • Verify extensions and tools are compatible with Code Server on SeaWulf.

Accessing Files & Collaboration

  • Sessions are private; there is no shareable link like MATLAB.
  • For collaboration, consider using Git or syncing code via scratch/project directories.

Example symlink to access scratch space from your home directory:

ln -s /gpfs/scratch/NetID /gpfs/home/NetID

Replace NetID with your actual NetID.