VS Code Server

VS Code Server with GPU support, AI tools, and data science integration for the CyVerse Discovery Environment.

Quick Launch

Version	Launch
latest

Features

AI Development Tools

• Claude Code - Anthropic AI coding assistant (claude)
• Gemini CLI - Google AI CLI (gemini)
• OpenAI Codex - OpenAI coding assistant (codex)
• Ollama - Run AI models locally
• MCP Server - Model Context Protocol filesystem server

GPU & Machine Learning

• NVIDIA CUDA 12.5 - GPU acceleration support
• NVIDIA Driver 535 - Graphics driver
• Miniconda & Mamba - Efficient Python package management

Development Tools

• GitHub CLI (gh) - Command-line tool for GitHub operations
• Git Credential Manager - Secure credential storage
• AWS CLI - Amazon Web Services CLI
• Globus Connect - Research data management

VS Code Extensions (Pre-installed)

• Python - Python language support
• Jupyter - Notebook support
• vscode-icons - File icons
• Claude Dev - AI assistant integration

CyVerse Integration

• GoCommands (gocmd) - CyVerse data transfer utilities
• S3FS-FUSE - Mount S3 buckets as filesystem
• OSN Support - Open Storage Network bucket mounting

System Utilities

• Monitoring - htop, glances, nvtop (GPU monitoring)
• Build tools - gcc, development essentials

Run Locally

# Pull the container
docker pull harbor.cyverse.org/vice/vscode:latest

# Run without GPU
docker run -it --rm -p 8080:8080 harbor.cyverse.org/vice/vscode:latest

# Run with GPU support
docker run -it --rm -p 8080:8080 --gpus all harbor.cyverse.org/vice/vscode:latest

Access VS Code at: http://localhost:8080

With mounted home directory

docker run -it --rm -v $HOME:/app --workdir /app -p 8080:8080 harbor.cyverse.org/vice/vscode:latest

CyVerse Integration

When running in the CyVerse Discovery Environment, this container automatically:

• Connects to your CyVerse Data Store
• Transfers your Git and SSH configurations
• Provides access to CyVerse's high-performance computing and storage

Build Your Own Container

FROM harbor.cyverse.org/vice/vscode:latest

# Add system packages
USER root
RUN apt-get update && apt-get install -y your-package

# Add Python packages
USER vscode
RUN mamba install -y -c conda-forge your-python-package

Container Architecture

This container implements several best practices:

• Robust user handling - Compatible with base image user/group setup
• Version-pinned dependencies - Better reproducibility
• Layer optimization - Minimized container size
• Built-in healthcheck - Container monitoring at /
• CUDA integration - GPU acceleration ready
• Pre-configured extensions - Ready for Python development

Environment Variables

Variable	Description	Default
GCM_CREDENTIAL_STORE	Git credential storage method	cache

Ports

Port	Service
8080	VS Code Server (HTTP)
8443	VS Code Server (HTTPS)

Resources

• CyVerse VICE Documentation
• Integrate Your Own Tools
• GoCommands Documentation
• CyVerse Learning Center