Research Users: Your Persistent Identity Across Cloud Environments¶

Research Users provide Research Users - your persistent identity that follows you across all your cloud research environments. No more worrying about different usernames, file permissions, or SSH keys when switching between Python, R, and other research templates.

Research user commands are available as prism user.

Quick Start¶

What Are Research Users?¶

Research Users give you a consistent identity across all your Prism instances: - Same username and user ID on every instance you launch - Persistent home directory that survives instance shutdowns - SSH keys that work everywhere without reconfiguration - File sharing that just works between different instances

The Problem Research Users Solve¶

Before Research Users:

# Python ML instance
ssh researcher@my-python-instance    # Different username each time
ls /home/researcher                   # Files lost when instance stops

# R Research instance
ssh rstudio@my-r-instance            # Different username again!
ls /home/rstudio                      # Can't access Python files

With Research Users:

# Python ML instance
ssh alice@my-python-instance         # Your research username
ls /efs/home/alice                    # Your persistent files

# R Research instance
ssh alice@my-r-instance              # Same username!
ls /efs/home/alice                    # Same files available instantly

How It Works: Dual User System¶

Prism uses a Dual User System that gives you the best of both worlds:

System Users (Template-Created)¶

• Purpose: Run services like Jupyter, RStudio, databases
• Examples: ubuntu, researcher, rstudio, rocky
• Characteristics: Created by templates, manage applications

Research Users (Your Persistent Identity)¶

• Purpose: Your files, SSH access, persistent work
• Examples: alice, bob, your-name
• Characteristics: Same across all instances, consistent file permissions

Real-World Example¶

When you launch a "Python Machine Learning" instance with research user alice:

Instance Setup:
├── ubuntu (system user)    ← SSH access, system management
├── researcher (system user) ← Runs Jupyter notebook server
└── alice (research user)   ← Your files, persistent home directory

You SSH in as alice, your files are in /efs/home/alice, and Jupyter runs as the researcher user but you can access it seamlessly.

Getting Started¶

1. Create Your Research User¶

# Create your research user
prism user create alice

# Generate SSH keys automatically
prism user ssh-key generate alice ed25519

2. Launch Instances with Research Users¶

# Launch with your research user
prism workspace launch "Python Machine Learning" my-analysis --research-user alice

# Later, launch R instance with same user
prism workspace launch "R Research Environment" my-analysis-r --research-user alice

3. Access Your Persistent Home Directory¶

# SSH into any instance
ssh alice@your-instance

# Your files are always in the same place
ls /efs/home/alice/
├── projects/          # Your research projects
├── .bashrc           # Your shell configuration
├── .ssh/             # Your SSH keys
└── research/         # Shared research data

Key Benefits¶

🔄 Cross-Template Compatibility¶

Work seamlessly across different research environments:

# Day 1: Python analysis
prism workspace launch python-ml analysis --research-user alice
ssh alice@analysis
cd /efs/home/alice/projects
python create_dataset.py  # Creates dataset.csv

# Day 2: R visualization
prism workspace launch r-research visualization --research-user alice
ssh alice@visualization
cd /efs/home/alice/projects
R -e "data <- read.csv('dataset.csv')"  # Same file, no copying!

📁 Persistent File Storage¶

Your files persist across instance lifecycles:

# Create files on any instance
echo "Important research data" > /efs/home/alice/results.txt

# Instance shutdown/hibernation → files preserved
# Launch new instance → files immediately available
ls /efs/home/alice/results.txt  ✅ Still there!

👥 Collaborative Research¶

Multiple researchers can share files with correct permissions:

# Alice creates shared project
alice@instance: mkdir /efs/shared/team-project
alice@instance: cp analysis.py /efs/shared/team-project/

# Bob accesses same project
bob@another-instance: ls /efs/shared/team-project/
analysis.py  ← Bob can read Alice's files

🔐 Unified SSH Access¶

One set of SSH keys works everywhere:

# Generate keys once
prism user ssh-key generate alice ed25519

# Use same key for all instances
ssh alice@python-instance    # Works
ssh alice@r-instance         # Works
ssh alice@rocky-instance     # Works

File Organization¶

Your Home Directory Structure¶

/efs/home/alice/              # Your persistent home
├── .bashrc                   # Your shell preferences
├── .ssh/                     # Your SSH keys
│   └── authorized_keys
├── projects/                 # Your research projects
│   ├── ml-analysis/
│   │   ├── data.csv
│   │   ├── analysis.py
│   │   └── results.ipynb
│   └── r-visualization/
│       ├── plots.R
│       └── figures/
└── research/                 # Shared research data
    └── datasets/

Shared Directories¶

/efs/shared/                  # Collaborative space
├── datasets/                 # Shared datasets
├── libraries/                # Shared code libraries
├── notebooks/                # Shared Jupyter notebooks
└── team-projects/            # Multi-user projects

SSH Key Management¶

Generate New Keys¶

# Generate Ed25519 key (recommended)
prism user ssh-key generate alice ed25519

# Or generate RSA key for compatibility
prism user ssh-key generate alice rsa

Import Existing Keys¶

# Import your existing public key
prism user ssh-key import alice ~/.ssh/id_rsa.pub "My laptop key"

List and Manage Keys¶

# List all SSH keys
prism user ssh-key list alice

# Delete a key
prism user ssh-key delete alice key-id

Working with Templates¶

Template Compatibility¶

Research users work with all Prism templates:

• ✅ Python Machine Learning: Research user + Jupyter service user
• ✅ R Research Environment: Research user + RStudio service user
• ✅ Rocky Linux Base: Research user + rocky system user
• ✅ Ubuntu Desktop: Research user + ubuntu system user
• ✅ Custom Templates: Research user + any template-defined users

How Templates Integrate (Phase 5A+ Template Integration)¶

🎉 NEW: Automatic Research User Creation

Templates can now automatically create and provision research users during launch! Use the --research-user flag with research-enabled templates:

# Automatic research user creation with new templates
prism workspace launch python-ml-research my-project --research-user alice
# ✅ Launches instance + creates 'alice' research user + provisions SSH keys + sets up EFS home

# Check template capabilities
prism templates info python-ml-research
# Shows research user integration features

Research-Enabled Templates:

Templates specify research user integration in their configuration:

# Example: templates/python-ml-research.yml
name: "Python ML Research (Research User Enabled)"
research_user:
  auto_create: true                       # Create research user automatically
  require_efs: true                       # Ensure EFS home directories
  efs_mount_point: "/efs"                 # EFS mount location
  efs_home_subdirectory: "research"       # Home structure: /efs/research/<username>
  install_ssh_keys: true                 # Generate and install SSH keys
  default_shell: "/bin/bash"              # Default shell for research users
  default_groups: ["research", "efs-users", "docker"]  # Research user groups
  user_integration:
    strategy: "dual_user"                 # System + research user architecture
    primary_user: "research"              # Research user is primary
    collaboration_enabled: true           # Multi-user collaboration

Template Integration Benefits: - One-Step Launch: Create instance + research user + SSH keys in single command - EFS Auto-Setup: Persistent home directories created automatically - Cross-Template Identity: Same research user works across all templates - Professional Display: Template info shows research user capabilities

Multi-User Collaboration¶

Setting Up Team Research¶

1. Each team member creates their research user:

   # Alice
   prism user create alice
   
   # Bob
   prism user create bob
   
   # Carol
   prism user create carol
   

2. Share EFS volumes across instances:

   # Create shared EFS volume
   prism volumes create team-research-data
   
   # Mount on all instances
   prism volumes mount team-research-data alice-instance
   prism volumes mount team-research-data bob-instance
   

3. Collaborate with consistent permissions:

   # Alice creates project (on alice-instance)
   mkdir /efs/shared/project-alpha
   echo "# Project Alpha" > /efs/shared/project-alpha/README.md
   
   # Bob contributes (on bob-instance)
   cd /efs/shared/project-alpha
   git clone https://github.com/team/analysis.git
   
   # Carol reviews (on carol-instance)
   cd /efs/shared/project-alpha/analysis
   jupyter notebook review.ipynb
   

Best Practices¶

File Organization¶

# Recommended structure
/efs/home/alice/
├── projects/          # Individual projects
│   ├── project-a/     # One directory per project
│   └── project-b/
├── scratch/           # Temporary work
├── archive/           # Completed projects
└── shared/            # Symlinks to shared directories
    ├── datasets → /efs/shared/datasets
    └── libraries → /efs/shared/libraries

Security¶

• Keep SSH keys secure: Research user SSH keys protect all your instances
• Use strong usernames: Choose usernames that don't conflict with system users
• Regular key rotation: Periodically generate new SSH keys
• Monitor access: Check which instances your research user is active on

Performance¶

• EFS caching: Files in /efs/home cached for better performance
• Local scratch space: Use /tmp or /home/ubuntu for temporary files
• Shared directories: Organize shared files to minimize EFS traffic

Backup¶

• EFS persistence: Your /efs/home directory is automatically persistent
• Regular snapshots: Consider EFS snapshots for important research data
• Version control: Use git for code, research notebooks, and documentation

Migration from Existing Setups¶

Migrating Files¶

If you have existing research files on old instances:

# From old instance (as ubuntu/researcher)
tar -czf ~/my-research-backup.tar.gz ~/my-research-files/

# Copy to new instance
scp my-research-backup.tar.gz alice@new-instance:/efs/home/alice/

# On new instance (as alice)
cd /efs/home/alice
tar -xzf my-research-backup.tar.gz

Migrating SSH Keys¶

# Import your existing SSH key
prism user ssh-key import alice ~/.ssh/id_rsa.pub "Migrated key"

# Or generate new keys and update GitHub/servers
prism user ssh-key generate alice ed25519
cat /efs/home/alice/.ssh/id_ed25519.pub  # Add to GitHub, servers

Troubleshooting¶

Common Issues¶

Q: I can't SSH into my instance with my research user

# Check SSH key is properly configured
prism user ssh-key list alice

# Verify user was provisioned
prism user status alice --instance my-instance

# Check SSH key permissions on instance
ssh ubuntu@my-instance "ls -la /efs/home/alice/.ssh/"

Q: My files disappeared after launching a new instance

# Check EFS mount
ssh alice@my-instance "mount | grep efs"

# Verify home directory
ssh alice@my-instance "ls -la /efs/home/"

# Check if EFS volume is mounted
prism volumes list

Q: File permissions are wrong

# Check research user UID consistency
ssh alice@instance1 "id alice"  # Should be same UID everywhere
ssh alice@instance2 "id alice"

# Fix permissions if needed
ssh alice@my-instance "sudo chown -R alice:research /efs/home/alice"

Getting Help¶

# Check research user status
prism user status alice

# List all research users
prism user list

# View detailed instance information
prism instances describe my-instance --show-users

What's Coming Next¶

These features are available in v0.35.3: - CLI Integration: prism user command suite

• GUI Support: Point-and-click research user management
• Template Enhancement: Templates with built-in research user support
• Advanced EFS: Automatic EFS volume creation and management

Future Enhancements: - Multi-Profile Collaboration: Share research users across Prism profiles - Advanced Policies: Institutional controls and governance - Usage Analytics: Track research user activity and resource usage - Globus Integration: Institutional authentication and data transfer

Summary¶

Research Users transform Prism from a single-instance tool into a collaborative research platform:

• 🎯 Consistent Identity: Same username and UID across all instances
• 💾 Persistent Storage: Files survive instance shutdowns and changes
• 🔑 Unified Access: One SSH key for all your research environments
• 🤝 Collaboration: Share files with teammates using consistent permissions
• 🔧 Template Flexibility: Works with any research environment template

Start using Research Users today to streamline your cloud research workflows!

Need Help? - 📚 Full documentation: Research User Architecture - 🐛 Report issues: GitHub Issues - 💬 Discuss: GitHub Discussions