VM Tool - Deployment Approaches

Two Deployment Methods

vm_tool supports two deployment approaches:

1. Simple Deployment (`deploy-docker`)

Use when: VM is already set up with Docker

vm_tool deploy-docker --host IP --user USER

Features:

Deploys docker-compose application
Health checks
Rollback support
Idempotent

GitHub Actions:

- name: Deploy
  run: vm_tool deploy-docker --host ${{ secrets.EC2_HOST }}

2. Full Setup (`run_cloud_setup`)

Use when: Fresh VM needs complete setup

from vm_tool.runner import SetupRunner, SetupRunnerConfig, SSHConfig

config = SetupRunnerConfig(
    github_username=os.environ.get('GITHUB_USERNAME'),
    github_token=os.environ.get('GITHUB_TOKEN'),
    github_project_url='https://github.com/user/repo.git',
    docker_compose_file_path='docker-compose.prod.yml',
)

runner = SetupRunner(config)
ssh_configs = [SSHConfig(...)]
runner.run_cloud_setup(ssh_configs)

Features:

Installs Docker
Clones GitHub repository
Sets up docker-compose
Configures environment

GitHub Actions:

- name: Deploy
  run: python scripts/deploy_to_vm.py

When to Use Each

Scenario	Method	Why
Fresh EC2 instance	Full Setup	Needs Docker installation
Existing server	Simple Deployment	Docker already installed
First deployment	Full Setup	Complete environment setup
Updates/redeploys	Simple Deployment	Faster, idempotent
Multiple VMs	Full Setup	Provision all at once
Single VM	Either	Both work

Migration Path

Start: Full Setup (first time)

runner.run_cloud_setup(ssh_configs)  # Installs everything

Later: Simple Deployment (updates)

vm_tool deploy-docker --profile prod  # Just deploy changes

Comparison

Feature	Simple Deployment	Full Setup
Docker installation	❌ (assumes installed)	✅
GitHub clone	❌ (manual)	✅
Deploy app	✅	✅
Health checks	✅	❌
Rollback	✅	❌
Idempotent	✅	⚠️
Speed	Fast	Slower
Use case	Updates	Initial setup

Recommendation

Best Practice:

Use Full Setup for initial VM provisioning
Use Simple Deployment for all subsequent updates

This gives you the best of both worlds!