Building Scalable Infrastructure with Terraform and VMware vSphere

🚀 Building Scalable Infrastructure with Terraform + vSphere

I recently completed a comprehensive and modular Terraform setup to provision and manage vSphere virtual machines across multiple environments — dev, uat, and prod — using best practices for reusability, maintainability, and automation.

📦 Structure Highlights:

Modular design: vm, disk, network, and customization modules

Environment-specific configurations using Terraform workspaces

Remote state management via S3 + DynamoDB for safe collaboration

Support for CI/CD pipelines (e.g., Jenkins) with secure credential handling

🧱 Key Practices Implemented:

✅ Modular & reusable codebase

✅ DRY variable and data management

✅ Dynamic resource scaling via for_each

✅ Clean separation of environments (environments/dev, uat, prod)

✅ Shell script to bootstrap Terraform workspaces

✅ Output structure ready for integration with ticketing/approval systems

---

📁 Overview: Project Directory Structure

The project is organized to support:

• Multi-environment deployments (dev, uat, prod)

• Reusability via modular design

• Scalability and automation using workspaces and CI/CD integration

---

🧱 Root-Level Files

README.md

• Project overview, usage instructions, and prerequisites.

• ✅ Best practice: Include how to use workspaces, module inputs, and secrets management.

versions.tf

• Specifies the required Terraform version and provider versions.

• ✅ Best practice: Lock versions to avoid unexpected breaking changes.

provider.tf

• Configures the vSphere provider, often referencing TF_VAR_* environment variables for security.

• ✅ Best practice: Don’t hardcode secrets; use Jenkins credentials or .tfvars files.

backend.tf

• Defines the remote backend (e.g., S3 + DynamoDB) for state storage and locking.

• ✅ Best practice: Use environment-specific state files with workspace key prefixes.

data.tf

• Centralized location for reusable data sources like datacenters, clusters, datastores, templates, and networks.

• ✅ Best practice: Keep this DRY and minimal; share common logic across modules.

locals.tf

• Derived values such as naming conventions, tag formats, or environment settings.

• ✅ Best practice: Store computed values to avoid repetition in resources and modules.

---

🌍 Environment-Specific Configuration: environments/

Each environment (dev, uat, prod) is self-contained and contains:

main.tf

• Entry point that calls reusable modules (vm, network, disk, customization).

• ✅ Best practice: Clearly define module sources and pass environment-specific variables.

variables.tf

• Declares all inputs required by this environment.

• ✅ Best practice: Document descriptions, types, and defaults where applicable.

terraform.tfvars

• Contains sensitive or environment-specific values (e.g., IPs, templates, naming prefixes).

• ✅ Best practice: Never commit secrets to version control.

data.tf & locals.tf

• Environment-specific data lookups (e.g., different datacenters).

• Local values derived for naming, tagging, or conditions based on the env.

---

⚙️ Reusable Modules: modules/

Each module is focused, loosely coupled, and reusable.

vm/

• Provisions the virtual machine.

• Inputs: CPU, memory, guest OS, template, boot options.

• Outputs: VM ID, IP, hostname.

network/

• Handles VM NIC assignment and port group attachment.

• Supports dynamic or static IP assignment.

disk/

• Manages one or more disks, sizes, and storage policies.

• ✅ Best practice: Support dynamic disk count using for_each.

customization/

• Handles guest OS customization: hostname, domain, DNS, post-deploy scripts.

Each module includes:

• main.tf: Resource definitions.

• variables.tf: Input schema.

• outputs.tf: Values passed back to root.

• locals.tf: Naming and tag formatting.

• data.tf: Module-specific data sources.

✅ Best practice: Use module versioning via Git tags or registries for stability.

---

🧰 Utility: workspace_setup/

create-workspaces.sh

• Script to automate creation of Terraform workspaces for dev, uat, and prod.

#!/bin/bash
for env in dev uat prod; do
  terraform workspace new $env || echo "Workspace $env already exists"
done

✅ Best practice: Run in CI/CD or during project bootstrap.

---

🚀 Best Practices Summary

Category Best Practice

🔒 Security Use secrets via environment variables or Jenkins credentials

🪜 Modularization Keep modules small, composable, and reusable

🚦 Environments Use isolated terraform.tfvars and workspaces

📦 State Management Use remote backend with locking (S3 + DynamoDB recommended)

✅ CI/CD Automate init, plan, apply, and approval workflows (e.g., via Jenkins)

🧪 Validation Use terraform validate and tflint pre-apply

📚 Documentation Keep README.md updated with module usage, variables, and outputs

---

If you're working with VMware + Terraform, happy to connect and share notes! 💬

#Terraform #vSphere #InfrastructureAsCode #DevOps #Automation #CI/CD #HashiCorp #VMware #TerraformModules