如何在Terraform中管理三个分环境的user data文件？

Got it, let's walk through how to solve this. You want your aws_instance resource to pick the right user-data script depending on the env variable you set—here are two clean ways to do it, depending on your Terraform version.

Modern Approach (Terraform 0.12+) #

Terraform 0.12 introduced the built-in templatefile function, which replaces the old template_file data resource and makes this much simpler. Here's how to implement it:

1. First, make sure your three user-data files (user_data_dev.sh, user_data_prod.sh, user_data_qa.sh) are in the same directory as your main Terraform configuration file (or adjust the path accordingly).
2. Update your aws_instance resource to dynamically reference the correct file using the env variable:

variable "env" {
  description = "choose env between dev | prod | qa "
  type        = string
  # Add validation to prevent invalid environment values
  validation {
    condition     = contains(["dev", "prod", "qa"], var.env)
    error_message = "Valid values for env are dev, prod, or qa."
  }
}

resource "aws_instance" "server" {
  ami           = "ami-123456"
  instance_type = "t2.medium"
  availability_zone = "us-east-1"
  
  # Dynamically select the user-data file based on the env variable
  user_data = templatefile(
    "${path.module}/user_data_${var.env}.sh",
    {} # No extra variables needed since your scripts are static
  )

  root_block_device {
    delete_on_termination = true
    volume_size           = var.volume_size
    volume_type           = var.volume_type
  }

  tags = {
    Name = "domain_testing"
  }
}

How This Works: #

• ${path.module} points to the directory of your current Terraform module, so it will reliably find your user-data files regardless of where you run Terraform from.
• We use string interpolation (${var.env}) to build the filename matching the selected environment.
• The validation block on the env variable ensures users can only input valid environment values, preventing typos that would break the file path.

Legacy Approach (Pre-Terraform 0.12) #

If you're stuck on an older Terraform version that doesn't support templatefile, you can use the template_file data resource with conditional logic:

variable "env" {
  description = "choose env between dev | prod | qa "
  type        = string
}

# Define a template file resource for each environment
data "template_file" "user_data_dev" {
  template = file("${path.module}/user_data_dev.sh")
}

data "template_file" "user_data_prod" {
  template = file("${path.module}/user_data_prod.sh")
}

data "template_file" "user_data_qa" {
  template = file("${path.module}/user_data_qa.sh")
}

resource "aws_instance" "server" {
  ami           = "ami-123456"
  instance_type = "t2.medium"
  availability_zone = "us-east-1"
  
  # Use ternary operators to select the correct rendered template
  user_data = var.env == "dev" ? data.template_file.user_data_dev.rendered :
              var.env == "prod" ? data.template_file.user_data_prod.rendered :
              data.template_file.user_data_qa.rendered

  root_block_device {
    delete_on_termination = true
    volume_size           = var.volume_size
    volume_type           = var.volume_type
  }

  tags = {
    Name = "domain_testing"
  }
}

Quick Note on Your Scripts: #

I noticed your user-data scripts have spaces around the equals sign in variable assignments (e.g., DOMAIN = "testing.dev.xxxx.com"). While bash allows this, it's more idiomatic and less error-prone to write them without spaces: DOMAIN="testing.dev.xxxx.com".

内容的提问来源于stack exchange，提问作者karthik