08-05-lifecycle

Terraform Meta Argument: lifecycle

lifecycle Meta Argument

• The lifecycle Meta Argument in Terraform is used to control specific aspects of how resources are managed during their lifecycle.

• The lifecycle Meta Argument provides fine-grained control over when and how Terraform should create, update, or delete resources.

• lifecycle Meta Argument options:

  1. create_before_destroy:

     • When set to true, this attribute indicates that Terraform should create a new resource before destroying the old one when it needs to replace the resource.
     • This can help minimize downtime during updates.
     • Default option is false

  2. prevent_destroy:

     • When set to true, this attribute prevents the resource from being destroyed or deleted.
     • This can be useful in protecting critical resources from accidental deletion.
     • Default option is false

  3. ignore_changes:

     • This attribute allows you to specify a list of attributes for which Terraform should ignore changes.
     • It's useful to prevent certain attributes from being updated during resource modifications.
     • Commonly used when you have some attributes that users are allowed to make changes on AWS console and you want terraform to ignore such changes
     • Example : minor version upgrade of RDS instance
     • Default option is false

• Example: create_before_destroy
  00_provider.tf

  terraform {
  required_providers {
      aws = {
          source = "hashicorp/aws"
          version = "~> 5.0"
      }
  }
  }
  
  provider "aws" {
      region = "us-east-1"
  
      default_tags {
      tags = {
          Terraform = "yes"
          Project = "terraform-learning"
          Owner = "Venkatesh"
      }
      }
  }

  01_ec2.tf

  resource "aws_instance" "myec2" {
      ami = "ami-0df435f331839b2d6"
      instance_type = "t2.micro"
      availability_zone = "us-east-1a"
      # availability_zone = "us-east-1b"
  
      tags = {
      Name = "Linux2023"
      }
  
      # lifecycle {
      #   create_before_destroy = true
      # }
  }

• Lets create an AWS EC2 without lifecycle arguments

  1. terraform init : Initialize terraform
  2. terraform validate : Validate terraform code
  3. terraform fmt : format terraform code
  4. terraform plan : Review the terraform plan
  5. terraform apply : Create Resources by terraform
  • Once terraform completes the execution you should be able to check on your AWS Console Resource created successfully

• Now lets try to change the Availability Zone of our EC2 Instance to without lifecycle arguments and observe the behavior

  resource "aws_instance" "myec2" {
      ami = "ami-0df435f331839b2d6"
      instance_type = "t2.micro"
      # availability_zone = "us-east-1a"
      availability_zone = "us-east-1b"
  
      tags = {
      Name = "Linux2023"
      }
  
      # lifecycle {
      #   create_before_destroy = true
      # }
  }

• Lets Execute Terraform commands to understand resource behavior

  1. terraform init : Initialize terraform
  2. terraform validate : Validate terraform code
  3. terraform fmt : format terraform code
  4. terraform plan : Review the terraform plan
  5. terraform apply : Create Resources by terraform

  • Once we run terraform apply , Terraform first destroys existing resources and than creates the resource

    • Example of terraform apply

      

      

      

      

• Now lets try to change the Availability Zone of our EC2 Instance to with lifecycle arguments and observe the behavior

  resource "aws_instance" "myec2" {
      ami = "ami-0df435f331839b2d6"
      instance_type = "t2.micro"
      availability_zone = "us-east-1a"
      # availability_zone = "us-east-1b"
  
      tags = {
      Name = "Linux2023"
      }
  
      lifecycle {
        create_before_destroy = true
      }
  }

  • Once we run terraform apply , Terraform first creates new resources first, and than destroys the existing resource

    • Example of terraform apply

      

      

      

• Example: prevent_destroy

  • lets use prevent_destroy argument and observe the behavior
    01_ec2.tf

    resource "aws_instance" "myec2" {
    ami               = "ami-0df435f331839b2d6"
    instance_type     = "t2.micro"
    availability_zone = "us-east-1a"
    # availability_zone = "us-east-1b"
    
    tags = {
        Name = "Linux2023"
    }
    
    lifecycle {
        prevent_destroy = true
    }
    }

  • Execute terraform destroy command and observe the behavior

    • Terraform shows Error: Instance cannot be destroyed

    

• Example: ignore_changes

  • Lets use the tags concept and understand how ignore_changes works

  • First we are going to manually add tag on our AWS EC2 using AWS console ( this tag was not part of terraform code)

    

  • Run terraform plan and see how terraform detects the changes that are made on AWS Console

  • Terraform detects the Manual changes made on AWS console and if we run terraform apply than terraform will remove the manually applied tags on AWS EC2

  • Now, lets add in meta argument ignore_changes for tags and see how terraform behaves.

    01_ec2.tf

    resource "aws_instance" "myec2" {
    ami               = "ami-0df435f331839b2d6"
    instance_type     = "t2.micro"
    availability_zone = "us-east-1a"
    # availability_zone = "us-east-1b"
    
    tags = {
        Name = "Linux2023"
    }
    
    lifecycle {
        ignore_changes = [ tags ]
    }
    }

  • Run terraform plan and see how terraform ignores the changes* that are made to tags AWS Console

• ignore_changes can additionally be used to ignore all the changes for particular resource with option, ignore_changes = all

  • Terraform will completely ignore any changes to all attributes of that resource when it attempts to update or modify it

  • Example:

    resource "aws_instance" "myec2" {
    ami               = "ami-0df435f331839b2d6"
    instance_type     = "t2.micro"
    availability_zone = "us-east-1a"
    # availability_zone = "us-east-1b"
    
    tags = {
        Name = "Linux2023"
    }
    
    lifecycle {
        ignore_changes = [ all ]
    }
    }

  • It's usually a better practice to specify the exact attributes you want to ignore changes for, rather than using "all" to have more fine-grained control over which attributes should be excluded from updates

  Cleanup

  6. terraform destroy : destroy or delete Resources, Cleanup the resources we created

     • After you type yes to terraform destroy prompt, terraform will start destroying resources

     • Once terraform completes the execution you should be able to check on your AWS Console resources are successfully deleted.

References :

The lifecycle Meta-Argument