terraform-provider-aws: aws_instance change to user_data no longer results in a destroy and re-create

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave “+1” or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform CLI and Terraform AWS Provider Version

Affected Resource(s)

• aws_instance

Terraform Configuration Files

Please include all Terraform configurations required to reproduce the bug. Bug reports without a functional reproduction may be closed without investigation.

resource "aws_instance" "tfe" {
  count                = 2
  ami                  = data.aws_ami.ubuntu.id
  instance_type        = "m5.xlarge"
  key_name             = aws_key_pair.main.key_name
  availability_zone    = data.aws_availability_zones.available.names[count.index]
  iam_instance_profile = aws_iam_instance_profile.tfe.id

  network_interface {
    network_interface_id = aws_network_interface.tfe[count.index].id
    device_index         = 0
  }

  root_block_device {
    volume_size = 100
    tags = {
      Name = "${var.friendly_name_prefix}-tfe-server-ebs-root-${count.index + 1}"
    }
  }

  user_data = (count.index == 0 && var.install_type != "tfe_manual" && var.install_type != "apache_hello_world") ? "${local.final_tfe_script}${local.tfe_script_setup_admin_user}" : local.final_tfe_script

  tags = {
    Name = "${var.friendly_name_prefix}-tfe-server-${count.index + 1}"
  }
}

Debug Output

Panic Output

Expected Behavior

Prior to version 4.0.0 a change to the user_data argument resulted in the resource being destroyed and recreated.

Actual Behavior

The resource is no longer destroyed and recreated when there is a change to user_data.

Steps to Reproduce

1. terraform apply

Important Factoids

The docs appear to have been updated to include this line ‘Updates to this field will trigger a stop/start of the EC2 instance’.

However there does not appear to be anything in the release notes about this being a breaking change, so thought I should flag it in case it is unintentional. This change causes significant impact on users that expect the previous behaviour. This is particularly the case for Terraform Cloud and Enterprise users with VCS integration who are unable to use the -replace command line flag.

Given that user_data scripts are only run the first time an EC2 instance is deployed, it seemed logical that a change would result in a recreation of the resource, otherwise user_data changes will not be applied which may cause confusion.

References