Removing key_name from launch configs does not trigger an updated launch config.

[hashibot]

This issue was originally opened by @Juberstine as hashicorp/terraform#15944. It was migrated here as a result of the provider split. The original body of the issue is below.

---

I needed to remove the key_name attached to an instance created by an ASG via a LC.

But running plan showed no change to be made.

Terraform Version

v0.9.5

Affected Resource(s)

Please list the resources as a list, for example:

aws_launch_configuration
Terraform Configuration Files

Actual code block post update. The only change made here was the removal of key_name:

resource "aws_launch_configuration" "triton-as-conf" {
  name_prefix   = "triton-${var.name}-launch-conf"
  image_id      = "${var.container_host_ami}"
  instance_type = "${var.container_host_instance_type}"

  security_groups = [
    "${aws_security_group.ssh.id}",
    "${aws_security_group.web.id}",
    "${aws_security_group.self_reference.id}",
  ]

  user_data            = "${data.template_file.user_data.rendered}"
  iam_instance_profile = "${aws_iam_instance_profile.container_host.name}"

  lifecycle {
    create_before_destroy = true
  }
}

All vars are set.

Inside the launch configs, the only code change was the removal of:

key_name = "${aws_key_pair.deployer.key_name}"

The above var was set but the concept here is that removing key_name does not trigger the LC to change.

Expected Behavior

LC should have been recreated without key_name.

Actual Behavior

Nothing.

Steps to Reproduce

Please list the steps required to reproduce the issue, for example:

terraform apply
Remove the key_name value from launch configs
terraform plan
terraform apply

Seems similar to the issue another user experience with EBS:

hashicorp/terraform#14826

[ChrisF987]

+1 with terraform v0.11.1

[george-angel]

We are seeing the same behaviour with aws_instance resource. Removing key_name or setting it to an empty string, terraform does not see change in configuration and results in no-op run.

As a workaround we taint the affected resources.

[MissStava]

+1 with terraform v0.11.11, provider.aws v1.60.0

[jvshahid]

We are also seeing the same issue with terraform v0.12.20 and aws provider version v2.49.0

[jwenz723]

+1 with terraform v0.12.23, provider.aws v2.52.0

[idubinskiy]

Looks like this is related to the Terraform behavior described in hashicorp/terraform#4501. key_name is configured as both Optional and Computed. I'm not sure why key_name was marked as Computed in the first place, but it's possible that simply removing that property will fix this issue.

[github-actions]

Marking this issue as stale due to inactivity. This helps our maintainers find and focus on the active issues. If this issue receives no comments in the next 30 days it will automatically be closed. Maintainers can also remove the stale label.

If this issue was automatically closed and you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thank you!

[george-angel]

remove stale

[aitorres]

This bug is still happening, is there any proposed fix on Terraform's side?

[jaffel-lc]

still happens on v1.3.4

[github-actions]

Marking this issue as stale due to inactivity. This helps our maintainers find and focus on the active issues. If this issue receives no comments in the next 30 days it will automatically be closed. Maintainers can also remove the stale label.

If this issue was automatically closed and you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thank you!