Added use case for availability_zones in autoscaling docs

[unixninja92]

Changes proposed in this pull request:

• Fix docs to include new use case for the availability_zones parameter in the auto scaling group resource

[bflad]

Hi @unixninja92 👋 Thanks for submitting this. Could you please provide us with some additional context about this note? e.g. relevant AWS documentation or an example Terraform configuration and error? Thanks.

[unixninja92]

yeah, when you try and create an auto scaling group based on a launch template that attaches a specific ENI via id, if you read the current docs you assume you need to set vpc_zone_identifier in the asg config rather than availability_zones becasue of the note that the latter is only for ec2-classic. When running

resource "aws_network_interface" "XXXXX" {
  subnet_id = "${data.terraform_remote_state.vpc.public-subnet}"
  security_groups = ["${aws_security_group.XXXXX.id}"]
  source_dest_check = false
}

resource "aws_launch_template" "XXXXX" {
    name = "XXXXX"
    ebs_optimized = true
    image_id = "ami-XXXXX"
    instance_type               = "m5.large"
    key_name = "XXXXX"
    network_interfaces {
      network_interface_id = "${aws_network_interface.XXXXX.id}"
    }
    user_data            = "${base64encode(local.XXXXX-userdata)}"
}
resource "aws_autoscaling_group" "XXXXX" {
  name                 = "XXXXX"
  launch_template {
    id = "${aws_launch_template.XXXXX.id}"
    version = "${aws_launch_template.XXXXX.latest_version}"
  }
  min_size             = 1
  desired_capacity     = 1
  max_size             = 1
  vpc_zone_identifier = ["${data.terraform_remote_state.vpc.public-subnet}"]
}

you get the following error:
Invalid details specified: You must use a valid fully-formed launch template. A network interface may not specify both a network interface ID and a subnet
This is because the subnet is selected based on where the ENI is and they don't want you defining this twice. When neither vpc_zone_identifier or availability_zones is specified you get an error saying you need to set one of them (don't have the exact error message on hand).

I opened a support ticket with aws about this as I was also having this issue with their web console. They confirmed that it was a bug in the web console, but gave me cli commands that work (which sets availability_zones instead of vpc_zone_identifier) and also noted that their cli docs do not currently cover this use case. I believe fixing this is now on their internal roadmap.

To summarize, switching
vpc_zone_identifier = ["${data.terraform_remote_state.vpc.public-subnet}"]
to
availability_zones = ["${data.aws_subnet.XXXXX-subnet.availability_zone}"]
makes it work but is not reflected in current docs.

[rflume]

They confirmed that it was a bug in the web console, but gave me cli commands that work [...]

Hey, could you provide that command maybe? I'm having the same issue currently but try not to use terraform to create the ASG.

[unixninja92]

They confirmed that it was a bug in the web console, but gave me cli commands that work [...]

Hey, could you provide that command maybe? I'm having the same issue currently but try not to use terraform to create the ASG.

aws autoscaling create-auto-scaling-group --auto-scaling-group-name <NAME> --min-size 1 --max-size 1 --launch-template LaunchTemplateId=<TemplateID> --availability-zones <az-name>

[github-actions]

This functionality has been released in v3.57.0 of the Terraform AWS Provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

For further feature requests or bug reports with this functionality, please create a new GitHub issue following the template. Thank you!

[github-actions]

I'm going to lock this pull request because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems related to this change, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.