task_definition version always known after apply

[TarekAS]

Hello,

While using this module, I get the following message every time I do terraform apply.

~ resource "aws_ecs_service" "atlantis" {
        cluster                            = "arn:aws:ecs:eu-west-1:[REDACTED]:cluster/atlantis"
        deployment_maximum_percent         = 200
        deployment_minimum_healthy_percent = 50
        desired_count                      = 1
        enable_ecs_managed_tags            = false
        health_check_grace_period_seconds  = 0
        iam_role                           = "aws-service-role"
        id                                 = "arn:aws:ecs:eu-west-1:[REDACTED]:service/atlantis"
        launch_type                        = "FARGATE"
        name                               = "atlantis"
        platform_version                   = "LATEST"
        propagate_tags                     = "NONE"
        scheduling_strategy                = "REPLICA"
        tags                               = {}
      ~ task_definition                    = "atlantis:2" -> (known after apply)

Is there any way to avoid this?

The following is my module usage:

module "atlantis" {
  source  = "terraform-aws-modules/atlantis/aws"
  version = "~> 2.0"

  name               = "atlantis"
  vpc_id             = "${aws_vpc.myvpc.id}"
  private_subnet_ids = ["${aws_subnet.private-1a.id}", "${aws_subnet.private-1b.id}", "${aws_subnet.private-1c.id}"]
  public_subnet_ids  = ["${aws_subnet.public-1a.id}", "${aws_subnet.public-1b.id}", "${aws_subnet.public-1c.id}"]

  certificate_arn       = "redacted"
  create_route53_record = false

  atlantis_github_user       = "${local.atlantis_github_username}"
  atlantis_github_user_token = "${local.atlantis_github_access_token}"
  atlantis_repo_whitelist    = ["redacted"]

  ssm_kms_key_arn = "redacted"
}

[jondunning]

Having the same issue

[jseiser]

Same issue present in my deployments as well.

[wreulicke]

I have same issue.

I seem that this issue is caused by this PR.
#62

[wreulicke]

@antonbabenko
What do you think?

[tomwj]

tomwj@22a4b64

Fix issue #72 task_definition known after apply

Using a data provider prevents terraform converging. Basically until the apply runs it doesn't have the data to know which revision is great. Forcing it to be updated on every apply. It doesn't have the data available since it's pulling from AWS when it has it locally already.

It does mean if you update the task_definition in the AWS console this will overwrite it. Which I believe is correct behaviour. e.g. I bumped the task_definition manually to 16 and it will roll it back.

    ~ task_definition                    = "atlantis:16" -> "atlantis:15"

https://www.terraform.io/docs/configuration/data-sources.html#data-resource-dependencies

[thyming]

does anyone have a workaround for this issue?

[rotarur]

I have the same issue. i think depends_on can be safely removed as it is referenced in another resource to be used.
I did it in my tests and the plan doesn't show any changes:

data "aws_ecs_task_definition" "atlantis" {
   task_definition = var.name
}

[rpdelaney]

Looks like this was fixed in #163

[github-actions]

I'm going to lock this issue 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 similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.