Using `cname_prefix` in Beanstalk environment always requires replacement of environment

[OblateSpheroid]

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

Have tried on several versions, most recently:

• Terraform v0.15.5 on linux_amd64
• hashicorp/aws v3.44.0

Affected Resource(s)

• aws_elastic_beanstalk_environment

Terraform Configuration Files

provider "aws" {
  region = "us-gov-west-1"
}

resource "aws_elastic_beanstalk_application" "tftest" {
  name        = "tf-test-name"
  description = "tf-test-desc"
}

resource "aws_elastic_beanstalk_environment" "tftest" {
  name                = "test-beanstalk-fix-myapp"
  cname_prefix        =  "test-beanstalk-fix-myapp"
  application         = aws_elastic_beanstalk_application.tftest.name
  solution_stack_name = "64bit Amazon Linux 2018.03 v2.25.0 running Multi-container Docker 19.03.13-ce (Generic)"
  setting {
    namespace = "aws:autoscaling:launchconfiguration"
    name      = "IamInstanceProfile"
    value     = "aws-elasticbeanstalk-ec2-role"
  }
}

Expected Behavior

terraform apply creates 2 resources: aws_elastic_beanstalk_application.tftest and aws_elastic_beanstalk_environment.test. A subsequent terraform plan outputs No changes. Your infrastructure matches the configuration.

Actual Behavior

A subsequent terraform plan gives:

aws_elastic_beanstalk_environment.tftest must be replaced
...
cname_prefix           = "test-beanstalk-fix-myapp" # forces replacement
...

Steps to Reproduce

1. terraform apply (correctly creates 2 resources)
2. terraform apply (incorrectly replaces aws_elastic_beanstalk_environment.tftest - very time consuming)

The only work-around I have found is to make cname_prefix null if the environment already exists. For example, if I amend the above code with:

variable "init" {
  type = bool
  default = true
}

resource "aws_elastic_beanstalk_environment" "tftest" {
  name                = "test-beanstalk-fix-myapp"
  cname_prefix        = var.init ? "test-beanstalk-fix-myapp" : null  # edited line
  application         = aws_elastic_beanstalk_application.tftest.name
  solution_stack_name = "64bit Amazon Linux 2018.03 v2.25.0 running Multi-container Docker 19.03.13-ce (Generic)"
  setting {
    namespace = "aws:autoscaling:launchconfiguration"
    name      = "IamInstanceProfile"
    value     = "aws-elasticbeanstalk-ec2-role"
  }
}

I can then run:

export TF_VAR_init=$(if [ -n "$(terraform state list | grep aws_elastic_beanstalk_environment.tftest)" ]; then echo false; else echo true; fi)
terraform apply

And it will show no changes needed. This seems like a hacky way around a simple problem of Terraform not recognizing the cname_prefix does not need to be updated.

Important Factoids

Testing this on a gov cloud account.

[florian0410]

Maybe linked but using the Beanstalk feature Swap environment URLs cause also the same behavior. https://docs.aws.amazon.com/fr_fr/elasticbeanstalk/latest/dg/using-features.CNAMESwap.html

Once the Swap is over, if you launch an apply terraform try to recreate the resource...

Too bad since we could do Blue green deployment easier if it worked.

[michaelsmoody]

Any chance this has a fix incoming any time soon?

It's unfortunately not possible to ignore ONLY that setting, at least that I've found (and opened a bug about, prior to finding this).