terraform destroy error when referencing a random_id resource in a module's locals block

[metmajer]

Terraform Version

Terraform v0.11.7
+ provider.aws v1.26.0
+ provider.random v1.3.1

Terraform Configuration Files

provider "aws" {}
provider "random" {}

resource "random_id" "random" {
  byte_length = 8
}

module "aws-ec2-instance" {
  source = "github.com/metmajer/terraform-aws-ec2"

  name = "${var.name}"
  uuid = "${random_id.random.hex}"
}

Debug Output

See https://gist.github.com/metmajer/11063a52fb166a1d228fe115b637da91#file-terraform-destroy-log.

Expected Behavior

terraform destroy should not complain about a nil reference in random_id.

Actual Behavior

In a newly initialized workspace, terraform destroy issues the following error:

Error: Error applying plan:

1 error(s) occurred:

* module.aws-ec2-instance.var.uuid: variable "random" is nil, but no error was reported

Steps to Reproduce

1. terraform init
2. terraform destroy

Additional Context

I could only reproduce the issue if ${random_id.random.hex} is passed on to a module AND if the module references the passed-on variable in a locals block. See ${var.uuid} in https://github.com/metmajer/terraform-aws-ec2/blob/master/main.tf.

If https://github.com/metmajer/terraform-aws-ec2/blob/master/main.tf was changed to https://github.com/metmajer/terraform-aws-ec2/blob/no-locals/main.tf, the error does not occur.

It appears that I needed to rm -rf terraform.tfstate* .terraform; terraform init before validating that the error does not occur with https://github.com/metmajer/terraform-aws-ec2/blob/no-locals/main.tf. Otherwise, the error persisted.

References

Related issues are:

• terraform destroy error for a local value from random_id resource #16190
• Unexpected nil variable on terraform destroy #7993

[metmajer]

@jbardin: is there any other information I can provide? I am wondering since this issue is currently blocking us in our project. Let me know if we can collaborate somehow!

[jbardin]

Hi @metmajer,

Thanks for following up. There's nothing else we need here, this is related to some known edge cases with destroy time dependencies. It's likely to be the same root cause as #18026, but I'll leave this open until I can verify. Once we can unfreeze the master branch after the 0.12 work is complete we can work on resolving these.

[Constantin07]

I'm having the same issue when trying to destroy.

Error: Error applying plan:

1 error occurred:

	* local.db_password: local.db_password: Resource 'random_id.uuid' does not have attribute 'hex' for variable 'random_id.uuid.hex'

Is there any workaround for it meanwhile ?

[tbugfinder]

@jbardin It still is failing with 0.11.14.

Hello,

This particular problem is no longer reproduced with Terraform 0.12. Users experiencing similar errors with 0.12 are not likely from the same root cause, and we ask that a new issue be opened in that case.

Since there is no more planned development for 0.11, we are going to close out this issue.

Thanks

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.