Error: Resource node has no configuration attached

[nickmhankins]

Terraform Version

Terraform v0.12.2

Debug Output

I have the trace output but would prefer not to post it here - even encrypted. It contains a lot of sensitive information. Is there an email address @hashicorp I can send it to instead?

Expected Behavior

Successful plan should have applied

Actual Behavior

Received error message for seemingly every resource in every module:

Error: Resource node has no configuration attached

The graph node for
module.aas_resources.module.aas_vm_avzone_creation.azurerm_virtual_machine_extension.oms_extension[0]
has no configuration attached to it. This suggests a bug in Terraform's apply
graph builder; please report it!

Error: Resource node has no configuration attached

The graph node for
module.hana_resources.module.hana_vm_avzone_creation.azurerm_virtual_machine_data_disk_attachment.disk_attachments[7]
has no configuration attached to it. This suggests a bug in Terraform's apply
graph builder; please report it!

Error: Resource node has no configuration attached

The graph node for
module.ci_resources.module.ci_vm_avzone_creation.azurerm_network_interface.nic[0]
has no configuration attached to it. This suggests a bug in Terraform's apply
graph builder; please report it!

Error: Resource node has no configuration attached

The graph node for
module.ci_resources.module.ci_vm_avzone_creation.azurerm_virtual_machine_data_disk_attachment.disk_attachments[3]
has no configuration attached to it. This suggests a bug in Terraform's apply
graph builder; please report it!

Steps to Reproduce

terraform init
terraform plan
terraform apply

Additional Context

We have around 4 layers of modules so it's hard to post a code snippet to describe the problem. The plan comes back green and it looks like everything it wants to create is what it should be creating. Everything worked as it should under TF 0.11.14, we are working through 0.12 changes so we can utilize the new version.

Workaround

I was able to workaround the issue by not using an output file from terraform plan and just doing an apply. Unfortunately that is not going to work for us because our pipeline automation doesn't work that way.

[word]

I came across a similar issue and it was due to an input variable having a wrong type set. It was:

variable "example" {                                                                                                                                                                    
  type  = "boolean"
}

changed it to:

variable "example" {                                                                                                                                                                    
  type  = bool
}

without the quotes. Not sure if that's the same issue but I had the exact same symptoms. I'd review your input types first as per: https://www.terraform.io/docs/configuration/variables.html

[word]

I can reproduce this reliably by specifying invalid type in 0.12.3:

variable "vpc_cidr" {
  type = invalidtype
}

plan works, apply with plan file fails:

<SNIP>
Error: orphan resource module.srv-vpc.module.vpc.aws_route.nat still has a non-empty state after apply; this is a bug in Terraform

Error: orphan resource module.srv-vpc.module.vpc.aws_internet_gateway.igw still has a non-empty state after apply; this is a bug in Terraform
<SNIP>

Apply without a plan file seems to work. This should error during plan and display a sensible message.

[dyegoe]

I don't know what exactly the steps that I did but I got the same errors and the same workaround worked for me.

[Esya]

We've had the same issue here after upgrading to 0.12 - and the problem was that we defined the same attribute twice on the same ressource :

resource "aws_alb_target_group" "instance_tg" {
  target_type = "ip"
  name        = "admin-${var.env_name}-tg"
  protocol    = "HTTP"
  port        = "80"
  vpc_id      = "${var.vpc_id}"

  health_check {
    path = "/"
  }
  target_type = "ip"

}

We removed the extra "target_type" and it works again

[sarneaud]

I've produced it in 0.12 with a duplicate locals section.

[dyegoe]

@sarneaud nice clue. Just to clarify, did you write your locals in one terraform and in a module (that you did)?

[sarneaud]

No, I have multiple locals in multiple modules and files and it's normally not a problem. As a part of some refactoring, one got duplicated verbatim.

It looks like a lot of mistakes like this are showing up as "no configuration attached" errors during apply in 0.12. Unfortunately, it can be pretty hard to debug.

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.