[v0.12.6] provider.aws v3.1.0 still shows "Invalid resource instance data in state"

[hacos]

Terraform Version

Terraform v0.12.6
+ provider.aws v3.1.0
+ provider.template v2.1.2

Note: seems like it might still be happening with Terraform v0.13.0? hashicorp/terraform-provider-aws#14431 (comment)

Terraform Configuration Files

Note: redacted for privacy purposes

resource "aws_instance" "standard_instance" {
  count                  = var.instance_count
  ami                    = var.instance_ami
  instance_type          = var.instance_type
  vpc_security_group_ids = var.vpc_security_group_ids
  subnet_id              = var.subnet_ids[count.index % length(var.subnet_ids)]
  key_name               = var.key_name
  iam_instance_profile   = var.iam_instance_profile
  ebs_optimized          = var.ebs_optimized

  user_data = var.instance_user_data

  root_block_device {
    volume_type = var.root_volume_type
    volume_size = var.root_volume_size
    encrypted   = var.encrypted
  }

  dynamic "ebs_block_device" {
    for_each = var.ebs_devices
    content {
      device_name = ebs_block_device.value.device_name
      volume_type = ebs_block_device.value.volume_type
      volume_size = ebs_block_device.value.volume_size
      iops        = lookup(ebs_block_device.value, "iops", null)
      encrypted   = var.encrypted
    }
  }

  dynamic "ephemeral_block_device" {
    for_each = var.ephemeral_devices
    content {
      device_name  = ephemeral_block_device.value.device_name
      virtual_name = ephemeral_block_device.value.virtual_name
    }
  }

  lifecycle {
    ignore_changes = [
      user_data,
      ami,
      ebs_block_device,
      ephemeral_block_device,
      subnet_id
    ]
  }
}

Debug Output

Need to review and scrub before posting

Crash Output

Expected Behavior

The plan to work. It should have replaced the instance because of a new variable. Interestingly enough, if I run just terraform plan instead of using -target= the plan outputs properly. Since these aws_instances are being replaced, I'd rather not take down the entire group of them all at once, hence the -target= flag.

Actual Behavior

Got an error similar to hashicorp/terraform-provider-aws#14431

Error: Invalid resource instance data in state

  on ../../../../base_modules/standard_instance/standard_instance.tf line 11:
  11: resource "aws_instance" "standard_instance" {

Instance module.REDACTED.aws_instance.standard_instance[3] data
could not be decoded from the state: unsupported attribute
"network_interface_id".

Steps to Reproduce

terraform init
terraform plan -target=module.REDACTED.aws_instance.standard_instance[3]

Additional Context

References

Was told by this comment to log an issue if it's still happening with Terraform v0.12.X (which I am on). hashicorp/terraform-provider-aws#14431 (comment)

[danieldreier]

Hi! Thanks for reporting this. Based on the error report you showed, I think this is probably a valid issue, and I'd like to reproduce it locally.

To do that, I have to be able to run this and run it on my workstation without inventing any details in order to be confident we're seeing the same behavior. As-is, I don't have your vars, so I'd have to invent some details. Can you please restate your reproduction case such that I can copy-paste it and run it locally? If it's easier to provide via a github repo rather than a copy-paste, you're welcome to make a PR against https://github.com/danieldreier/terraform-issue-reproductions to share it, or provide your own GH repo.

[hacos]

Hi! I can try but I believe the issue isn't with the variables but with the state itself. In any case, here's the variables.tf and terraform.tfvars file:

variable "instance_ami" {}
variable "instance_type" {}
variable "name" {}
variable "vpc_security_group_ids" {
  type = list(string)
}
variable "subnet_ids" {
  type = list(string)
}
variable "instance_user_data" {
  default = ""
}
variable "key_name" {
}
variable "iam_instance_profile" {
  default = ""
}
variable "instance_count" {
  default = 1
}
variable "cluster_tag" {
}
variable "deployment_tag" {
}
variable "ebs_optimized" {
  default = false
}
variable "owner_tag" {
  default = "InfraShared"
}
variable "ebs_devices" {
  type    = list(map(string))
  default = []
}
variable "ephemeral_devices" {
  type    = list(map(string))
  default = []
}
variable "encrypted" {
  default = true
}
variable "root_volume_type" {
  default = "gp2"
}
variable "root_volume_size" {
  default = 64
}
variable "extra_tags" {
  type    = map(string)
  default = {}
}

instance_ami   = "REDACTED"
instance_type  = "t3.medium"
name           = "terraform-bug-25827"
key_name       = "REDACTED"
cluster_tag    = "terraform-bug"
deployment_tag = "terraform"

[Craigfis]

I'm also hitting this error (with Terraform 0.13.0). In my case it is complaining because my state file contains the obsolete API gateway attribute "request_parameters_in_json".

[mshkrebtan]

Editing the state file and removing the unsupported attributes helped me.

• Pull the state from the remote location:

  terraform state pull > tf-state.json

• Edit the state file: remove the unsupported attributes and increment the serial number.

• Push the state to the remote location:

  terraform state push tf-state.json

• Run terraform plan again.

[Mikescops]

Scary but it works 😨

[mildwonkey]

👋 Hi! This is a duplicate of #25752 - this issue will be fixed in the next terraform release. Thanks!

[ghost]

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.