AWS EIP does not have attribute ID

[elliottsam]

Hi,
When trying to associate an EIP with a AWS NAT gateway, I am seeing the following error.

02:02:03 Error applying plan:
02:02:03 
02:02:03 1 error(s) occurred:
02:02:03 
02:02:03 * Resource 'aws_eip.nat' does not have attribute 'id' for variable 'aws_eip.nat.id'

This error occurs randomly and works normally 75% of the time, the logs show that the EIP was created before the NAT gateway.

[brikis98]

Same problem here. As an added complexity, I'm using the count attribute on my aws_eip:

resource "aws_eip" "nat" {
    count = "${var.num_nat_gateways}"
    vpc = true
    depends_on = ["aws_internet_gateway.main"]
}

resource "aws_nat_gateway" "nat" {
    count = "${var.num_nat_gateways}"
    allocation_id = "${element(aws_eip.nat.*.id, count.index)}"
    subnet_id = "${element(aws_subnet.public.*.id, count.index)}"
    depends_on = ["aws_internet_gateway.main"]
}

The error I get intermittently is:

Resource 'aws_eip.nat' does not have attribute 'id' for variable 'aws_eip.nat.*.id'

[brikis98]

May be related to #4944. I'm trying the depends_on workaround now too.

[c4milo]

Even using depends_on, I have experienced sometimes AWS just taking more time than usual to replicate changes made by Terraform, causing this sort of errors to show up once in a while. Waiting for a little while and re-running terraform apply has worked for me.

[skarungan]

Got the same issue, while using different state-files
terraform apply -state=dev.tfstate ---> Runs fine and succeeds
terraform apply -state=qa.tfstate ----> Throws the following error

Error applying plan:

1 error(s) occurred:

* Resource 'aws_eip.mgmt_nat_gw' does not have attribute 'id' for variable 'aws_eip.mgmt_nat_gw.id'

Re-running the command worked though

[NeckBeardPrince]

+1

[brikis98]

Update: the depends_on workaround in #4944 does not work.

[cicdteam]

Meet the same issue very often last days (especially when terraforming by CircleCI build process)

[carlossg]

similar error is in #6991 with detailed logs

[hugoduncan]

I'm seeing this too, when running terraform plan where the resources don't already exist. Adding a depends_on doesn't seem to help.

[hugoduncan]

Running TF_VAR_availability_zones=us-east-1b,us-east-1e terraform plan on the below, when none of the resources exist on AWS, gives aws_nat_gateway.gateway #0: "allocation_id": required field is not set

variable aws_region {
  default = "us-east-1"
}

provider "aws" {
  region = "${var.aws_region}"
}

variable "availability_zones" {
  description = "Comma separated list of availability zones"
}

variable "cidrs" {
  default = {
    us-east-1b = "10.0.4.0/24"
    us-east-1d = "10.0.5.0/24"
    us-east-1e = "10.0.6.0/24"
  }
}

resource "aws_vpc" "vpc" {
  cidr_block = "10.0.0.0/16"
}

resource "aws_subnet" "public" {
  count = "${length(split(\",\", var.availability_zones))}"
  vpc_id = "${aws_vpc.vpc.id}"
  cidr_block = "${lookup(var.cidrs, element(split(\",\",var.availability_zones), count.index))}"
  availability_zone = "${element(split(\",\",var.availability_zones), count.index)}"
}

resource "aws_nat_gateway" "gateway" {
  count = "${length(split(\",\", var.availability_zones))}"
  subnet_id = "${element(aws_subnet.public.id, count.index)}"
}

[hugoduncan]

It looks like using subnet_id = "${element(aws_subnet.public.*.id, count.index)}" fixes the issue.

[NeckBeardPrince]

Any news on getting this patched and fixed?

[pawelprazak]

Still an issue in Terraform v0.7.5

[bplunkert]

I am also experiencing this issue on v0.7.9.

[bodgit]

I'm seeing something similar here when I try and use the (Allocation) ID of the EIP resource. I have the following:

resource "aws_eip" "haproxy" {
  vpc = true
}

The Allocation ID is needed by keepalived configuration so that whichever host in the cluster is the master can associate the EIP to itself (disassociating it first from a considered-dead peer if necessary) so I'm trying to pass the ID in a template_file resource:

data "template_file" "haproxy" {
  count = "${var.count}"
  template = "${file("template.tpl")}"
  vars {
    ...
    allocation_id = "${aws_eip.haproxy.id}"
    ...
  }
}

When I try and apply this I get the following error:

aws_eip.haproxy: Creating...
  allocation_id:     "" => "<computed>"
  association_id:    "" => "<computed>"
  domain:            "" => "<computed>"
  instance:          "" => "<computed>"
  network_interface: "" => "<computed>"
  private_ip:        "" => "<computed>"
  public_ip:         "" => "<computed>"
  vpc:               "" => "true"
...

* Resource 'aws_eip.haproxy' does not have attribute 'id' for variable 'aws_eip.haproxy.id'

Terraform has created the EIP resource but it didn't make it into the state file so if I re-apply I end up with multiple "orphaned" EIP resources in AWS. If I remove the dependency from the data template then everything works but I don't have the ID in the templates. If I use something else like the public_ip attribute then this works fine, it seems to be just the id attribute that breaks this.

I'm seeing this with 0.8.1.

[Ninir]

Hey folks,

Since it seems to be an intermittent issue: Is anyone still having the issue?
Would like to close it in the AWS provider if it does not exist anymore! Thanks :)

[c4milo]

@Ninir, nop, I would close it and open a new one if it comes up again.

[Ninir]

Thanks for letting me know @c4milo, cheers!

[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.