[bugfix] aws/eip: Don't save instance ID if association fails

[radeksimko]

This is allowing clean destroy in cases when no internet GW is available in VPC and corrupted state would normally appear.

How to replicate

resource "aws_vpc" "default" {
  cidr_block = "10.10.0.0/16"
}

resource "aws_subnet" "public" {
  vpc_id = "${aws_vpc.default.id}"
  cidr_block = "10.10.1.0/24"
  availability_zone = "us-east-1c"
}

resource "aws_instance" "web" {
  ami = "ami-323b195a"
  availability_zone = "us-east-1c"
  instance_type = "t2.small"
  key_name = "coreos-test"
  subnet_id = "${aws_subnet.public.id}"
  associate_public_ip_address = true
}

resource "aws_eip" "lb" {
  instance = "${aws_instance.web.id}"
  vpc = true
}

Output (that's still valid):

* Failure associating instances: Gateway.NotAttached: Network vpc-7c9abd19 is not attached to any internet gateway

destroy:

* MissingParameter: Either public IP or association id must be specified

The instance ID is saved to state file even though the API request ends up in error, which then results in trying to remove association which doesn't actually exist = corrupted state.

[mitchellh]

LGTM

[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.