template_cloudinit_config adding headers in rendered user-data

[jonaskint]

Terraform Version

Terraform v0.8.7

Affected Resource(s)

• template_cloudinit_config

Terraform Configuration Files

data "template_file" "controller" {
  template = "${file("${path.module}/template/controller_user_data.yml")}"

  vars {
    CA_PEM               = "${base64encode(var.ca-key)}"
    CA_CERT              = "${base64encode(var.ca-cert)}"
    ETCD_ENDPOINT        = "https://${var.etcd_endpoint}:2379"
    ETCD_DNS             = "${var.etcd_endpoint}"
    POD_NETWORK          = "172.16.0.0/16"
    K8S_VER              = "v1.5.2_coreos.0"
    HYPERKUBE_IMAGE_REPO = "quay.io/coreos/hyperkube"
    DNS_SERVICE_IP       = "172.17.0.10"
    SERVICE_IP_RANGE     = "172.17.0.0/24"
    NETWORK_PLUGIN       = "cni"
    REGION               = "${var.region}"
  }
}

data "template_cloudinit_config" "controller" {
  gzip          = false
  base64_encode = false

  part {
    content = "${data.template_file.controller.rendered}"
  }
}

resource "aws_instance" "controller" {
  ami           = "${var.ami}"
  instance_type = "${var.instance_type}"
  count         = "${var.count}"

  subnet_id = "${element(var.subnet_ids, count.index)}"

  # Prevent terraform booboos where the etcd cluster gets deleted. Now we need to delete them manually.
  //  disable_api_termination              = "true"
  disable_api_termination = "false"

  instance_initiated_shutdown_behavior = "stop"

  key_name             = "${var.key_name}"
  user_data            = "${data.template_cloudinit_config.controller.rendered}"
  security_groups      = ["${aws_security_group.controller.id}"]
  iam_instance_profile = "${aws_iam_instance_profile.controller.id}"

  root_block_device {
    volume_size           = 40
    volume_type           = "gp2"
    delete_on_termination = true
  }

  tags {
    Name        = "${var.name}-controller${count.index}"
    Environment = "${var.environment}"
    Region      = "${var.region}"
    Os          = "ContainerLinux"
  }

  lifecycle {
    create_before_destroy = true
    ignore_changes        = ["security_groups", "user_data"]
  }
}

Expected Behavior

Content of user-data through AWS console

#cloudconfig
...

Actual Behavior

Actual content of user-data through AWS console

Content-Type: multipart/mixed; boundary="MIMEBOUNDARY"
MIME-Version: 1.0
--MIMEBOUNDARY
Content-Transfer-Encoding: 7bit
Content-Type: text/plain
Mime-Version: 1.0

#cloud-config
...

Steps to Reproduce

Had an oversized user-data file that was being used through a simple 'template_file' data resource and because of the AWS limit in the api, I started using 'template_cloudinit_config' data resource to make it more compact. After the apply, I noticed the node was stuck during initialisation and I could not retrieve the user-data through the AWS console.

So I did a second apply with gzip and base64 turned off and the half of my user-data file in order to trouble shoot the 'template_cloudinit_config' resource and noticed the extra headers in the user-data through the AWS console (vm didn't come up either and was stuck during init).

Important Factoids

AWS ec2 instance

[jonaskint]

I've been reading through the code from builtin/providers/template/datasource_clopudinit_config.go and I was wondering why the Content-Type is being set 2 times. One time in the multipart writer and one time in the headers. Is this related?

[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.