Templates not working with latest master branch

[mikeyhill]

Hey, I just updated TF to pull in some changes and noticed that all of our templates stopped working. If I move back to a previous release all is well. It's been quite a while (a week?) since I updated TF so I don't have any info about which commit this stops working at but for now have reverted back.

  * Error configuring template: connection is shut down
  * Error configuring template: connection is shut down
  * Error configuring aws: connection is shut down

Occasionally I'll get multiple lines the same as above. Is this an issue with TF or have the templates changed? I don't see any differences in the documentation.

version: Terraform v0.6.0-dev (c1bb9116ad1899209bc87eae6122b478aee4d839)

[phinze]

Thanks for the report. It could be related to #2406 - will take a look.

[mitchellh]

Can you show us your configuration? Or a debug log? That'll help debug this.

[mikeyhill]

Yes, was just getting that - the debug log is so large it's crashing my browser on gists so looking for another way to post it.

[mikeyhill]

https://gist.github.com/mikeyhill/eadb407027b3e2bd5833

[mikeyhill]

    resource "aws_elb" "admin-elb" {

      name            = "${var.env}-admin-elb"
      subnets         = ["${var.admin_subnet_1_id}","${var.admin_subnet_2_id}"]
      security_groups = ["${aws_security_group.admin-elb-security.id}"]
      internal        = false

      listener {
        instance_port     = 80
        instance_protocol = "http"
        lb_port           = 80
        lb_protocol       = "http"
      }

      listener {
        instance_port     = 443
        instance_protocol = "http"
        lb_port           = 443
        lb_protocol       = "http"
      }

      health_check {
        healthy_threshold   = 2
        unhealthy_threshold = 5
        timeout             = 3
        target              = "HTTP:80/index.php"
        interval            = 20
      }

      cross_zone_load_balancing   = true
      idle_timeout                = 400
      connection_draining         = true
      connection_draining_timeout = 400

    }

    resource "aws_lb_cookie_stickiness_policy" "admin-sticky-cookie" {

          name          = "${var.env}-admin-sticky-cookie"
          load_balancer = "${aws_elb.admin-elb.id}"
          lb_port       = 80
          cookie_expiration_period = 600

    }

    resource "template_file" "admin-cloud-init" {
        filename = "cloud-config/elb-init.yml"
        vars {
            recipe          = "server-admin::runtime"
            chef_dns        = "${var.chef_dns}"
            chef_validator  = "${var.chef_validator}"
            chef_validator_key = "${file(\"validators/${var.env}-validator.pem\")}"
            datacenter      = "${var.dns_region}"
            env             = "${var.env}"
            elb             = "${aws_elb.admin-elb.name}"
            proxy_elb       = ""
        }
    }

    resource "aws_launch_configuration" "admin-launch-config" {

        name                    = "${var.env}-admin-launch-config"
        image_id                = "${var.ami}"
        instance_type           = "${var.admin_instance_type}"
        user_data               = "${template_file.admin-cloud-init.rendered}"
        iam_instance_profile    = "${aws_iam_instance_profile.admin-server-profile.name}"
        security_groups         = ["${aws_security_group.admin-security.id}"]
        key_name                = "${var.provision_key}"

        associate_public_ip_address = false

        root_block_device = {
          volume_type = "gp2"
          volume_size = "100"
          delete_on_termination = true
        }
        ephemeral_block_device = {
              device_name = "/dev/sdb"
              virtual_name = "ephemeral0"
        }
        ephemeral_block_device = {
              device_name = "/dev/sdc"
              virtual_name = "ephemeral1"
        }
    }

    resource "aws_autoscaling_group" "admin-autoscaling-group" {

      availability_zones        = ["${var.az_1}","${var.az_2}"]
      vpc_zone_identifier       = ["${var.admin_subnet_1_id}","${var.admin_subnet_2_id}"]
      name                      = "${var.env}-admin-autoscaling-group"
      max_size                  = "${var.admin_max_size}"
      min_size                  = "${var.admin_min_size}"
      health_check_grace_period = 300
      load_balancers          = ["${aws_elb.admin-elb.name}"]
      health_check_type         = "EC2"
      desired_capacity          = "${var.admin_desired_capacity}"
      force_delete              = true
      launch_configuration      = "${aws_launch_configuration.admin-launch-config.name}"

      tag {
        key                 = "Name"
        value               = "${var.env}-admin-node"
        propagate_at_launch = true
      }
      tag {
        key                 = "Environment"
        value               = "${var.env}"
        propagate_at_launch = true
      }
      tag {
        key                 = "Role"
        value               = "admin"
        propagate_at_launch = true
      }
      tag {
        key                 = "Group"
        value               = "application"
        propagate_at_launch = true
      }
    }

[mikeyhill]

following up on this, looks like the issue first appeared in 3815122 but not sure if this is intended or not.

[lamdor]

I also ran into this problem when I upgraded to terraform master this morning. I was able to do a git bisect on terraform against our configuration and found that 0b1dbf3 was the commit that broke it. We have 5 different providers in our configuration and it doesn't seem to be selective or predictable in which provider it has connection problems with. Sometimes it's multiple.

I've reverted to the commit before 0b1dbf3, and can verify that things work great as usual.

[phinze]

I'm working on reproing this now. The config from @mikeyhill has a bunch of dependencies I'm trying to meet - @rubbish if you happen to have a simpler repro case let me know.

[lamdor]

@phinze So it's happening when I'm doing a targeted plan against a certain module. I'll gist up an example. One sec.

[lamdor]

@phinze Here's an example

in main.tf:

provider "aws" {
  region = "us-east-1"
}

module "test" {
  source = "./module"
}

module "test2" {
  source = "./module"
}

and in module/main.tf:

resource "aws_route53_zone" "test" {
  name = "test"
}

When I run terraform plan --target=module.test, I sporadically get:

$ terraform plan --target=module.test
Error configuring: 1 error(s) occurred:

* Error configuring aws: connection is shut down

But sometimes it works.

Sometimes it comes back with:

$ terraform plan --target=module.test
There are warnings and/or errors related to your configuration. Please
fix these before continuing.

Errors:

  * module.test2.provider.aws: connection is shut down

[lamdor]

Also, when doing a full terraform plan, I've not been able to get the connection errors.

[phinze]

@rubbish perfect thank you - on it now

[mikeyhill]

@phinze - thanks! apologies on the useless config sample above, I'll keep that in mind that you need a working example rather than the actual configs - I'm still a bit green with this.

[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.