vsphere_virtual_machine.vm does not have attribute 'default_ip_address'

[den-is]

After upgrading vCenter from 5.5 to 6.5 some optional steps stopped working.
In this case terraform plan is trying to get VM IP before VM is actually created.

Versions

❯ terraform -v
Terraform v0.11.3
+ provider.vsphere v1.3.2

Affected Resource(s)

❯ terraform plan
Refreshing Terraform state in-memory prior to plan...
The refreshed state will be used to calculate this plan, but will not be
persisted to local or remote state storage.

data.vsphere_datacenter.dc: Refreshing state...
data.vsphere_resource_pool.rpool: Refreshing state...
data.vsphere_datastore.ds: Refreshing state...
data.vsphere_virtual_machine.template: Refreshing state...
data.vsphere_network.net: Refreshing state...
------------------------------------------------------------------------
Error: Error running plan: 1 error(s) occurred:
* module.tftest.output.ip_addresses: Resource 'vsphere_virtual_machine.vm' does not have attribute 'default_ip_address' for variable 'vsphere_virtual_machine.vm.*.default_ip_address'

outputs.tf

output "ip_addresses" {
 value = "${vsphere_virtual_machine.vm.*.default_ip_address}"
}

[vancluever]

@den-is this is usually due to an error higher up the dependency chain. Can you try removing the output and planning again and see if you can plan successfully?

Thanks!

[den-is]

@vancluever yes of course when I'm removing it everything works fine.
I've created TF module which contains some default outputs declarations for future use.
with 5.5 it was working

[vancluever]

@den-is - I think I see the error in your output actually:

module.tftest.output.ip_addresses: Resource 'vsphere_virtual_machine.vm' does not have attribute 'default_ip_address' for variable 'vsphere_virtual_machine.vm.*.default_ip_address'

The resource address implies that there is only one VM (ie: no count specified), so globbing with * won't work here. Try vsphere_virtual_machine.vm.default_ip_address.

[den-is]

hmmmm can't get it...
if what am I doing wrong
this is my module initialization:

module "tftest" {
  source           = "./vmwareprod"
  servers          = 2
  hostname_prefix  = "tftest"
  domain           = "int.example.com"
  template         = "template-centos-7.4"
  cluster          = "prod"
  datastore        = "VM_SATA_NFS_Linux_01"
  ipaddresses      = ["10.55.1.99", "10.55.1.100"]
  network          = "int"
  mask             = 23
  gateway          = "10.55.1.1"
  dns              = ["10.55.1.17", "10.55.1.18"]
}

module has this outputs declaration:

output "ip_addresses" {
 value = "${vsphere_virtual_machine.vm.*.default_ip_address}"
}

external to module outputs declaration, what user sees:

output "vm_IPs" {
  value = "${module.tftest.ip_addresses}"
}

there is count in module initialization, in this case 2 - doesn't work. same with when count is 1.

[kevinskim93]

I am running into a similar issue. I am running

output "ip" { value = "${vsphere_virtual_machine.vm.default_ip_address}" }

and I am seeing

Error: Error running plan: 1 error(s) occurred:

* output.ip: Resource 'vsphere_virtual_machine.vm' does not have attribute 'default_ip_address' for variable 'vsphere_virtual_machine.vm.default_ip_address'

[vancluever]

Hey @den-is and @kevinskim93, thanks for the input. I've changed this to a bug. I can't reproduce it over here with a basic configuration, unfortunately, so it might be hard to track down on this end.

I am going to attempt the upgrade path to see if it's easy to re-produce that way.

[astraios]

I have the same issue:

module.instance.output.ip: Resource 'vsphere_virtual_machine.instance' does not have attribute 'default_ip_address' for variable 'vsphere_virtual_machine.instance.default_ip_address'

I wrote myself a terraform module that create a vSphere VM.

I tested it with vSphere 6.0 and vSphere provider 1.3.2 and 1.3.3
terraform version is 0.11.5.

[dmilind]

I am getting same issue.
I have this block in .tf file
output "ip" { value = "${vsphere_virtual_machine.vm.*.default_ip_address}" }
error that I am getting is
Error: vsphere_virtual_machine.vm: vsphere_virtual_machine.vm: self reference not allowed: "vsphere_virtual_machine.vm.*.default_ip_address"
any clue ?

[422158]

any progress on this... this bug is compromising our automated destroy of environments

[sholland]

Seeing this issue as well, anyway to make this a priority?

[spstratis]

Another issue that I think is related to this is that I'm not even able to run a remote-exec provisioner if the host IP is assigned by DHCP. It doesn't seem like I can actually access the default_ip_address from TF. Here is an example of my provisioner where I run into the issue...

  provisioner "remote-exec" {
    inline = [
      "sudo apt-get update -y",
      "sudo apt-get install -y nginx"
    ]

    connection {
      host = "${self.default_ip_address}"
      type = "ssh"
      user = "ubuntu"
      private_key = "${file("files/test-prod.pem")}"
    }
  }

[erickellerek1]

@spstratis any hint about this issue? did you achieve to get the host ip in another way?
wait_for_guest_net_routable = true did the trick for me...

[hashibot]

This issue has been open 180 days with no activity. If this issue is reproducible with the latest version of the provider and with Terraform 0.12, please comment. Otherwise this issue will be closed in 30 days.

[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 feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. If you feel I made an error 🤖 🙉 , please reach out to my human friends 👉 hashibot-feedback@hashicorp.com. Thanks!