azurerm_virtual_machine does not update state file if disks are renumbered

[cchildress]

Terraform Version

0.7.3 (waiting for 0.7.6 to come out due to [GH-9122]

Affected Resource(s)

azurerm_virtual_machine

Terraform Configuration Files

Before:

  storage_data_disk {
    name = "${var.node_name}_data_disk_premium_01"
    vhd_uri = "${var.node_data_storage_account_premium_blob_endpoint}${azurerm_storage_container.node_container_data_premium.name}/${var.node_name}_data_premium_01.vhd"
    create_option = "Empty"
    disk_size_gb = 1023
    lun = 0
  }
  delete_data_disks_on_termination = true

After:

  storage_data_disk {
    name = "${var.node_name}_data_disk_standard_01"
    vhd_uri = "${var.node_data_storage_account_standard_blob_endpoint}${azurerm_storage_container.node_container_data_standard.name}/${var.node_name}_data_standard_01.vhd"
    create_option = "Empty"
    disk_size_gb = 1023
    lun = 0
  }
  storage_data_disk {
    name = "${var.node_name}_data_disk_premium_01"
    vhd_uri = "${var.node_data_storage_account_premium_blob_endpoint}${azurerm_storage_container.node_container_data_premium.name}/${var.node_name}_data_premium_01.vhd"
    create_option = "Empty"
    disk_size_gb = 1023
    lun = 32
  }
  delete_data_disks_on_termination = true

Debug Output

~ module.<some_server>.azurerm_virtual_machine.node
    storage_data_disk.0.lun:     "32" => "0"
    storage_data_disk.0.name:    "<some_server>_data_disk_premium_01" => "<some_server>_data_disk_standard_01"
    storage_data_disk.0.vhd_uri: "<foo>" => "<bar>"
    storage_data_disk.1.lun:     "0" => "32"
    storage_data_disk.1.name:    "<some_server>_data_disk_standard_01" => "<some_server>_data_disk_premium_01"
    storage_data_disk.1.vhd_uri: "<bar>" => "<foo>"

Expected Behavior

This change is successfully applied to Azure, but does not appear to be recored properly in the state file.

Actual Behavior

This shows up as a change to be made in every following terraform plan

Steps to Reproduce

1. Set up a VM using the before layout.
2. Change the Terraform config to match the after.
3. Run a terraform apply and confirm you see the changes in the Azure portal
4. Run terraform plan and the changes should be listed as pending still.

Important Factoids

I tried just editing the state file to see if it was something that could be manually corrected. Maybe I missed a spot, but I could never get it to see the change as completed.

[cchildress]

I went through the state file and edited the order of the disks. Now I've found that if I run terraform plan -refresh=false Terraform does not think it needs to apply any changes to the disks on the virtual machine (which is what I would expect). If I run terraform plan and allow it to do the refresh it does think the disk changes need to be re-applied. This suggests to me that the out of order information is actually coming from Azure and not the local statefile.

This is a little odd because the lun layout and disk ordering in the Azure web UI is set properly.

[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.