Destroy stop-vm continues to remove-vm before vms have stopped.

[timvia]

When running a destroy our vms get stuck in a shutting down state. They never complete the shutdown and fail.

We would like the ability to use the Turnoff flag when running a destroy.

This would then run "Stop-VM -Name TestVM -TurnOff" rather than "Stop-VM -Name VM1 -Force"

As this is a destroy task we do not care about a clean shutdown.

Terraform Version

1.9.8

Expected Behavior

When destroying a vm the vm should shutdown quickly and forcefully.

Actual Behavior

VMs runnning K8s on linux (rhel 8/9) change state to shutting down and hang there until it fails to shutdown.

Terraform error message is

│ Remove-VM : Deleting 'cluster-1' failed.
│ The operation cannot be performed while the object is in its current state.
│ At C:\Users\proteanautomation\AppData\Local\Temp\shell-terraform-2nkPyhRSQesox83fe8cgpPoodEv.ps1:3 char:106
│ + ... ?{$_.Name -eq 'pipeline-dp-performance-cluster-1'} | Remove-VM -force

This is due to the state of the VM still being in a shutting down state.

To work around this we are manually going to hyperv and turning off the vms before running the terraform. Not ideal for automation though

[timvia]

Using a null resource ro issue the turn off command to work around this issue.

resource "null_resource" "force_shutdown_vm_on_destory" {
  triggers = {
    vm_names = hyperv_machine_instance.default.name
    host     = var.hyperv_host_name
    user     = var.hyperv_user
    password = var.hyperv_password
  }
  provisioner "local-exec" {
    when    = destroy
    command = <<EOT
      curl -k -u "${self.triggers.user}:${self.triggers.password}" \
      -H "Content-Type: application/json" \
      -d '{"PowerShell":"Stop-VM -Name ${self.triggers.vm_names} -TurnOff -Force; while((Get-VM -Name ${self.triggers.vm_names}).State -ne \"Off\") { Start-Sleep -s 2 }}"}' \
      https://${self.triggers.host}:5986/wsman
    EOT
  }
}