Skip ipv6 and ipv4 if os-type="cloud-init" and no IP configured.

[TobiPeterG]

This fixes #1413

I will test the changes now and see if it works as expected.

[TobiPeterG]

Tested it, this seems to fix the issue. Now the provider won't wait for the VM to come up with cloud-init.

However, I noticed that a destroy action might not work and also cause a 500 error when the guest agent is not running in the VM yet. Maybe the error message during creation or destruction with agent = 1 should be extended to not only show the 500 error, but also show that the guest agent is enabled but might not be running yet on the VM.

[TobiPeterG]

Hmm I installed the guest agent on the VM and tried to delete it but I got a 500 error and the VM is not destroyed. Will investigate

[TobiPeterG]

Curious, on the server itself I don't see a 500 error.
It only made two calls:
::ffff:10.124.31.44 - OpenTofu@pve!OpenTofu [02/10/2025:11:44:28 +0000] "GET /api2/json/nodes/cx39/qemu/198/pending HTTP/1.1" 200 545
::ffff:10.124.31.44 - OpenTofu@pve!OpenTofu [02/10/2025:11:44:28 +0000] "GET /api2/json/nodes/cx39/qemu/198/status/current HTTP/1.1" 200 850

[TobiPeterG]

The 500 error also occurs with your current master version, so I don't think I caused it. Can this be confirmed?

My main.tf that I used for my provider version and your master version:

provider "proxmox" {
  pm_api_url      = "OURAPIURL"
  pm_api_token_id = "OpenTofu@pve!OpenTofu"
  pm_api_token_secret = "OURTOKEN"
  pm_tls_insecure = true
}

resource "proxmox_vm_qemu" "vm" {
  name         = "local-test"
  protection   = false
  target_nodes = ["OURNODENAME"]
  memory       = 4096
  scsihw       = "virtio-scsi-pci"
  clone        = "OURTEMPLATE"
  os_type      = "cloud-init"
  pool         = "OURPOOL"
  agent        = 1

  cpu {
    cores   = 4
    sockets = 1
    type    = "host"
  }

  network {
      id      = 0
      model   = "virtio"
      bridge  = "vmbr0"
      tag     = OURVLAN
  }

  disks {
    scsi {
      scsi0 {
        disk {
          size    = 40
          storage = "OURSTORAGE"
        }
      }
    }
    ide {
      ide2 {
        cdrom {
          iso = "OURSTORAGEPOOL:iso/OURISONAME.iso"
        }
      }
    }
  }
  lifecycle {
    ignore_changes = [
      agent_timeout,
      disks[0].ide[0].ide2[0].cdrom[0].iso,
      target_node,
      target_nodes,
      agent
    ]
  }
}

With this setup I also confirmed that VM creation works without problems with my provider and fails with yours.
VM deletion fails with both versions, but I'm very sure that's another issue.