proxmox_virtual_environment_vm clonevm has changes and restart vm

[SkipTam]

Describe the bug
Tested with the example template for clone-vm. Created a new resource and run again. It is trying to delete and change things. It does not delete or change anything it is trying to but it restart the vm.

To Reproduce
Steps to reproduce the behavior:

1. Create a resource from with clone-vm template. Adjusted to me
2. Run tofu apply
3. No error is shown
4. Run tofu apply
5. See changes that would be made
6. No changes are made but vm restarts

Please also provide a minimal Terraform configuration that reproduces the issue.

resource "proxmox_virtual_environment_vm" "terra-test" {
  name      = "terra-test"
  node_name = "pve"
  tags        = ["terraform", "adguard", "podman"]

  clone {
    vm_id = 1234
  }

  agent {
    enabled = true
  }
    memory {
    dedicated = 2048
  }
  
  initialization {
    ip_config {
      ipv4 {
        address = "10.112.10.36/24"
        gateway = "10.112.10.1"
      }
      ipv6 {
        address = "auto"
      }
    }
  }
}

and the output of terraform|tofu apply.

 ~ resource "proxmox_virtual_environment_vm" "terra-test" {
        id                                   = "101"
      ~ ipv4_addresses                       = [
          - [
              - "127.0.0.1",
            ],
          - [
              - "10.112.10.36",
            ],
          - [
              - "10.88.0.1",
            ],
          - [],
        ] -> (known after apply)
      ~ ipv6_addresses                       = [
          - [
              - "::1",
            ],
          - [
              - "THIS IS A IPV6",
              - "THIS IS A IPV6",
              - "fe80::THIS IS A IPV6",
            ],
          - [
              - "fe80::483a:b7ff:fea2:f732",
            ],
          - [
              - "fe80::9096:c3ff:fe6c:a68f",
            ],
        ] -> (known after apply)
        name                                 = "terra-test"
      ~ network_interface_names              = [
          - "lo",
          - "eth0",
          - "podman0",
          - "veth0",
        ] -> (known after apply)
        tags                                 = [
            "adguard",
            "podman",
            "terraform",
        ]
        # (26 unchanged attributes hidden)

      ~ initialization {
            # (2 unchanged attributes hidden)

          - user_account {
              - keys     = [
                  - "ssh-ed25519  KEYYYY ",
                ] -> null
              - password = (sensitive value) -> null
              - username = "test" -> null
            }

            # (1 unchanged block hidden)
        }

        # (5 unchanged blocks hidden)
    }

Expected behavior
No changes indicated and no reboot of the vm.

Additional context
I was thinking about adding everything instead of having it in the template. But this wouldn't really solve the problem, of deleting lo and 127.0.0.1. Not sure if i maybe have forgotten some flag or option somewhere.

If you need any more logs or something to try to fix, il be glad to help.

• Single or clustered Proxmox:
• Proxmox version: 9.0.10
• Provider version (ideally it should be the latest version): 0.87.0
• Terraform/OpenTofu version: OpenTofu v1.10.7
• OS (where you run Terraform/OpenTofu from): Arch Linux
• Debug logs (TF_LOG=DEBUG terraform apply):

2025-11-25T22:21:31.120+0100 [WARN]  Provider "registry.opentofu.org/bpg/proxmox" produced an unexpected new value for proxmox_virtual_environment_vm.terra-test during refresh.
      - .network_interface_names: new element 2 has appeared
      - .network_interface_names: new element 3 has appeared
      - .mac_addresses: new element 2 has appeared
      - .mac_addresses: new element 3 has appeared
      - .ipv4_addresses: new element 2 has appeared
      - .ipv4_addresses: new element 3 has appeared
      - .ipv6_addresses[1][0]: was cty.StringVal("fe80::THIS IS A IPV6"), but now cty.StringVal("THIS IS A IPV6")
      - .ipv6_addresses[1]: new element 1 has appeared
      - .ipv6_addresses[1]: new element 2 has appeared
      - .ipv6_addresses: new element 2 has appeared
      - .ipv6_addresses: new element 3 has appeared
2025-11-25T22:21:31.158+0100 [WARN]  Provider "provider[\"registry.opentofu.org/bpg/proxmox\"]" produced an invalid plan for proxmox_virtual_environment_vm.terra-test, but we are tolerating it because it is using the legacy plugin SDK.
    The following problems may be the cause of any confusing errors from downstream operations:
      - .timeout_create: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .reboot: planned value cty.False for a non-computed attribute
      - .stop_on_destroy: planned value cty.False for a non-computed attribute
      - .timeout_shutdown_vm: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .timeout_stop_vm: planned value cty.NumberIntVal(300) for a non-computed attribute
      - .pool_id: planned value cty.StringVal("") for a non-computed attribute
      - .timeout_move_disk: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .bios: planned value cty.StringVal("seabios") for a non-computed attribute
      - .keyboard_layout: planned value cty.StringVal("en-us") for a non-computed attribute
      - .purge_on_destroy: planned value cty.True for a non-computed attribute
      - .reboot_after_update: planned value cty.True for a non-computed attribute
      - .started: planned value cty.True for a non-computed attribute
      - .timeout_clone: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .timeout_migrate: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .delete_unreferenced_disks_on_destroy: planned value cty.True for a non-computed attribute
      - .on_boot: planned value cty.True for a non-computed attribute
      - .protection: planned value cty.False for a non-computed attribute
      - .template: planned value cty.False for a non-computed attribute
      - .timeout_reboot: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .tablet_device: planned value cty.True for a non-computed attribute
      - .timeout_start_vm: planned value cty.NumberIntVal(1800) for a non-computed attribute
      - .acpi: planned value cty.True for a non-computed attribute
      - .migrate: planned value cty.False for a non-computed attribute
      - .scsi_hardware: planned value cty.StringVal("virtio-scsi-pci") for a non-computed attribute
      - .agent[0].timeout: planned value cty.StringVal("15m") for a non-computed attribute
      - .agent[0].trim: planned value cty.False for a non-computed attribute
      - .agent[0].type: planned value cty.StringVal("virtio") for a non-computed attribute
      - .memory[0].floating: planned value cty.NumberIntVal(0) for a non-computed attribute
      - .memory[0].hugepages: planned value cty.StringVal("") for a non-computed attribute
      - .memory[0].keep_hugepages: planned value cty.False for a non-computed attribute
      - .memory[0].shared: planned value cty.NumberIntVal(0) for a non-computed attribute
      - .initialization[0].datastore_id: planned value cty.StringVal("local-lvm") for a non-computed attribute
      - .initialization[0].interface: planned value cty.StringVal("ide0") for a non-computed attribute
      - .initialization[0].ip_config[0].ipv6[0].gateway: planned value cty.StringVal("") for a non-computed attribute
      - .vga: block count in plan (1) disagrees with count in config (0)
      - .clone[0].retries: planned value cty.NumberIntVal(1) for a non-computed attribute
      - .clone[0].datastore_id: planned value cty.StringVal("") for a non-computed attribute
      - .clone[0].full: planned value cty.True for a non-computed attribute
      - .clone[0].node_name: planned value cty.StringVal("") for a non-computed attribute
      - .network_device: block count in plan (1) disagrees with count in config (0)

[SkipTam]

After further investigation with a friend, we have tested a couple of things. It seems that, if you change something in the "high level", this error occurs. For example the ip address in cloud init, but do not add the adapter in the config, it will try to change it. Im guessing this is normal terraform. This is my first adventure into terraform, so this might be a desired state.