### Terraform Version ```shell terraform version 1.5.7 provider[registry.terraform.io/hashicorp/vsphere] 2.3.1 vsphere 7.0.3-01900 open-vm-tools 2:12.2.0-1+deb12u2 govc 0.29.0 ``` ### Terraform Configuration Files ``` disk { label = "linux_system" size = local.system_disk.size thin_provisioned = local.system_disk.thin_provisioned keep_on_remove = local.system_disk.keep_on_remove eagerly_scrub = data.vsphere_virtual_machine.TEMPLATE.disks[0].eagerly_scrub } dynamic "disk" { for_each = var.optional_disks content { size = disk.value.size label = "linux_disk-${disk.key}" unit_number = disk.key + 1 thin_provisioned = disk.value.thin_provisioned keep_on_remove = disk.value.keep_on_remove } } ``` A sample VM was created with : - a 45G system disk, from a VM template (repartitionned using exec, upon first boot to extend existing partitions and create new ones) - a second 5G optional disk (same exec/paritionning treatment on first boot) - a third 10G optional disk (same exec/paritionning treatment on first boot) ### Debug Output The desired result should be : - sda should be the 45G disk - sdb the 5G disk - sdc the 10G disk But sometimes, the detected disk order is any variation of the following ``` # lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 5G 0 disk sdb 8:16 0 45G 0 disk ├─sdb1 8:17 0 487M 0 part /boot ├─sdb2 8:18 0 1K 0 part └─sdb5 8:21 0 9.5G 0 part ├─system-swap_1 254:0 0 976M 0 lvm [SWAP] ├─system-root 254:1 0 976M 0 lvm / ├─system-tmp 254:2 0 488M 0 lvm /tmp ├─system-srv 254:3 0 488M 0 lvm /srv ├─system-home 254:4 0 976M 0 lvm /home ├─system-var 254:5 0 976M 0 lvm /var ├─system-var+log 254:6 0 488M 0 lvm /var/log ├─system-var+tmp 254:7 0 488M 0 lvm /var/tmp ├─system-opt 254:8 0 976M 0 lvm /opt └─system-usr 254:9 0 2.8G 0 lvm /usr sdc 8:32 0 10G 0 disk └─sdc1 8:33 0 10G 0 part └─vg_sdc1-lv_varoptfobar 254:10 0 10G 0 lvm /var/opt/fobar ``` ### Expected Behavior The order of disks **should** stay the same Disk the order **must** be the order of declared disks in the resource ### Actual Behavior Power cycling the created VM in a loop gives the following filtered lsblk output : ``` ===== variation NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 10G 0 disk sdb 8:16 0 5G 0 disk sdc 8:32 0 45G 0 disk ===== variation NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 5G 0 disk sdb 8:16 0 45G 0 disk sdc 8:32 0 10G 0 disk ===== The most often seen (and desired) order NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 45G 0 disk sdb 8:16 0 5G 0 disk sdc 8:32 0 10G 0 disk ===== variation NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 45G 0 disk sdb 8:16 0 10G 0 disk sdc 8:32 0 5G 0 disk ===== variation NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 5G 0 disk sdb 8:16 0 10G 0 disk sdc 8:32 0 45G 0 disk ===== variation NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 10G 0 disk sdb 8:16 0 45G 0 disk sdc 8:32 0 5G 0 disk ``` Which shows that any disk order combination can actually happend on **power on** ! Note : once the VM is powered on, the disk order stays the same during reboots. It only may change on VM power cycle (on => off => on) ### Steps to Reproduce 1. create a vsphere vm with multiple disks 2. power off vm, wait, then power on 3. check disk order in guest 4. goto 2 ### Additional Context I have read the following links : - https://github.com/hashicorp/terraform/issues/10225 - https://knowledge.broadcom.com/external/article?legacyId=2011654 - https://www.codyhosterman.com/2020/03/whats-new-in-vsphere-7-0-storage-part-ii-guestinfo-virtualdiskmapping/ So far : - i could not fathom why or how this happens. - i could not find a way to get the order correctly everytime. And i have no clue how to solve this. ### References - #10225 ### Generative AI / LLM assisted development? _No response_