Old LVM PV headers can cause conversion hang

[ispcolohost]

If you've converting a system that began as CentOS 7, converted to Alma 8, now converting to Alma 9, you may run into this:

[    2.718678] dracut-initqueue[687]:   WARNING: PV /dev/sda2 in VG centos_var is using an old PV header, modify the VG to update.
[    3.405633] upgrade[731]: starting upgrade hook
[    3.416906] upgrade[731]: /bin/upgrade: line 20: /sysroot/var/tmp/system-upgrade.state: No such file or directory
[    3.417550] upgrade[731]: Storage initialisation: Attempt 0 of 11. Wait 15 seconds.
[   18.432834] upgrade[776]:   WARNING: PV /dev/sda2 in VG centos_var is using an old PV header, modify the VG to update.
[   18.471917] upgrade[776]:   4 logical volume(s) in volume group "centos_var" now active

and then it hangs forever. I gave it 12 hours before breaking out. The system still lets you boot back to the Alma 8 kernel instead of the leapp upgrade. I upgraded that volume header via:

# vgck --updatemetadata centos_var
  WARNING: PV /dev/sda2 in VG centos_var is using an old PV header, modify the VG to update.
  WARNING: updating PV header on /dev/sda2 for VG centos_var.

Executing vgck with no arguments will reveal this issue, so the affected VG's can be updated:

# vgck
  WARNING: PV /dev/sda2 in VG centos_var is using an old PV header, modify the VG to update.

After correction it allowed the conversion to occur without hanging. It ultimately still failed from Plesk's perspective due to the other ongoing issues, but the system did end up as Alma 9 so I'll add this to my own pre-upgrade script and keep working on the other issues.