Mixed configuration UEFI/GRUB in Boot-Tool after disk replacement on non-uefi system

[MH_MUC]

Hi,
I have a server running pv 7.3 with a zfs raid1. I had to replace one disk.
I followed the https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_zfs_administration guide for replacing a bootable disk.

efibootmgr -v
EFI variables are not supported on this system.

My problem: After running the promox-boot-tool I have the following status:

Re-executing '/usr/sbin/proxmox-boot-tool' in new private mount namespace..
System currently booted with legacy bios
3E7D-B29C is configured with: uefi (versions: 5.15.83-1-pve, 5.15.85-1-pve, 5.4.195-1-pve)
A667-3C46 is configured with: grub (versions: 5.15.83-1-pve, 5.15.85-1-pve, 5.4.195-1-pve)

*29C is the old device that the system was booted from after replacement.

This seems to be messed up. Shouldn't the old drive also be configured as grub?
How can I make sure the system is bootable from the new disk? I would like to replace the other disk with a bigger one.

Thank you for you help!

 

[mr44er]

Shouldn't the old drive also be configured as grub?

No, both disks should be configured with uefi (uefi and grub partitions should co-exist, but uefi preferred).

EFI variables are not supported on this system.

I'm guessing the server is not bootet with uefi in the first place. Recheck that the server boots with uefi (not legacy or csm!) and be sure to boot from 3E7D-B29C. After boot check again proxmox-boot-tool status and efibootmgr -v. If it is bootet with uefi and 3E7D-B29C, kill A667-3C46 and repeat the steps from the guide.

Be careful and doubletriple-check the correct disk!

 

[MH_MUC]

Hi,.
Thank you for the quick reply.
The server is pretty old. I don't think the BIOS is supporting uefi. Besides it is in a Datacenter so it is kind of hard to change bios settings without an urgent need.

Maybe I have to reconfigure the existing drive.
I'll have to check when the server was installed. I think it was after 6.4 when the changes to the bootloader were implemented.

 

[mr44er]

Ah, this wasn't clear to me.

Besides it is in a Datacenter so it is kind of hard to change bios settings without an urgent need.

Mhm ok, and no remote access if testing to boot from the new disk fails is also problematic?

You can check with the GUI->disks. The new disk should have BIOS boot, EFI and ZFS partitions and the old only BIOS and ZFS.

But imho reboot should be tested on both after replacement. I would not want to find that out (if not) when the other disk fails in the near future unexpected.

 

[MH_MUC]

The old disk has the exact same layout as the new one as I copied the partition table as described in the manual.

Actually I got it solved now by ordering a kvm to make sure the server stays up and running. The server wasn't booted in UEFI-Mode for whatever reason. So I just booted in UEFI mode thorugh the BIOS boot order. Thereafter I was able to reinit the "old" disk.
Now proxmox boot tool shows

Re-executing '/usr/sbin/proxmox-boot-tool' in new private mount namespace..
System currently booted with uefi
ls: cannot access '/var/tmp/espmounts/3E7D-B29C/vmlinuz-*': No such file or directory
3E7D-B29C is configured with: uefi (versions: 5.15.83-1-pve, 5.15.85-1-pve, 5.4.195-1-pve), grub (versions: )
A667-3C46 is configured with: uefi (versions: 5.15.83-1-pve, 5.15.85-1-pve, 5.4.195-1-pve), grub (versions: 5.15.83-1-pve, 5.15.85-1-pve, 5.4.195-1-pve)

and both disks are bootable.
I don't know where the ls ... error is coming from, but the "old" disk will be replaced by a bigger one shortly so it doesn't matter.

Thank you for your help!

 

[mr44er]

Ah, perfect. UEFI works, not that old as expected

I don't know where the ls ... error is coming from, but the "old" disk will be replaced by a bigger one shortly so it doesn't matter.

Yes, that should resolve itself then. Or the next proxmox-boot-tool refresh, proxmox-boot-tool clean.