grub rescue lvmid not found

[misterbreen]

Hi. I've very new to proxmox, and following some hubris, I seem to have shot myself in the foot.
So, in the last week:

1. Found an old Dell with 2x250gb storage drives.
2. Downloaded and installed the latest Proxmox VE.
3. Installed and combined the two drives into one 500gb volume
4. (I should have perhaps attempted a reboot of the system at this point but didnt).
5. Went on to install/create 3 new VMs.
6. (all is going well).
7. Followed some internet instructions to create a pve-no-subscription.list and did a dist-upgrade.
8. (didnt pay a lot of attention to the console output).
9. shutdown the 3 VMs and typed 'reboot'
10. grub then reports.... 'grub rescue lvmid ...... not found'
11. (sigh - I shouldnt have performed step 7).
12. with the original install media I can select 'rescue' and proxmox boots just fine and I can start VMs again.
13. without the 'rescue' selection on boot, I just get the original grub lvmid error when I start the Dell.

Feels like there should be a quick-ish fix for this. Any help?

I'll be a bit more careful in the future...

 

[fiona]

Hi,
you might've run into this bug. Please follow the steps described here and if there is an error when you use update-grub, change your LVM as suggested here (I think e.g. creating a new LV might also count as a change) and try again with update-grub.

 

[misterbreen]

So, I did attempt to follow your instructions, with no joy (perhaps someone more experienced with Proxmox could have sorted it).
The next step was a reinstall of proxmox - but the grub error has returned

I've a Dell with two 250gb drives.

1. I installed to /dev/sda
2. Post install:
   1. In the GUI, I delete local-lvm (suggested 15m into this video https://www.youtube.com/watch?v=_u8qTN3cCnQ)
      1. This apparently gives me more space for VMs
   2. Then in the shell.
      1. lvremove /dev/pve/data
      2. fdisk /dev/sdb
         1. .... etc ...
      3. vgextend pve /dev/sdb1
      4. lvresize -l +100%FREE /dev/pve/root
      5. resize2fs /dev/mapper/pve-root
3. I now have a single ~500gb storage for VMs.

I can now perform a reboot and I get the original grub error

Am I doing something wrong here? Or is my error due to the bug (in post #2), which I can't seem to fix.

Any tips on how I can setup my proxmox so that I can reboot after install (after I combine my drives into a single storage volume)?
Or should I wait for the next version of Proxmox VE (when the bug should be fixed).

Thanks.

 

[fiona]

Regardless if you are affected by the bug or not, you might need to run update-grub after adding the second disk/resizing the root file system.

If you are affected by the bug, update-grub should throw errors, and so should grub-probe -d /dev/mapper/pve-root -t fs.

 

[misterbreen]

Hi.
Went through the install and 5 steps above again.
Added a step 6 of the update-grub and still get the same error

 

[fiona]

I was able to reproduce the issue. You need to (also?) run grub-install /dev/sda.

 

[misterbreen]

No - but that will be my next attempt
Thanks for the support.
I'll update you after that.
Thanks

 

[misterbreen]

Result!
The update-grub and grub-install did the trick.
Thanks for your help.

 

[JL17]

Hi,
you might've run into this bug. Please follow the steps described here and if there is an error when you use update-grub, change your LVM as suggested here (I think e.g. creating a new LV might also count as a change) and try again with update-grub.

Hi fiona, this grub bug yo mention persists on proxmox v8?
Is there any workaround to avoid the bug?

Thanx in advance!

 

[fiona]

Hi,

Hi fiona, this grub bug yo mention persists on proxmox v8?

it should be fixed in Proxmox VE 8/Debian 12 with grub2(-common) >= 2.06-8.1. From the changelog:

grub2 (2.06-8.1) experimental; urgency=medium

  * Non-maintainer upload.
  * Fix an issue where a logical volume rename would lead grub to fail to
    boot (Closes: #987008)

 -- Antoine Beaupré <anarcat@debian.org>  Sat, 25 Feb 2023 15:16:55 -0500

Is there any workaround to avoid the bug?

While it is rare, I think the only way would be to not create/destroy/resize/rename LVM volumes. Or always check with grub-probe after such an operation, that it's still safe to reboot.

 

[JL17]

Thanx for the info Fiona!
Very usefull

 

[troykelly]

Hi,

it should be fixed in Proxmox VE 8/Debian 12 with grub2(-common) >= 2.06-8.1. From the changelog:

grub2 (2.06-8.1) experimental; urgency=medium

  * Non-maintainer upload.
  * Fix an issue where a logical volume rename would lead grub to fail to
    boot (Closes: #987008)

 -- Antoine Beaupré <anarcat@debian.org>  Sat, 25 Feb 2023 15:16:55 -0500

While it is rare, I think the only way would be to not create/destroy/resize/rename LVM volumes. Or always check with grub-probe after such an operation, that it's still safe to reboot.

This issue still exists with the current install iso (as of today) using `proxmox-ve_8.0-2.iso`

 

[fiona]

Hi,

This issue still exists with the current install iso (as of today) using `proxmox-ve_8.0-2.iso`

the ISO ships with a more recent version of GRUB, i.e. 2.06-13, which does include the fix mentioned above. Please describe your issue in detail, it might be something else. What is the exact error you get? After what steps?

 

[troykelly]

Thanks @fiona

The error is: (screenshot)
```text
Welcome to GRUB!

error: disk 'lvmid/i2s3IL-GIe3-0COx-ENc6-vKwA-aJP5-QcUNCb/6V3eDM-XdD8-nYGx-cpyK-K96d-u0Dp-Dzfvhv' not found.
grub rescue> $$
```

Clean install of the current iso. (recording)

 

[fiona]

Thanks @fiona

The error is: (screenshot)
```text
Welcome to GRUB!

error: disk 'lvmid/i2s3IL-GIe3-0COx-ENc6-vKwA-aJP5-QcUNCb/6V3eDM-XdD8-nYGx-cpyK-K96d-u0Dp-Dzfvhv' not found.
grub rescue> $$
```

Clean install of the current iso. (recording)

Since it happens directly after install and not at some later time when changes to the LVM metadata happen, I would suspect it to be a different issue. Can you install an other OS successfully? E.g. you can try going with https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm