Grub rescue | error: checksum verification failed.

mathd8406

New Member
Mar 24, 2019
2
0
1
34
Hello,

I have been installing Proxmox for a week, after restarting Proxmox it remains stuck on the grub at startup, see screenshot added.

====================================================================
Error message : "error: checksum verification failed.
Entering recue mode ...
grub rescue > "
====================================================================

For information my hdd configuration is as follows : zfs (raidz -1).

Do you have any idea.

Thank you for your help.
 

Attachments

Last edited:

wolfgang

Proxmox Staff Member
Staff member
Oct 1, 2014
4,763
316
83
Hi,

what FS do you use as rootfs.
What have you changed/done?
 

mathd8406

New Member
Mar 24, 2019
2
0
1
34
Hello Wolfgang,

I'm not sure I understand the question.
I use as FS zfs.
What is rootfs? how to obtain this information?
No significant changes, just a reboot.

Thanks you for your help.

Sincerely,
 

Manfred Heubach

New Member
Nov 14, 2016
2
4
3
47
I had the same problem on two hosts lately.

grub rescue with debug=zfs and ls (hd0,gpt2) shows fletcher4 checksum errors.

I solved this by doing this from the proxmox installer (start installer in debug mode -> CTRL+D -> Accept License Agreement -> Abort):

zpool import -N -R /mnt rpool
zfs snapshot rpool/ROOT@move
zfs snapshot rpool/ROOT/pve-1@move
zfs create rpool/move
zfs rename rpool/ROOT rpool/move/ROOT
zfs send -v -R rpool/move/ROOT@move|zfs receive rpool/ROOT
zfs set canmount=off rpool/move/ROOT/pve-1
zpool export rpool

Just rebooted and all was fine again.

I have however no idea what causes the problem in the first place. A scrub doesn't find any invalid checksums. Only thing both hosts have in common is that they have been installed in december 2018 and january 2019. And yes - both are really cheap systems (Intel Atom, non ECC ram (though a 24h memtest ran without errors)). Both system where frozen and booted directly into grub rescue after being reset.

I hope this small solution helps everybody who runs into this problem.

Best regards
Manfred
 

exochris7

New Member
Mar 20, 2018
3
0
1
48
I had the same problem on two hosts lately.

grub rescue with debug=zfs and ls (hd0,gpt2) shows fletcher4 checksum errors.

I solved this by doing this from the proxmox installer (start installer in debug mode -> CTRL+D -> Accept License Agreement -> Abort):

zpool import -N -R /mnt rpool
zfs snapshot rpool/ROOT@move
zfs snapshot rpool/ROOT/pve-1@move
zfs create rpool/move
zfs rename rpool/ROOT rpool/move/ROOT
zfs send -v -R rpool/move/ROOT@move|zfs receive rpool/ROOT
zfs set canmount=off rpool/move/ROOT/pve-1
zpool export rpool

Just rebooted and all was fine again.

I have however no idea what causes the problem in the first place. A scrub doesn't find any invalid checksums. Only thing both hosts have in common is that they have been installed in december 2018 and january 2019. And yes - both are really cheap systems (Intel Atom, non ECC ram (though a 24h memtest ran without errors)). Both system where frozen and booted directly into grub rescue after being reset.

I hope this small solution helps everybody who runs into this problem.

Best regards
Manfred
Hi, same issue with one HP DL360e installed in the same period.
Your procedure save my install and lot of time.
Rgds
 

Beatus

New Member
Jun 11, 2019
1
0
1
53
I had the same problem on two hosts lately.

grub rescue with debug=zfs and ls (hd0,gpt2) shows fletcher4 checksum errors.

I solved this by doing this from the proxmox installer (start installer in debug mode -> CTRL+D -> Accept License Agreement -> Abort):

zpool import -N -R /mnt rpool
zfs snapshot rpool/ROOT@move
zfs snapshot rpool/ROOT/pve-1@move
zfs create rpool/move
zfs rename rpool/ROOT rpool/move/ROOT
zfs send -v -R rpool/move/ROOT@move|zfs receive rpool/ROOT
zfs set canmount=off rpool/move/ROOT/pve-1
zpool export rpool

Just rebooted and all was fine again.

I have however no idea what causes the problem in the first place. A scrub doesn't find any invalid checksums. Only thing both hosts have in common is that they have been installed in december 2018 and january 2019. And yes - both are really cheap systems (Intel Atom, non ECC ram (though a 24h memtest ran without errors)). Both system where frozen and booted directly into grub rescue after being reset.

I hope this small solution helps everybody who runs into this problem.

Best regards
Manfred
Hi! Thanks Manfred! Came accross this grub_rescue/fletcher4 boot issue on an old HP G8 and, while unable to access the pool from a proxmox installer console (complained about missing devices, damaged pool... not just one or two, but all the disks!) I was able to follow your steps booting a debian live cd.

Had previously attempted many other ways of getting the thing to boot; as it seemed to be a grub issue, but apparently not... very grateful to you for your post!

Just in case it's useful to anyone, my setup was a standard debian stretch install, with 4 x 3TB drives in two striped mirrors, zfs root, later turned into proxmox. Worked perfectly for some 6 months until a maintenance reboot...

Cheers!
 
I had the same problem on two hosts lately.

grub rescue with debug=zfs and ls (hd0,gpt2) shows fletcher4 checksum errors.

I solved this by doing this from the proxmox installer (start installer in debug mode -> CTRL+D -> Accept License Agreement -> Abort):

zpool import -N -R /mnt rpool
zfs snapshot rpool/ROOT@move
zfs snapshot rpool/ROOT/pve-1@move
zfs create rpool/move
zfs rename rpool/ROOT rpool/move/ROOT
zfs send -v -R rpool/move/ROOT@move|zfs receive rpool/ROOT
zfs set canmount=off rpool/move/ROOT/pve-1
zpool export rpool

Just rebooted and all was fine again.

I have however no idea what causes the problem in the first place. A scrub doesn't find any invalid checksums. Only thing both hosts have in common is that they have been installed in december 2018 and january 2019. And yes - both are really cheap systems (Intel Atom, non ECC ram (though a 24h memtest ran without errors)). Both system where frozen and booted directly into grub rescue after being reset.

I hope this small solution helps everybody who runs into this problem.

Best regards
Manfred
I can see everyone is saying it is possible to tshoot zfs with Proxmox ISO debugging mode, when I booted from it, it says zpool command not found
 

John Morrison

Member
Feb 26, 2018
35
4
8
Sweden
I had the same problem on two hosts lately.

grub rescue with debug=zfs and ls (hd0,gpt2) shows fletcher4 checksum errors.

I solved this by doing this from the proxmox installer (start installer in debug mode -> CTRL+D -> Accept License Agreement -> Abort):

zpool import -N -R /mnt rpool
zfs snapshot rpool/ROOT@move
zfs snapshot rpool/ROOT/pve-1@move
zfs create rpool/move
zfs rename rpool/ROOT rpool/move/ROOT
zfs send -v -R rpool/move/ROOT@move|zfs receive rpool/ROOT
zfs set canmount=off rpool/move/ROOT/pve-1
zpool export rpool

Just rebooted and all was fine again.

I have however no idea what causes the problem in the first place. A scrub doesn't find any invalid checksums. Only thing both hosts have in common is that they have been installed in december 2018 and january 2019. And yes - both are really cheap systems (Intel Atom, non ECC ram (though a 24h memtest ran without errors)). Both system where frozen and booted directly into grub rescue after being reset.

I hope this small solution helps everybody who runs into this problem.

Best regards
Manfred
Will this still work with the rpool/data being untouched? ie no snapshot?
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE and Proxmox Mail Gateway. We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get your own in 60 seconds.

Buy now!