Grub 2:2.06.r322.gd9b4638c5-1 won't boot and goes straight to the BIOS after update

General system
21

I thought it’s my fault I did something wrong thats why I got broken angel - (my pc)

1 Like
22

i updated grub this morning but haven’t restarted yet so not sure if i was affected

23

Perhaps :enos: should follow the example of a certain other Arch based distro, and hold back updates for a couple of weeks, then release them in a bundle?

[Locks self in nuclear bunker :radioactive: to avoid missile strikes :rocketa:]

2 Likes
24

this bug doesn’t seem to reported on Arch bugreports

1 Like
25

I didn’t update the system until now. Shall I block GRUB in pacman before running pacman and wait for a fix?

1 Like
26

No :frowning: I was looking there just now. I’m in the fortunate position that I’ve downloaded the grub update with pacman -Syuw but not yet installed it. Given that I haven’t (yet!) experienced the bug, I didn’t feel best placed to report it.

27

Be a aware, be very aware… :imp: :stuck_out_tongue_winking_eye:

2 Likes
28

Same problem here, chroot and ran grub-mkconfig -o /boot/grub/grub.cfg didn’t work
Had to chroot, then downgrade grub and its working now.

5 Likes
29

I just restarted, no problem here, i have legacy bios

30

No issues on my end:

Lenovo Yoga Slim 7 | UEFI boot | AMD CPU/iGPU

Grub version: 2:2.06.r322.gd9b4638c5-1

31

First downgrade and then block.

1 Like
32

Step-by-step guide to downgrade grub:

Preparations: You need a bootable USB stick, see here.
Boot from the Stick
.
Open a Terminal and check your Disks:

$> fdisk -l

fdisk lists some entries. You need to identify the system disk with the defective grub, like:
/dev/sda2 618496 481648510 481030015 229,4G Linux-Filesystem

Prepare a mount-Target:

$> sudo mkdir /mnt/rescue

Mount the system-Drive into /mnt/rescue:

$> sudo mount /dev/sdaN /mnt/rescue

where sdaN is your drive that you identified above with fdisk (in my example above, it is sda2).

Mount more directories for a clean chroot:

$> sudo mount -t proc proc /mnt/rescue/proc
$> sudo mount -t sysfs sys /mnt/rescue/sys
$> sudo mount -o bind /dev /mnt/rescue/dev

And now change the environment with chroot:

$> chroot /mnt/rescue

Now you are in the environment to downgrade grub. To do this, list the pacman packages that are already installed, here we will limit ourselves to grub:

$> ls -l /var/cache/pacman/pkg/grub*
[...many files...]

i used grub-2:2.06.r297

downgrade grub with pacman:

$> sudo pacman -U /var/cache/pacman/pkg/grub-2:2.06.r297.g0c6c1aff2-1-x86_64.pkg.tar.zst

I got an error message because a file was not found, but the downgrade itself worked.
Leave the chroot-Environment:

$> exit

and unmount all:

$> sudo umount /mnt/rescue/dev
$> sudo umount /mnt/rescue/sys
$> sudo umount /mnt/rescue/proc
$> sudo umount /mnt/rescue

Restart and check if it works.
That should have been it.

Edit:
Please note the hints from dalto below!

7 Likes
33

A few things to be clear to others reading this:

  • You need modified instructions if you are using luks encryption, btrfs or lvm.
  • It is much simpler to use arch-chroot instead of manually creating a chroot environment and then having to unroll it all afterwards.
8 Likes
34

Just a tip: Most EndeavourOS installs ship with downgrade by default, if not you can fetch it from our repo. :wink:

2 Likes
35

I only update on Fridays :wink:

3 Likes
36

Using BTRFS, I took the following approach:

Your partition layout and device may differ 
sudo mount -o subvol=@ /dev/sda2 /mnt
sudo mount -o subvol=@log /dev/sda2 /mnt/var/log
sudo mount -o subvol=@cache /dev/sda2 /mnt/var/cache
sudo mount -o subvol=@home /dev/sda2 /mnt/home

sudo mount /dev/sda1 /mnt/boot/efi

sudo arch-chroot /mnt

sudo downgrade grub

Chose the package: 2:2.06.r297.g0c6c1aff2-1

Reboot.

7 Likes
37

It won’t hurt anything if you do, but you usually don’t need to mount /home when you chroot.

2 Likes
38

Thanks for the hints, I am not so Linux-experienced yet… I did not know ‘arch-chroot’ at all.

1 Like
39

Noticed the same, no reports at archlinux forum…
Can this be caused by changes tailored to EOS and/or Calamares??
Maybe @Bryanpwo, @dalto, @joekamprad can check this for us?

edit: I updated one of my systems just to check, it is borked :frowning:
It goes direct to hardware UEFI settings and no boot at all :frowning:

40

I’m a member of this club too… Followed your guide to repair and was successful. A big thank you!