Bug 21583

*** Bug 21573 has been marked as a duplicate of this bug. ***

CC: (none) => Kn4ia

I can confirm this bug. Here is what I did:

In mcc I choose « Configure the boot of the system» - Then I get the message «No boot loader found. Creation of a new configuration.» – I press OK – Then I chose «grub2 graphic mode» and «System partition EFI» (there is no other option in the drop-down menu) – I press OK – Afterwards I configure the booting of the system – I press OK – Then it says that it is installing the boot loader and after a few seconds there is the above error message of «drakboot».

This happen on a system with a fresh install of mageia 6 in July. There haven't been any problems booting the system until very recently. Since then I can't configure the boot loader anymore.

Output of ll /sys/firmware:
total 0
drwxr-xr-x  6 root root 0 aoû 26 20:11 ./
dr-xr-xr-x 13 root root 0 aoû 26 20:11 ../
drwxr-xr-x  6 root root 0 aoû 26 18:11 acpi/
drwxr-xr-x  3 root root 0 aoû 26 18:11 dmi/
drwxr-xr-x  6 root root 0 aoû 26 20:11 efi/
drwxr-xr-x 24 root root 0 aoû 26 20:25 memmap/

Output of ll /boot/EFI
total 0
drwxr-xr-x 1 root root   0 mai 26 22:33 ./
drwxr-xr-x 1 root root 754 aoû 26 20:18 ../

I see the bugs 20023, 21573 and 19014 which concern all the same error message. My problem could also be related to one of these bugs.

CC: (none) => ed1

Please provide the output of:

df

and

su
gdisk -l /dev/sda

if you have more than one drive then also the output from /dev/sdb etc.

Thanks

Output of df
Sys. de fichiers blocs de 1K   Utilisé Disponible Uti% Monté sur
devtmpfs             8149436         0    8149436   0% /dev
tmpfs                8155300         0    8155300   0% /dev/shm
tmpfs                8155300      1232    8154068   1% /run
/dev/nvme0n1p6      23552000   9829024   13452544  43% /
tmpfs                8155300         0    8155300   0% /sys/fs/cgroup
tmpfs                8155300       872    8154428   1% /tmp
/dev/nvme0n1p12    226928640 140296596   85911020  63% /media/data
/dev/nvme0n1p11    107520000  78622216   28182648  74% /home
tmpfs                1631060         0    1631060   0% /run/user/983
tmpfs                1631060        24    1631036   1% /run

Output of gdisk -l /dev/sda
GPT fdisk (gdisk) version 1.0.1

Problem opening /dev/sda for reading! Error is 2.
The specified file does not exist!

Please also:

cat /etc/fstab

and:

su
gdisk -l /dev/nvme0n1

thanks

Output of cat /etc/fstab
/dev/nvme0n1p6 / btrfs noatime 1 1
/dev/nvme0n1p1 /boot/EFI vfat umask=000,iocharset=utf8,noauto 0 0
/dev/nvme0n1p11 /home btrfs noatime 1 2
/dev/nvme0n1p12 /media/data btrfs noatime 1 2
none /proc proc defaults 0 0
/dev/nvme0n1p5 swap swap defaults 0 0

Output ofgdisk -l /dev/nvme0n1
GPT fdisk (gdisk) version 1.0.1

Partition table scan:
  MBR: protective
  BSD: not present
  APM: not present
  GPT: present

Found valid GPT with protective MBR; using GPT.
Disk /dev/nvme0n1: 1000215216 sectors, 476.9 GiB
Logical sector size: 512 bytes
Disk identifier (GUID): A4D03F7C-87F1-45E8-BDBD-4615D9C3B378
Partition table holds up to 128 entries
First usable sector is 34, last usable sector is 1000215182
Partitions will be aligned on 2048-sector boundaries
Total free space is 2669 sectors (1.3 MiB)

Number  Start (sector)    End (sector)  Size       Code  Name
   1            2048          534527   260.0 MiB   EF00  EFI system partition
   2          534528          567295   16.0 MiB    0C01  Microsoft reserved ...
   3          567296       123447295   58.6 GiB    0700  Basic data partition
   4       999192576      1000214527   499.0 MiB   2700  Basic data partition
   5       966424576       999192575   15.6 GiB    8200  swap
   6       123447296       170551295   22.5 GiB    8300  linux1
   7       170551296       213559295   20.5 GiB    8300  linux2
   8       213559296       244279295   14.6 GiB    8300  linux3
   9       244279296       272951295   13.7 GiB    8300  linux4
  10       272951296       297527295   11.7 GiB    8300  linux5
  11       297527296       512567295   102.5 GiB   8300  home
  12       512567296       966424575   216.4 GiB   8300  data

Looks like /boot/EFI is failing to mount from fstab on boot.

Maybe an issue related to nvme.

Could you *attach* the boot.txt output file from:

journalctl -xb > boot.txt

Thanks

Created attachment 9641 [details]
Output of journalctl -xb

There is no mention of nvme0n1p1 in the log so I have no idea what is happening here.

Can you try to mount it manually at /boot/EFI and look for errors?

I'm not a developer and don't know how to do what you suggest. 

When I look at the partitions in mcc I found this information about the first partition:

Point de montage : /boot/EFI
Périphérique : nvme0n1p1
Nom du volume : SYSTEM
Type : EFI System Partition
Taille : 260Mio (0% du disque dur)

Maybe this might be helpful for you.

(In reply to Béat E from comment #11)
> I'm not a developer and don't know how to do what you suggest. 
> 
> When I look at the partitions in mcc I found this information about the
> first partition:
> 
> Point de montage : /boot/EFI
> Périphérique : nvme0n1p1
> Nom du volume : SYSTEM
> Type : EFI System Partition
> Taille : 260Mio (0% du disque dur)
> 
> Maybe this might be helpful for you.

Thanks,
Try this and report back any errors:

su
mount /dev/nvme0n1p1 /boot/EFI

Also, are you sure that you are still booting in UEFI mode?

mount /dev/nvme0n1p1 /boot/EFI doesn't create any error.

I don't know whether I am booting in UEFI mode. In BIOS secure boot control is disabled.

Ah OK, so if the output of:

df

shows that /dev/nvme0n1p1  is indeed now mounted at /boot/EFI then try running grub2-install and see if you get the error now:

su
grub2-install


Secure boot has nothing to do with UEFI except that UEFI is required for secure boot. Don't worry about that just now.

Yes, /dev/nvme0n1p1 is mounted at /boot/EFI. 

grub2-install works now without error and mageia boots with grub2 without any problem.

OK so the issue is not with grub2 but with the fact that /dev/nvme0n1p1 is not mounted at boot from the current fstab entry.

Now you have rebooted, check that it is now not mounted again (using df) and assuming it is not, manually mount it again as before.

Now use 'mcc -> boot -> set up boot system' to again re-install grub2 and check that too works without error.

If that goes OK then we can start looking at the real issue.

Please test this:
Run:
su
blkid /dev/nvme0n1p1

Now use the UUID= value found above in the next command (replace the XXXXXXX with the actual UUID.

su
sed -i 's:/dev/nvme0n1p1:UUID=XXXXXXX:' /etc/fstab

Now assuming no errors reboot and check whether /dev/nvme0n1p1 is mounted.

I did what you wrote. All worked fine. 

But now I cannot boot into the mageia system anymore. I am not very happy with this of course. What can I do to fix this?

Where does boot fail, and what do you see?
Do you have a live mageia (or another linux system on the machine) that you can boot into?
Can you boot directly from the UEFI bios firmware using maybe F12 (or similar) and selecting the Mageia entry there?

Created attachment 9645 [details]
Messages during boot

Created attachment 9646 [details]
Messages during boot - detail

I added the last messages before the boot stopped. The second image is just from the messages that weren't OK.

If you can get into recovery mode (from the grub2 menu) or emergency mode (which is offered in your last screenshot) then by entering your root password you may be able to use:

mc

to navigate to /etc/fstab and edit the entry for /boot/EFI back to it's original.
While there check that the changes were as below. Double check the UUID by using blkid again.

To test I think you changed the line (with the sed command) to:
UUID=1C36-B3F12 /boot/EFI vfat umask=000,iocharset=utf8,noauto 0 0

It was originally:
/dev/nvme0n1p1 /boot/EFI vfat umask=000,iocharset=utf8,noauto 0 0

Then see if it re-boots - changing this on a regular system (i.e. without nvme drive) does not cause any issues, either /dev/..., UUID or LABEL all work equally well in /etc/fstab.

I changed the mentioned line in fstab. This didn't help. There were two other lines that had the UUID:
UUID=1C36-B3F11 /home btrfs noatime 1 2
UUID=1C36-B3F12 /media/data btrfs noatime 1 2

I changed them back to
/dev/nvme0n1p11 /home btrfs noatime 1 2
/dev/nvme0n1p12 /media/data btrfs noatime 1 2

Now boot works properly again.

Re-reading this report from the top I'm wondering if this issue started after the last kernel update.

The last kernel update to Mageia 6 was around 19 August, does that sound like the right time?

If you get this to boot again, and assuming that you still have the previous kernels installed, please try booting the older ones and for each one check if /boot/EFI gets mounted.

(In reply to Béat E from comment #24)
> I changed the mentioned line in fstab. This didn't help. There were two
> other lines that had the UUID:
> UUID=1C36-B3F11 /home btrfs noatime 1 2
> UUID=1C36-B3F12 /media/data btrfs noatime 1 2
> 
> I changed them back to
> /dev/nvme0n1p11 /home btrfs noatime 1 2
> /dev/nvme0n1p12 /media/data btrfs noatime 1 2
> 
> Now boot works properly again.

Phew! Well spotted. My mistake of course, nvme0n1p11 contains nvme0n1p1 !!

So it's now booting OK with the UUID=.. for /boot/EFI ?

If so is /boot/EFI mounted at boot?

I tried the older kernels (4.9.40 and 4.9.35) and in mcc I see the following:

Point de montage : /boot/EFI
Périphérique : nvme0n1p1
Nom du volume : SYSTEM
Type : EFI System Partition
Taille : 260Mio (0% du disque dur)

In both cases this partition was not mounted.

Thanks for testing.
I am changing the summary of this bug and raising severity.
I am out of ideas but I am sure someone with more knowledge will pitch in.

Not yet seen any replies from the OP or Steve Malich from #12

Summary: drakboot crashed (grub2-install failed: Installing for x86_64-efi platform. grub2-install: error: cannot find EFI directory.) => /boot/EFI fails to mount from fstab on boot with NVMe drive
Severity: normal => major

And with the most recent kernel /boot/EFI isn't mounted neither. When I try to configure the boot process in mcc is say something like "Boot loader not found. Creation of a new configuration". I stop there because if I go further I will break the boot process again.

To get /boot/EFI mounted during boot, the noauto flag must be deleted in fstab.
The question is how it got in there in the first place.

CC: (none) => gm2.asp

If the noauto flag is the problem I must admit that I have set it so that this partition doesn't appear in the file manager. This means that this isn't a problem of mageia, but of my personal setting.