Bug 29933

Summary: drakboot cannot create bootmanager configuration
Product: Mageia Reporter: Markus Mertens <Mertens.MPT>
Component: Release (media or process)Assignee: Mageia Bug Squad <bugsquad>
Status: RESOLVED INVALID QA Contact:
Severity: major    
Priority: Normal CC: davidwhodgins, lewyssmith, mageia
Version: 8   
Target Milestone: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Source RPM: CVE:
Status comment:
Attachments: lspcidrake -v
blkid
journalctl output of last three days

Description Markus Mertens 2022-01-24 11:30:41 CET
Description of problem:

I tried to configure the bootmanager with mcc. First, mcc did not find any configuration and decided to create a new one. But when I finished with the dialog boxes it crashed.


How reproducible:

It happens always.


Steps to Reproduce:

1. Start mcc

2. Click on "Bootmanager konfigurieren" -> mcc will tell you that there is no configuration

3. It does not matter if you make changes or not. Once you finish drakboot will crash:

Das Programm "drakboot" ist mit folgender Fehlermeldung abgestürzt:

  grub2-install failed: x86_64-efi wird für Ihre Plattform installiert.
  File descriptor 9 (/dev/urandom) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 10 (/dev/urandom) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 21 (/proc/meminfo) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 22 (/proc/zoneinfo) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 23 (/proc/8297/cgroup) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 24 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.current) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 25 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.max) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 26 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.high) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 9 (/dev/urandom) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 10 (/dev/urandom) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 21 (/proc/meminfo) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 22 (/proc/zoneinfo) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 23 (/proc/8297/cgroup) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 24 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.current) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 25 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.max) leaked on vgs invocation. Parent PID 14743: grub2-install
  File descriptor 26 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.high) leaked on vgs invocation. Parent PID 14743: grub2-install
  grub2-install: Fehler: »/boot/efi/EFI/mageia/grubx64.efi« kann nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden.
  	...propagated at /usr/lib/libDrakX/any.pm line 278.
  	...propagated at /usr/libexec/drakboot line 49.
  Perl's trace:
  drakbug::bug_handler() called from /usr/libexec/drakboot:49

Benutztes Theme: Adwaita

Um einen Fehlerbericht abzugeben, klicken Sie auf den Knopf „Berichten“.
Dies wird ein Webbrowser-Fenster auf Bugzilla öffnen, in welchem Sie das Formular ausfüllen müssen. Die oben dargestellten Informationen werden auf den Server übertragen.
Es wäre nützlich, wenn Sie Ihrem Bericht die Ausgabe der folgenden Kommandos anfügen würden: 'lspcidrake -v', 'blkid'.
Sie sollten auch die folgenden Dateien hinzufügen: /etc/modprobe.conf, /etc/fstab, /boot/grub/menu.lst, /boot/grub/devices.map sowohl als auch /etc/lilo.conf.
Comment 1 Markus Mertens 2022-01-24 11:32:17 CET
Created attachment 13102 [details]
lspcidrake -v
Comment 2 Markus Mertens 2022-01-24 11:32:47 CET
Created attachment 13103 [details]
blkid
Comment 3 Lewis Smith 2022-01-27 22:03:38 CET
Thank you for the report.
It looks as if you have both LVM and raid?
Can you please describe the boot mechanism before you ran drakboot? And what you were trying to achieve/define? "mcc did not find any configuration" is strange. Had you defined booting at installation of Mageia?

Perhaps attach screenshots (window) at the start, and for each subsequent stage before the crash.

Can you please also post the output from:
 $ inxi -MSdopLxx

CC: (none) => lewyssmith

Comment 4 Markus Mertens 2022-03-29 12:50:23 CEST
I have two 1.8 TB nvme disks (nvme0n1, nvme1n1) with two partitions each:

first partition (nvme*n1p1) = 1 GB (EFI System Partition)
second partition (nvme*n1p2) = 1.8 TB (Linux RAID)

On the RAID I installed LVM for /home, /opt, /var and /.

The system boots and I can work with it. I can list the contents of /boot/EFI but the bootmanager seems to have problems reading /boot/EFI.



[root@cortex ~]# inxi -MSdopLxx
System:    Host: cortex Kernel: 5.15.28-desktop-1.mga8 x86_64 bits: 64 compiler: gcc v: 10.3.0 Console: tty 0 wm: kwin_x11 
           DM: GDM, LightDM, SDDM Distro: Mageia 8 mga8 
Machine:   Type: Desktop System: FUJITSU product: CELSIUS R970power v: N/A serial: EIAJ001384 Chassis: type: 3 v: CR970PF 
           serial: EIAJ001384 
           Mobo: FUJITSU model: D3488-A2 v: S26361-D3488-A2 serial: 95251033 UEFI: FUJITSU // American Megatrends 
           v: 5.0.0.14 R1.9.0 for D3488-A2x date: 07/19/2021 
Logical:   Device-1: VG: vg0 type: LVM2 size: 1.86 TiB free: 1010.6 GiB 
           LV-1: lv_home type: linear dm: dm-3 size: 512 GiB 
           Components: c-1: md0 pp-1: nvme0n1p2 pp-2: nvme1n1p2 
           LV-2: lv_opt type: linear dm: dm-2 size: 128 GiB 
           Components: c-1: md0 pp-1: nvme0n1p2 pp-2: nvme1n1p2 
           LV-3: lv_root type: linear dm: dm-0 size: 128 GiB 
           Components: c-1: md0 pp-1: nvme0n1p2 pp-2: nvme1n1p2 
           LV-4: lv_var type: linear dm: dm-1 size: 128 GiB 
           Components: c-1: md0 pp-1: nvme0n1p2 pp-2: nvme1n1p2 
Drives:    Local Storage: total: raw: 10.27 TiB usable: 8.41 TiB lvm-free: 1010.6 GiB used: 145.83 GiB (1.7%) 
           ID-1: /dev/nvme0n1 vendor: Micron model: MTFDHBA2T0TDV-1AZ1AABFA size: 1.86 TiB speed: 31.6 Gb/s lanes: 4 
           serial: 21052D45E452 
           ID-2: /dev/nvme1n1 vendor: Micron model: MTFDHBA2T0TDV-1AZ1AABFA size: 1.86 TiB speed: 31.6 Gb/s lanes: 4 
           serial: 21052D45E45D 
           ID-3: /dev/sda model: PRAID EP400i size: 6.55 TiB serial: N/A 
           Optical-1: /dev/sr0 vendor: HL-DT-ST model: BD-RE BU40N rev: PH07 dev-links: cdrom,cdrw,dvd,dvdrw 
           Features: speed: 24 multisession: yes audio: yes dvd: yes rw: cd-r,cd-rw,dvd-r,dvd-ram state: running 
Partition: ID-1: / size: 124.99 GiB used: 21.8 GiB (17.4%) fs: ext4 dev: /dev/dm-0 mapped: vg0-lv_root 
           ID-2: /boot/EFI size: 1021 MiB used: 75.6 MiB (7.4%) fs: vfat dev: /dev/nvme1n1p1 
           ID-3: /data size: 6.48 TiB used: 46.32 GiB (0.7%) fs: xfs dev: /dev/sda2 
           ID-4: /home size: 502.96 GiB used: 21.75 GiB (4.3%) fs: ext4 dev: /dev/dm-3 mapped: vg0-lv_home 
           ID-5: /opt size: 124.99 GiB used: 30.42 GiB (24.3%) fs: ext4 dev: /dev/dm-2 mapped: vg0-lv_opt 
           ID-6: /var size: 124.99 GiB used: 25.47 GiB (20.4%) fs: ext4 dev: /dev/dm-1 mapped: vg0-lv_var 
           ID-7: swap-1 size: 64 GiB used: N/A (0.0%) fs: swap priority: -2 dev: /dev/sda1 
Unmounted: ID-1: /dev/nvme0n1p1 size: 1 GiB fs: vfat label: SYSTEM uuid: F216-55AA 
[root@cortex ~]#
Comment 5 Dave Hodgins 2022-03-29 18:50:43 CEST
Why is /dev/nvme0n1p1 unmounted?

CC: (none) => davidwhodgins

Comment 6 Markus Mertens 2022-03-30 13:44:09 CEST
I don't know. According to my fstab it should be mounted:


[mer@cortex ~]$ mount | grep EFI
/dev/nvme1n1p1 on /boot/EFI type vfat (rw,relatime,fmask=0000,dmask=0000,allow_utime=0022,codepage=437,iocharset=utf8,shortname=mixed,utf8,errors=remount-ro)
[mer@cortex ~]$ cat /etc/fstab
/dev/vg0/lv_root / ext4 relatime,acl 1 1
# Entry for /dev/nvme0n1p1 :
UUID=F216-55AA /boot/EFI vfat iocharset=utf8,umask=000 0 0
# Entry for /dev/sda2 :
UUID=59b64bfb-49f7-4998-8ed6-fc940e20f0f6 /data xfs relatime 1 2
/dev/vg0/lv_home /home ext4 relatime,acl 1 2
cortex1:/mnt/analyses /mnt/analyses nfs wsize=8192,nosuid,rsize=8192,soft 0 0
cortex1:/mnt/imagepool /mnt/imagepool nfs wsize=8192,rsize=8192,nosuid,soft 0 0
/dev/vg0/lv_opt /opt ext4 relatime,acl 1 2
none /proc proc defaults 0 0
/dev/vg0/lv_var /var ext4 relatime,acl 1 2
# Entry for /dev/sda1 :
UUID=9361b318-0bcb-4538-b6ae-cd1129ede156 swap swap defaults 0 0
[mer@cortex ~]$
Comment 7 Markus Mertens 2022-03-30 14:05:25 CEST
Some additional information:

[root@cortex mer]#  efibootmgr -v | grep mageia
Boot0001* mageia        HD(1,GPT,82418868-d02c-4cbc-85be-35753ed578f9,0x800,0x200800)/File(\EFI\MAGEIA\GRUBX64.EFI)
Boot000F* mageia2       HD(1,GPT,a2629746-e386-47af-ba90-e3be965734f5,0x800,0x200800)/File(\EFI\MAGEIA\GRUBX64.EFI)
[root@cortex mer]# blkid | grep nvme
/dev/nvme0n1p1: LABEL_FATBOOT="SYSTEM" LABEL="SYSTEM" UUID="F216-55AA" BLOCK_SIZE="512" TYPE="vfat" PARTLABEL="COB" PARTUUID="82418868-d02c-4cbc-85be-35753ed578f9"
/dev/nvme0n1p2: UUID="5bf7cd42-f0b4-348a-8377-1ea7302cfa2c" UUID_SUB="fd7571b9-a5d5-54cd-d473-b735a6541526" LABEL="localhost:0" TYPE="linux_raid_member" PARTUUID="302427ea-45c8-4731-b0ce-82dae79646ee"
/dev/nvme1n1p1: LABEL_FATBOOT="SYSTEM" LABEL="SYSTEM" UUID="F216-55AA" BLOCK_SIZE="512" TYPE="vfat" PARTUUID="a2629746-e386-47af-ba90-e3be965734f5"
/dev/nvme1n1p2: UUID="5bf7cd42-f0b4-348a-8377-1ea7302cfa2c" UUID_SUB="79683ffe-3b6e-e065-73a1-8e4a12438ff8" LABEL="localhost:0" TYPE="linux_raid_member" PARTUUID="d171863e-4d3d-4451-8281-241cde6ab7e6"
[root@cortex mer]#
Comment 8 Dave Hodgins 2022-03-30 17:09:00 CEST
What's the outout of "journalctl --no-h |grep -e nvme0n1p1 -e F216-55AA -e /EFI"?
Comment 9 Dave Hodgins 2022-03-30 17:17:45 CEST
Also, /dev/nvme0n1p1 and /dev/nvme1n1p1 have duplicate uuids although the partuuids are unique. Any mount or boot parameter referencing them may get
either one. A Universally unique identifier must be UNIQUE.
Comment 10 Markus Mertens 2022-03-31 08:29:39 CEST
Created attachment 13211 [details]
journalctl output of last three days
Comment 11 Dave Hodgins 2022-03-31 21:58:15 CEST
With /dev/nvme0n1p1 mounted, what's the output of "df -h" and "df -hi"?
Comment 12 Dave Hodgins 2022-03-31 22:08:30 CEST
Never mind the df output, it's in the inxi output. Still looking through the
journalctl output.
Comment 13 Dave Hodgins 2022-03-31 22:10:37 CEST
Does running (as root) "update-grub2" work?
Comment 14 Dave Hodgins 2022-03-31 22:48:05 CEST Comment hidden (obsolete)
Comment 15 Dave Hodgins 2022-03-31 22:50:00 CEST
Also, the partitioning is a mess. Having two partitions using the same
mount point, and non-unique "unique" identifiers makes which partition
gets written to, or read from, un-predictable.
There's nvme1n1p1 and nvme0n1p1, both with UUID=F216-55AA.

Both are getting mounted to /boot/EFI at different times.

Which one is the UEFI system set to actually use?

While it's fine to backup the one that's in use, the backup must not have
the same "uuid" (aka VOLUME-ID for vfat file systems).
Comment 16 Markus Mertens 2022-04-04 21:46:14 CEST
The UEFI system should use nvme0n1p1. How can I change the UUID of the other disk (nvme1n1p1)?
Comment 17 Dave Hodgins 2022-04-04 22:42:54 CEST
Boot a live iso, to ensure it's not currently mounted or in use.

If you want, you can temporarily mount it, copy it's contents to a safe
location, and unmount it before reformatting it, then mount it again, and
copy the contents back.

Don't use diskdrake to reformat it, as it will keep the old uuid.
To reformat it, use "mkfs.vfat /dev/nvme1n1p1" (as root).

Mount the root file system and ensure it's /etc/fstab does either does not
reference the partition, or has it's new uuid before rebooting into the
installed system. Use the blkid command to determine the newly generated
uuid.

You can probably do this from the running system, provided you can ensure that
file system is not in use (partition not mounted). I'd do it from a live iso
so you know you can use that to fix things if anything does go wrong.
Comment 18 Markus Mertens 2022-04-05 08:23:26 CEST
[root@cortex ~]# blkid | grep SYSTEM
/dev/nvme0n1p1: LABEL_FATBOOT="SYSTEM" LABEL="SYSTEM" UUID="2BAF-216B" BLOCK_SIZE="512" TYPE="vfat" PARTLABEL="COB" PARTUUID="82418868-d02c-4cbc-85be-35753ed578f9"
/dev/nvme1n1p1: LABEL_FATBOOT="SYSTEM" LABEL="SYSTEM" UUID="F216-55AA" BLOCK_SIZE="512" TYPE="vfat" PARTUUID="a2629746-e386-47af-ba90-e3be965734f5"

[root@cortex ~]# mount | grep nvme
/dev/nvme0n1p1 on /boot/EFI type vfat (rw,relatime,fmask=0000,dmask=0000,allow_utime=0022,codepage=437,iocharset=utf8,shortname=mixed,utf8,errors=remount-ro)

[root@cortex ~]# cat /etc/fstab | grep UUID
UUID=2BAF-216B /boot/EFI vfat iocharset=utf8,umask=000 0 0
UUID=59b64bfb-49f7-4998-8ed6-fc940e20f0f6 /data xfs relatime 1 2
UUID=9361b318-0bcb-4538-b6ae-cd1129ede156 swap swap defaults 0 0

[root@cortex ~]# update-grub2
GRUB-Konfigurationsdatei wird erstellt …
Thema gefunden: /boot/grub2/themes/maggy/theme.txt
Linux-Abbild gefunden: /boot/vmlinuz-5.15.32-desktop-1.mga8
initrd-Abbild gefunden: /boot/initrd-5.15.32-desktop-1.mga8.img
Linux-Abbild gefunden: /boot/vmlinuz-5.15.28-desktop-1.mga8
initrd-Abbild gefunden: /boot/initrd-5.15.28-desktop-1.mga8.img
Linux-Abbild gefunden: /boot/vmlinuz-5.15.25-desktop-1.mga8
initrd-Abbild gefunden: /boot/initrd-5.15.25-desktop-1.mga8.img
Linux-Abbild gefunden: /boot/vmlinuz-5.15.6-desktop-2.mga8
initrd-Abbild gefunden: /boot/initrd-5.15.6-desktop-2.mga8.img
Linux-Abbild gefunden: /boot/vmlinuz-desktop
initrd-Abbild gefunden: /boot/initrd-desktop.img
Bootmenü-Eintrag für UEFI-Firmware-Einstellungen wird hinzugefügt …
abgeschlossen
[root@cortex ~]# 



But still mcc cannot configure the bootmanager


  grub2-install failed: x86_64-efi wird für Ihre Plattform installiert.
  File descriptor 9 (/dev/urandom) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 10 (/dev/urandom) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 21 (/proc/meminfo) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 22 (/proc/zoneinfo) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 23 (/proc/6483/cgroup) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 24 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.current) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 25 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.max) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 26 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.high) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 9 (/dev/urandom) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 10 (/dev/urandom) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 21 (/proc/meminfo) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 22 (/proc/zoneinfo) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 23 (/proc/6483/cgroup) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 24 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.current) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 25 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.max) leaked on vgs invocation. Parent PID 53350: grub2-install
  File descriptor 26 (/sys/fs/cgroup/user.slice/user-1000.slice/session-c2.scope/memory.high) leaked on vgs invocation. Parent PID 53350: grub2-install
  grub2-install: Fehler: »/boot/efi/EFI/mageia/grubx64.efi« kann nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden.
  	...propagated at /usr/lib/libDrakX/any.pm line 278.
  	...propagated at /usr/libexec/drakboot line 49.
  Perl's trace:
  drakbug::bug_handler() called from /usr/libexec/drakboot:49
Comment 19 Markus Mertens 2022-04-05 09:41:34 CEST
This is strange: grub2 looks for "grubx64.efi" in /boot/efi/EFI, but the file is located in /boot/EFI/efi:

[root@cortex mer]# grub2-install 
x86_64-efi wird für Ihre Plattform installiert.
grub2-install: Fehler: »/boot/efi/EFI/mageia/grubx64.efi« kann nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden.

[root@cortex mer]# find /boot -name grubx64.efi
/boot/EFI/efi/mageia/grubx64.efi
[root@cortex mer]# 

How can I correct this?
Comment 20 Dave Hodgins 2022-04-05 19:26:48 CEST
Checking on my uefi system ...
# ls -l /boot|grep EFI
drwxrwxrwx 3 root root     4096 Dec 31  1969 EFI/
lrwxrwxrwx 1 root root        3 May  9  2021 efi -> EFI/

Does the symlink exist on that system? It should have been created during
the installation of Mageia. If it's missing, re-add it with (as root)
# ln -s /boot/EFI /boot/efi
Comment 21 Markus Mertens 2022-04-06 13:35:41 CEST
I just re-added the link but it does not solve the problem. Now the "missing" file grubx64.efi exists in the followeing two paths:

/boot/efi/efi/mageia/
/boot/EFI/efi/mageia/

But drakboot/grub2-install look in:

/boot/efi/EFI/mageia/


  grub2-install failed: x86_64-efi wird für Ihre Plattform installiert.
  grub2-install: Fehler: »/boot/efi/EFI/mageia/grubx64.efi« kann nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden.
  	...propagated at /usr/lib/libDrakX/any.pm line 278.
  	...propagated at /usr/libexec/drakboot line 49.
  Perl's trace:
  drakbug::bug_handler() called from /usr/libexec/drakboot:49

  
[root@cortex boot]# grub2-install 
x86_64-efi wird für Ihre Plattform installiert.
grub2-install: Fehler: »/boot/efi/EFI/mageia/grubx64.efi« kann nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden.
[root@cortex boot]#
Comment 22 Dave Hodgins 2022-04-06 17:32:03 CEST
What's the output of "grep -i /boot/efi /etc/fstab" and
"grep -i /boot/efi /proc/mounts"?
Comment 23 Dave Hodgins 2022-04-06 18:15:18 CEST
Removing sysadmin-bugs from the cc list.

CC: sysadmin-bugs => (none)

Comment 24 Markus Mertens 2022-04-07 12:08:39 CEST
[root@cortex mer]# grep -i /boot/efi /etc/fstab
UUID=2BAF-216B /boot/EFI vfat iocharset=utf8,umask=000 0 0
[root@cortex mer]# grep -i /boot/efi /proc/mounts
/dev/nvme0n1p1 /boot/EFI vfat rw,relatime,fmask=0000,dmask=0000,allow_utime=0022,codepage=437,iocharset=utf8,shortname=mixed,utf8,errors=remount-ro 0 0
[root@cortex mer]#
Comment 25 Thomas Backlund 2022-04-07 15:12:39 CEST
(In reply to Markus Mertens from comment #21)
> I just re-added the link but it does not solve the problem. Now the
> "missing" file grubx64.efi exists in the followeing two paths:
> 
> /boot/efi/efi/mageia/
> /boot/EFI/efi/mageia/
> 
> But drakboot/grub2-install look in:
> 
> /boot/efi/EFI/mageia/
> 

Because that's according to spec.


the "second "efi" in path" _must_ be all caps

so 

/boot/efi/EFI/mageia

is correct
Comment 26 Dave Hodgins 2022-04-07 18:18:27 CEST
(In reply to Markus Mertens from comment #24)
> [root@cortex mer]# grep -i /boot/efi /etc/fstab
> UUID=2BAF-216B /boot/EFI vfat iocharset=utf8,umask=000 0 0
> [root@cortex mer]# grep -i /boot/efi /proc/mounts
> /dev/nvme0n1p1 /boot/EFI vfat
> rw,relatime,fmask=0000,dmask=0000,allow_utime=0022,codepage=437,
> iocharset=utf8,shortname=mixed,utf8,errors=remount-ro 0 0
> [root@cortex mer]#

Change the fstab entry to
UUID=2BAF-216B /boot/EFI vfat defaults,umask=000 0 0
then reboot
Comment 27 Markus Mertens 2022-04-08 09:05:38 CEST
This did not help. The system did not boot:

[root@cortex boot]# journalctl | grep Unrecognized
Apr 08 08:47:15 cortex kernel: FAT-fs (nvme0n1p1): Unrecognized mount option "default" or missing value
[root@cortex boot]# 

It is also starnge that I cannot create the directory "efi/EFI" because it exists. But when I try to list its content I am told that it is not found:

[root@cortex boot]# mkdir -p efi/EFI
mkdir: das Verzeichnis „efi/EFI“ kann nicht angelegt werden: Die Datei existiert bereits
[root@cortex boot]# ls efi/EFI
ls: Zugriff auf 'efi/EFI' nicht möglich: Datei oder Verzeichnis nicht gefunden
[root@cortex boot]#
Comment 28 Martin Whitaker 2022-04-08 10:10:13 CEST
(In reply to Markus Mertens from comment #27)
> It is also starnge that I cannot create the directory "efi/EFI" because it
> exists. But when I try to list its content I am told that it is not found:

This is because it is a FAT filesystem. The actual directory name is still a 8+3 character DOS filename, which is case insensitive. If you accessed that directory when running Windows, you could use either "efi" or "EFI". But for reasons I don't know, Linux only allows you to access it via the exact name you used when you created it.

Try

  mv efi/efi efi/tmp
  mv efi/tmp efi/EFI

CC: (none) => mageia

Comment 29 Markus Mertens 2022-04-08 11:34:17 CEST
Great! This is working. Even after a reboot drakconf finds a configuration and can write a new one.
Comment 30 Dave Hodgins 2022-04-08 16:27:01 CEST
(In reply to Markus Mertens from comment #27)
> This did not help. The system did not boot:
> 
> [root@cortex boot]# journalctl | grep Unrecognized
> Apr 08 08:47:15 cortex kernel: FAT-fs (nvme0n1p1): Unrecognized mount option
> "default" or missing value

copy/paste

UUID=2BAF-216B /boot/EFI vfat defaults,umask=000 0 0


Note that "default" is not the same as the above "defaults". The missing s
matters.

Glad it's working. I'm aware vfat uses case insensitive 8.3 names. I was
thinking replacing "iocharset=utf8" with "defaults" would make the difference.
Comment 31 Dave Hodgins 2022-04-08 16:28:23 CEST
Closing

Resolution: (none) => INVALID
Status: NEW => RESOLVED