After 3 Months I made again an update and the system does not reboot anymore.
Huge list of errors starting with failed to mount huge page file system.
The problem is described here:
But unfortunately I did not understand how to resolve it. I need to use old 5.4 kernel.
Can someone give advise how to get that system run again?
Best regards
Andreas
Hello @New2Mabox ,
Have you tried the method given here →
Sorry to say, but it is too high level.
I am able to boot with a stick and see the volume with the Mabox Linux now.
Should I open a terminal and just type ‘manjaro-chroot’ now ?
What means ‘login to my installation’ ?
manjaro-chroot -a
If manjaro-chroot is not available install it with:
yay -S manjaro-tools-base
(root password in mabox live is: mabox)
The most important question is: can you edit text files on the command line?
vim? nano?
If not, your chances to fix this yourself are slim 
I use nano etc/default/grub
I need to add two lines : SYSTEMD_CGROUP_ENABLE_LEGACY_FORCE=1 systemd.unified_cgroup_hierarchy=0" and remove the " after quiet udev.log_priority=3
right?
I have no idea…
Try what user on Manjaro forum suggested 
Good luck
sudo grub-install
fatal library error, lookup self
Installing for i386-pc platform
grub -install:error: install device isn’t specified
Do I need to specify the target /dev/sda2 ?
With path /dev/sda2 I get an error: will not proceed with blocklists
This is after warning that ext2 does not support embedding, so blocklists have to be used
I assume you did have GRUB? If so, you could have chosen for Advanced Option for Mabox in the GRUB menu and could have chosen the previous kernel (which worked). But maybe it’s not even the kernel but something else indeed
This is the background:
But I do not understand any of it.
Hi @New2Mabox ,
Suggestion: What about installing another kernel and try to boot from that one.
Detail from post:
boot into manjaro live usb, connect to internet, and chroot:
manjaro-chroot -a
install kernels:
pacman -S linux515 linux60 etc
if there are no errors, exit chroot:
exit
reboot
Command to see all kernels:
mhwd-kernel -l
available kernels:
good luck.
Thanks, but I think I need to stay with 5.4 kernel, it is an old Dell D620 with NVIDIA graphics. I was happy that Manjaro/Mabox was able to run it. Maybe I generate different live USB sticks and test other configurations (with my desktop pc) if I can find options - but this is time consuming. I was not able to update grub with that legacy force option. This might be a problem with my configuration, I have Boot-Us in MBR and GRUB is in /dev/sda2, which is EXT4. So I do not understand that error about ext2 and blocklists.
The problem with kernel 5.4 and cgroups is also strange, as the running system uses cgroup2:
mount | grep cgroup2
cgroup2 on /sys/fs/cgroup type cgroup2 (rw,nosuid,nodev,noexec,relatime,nsdelegate)
I could check this, since I could boot old winxp and play back a backup from June. Before going back, I ziped the home folder of the corrupt system. (and I made an image of that system).
To update the system now, what do I need to exclude in the update process? Or do I just need to edit etc/default/grub before starting the update?
There is no harm to try another kernel.
Accually, it is a good practise, for these kind of situations, to install more than one kernel versions, in case one is not bootin.
If another kernel is bootin, that is at least something.
From there you can search futher why kernel 5.4 is not working.
And maybe another kernel is just fine for your old hardware.
Since the problem is related to systemd, is there a method to downgrade this component in the installation with a usb linux live after manjaro-chroot -a ?
I think I need 255.7-1 version, since I found this in manjaro forum:
I did make the downgrade and everything is working now with systemd 255.7-1.
Boot live linux 5.4
sudo manjaro-chroot -a
sudo yay downgrade
sudo downgrade ‘systemd=255.7-1’ ‘systemd-sysvcompat=255.7-1’ ‘systemd-libs=255.7-1’ ‘lib32-systemd=255.7-1’