Automatically removing orphans could lead to problems, list them out and do it manually, it’s only a bit of time but could potentially save you time.
I’m getting a systemd error on boot after the update:
Failed to insert module 'nvidia_uvm': No such device
System continues on afterwards and finishes booting.
For background, I am using vfio-pci to stub out my nvidia card for GPU passthrough to a VM. I don’t need to load any drivers for it. lspci -nnk shows these drivers are correctly in-use for the video card.
EDIT: I simply added nvidia, nvidia-drm, and nvidia-uvm to the blacklist, restarted systemd-modules-load successfully.
Also for some feedback on the LUKS encrypted drive solution, I ran update-grub and grub-install after adding the cryptodisk module to the grub config. Didn’t rerun mkinitcpio however since it didn’t make sense to me do so (the pacman hooks already run this, and the grub config doesn’t affect the kernel images). Rebooted fine save for the one error message.
May I make a suggestion for the partition scheme?
It is very nice that Manjaro includes Plymouth again. Plymouth CAN be particularly useful if the root partition is encrypted - because the graphical password prompt has a few advantages over the simple text input. (It also just looks better…)
In order for the graphical password prompt to work with EFI we need actually an unencrypted /boot-partition in addition the unencrypted /boot/efi and /root partitions.
I know there is supposed to be a potential security vulnerability with an unencrypted /boot. I’m not sure what it is, but other otherwise professional distributions (Fedora, Ubuntu, …) do this as well. So maybe this would be a nice addition to the installer for Manjaro?
I read the comments about Grub and blocked it for now. Than the update went smooth.
Thanks to that Great 
Community and the Manjaro Team.
It’s for sure a great job of the team to deal with all that hardware out there.
A Backup on a 1TB SSD before major updates is done in 15 minutes, no big deal at all and Timeshift should be active for the root partition anyway!
Among the Arch forks, I still believe that Manjaro is the best choice concerning “user friendliness”.
Long live Manjaro 
3 Likes
I am still having an issue with grub. I followed the solution for grub with LUKS and am left with a grub_debug_malloc not found error and dropped into grub rescue.
I followed solutions from
and
https://bbs.archlinux.org/viewtopic.php?id=281512
Decided to revert to pre 3.31 until this can be solved.
I had the issue with grub that went into rescue mode, but I was able to start the live ISO and chroot and fix the grub installation.
Update /etc/default/grub config to add cryptodisk module
GRUB_PRELOAD_MODULES="cryptodisk part_gpt part_msdos"
Then apply changes
sudo update-grub
sudo mkinitcpio -P
sudo grub-install --target=x86_64-efi --efi-directory=/boot --bootloader-id=manjaro # as I use UEFI and my ESP is /boot (where my EFI folder is located i.e: /boot/EFI)
After fixing the boot, I had to edit the desktops in edit mode → manage desktops and panels.
Somehow I had several panels in one desktop, and the secondary was swapped with a fresh config.
Also it seems that Wayland still have issues with nvidia on 5.27, so I’ll keep x11 for now
Noticed a new issue after the update: when I disconnect my displayport monitor, plasmashell segfaults and is restarted by systemd. System info:
Operating System: Manjaro Linux
KDE Plasma Version: 5.27.3
KDE Frameworks Version: 5.104.0
Qt Version: 5.15.8
Kernel Version: 5.10.176-2-MANJARO (64-bit)
Graphics Platform: X11
Processors: 12 × Intel® Core™ i5-10600K CPU @ 4.10GHz
Memory: 31.2 GiB of RAM
Graphics Processor: Mesa Intel® UHD Graphics 630
Manufacturer: Micro-Star International Co., Ltd.
Product Name: MS-7C75
System Version: 1.0
Before plasmashell crashes I see this in the journal logs:
kwin_x11[1460]: kwin_core: Parse error in tiles configuration for monitor "5b7a5ba4-43cc-5910-8e98-f358f9d54b43" : "illegal value" Creating default setup
I have 1 DP monitor, and 1 HDMI. No crashes if I turn off the HDMI monitor, only happens with displayport. Also occurs if I switch the DP monitor to another input.
EDIT: SOLVED. I added an entry to ~/.config/kwinrc with the “ghost” monitor uuid:
[Tiling][5b7a5ba4-43cc-5910-8e98-f358f9d54b43]
tiles={"layoutDirection":"horizontal","tiles":[{"width":0.25},{"width":0.5},{"width":0.25}]}
No more segfaults!
Thanks the the heads up for grub, and for bringing KDE Plasma 5.27 to stable!
Took care of all my “pacnew” files (pacman & grub) post update (before reboot), sudo update-grub, sudo mkinitcpio -P for good measure, rebooted, and all was well!
1 Like
Ditto! After sending it to the background, things have finally stopped looking far too tiny. 
Most people notice the error message that the module cryptodisk is not found and can press any key to continue the boot process. Others have a broken grub bootloader installation, either following the steps wrong or other reasons.
So here again:
/etc/default/grubneeds to been checked forGRUB_PRELOAD_MODULES="cryptodisk part_gpt part_msdos"andGRUB_ENABLE_CRYPTODISK=y- grub menu needs to been updated via
update-grub
That should be the only thing you may need. Installing grub to MBR/UEFI always should be the last resort unless you want to have security updates or the latest grub version on your system. If your system however has a broken bootloader installation you may need to restore it. Read the linked discussion thread for more info about further solutions.
Well, the solution you posted is still confusing as the issue actually is only a warning that the cryptodisk module is not found and people can still boot. Then they start to do stupid things to get that fixed which is “ain’t broken”. As stated above the /etc/default/grub needs to been “fixed” and the menu updated. Calling mkinitcpio makes no sense here as it only recreates the kernels and calls update-grub in the end. Also calling grub-install should be a last resort based on known reasons. Also the issue what we have with grub in general should be clear by now.
3 Likes
I wonder why Pamac looks completely different in X11 Plasma session vs in Wayland Plasma?
PS. I had a problem finally switching from X11 to Wayland and changing scaling in KDE: the window of Pamac became 4 times bigger than my screen, so I had to do some tricks to make it smaller.
Pama is still a GTK3 or GTK4 application. Depending which version you have installed. You can also check Discover and Octopi if you want more native looking Plasma Apps.
So, this update kinda made the XMPP Client Programm “Dino” like… white? Beforehand it was dark like the rest of my system - Settings are Breath Dark for native and Breeze for GTK Programms. Funny thing is the output of neofetch:
It does not mention Breath Dark in anyway, I think it did in the past - maybe wrong memory…
What is wrong here and why did the update trigger this?
The QT alternative didn’t really take off and isn’t really a big payoff for the effort I assume, but GTK is fine… and with native Breeze or Lightly theming (follows colour scheme) there’s not really much of a problem.
CLI sucks a bit because it’s not inclusive… for example, I had issues with shortwave, sorted in the Flatpak
Flatpak only shows in GUI.
well write down on how we should improve the CLI experience of pamac. You can look if others already mentioned stuff here: Issues · Applications / pamac · GitLab
1 Like
Two more workarounds, at least here on XFCE …
WIN + pamac-manager + Enter
WIN + add + Enter (matches Add/Remove Software)
1 Like
Reading as much as I could carefully before upgrading, and not wanted to be locked out of my system (which uses LUKS) I put cryptodisk and …CRYPTODISK=y into the grub config file. The update ran fine, and I don’t need to update-grub as that’s done by the update.
When rebooting, I still get the ‘Crypto disk not found’ error, but it continued booting. Do I need to reinstall grub? I notice that /boot/efi/EFI/boot/bootx64.efi has the same date that I first installed Manjaro.
regarding xsettingsd and hidpi and GTK: Thanks for this. Does it require configuration?
edit: No configuration needed. Install and reboot!
A small criticism and a suggestion for improvement.
With a friend who does not know so much about Linus, the update failed because he had a full hard disk.
Now nothing works and I do not have the time to fix the system. I will reinstall the system.
Why, manjaro does not recognize the available disk space and does not give an error message before the update that the disk space is too small? It shouldn’t even start downloading files unless there is at least 10GB of space.
Should implement sometime.
1 Like