I just noticed, that CUDA is broken with this update. Error: ptxas application ptx input, line 9; fatal : Unsupported .version 7.3; current version is '7.2'
Installed components: community/cuda 11.3.0-1 [installed] local/linux54-nvidia 460.80-1 (linux54-extramodules)
Selected linux59 and linux59-nvidia for deletion, selected linux510 and linux510-nvidia for installation and selected nvidia-utils for update. Transaction then runs with no errors.
After rebooting laptop, full system update runs as normal.
I had exactly same issue, updated with warning that nvidia dependancy breaks, after reboot black screen.
Now Im doing timeshift restore from tty (glad it works), but it takes so much time to restore, frequently throwing rsync errors Input/output (probably the reason is my old hdd), but it has no progress bar and no sign of how much is already restored and how much still have to be done. I’ve been running timeshift for 20+ minutes now and have no idea how much more it will be restoring all the data.
But I believe the issue was completely same - nvidia drivers, since I have nvidia 960 which has two display ports. So black screen from stable version - is not something good
Hope such things won’t be put in stable updates anymore.
Again my multiseat-setup was broken after this update. I do use nvidia-cards, but with noveau driver. 2 seats (each with 4k-resolution)
cat /var/log/Xorg.0.log [1]
[ 7.268]
X.Org X Server 1.20.11
X Protocol Version 11, Revision 0
[ 7.268] Build Operating System: Linux Manjaro Linux
[ 7.268] Current Operating System: Linux manjaro18 5.12.2-1-MANJARO #1 SMP PREEMPT Fri May 7 17:53:15 UTC 2021 x86_64
[ 7.268] Kernel command line: BOOT_IMAGE=/@/boot/vmlinuz-5.12-x86_64 root=UUID=3487ba3d-1cba-4cdc-a043-c420ebca2aca rw rootflags=subvol=@ bootsplash.bootfile=bootsplash-themes/manjaro/bootsplash modprobe.blacklist=ccp udev.log_priority=3 audit=0 sysrq_always_enabled=1
[ 7.268] Build Date: 13 April 2021 04:11:08PM
[ 7.268]
[ 7.269] Current version of pixman: 0.40.0
[ 7.269] Before reporting problems, check http://wiki.x.org
to make sure that you have the latest version.
[ 7.269] Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[ 7.269] (==) Log file: "/var/log/Xorg.0.log", Time: Sun May 30 18:46:30 2021
[ 7.269] (==) Using config directory: "/etc/X11/xorg.conf.d"
[ 7.269] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[ 7.269] (EE) No ServerLayout section called "seat0"
[ 7.269] (EE) Unable to determine the screen layout
[ 7.269] (EE) Error parsing the config file
[ 7.269] (EE)
Fatal server error:
[ 7.269] (EE) no screens found(EE)
[ 7.269] (EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
[ 7.269] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 7.269] (EE)
[ 7.269] (EE) Server terminated with error (1). Closing log file.
Lightdm refused to start
because x11 did not start
because seat0 was unknown (EE) No ServerLayout section called "seat0"
because there is a valid ServerLayout “layout” in /usr/share/X11/xorg.conf.d/10-amdgpu-nvidia-drm-outputclass.conf and in /usr/share/X11/xorg.conf.d/10-amdgpu-nvidia-drm-outputclass.conf
This prevents the seat management from loginctl to work properly. If there is no conf-File at all, seat-management will work out of the box. if there is any conf-File with an Section “ServerLayout” named “layout” this will not work any more, butt expect to have also a conf file with seat0 and seat1 (but i don´t have any)
After editing these files and adding a # in front of every line, seat management does work again
I’ve tested many things:
I rolled back with timeshift, it didn’t start either - black screen, same issue.
I switched monitor from DVI to HDMI and it started working (rolled back to the state before update)
Then I tried to update and the issue was that I have kernel 5.9 which is unsupported. On the update it removes linux59-nvidia and I end up with same black screen.
I tried to change kernel before updating but it fails because it breaks nvidia drivers dependency, because new kernels require new nvidia drivers.
I have no idea so far how to fix it, but I manually updated everything except nvidia drivers.
So far so good, but I’m kinda stuck with old and unsupported 5.9 kernel.
Hello, @poplach
Have you tried to install another kernel alongside 5.9, say 5.10 ??
1st Manjaro Kernels - Manjaro → install the 5.10 kernel its LTS sudo mhwd-kernel -i linux510 or use the GUI Tool. With th GUI its easier.
2nd I think you should manually install the appropriate drivers nvidia drivers that go with kernel 5.10 and your system.
3rd Boot to kernel 5.10. If everything is oki, you can remove kernel 5.9.
This is what I was trying to do.
The problem is if I do try install any kernel before last update - it can’t find old nvidia drivers, all mirrors result in 404 not found.
But if I try after the update, all current kernels require latest nvidia drivers and they come as a dependency of any kernel I try to install.
I was trying that via ‘Manjaro settings’ GUI, I’m not sure if I can do it manually, because it runs some hooks, rebuilds grub menu and does many other things I don’t know.
I guess I will post my recent experience and see it might help. I am running stable gnome for a few years, amd cpu with nvidia 1080 video card. I was running the nvidia drivers and X w/ gnome, not Wayland. Black screen forced me to install free video drivers and the only way back to my desktop was to use Wayland. I get a black screen trying to log into gnome on X.
Hopefully the next update will set things right.
So I decided to live on the edge and run a blind update after having had done no maintenance for a month or so.
Turns out reading all 4 forum threads in advance wouldn’t have helped me deal with the actual issue anyway.
Namely my boot failed but NOT due to the usual.
The following note in /etc/default/grub rather downplays the issue:
# Uncomment to ensure that the root filesystem is mounted read-only so that
# systemd-fsck can run the check automatically. We use 'fsck' by default, which
# needs 'rw' as boot parameter, to avoid delay in boot-time. 'fsck' needs to be
# removed from 'mkinitcpio.conf' to make 'systemd-fsck' work.
# See also Arch-Wiki: https://wiki.archlinux.org/index.php/Fsck#Boot_time_checking
GRUB_ROOT_FS_RO=true
It’s not a “delay” but rather BREAKAGE.
Long story short, I had to chroot and remove the fsck HOOK from /etc/mkinitcpio.conf and rebuild the images with mkinitcpio -P,
Which was uncool™.
I don’t really care that 2 subsystems try to do the same job and step on each other’s toes. This is still no reason to make a system unbootable.
Another update and another system breaking bug most likely caused by the nvidia drivers. So another day wasted trying to get the OS to boot up and work normally. Or I might wait for the next “stable” update and pray it works …
Thankfully Timeshift exists. Better use outdated system than not having an OS at all.
Finally got around updating and everything went fine and no immediate issues after reboot. Connected my hi-res (4K) monitor through HDMI and no issues as well.
My graphics setup seems to run fine with the latest update and nvidia drivers: