I’ve successfully installed the latest Nvidia beta driver (v545.23.06) a couple of days ago. All went fine during the installation.
However, I prefer Wayland over X11 and the Wayland experience with this beta driver was much worse than with the production branch driver, so I decided to uninstall the beta driver and reinstall the latest nvidia production driver (v535.XX.XX).
Now my issue starts. The script with mhwd fails to install the driver and I don’t know how to solve it. Please see the outputs:
sudo mhwd -a pci nonfree 0300 ✔ 10s
> Using config 'video-nvidia' for device: 0000:2d:00.0 (0300:10de:2204) Display controller nVidia Corporation GA102 [GeForce RTX 3090]
> Installing video-nvidia...
Sourcing /etc/mhwd-x86_64.conf
Has lib32 support: true
Sourcing /var/lib/mhwd/db/pci/graphic_drivers/nvidia/MHWDCONFIG
Processing classid: 0300
Sourcing /var/lib/mhwd/scripts/include/0300
Processing classid: 0302
:: Paketdatenbanken werden synchronisiert …
core wird heruntergeladen …
extra wird heruntergeladen …
community wird heruntergeladen …
multilib wird heruntergeladen …
Abhängigkeiten werden aufgelöst …
Nach in Konflikt stehenden Paketen wird gesucht …
Warnung: Abhängigkeits-Zyklus entdeckt:
Warnung: eglexternalplatform wird vor seiner Abhängigkeit nvidia-utils installiert werden
Pakete (7) egl-wayland-2:1.1.13-1 eglexternalplatform-1.1-2 lib32-nvidia-utils-535.113.01-1 libxnvctrl-535.113.01-1 linux65-nvidia-535.113.01-10 nvidia-settings-535.113.01-1 nvidia-utils-535.113.01-2
Gesamtgröße der installierten Pakete: 939,49 MiB
:: Installation fortsetzen? [J/n]
Schlüsselbund wird geprüft …
Paketintegrität wird geprüft …
Paket-Dateien werden geladen …
Auf Dateikonflikte wird geprüft …
Fehler: Der Vorgang konnte nicht durchgeführt werden (In Konflikt stehende Dateien)
egl-wayland: /usr/lib/libnvidia-egl-wayland.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libEGL_nvidia.so.0 existiert im Dateisystem
nvidia-utils: /usr/lib/libGLESv1_CM_nvidia.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libGLESv2_nvidia.so.2 existiert im Dateisystem
nvidia-utils: /usr/lib/libGLX_nvidia.so.0 existiert im Dateisystem
nvidia-utils: /usr/lib/libcuda.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libcudadebugger.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvcuvid.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-allocator.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-cfg.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-encode.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-fbc.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-ml.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-ngx.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-nvvm.so.4 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-opticalflow.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvidia-ptxjitcompiler.so.1 existiert im Dateisystem
nvidia-utils: /usr/lib/libnvoptix.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libEGL_nvidia.so.0 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libGLESv1_CM_nvidia.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libGLESv2_nvidia.so.2 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libGLX_nvidia.so.0 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libcuda.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvcuvid.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-allocator.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-encode.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-fbc.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-ml.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-opticalflow.so.1 existiert im Dateisystem
lib32-nvidia-utils: /usr/lib32/libnvidia-ptxjitcompiler.so.1 existiert im Dateisystem
Fehler sind aufgetreten, keine Pakete wurden aktualisiert.
Error: pacman failed!
Error: script failed!
If anybody could help me to get out of my misery I’d be so thankful
Have a great day boys!
Oh and this is how I successfully removed the previous driver:
When posting terminal output, copy the output and paste it here, wrapped in three (3) backticks, beforeANDafter the pasted text. Like this:
```
pasted text
```
Or three (3) tilde signs, like this:
~~~
pasted text
~~~
This will just cause it to be rendered like this:
Sed
sollicitudin dolor
eget nisl elit id
condimentum
arcu erat varius
cursus sem quis eros.
Instead of like this:
Sed sollicitudin dolor eget nisl elit id condimentum arcu erat varius cursus sem quis eros.
Alternatively, paste the text you wish to format as terminal output, select all pasted text, and click the </> button on the taskbar. This will indent the whole pasted section with one TAB, causing it to render the same way as described above.
Thereby increasing legibility thus making it easier for those trying to provide assistance.
For more information, please see:
Additionally
If your language isn’t English, please prepend any and all terminal commands with LC_ALL=C. For example:
LC_ALL=C bluetoothctl
This will just cause the terminal output to be in English, making it easier to understand and debug.
Personally, I think this is the command to use to overwrite all conflicts:
I have only installed the latest Beta drivers with the .run file that Nvidia officially provides. Then I just uninstalled them and cannot reinstall stable drivers now with mhwd.
Thank you. Using the official Nvidia .run installer is generally not recommended, however, I’ve had no issue with them over the years (in other distributions). It seems to be that the official drivers were not uninstalled correctly, leaving many files behind in its wake. If that’s the case, someone else will need to help with that. Did you reboot after installing/uninstalling?
The output from the following might also be helpful:
Yeah that definitely might be the case, as I didn’t reboot yet, because after uninstalling the beta drivers I wasn’t able to reinstall any other drivers. So if I was to reboot now I’d boot a system with no graphics drivers, am i right?
LC_ALL=C sudo mhwd -li INT ✘ 36s
Warning: No installed PCI configs!
Warning: No installed USB configs!
LC_ALL=C sudo mhwd -l ✔
> 0000:2d:00.0 (0300:10de:2204) Display controller nVidia Corporation:
--------------------------------------------------------------------------------
NAME VERSION FREEDRIVER TYPE
--------------------------------------------------------------------------------
video-nvidia 2023.03.23 false PCI
video-nvidia-470xx 2023.03.23 false PCI
video-linux 2018.05.04 true PCI
video-modesetting 2020.01.13 true PCI
video-vesa 2017.03.12 true PCI
> 0000:27:00.0 (0200:10ec:8168) Network controller Realtek Semiconductor Co., Ltd.:
--------------------------------------------------------------------------------
NAME VERSION FREEDRIVER TYPE
--------------------------------------------------------------------------------
network-r8168 2023.09.12 true PCI
LC_ALL=C sudo inxi -G ✔
Graphics:
Device-1: NVIDIA GA102 [GeForce RTX 3090] driver: nvidia v: 545.23.06
Device-2: Logitech BRIO 4K Stream Edition
driver: hid-generic,snd-usb-audio,usbhid,uvcvideo type: USB
Display: server: X.Org v: 23.2.2 with: Xwayland v: 23.2.2 driver: X:
loaded: nvidia gpu: nvidia,nvidia-nvswitch resolution: 1: 1920x1080~60Hz
2: 1920x1080~60Hz
API: EGL v: 1.5 drivers: nvidia platforms: gbm
API: OpenGL v: 4.6.0 vendor: nvidia v: 545.23.06 renderer: NVIDIA GeForce
RTX 3090/PCIe/SSE2
API: Vulkan v: 1.3.269 drivers: nvidia surfaces: xcb,xlib
LC_ALL=C journalctl -xe ✔
-- Subject: A start job for unit systemd-coredump@1-40862-0.service has finished successfully
-- Defined-By: systemd
-- Support: https://forum.manjaro.org/c/support
--
-- A start job for unit systemd-coredump@1-40862-0.service has finished successfully.
--
-- The job identifier is 5490.
Oct 26 14:55:43 PowerTower systemd-coredump[40863]: Resource limits disable core dumping for process 40860 (eglinfo).
Oct 26 14:55:43 PowerTower systemd-coredump[40863]: [LNK] Process 40860 (eglinfo) of user 0 dumped core.
-- Subject: Process 40860 (eglinfo) dumped core
-- Defined-By: systemd
-- Support: https://forum.manjaro.org/c/support
-- Documentation: man:core(5)
--
-- Process 40860 (eglinfo) crashed and dumped core.
--
-- This usually indicates a programming error in the crashing program and
-- should be reported to its vendor as a bug.
Oct 26 14:55:43 PowerTower systemd[1]: systemd-coredump@1-40862-0.service: Deactivated successfully.
-- Subject: Unit succeeded
-- Defined-By: systemd
-- Support: https://forum.manjaro.org/c/support
--
-- The unit systemd-coredump@1-40862-0.service has successfully entered the 'dead' state.
Oct 26 14:55:43 PowerTower sudo[40832]: pam_unix(sudo:session): session closed for user root
Oct 26 14:55:49 PowerTower kernel: [UFW BLOCK] IN=enp39s0 OUT= MAC=01:00:5e:00:00:01:98:9b:cb:10:a2:48:08:00 SRC=192.168.178.
Also, I thought it might help to install nvidia-utils manually as there was an error referring to it previously, now that’s the output:
LC_ALL=C sudo mhwd --force --auto pci nonfree 0300 1 ✘
> Using config 'video-nvidia' for device: 0000:2d:00.0 (0300:10de:2204) Display controller nVidia Corporation GA102 [GeForce RTX 3090]
> Installing video-nvidia...
Sourcing /etc/mhwd-x86_64.conf
Has lib32 support: true
Sourcing /var/lib/mhwd/db/pci/graphic_drivers/nvidia/MHWDCONFIG
Processing classid: 0300
Sourcing /var/lib/mhwd/scripts/include/0300
Processing classid: 0302
:: Synchronizing package databases...
core downloading...
extra downloading...
community downloading...
multilib downloading...
warning: nvidia-utils-535.113.01-2 is up to date -- skipping
warning: nvidia-settings-535.113.01-1 is up to date -- skipping
warning: libxnvctrl-535.113.01-1 is up to date -- skipping
resolving dependencies...
looking for conflicting packages...
Packages (2) lib32-nvidia-utils-535.113.01-1 linux65-nvidia-535.113.01-10
Total Installed Size: 263.99 MiB
:: Proceed with installation? [Y/n]
checking keyring...
checking package integrity...
loading package files...
checking for file conflicts...
error: failed to commit transaction (conflicting files)
lib32-nvidia-utils: /usr/lib32/libEGL_nvidia.so.0 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libGLESv1_CM_nvidia.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libGLESv2_nvidia.so.2 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libGLX_nvidia.so.0 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libcuda.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvcuvid.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-allocator.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-encode.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-fbc.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-ml.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-opticalflow.so.1 exists in filesystem
lib32-nvidia-utils: /usr/lib32/libnvidia-ptxjitcompiler.so.1 exists in filesystem
Errors occurred, no packages were upgraded.
Error: pacman failed!
Error: script failed!
That worked! I’m going to reboot now, thank you so much!!
However, I’m wondering a bit why it still shows the previous (beta) driver I installed with the .run file? Or is this going to be different after a reboot?
That’s true, but I still have some “weird” issues going on. The Wayland session was buggy as hell, barely usable and on X11 I have some weird graphical artifacts.
Wayland in Gnome, and on Nvidia: seems much as you describe; from comments I’ve noticed. I don’t currently use Gnome or Nvidia, but on KDE, Wayland performs admirably; albeit with the occasional hiccup.
The artifacts you mention; do you mean screen tearing or something else?
The error you encountered “ Failed to initialize NVML: Driver/library version mismatch” happens when you have different versions of the NVIDIA driver installed in your system at the same time, or when the driver has been updated without restarting the machine afterward.
And considering you have rebooted, apparently, I suspect that driver you had installed is causing kind of a lot of problems. This is a good example of why you should only use mhwd and not proprietary files/scripts/whatever else there is…
I know, as I’m using Wayland on KDE for years myself without any major issues and with an NVIDIA card. That’s exactly my point, that it became sluggish as hell now after rebooting.
I think the problem is that I have leftovers from my previous driver installation with the .run file. Now I’d just need to know how I can purge everything Nvidia related and start with a clean install
Probably yeah However, the drivers from NVIDIA I installed with the .run file worked mostly fine. So no performance issues or anything like that. Were just much more Wayland specific issues I had than with the stable driver.
LC_ALL=C pamac search --installed nvidia ✔ 7s
egl-wayland 2:1.1.13-1 extra
EGLStream-based Wayland external platform
opencl-nvidia 535.113.01-2 extra
OpenCL implemention for NVIDIA
mhwd-nvidia-470xx 470.199.02-4 extra
MHWD module-ids for nvidia 470.199.02
mhwd-nvidia-390xx 390.157-8 extra
MHWD module-ids for nvidia 390.157
mhwd-nvidia 535.113.01-2 extra
MHWD module-ids for nvidia 535.113.01
linux65-nvidia 535.113.01-10 extra
NVIDIA drivers for linux
libnvidia-container1 1.13.5-1
NVIDIA container runtime library
libnvidia-container-tools 1.13.5-1 AUR
NVIDIA container runtime library
libnvidia-container 1.13.5-1 AUR
NVIDIA container runtime library
lib32-nvidia-utils 535.113.01-1 multilib
NVIDIA drivers utilities (32-bit)
nvidia-utils 535.113.01-2 extra
NVIDIA drivers utilities
nvidia-settings 535.113.01-1 extra
Tool for configuring the NVIDIA graphics driver
nvidia-docker2 1.13.5-1
NVIDIA container runtime library
nvidia-container-toolkit-base 1.13.5-1
NVIDIA container runtime library
nvidia-container-toolkit 1.13.5-1 AUR
NVIDIA container runtime library
nvidia-container-runtime 1.13.5-1 AUR
NVIDIA container runtime library
I gotta leave for an appointment. I hope you guys can help me sort this out I’ll be back in an hour