After daily automatic installation of all available updated packets, my system did no longer boot into graphics mode. Problem is, that the latest NVIDIA Kernel module is made for the last but one kernel release. After booting the last but one kernel, system starts into the graphics mode again. Would it be possible to impose an additional dependency in the kernel package, which prevents my automatic installation process from demageing my system?
It got me to today; thankfully snapper to the rescue and I just rolled back and then pinned the running kernel and then ran the rest of the updates and it went fine. I’m on kernel:
uname -a
Linux localhost.localdomain 5.7.11-1-default
I locked these packages:
| Name | Type | Repository
–±---------------------±--------±----------
1 | kernel-default | package | (any)
2 | kernel-default-devel | package | (any)
3 | kernel-devel | package | (any)
4 | kernel-macros | package | (any)
And, I’m using the nvidia driver from nvidia: x11-video-nvidiaG04
S | Name | Summary | Type
—±---------------------------±----------------------------------------------------------------------±----------
i | kernel-firmware-nvidia | Kernel firmware files for Nvidia Tegra and graphics drivers | package
| libnvidia-container | NVIDIA container runtime library | srcpackage
| libnvidia-container-devel | NVIDIA container runtime library (development files) | package
| libnvidia-container-static | NVIDIA container runtime library (static library) | package
| libnvidia-container-tools | NVIDIA container runtime library (command-line tools) | package
| libnvidia-container1 | NVIDIA container runtime library | package
i | nvidia-computeG04 | NVIDIA driver for computing with GPGPU | package
| nvidia-computeG05 | NVIDIA driver for computing with GPGPU | package
| nvidia-container-toolkit | NVIDIA Container Toolkit | srcpackage
| nvidia-container-toolkit | NVIDIA Container Toolkit | package
i | nvidia-gfxG04-kmp-default | NVIDIA graphics driver kernel module for GeForce 400 series and newer | package
| nvidia-gfxG04-kmp-pae | NVIDIA graphics driver kernel module for GeForce 400 series and newer | package
| nvidia-gfxG05-kmp-default | NVIDIA graphics driver kernel module for GeForce 600 series and newer | package
i | nvidia-glG04 | NVIDIA OpenGL libraries for OpenGL acceleration | package
| nvidia-glG05 | NVIDIA OpenGL libraries for OpenGL acceleration | package
| nvidia-texture-tools | NVIDIA Texture Tools | srcpackage
| nvidia-texture-tools | NVIDIA Texture Tools | package
| pcp-pmda-nvidia-gpu | Performance Co-Pilot (PCP) metrics for the Nvidia GPU | package
i+ | x11-video-nvidiaG04 | NVIDIA graphics driver for GeForce 400 series and newer | package
| x11-video-nvidiaG05 | NVIDIA graphics driver for GeForce 600 series and newer | package
- I noticed the module was not loaded via lsmod | grep nv…so I tried to reinstall the nvidia x11-video-nvidiaG04 package while in console mode running on the newer kernel and it still did not work. As stated above I just rolled back and pinned the old kernel and updated all the other stuff. My question is, I’m new to tumbleweed…is there anything we’re suppose to do after a new kernel drops and the system updates. Does zypper reinstall the nvidia drovers and compile them for the new kernel? Is this issue just a unsupported kernel at this time and we need to wait for nvidia to release a new driver? To be clear, what’s the procedure for updating the system kernel if you’re running a nvidia driver?
Hi
AFAIK the G04 and kernel 5.8+ needs patching to work, just need to wait for the maintainer to fix/push to the repository…
Systems with an NVIDIA card in place have a dependency with the kernel, as the NVIDIA driver must integrate into the kernel. So, why not implementing a dependency in the kernel package? Why waiting for the NVIDIA maintainer to complete the patching, which might take its time. During that time periode, all NVIDIA systems as damaged. This is quite bad promotion for Tumbleweed.
The Nvidia driver is not opensource.
If you have a newer card, you can use the G05 (450) driver.
It works without patching.
Also you can build the driver by yourselve using the run-file.
see:
uname -a && nvidia-settings -v
Linux linux64 5.8.1-3.g846658e-default #1 SMP Fri Aug 14 08:02:13 UTC 2020 (846658e) x86_64 x86_64 x86_64 GNU/Linux
nvidia-settings: version 450.57
The NVIDIA X Server Settings tool.
This program is used to configure the NVIDIA Linux graphics driver.
For more detail, please see the nvidia-settings(1) man page.
Find answers for all your questions here: https://lists.opensuse.org/opensuse-factory/2020-04/msg00097.html
The situation in Tumbleweed it not specific to NVIDIA, but for every out-of-tree kernel module, like VMWare, VirtualBox as well.
NVIDIA also doesn’t play nice with Wayland, and possibly with efforts to bring hardware video decoding in Firefox. NVIDIA is the one with power to change the status quo.
I happen to have a GeForce GT 710. If this card is supported by the G05 driver, what would I have to do in order to switch the driver?
Hi
It is, but I use the hard way to install (along with cuda/cudann);
(Main System)
pinxi -SGxx
System: Host: grover Kernel: 5.8.0-1-default x86_64 bits: 64 compiler: gcc v: 10.2.1 Desktop: GNOME 3.36.5 tk: GTK 3.24.22
wm: gnome-shell dm: GDM Distro: openSUSE Tumbleweed 20200815
Graphics: Device-1: Intel Xeon E3-1200 v2/3rd Gen Core processor Graphics driver: i915 v: kernel bus ID: 00:02.0
chip ID: 8086:016a
Device-2: NVIDIA GK208B [GeForce GT 710] vendor: ZOTAC driver: nvidia v: 450.57 bus ID: 02:00.0 chip ID: 10de:128b
Device-3: NVIDIA GK208B [GeForce GT 710] vendor: ZOTAC driver: nvidia v: 450.57 bus ID: 05:00.0 chip ID: 10de:128b
Device-4: NVIDIA GK208B [GeForce GT 710] vendor: ZOTAC driver: nvidia v: 450.57 bus ID: 06:00.0 chip ID: 10de:128b
Display: x11 server: X.Org 1.20.8 compositor: gnome-shell driver: modesetting,nvidia unloaded: fbdev,vesa
alternate: intel,nouveau,nv resolution: 1: 1920x1080~60Hz 2: 1920x1080 s-dpi: 96
OpenGL: renderer: Mesa DRI Intel HD Graphics P4000 (IVB GT2) v: 4.2 Mesa 20.1.4 compat-v: 3.0 direct render:
(QEMU System)
pinxi -SGxx
System: Host: grover-ostw Kernel: 5.8.0-1-default x86_64 bits: 64 compiler: gcc v: 10.2.1 Desktop: GNOME 3.36.5
tk: GTK 3.24.22 wm: gnome-shell dm: GDM Distro: openSUSE Tumbleweed 20200815
Graphics: Device-1: NVIDIA GK208B [GeForce GT 710] vendor: ZOTAC driver: nvidia v: 450.57 bus ID: 00:03.0 chip ID: 10de:128b
Display: x11 server: X.Org 1.20.8 compositor: gnome-shell driver: nvidia unloaded: fbdev,modesetting,nouveau,vesa
alternate: nv resolution: 1920x1080~60Hz s-dpi: 96
OpenGL: renderer: GeForce GT 710/PCIe/SSE2 v: 4.6.0 NVIDIA 450.57 direct render: Yes
How do you roll back? I face the same problem as OP and now I am stuck without the GUI. Can please share how to use CLI? Am a newbie here.
you can download Nvidia driver instlaller from nvidia site (.run file)
and install it by hands
its very simple, you need kernel source installed(package) and in boot menu(grub menu) edit kernel boot options and add 3 to it, then boot and launch nvidia.run file from root
to roll-back:
-
check grub boot options(on boot advance-menu select) is there previous kernel version, it should be there(I do not know if minor kernel updates saved there)
-
I think you can try use Yast2 in console mode, jsut launch yast2 command from root after booting
and look what version of Kernel and Nvidia driver in package manager you can install is there previous version, if so then install it -
some complex rollback whole file system, depends of file system you use
I have suddenly started have an issue also where my opensuse tumbleweed wont identify my nvidia card.I’m currently using a 560ti temporarily which uses the G04 driver ,but for some reason out of the blue it wont identify the card any-more. I tried reinstalling opensuse i am using the community repo drivers , i also reset my bios just in case some configuration was causing the problem but the problem persists. However on ubuntu its working fine therefore im inclined to think a recent update must have caused the issues which is also present in the latest build.
I have suddenly started have an issue also where my opensuse tumbleweed wont identify my nvidia card.I’m currently using a 560ti temporarily which uses the G04 driver ,but for some reason out of the blue it wont identify the card any-more. I tried reinstalling opensuse i am using the community repo drivers , i also reset my bios just in case some configuration was causing the problem but the problem persists. However on ubuntu its working fine therefore im inclined to think a recent update must have caused the issues which is also present in the latest build.
We are speaking about openSUSE Tumbleweed?
You have read the thread?
Than you should know, that the nvidia driver (390 or G04) does not work anymore with kernel 5.8.
But I think, a new one is up to the Repo.
And it still doesn’t work, today I got fooled too with the update to the Kernel5.8.2 the kernel5.7.11 that worked with the Nvidia drivers disappeared, and now I’m with the Mesa drivers, not even the nouveau ones work
I guess this advice is a little late.
I’m updating at the moment. But the 5.7.11 kernel won’t disappear (I hope).
You can edit (as root) the file “/etc/zypp/zypp.conf”
Look for the “multiversion.kernels” line
I have:
multiversion.kernels = oldest,latest,latest-1,running
So kernel 5.7.11 won’t disappear, because it is the oldest and I am protecting that. At some time – probably when the first 5.9 kernel arrives, I will manually remove that kernel and a more recent one will become the oldest.
If you are using Nvidia you really need to do something similar.
Do you:
Uninstall all!! nvidia drivers?
zypper se -si nvidia
do not have any nomodeset command in your grub bootline?
cat /proc/cmdline
or any blacklist-file:
grep -i 'blacklist nouveau' /etc/modprobe.d/*
Unfortunately it is late now, I did it now, attention, I also had problems with sddm that fails to start after the update I had to change loginmanager using lightdm
enziosavio@Linux-Area51:~> zypper se -si nvidia
Caricamento dati del repository in corso…
Lettura dei pacchetti installati in corso…
S | Nome | Tipo | Versione | Arch. | Repository
—±-----------------------±----------±-------------±-------±---------------------------
i+ | kernel-firmware-nvidia | pacchetto | 20200807-1.2 | noarch | Repository principale (OSS)
i+ | nvidia-texture-tools | pacchetto | 2.0.8-3.13 | x86_64 | Repository principale (OSS)
enziosavio@Linux-Area51:~>
enziosavio@Linux-Area51:~> cat /proc/cmdline
BOOT_IMAGE=/vmlinuz-5.8.2-1-default root=UUID=b338257d-c91d-4822-adb4-2d69c0a16a96 splash=silent quiet showopts mitigations=auto
enziosavio@Linux-Area51:~>
enziosavio@Linux-Area51:~> grep -i ‘blacklist nouveau’ /etc/modprobe.d/*
/etc/modprobe.d/nvidia.conf:blacklist nouveau
/etc/modprobe.d/nvidia-installer-disable-nouveau.conf:blacklist nouveau
enziosavio@Linux-Area51:~>
Ok I corrected the configuration, and now load the Nouveau
Linux-Area51:/home/enziosavio # hwinfo --gfxcard
16: PCI 100.0: 0300 VGA compatible controller (VGA)
[Created at pci.386]
Unique ID: VCu0.LRIsrTxKBxC
Parent ID: _Znp.+eZipJdga28
SysFS ID: /devices/pci0000:00/0000:00:02.0/0000:01:00.0
SysFS BusID: 0000:01:00.0
Hardware Class: graphics card
Model: "nVidia GF119 [GeForce GT 610]"
Vendor: pci 0x10de "nVidia Corporation"
Device: pci 0x104a "GF119 [GeForce GT 610]"
SubVendor: pci 0x3842 "eVga.com. Corp."
SubDevice: pci 0x2615
Revision: 0xa1
Driver: "nouveau"
Driver Modules: "nouveau"
Memory Range: 0xfd000000-0xfdffffff (rw,non-prefetchable)
Memory Range: 0xf0000000-0xf7ffffff (ro,non-prefetchable)
Memory Range: 0xf8000000-0xf9ffffff (ro,non-prefetchable)
I/O Ports: 0xdc00-0xdc7f (rw)
Memory Range: 0x000c0000-0x000dffff (rw,non-prefetchable,disabled)
IRQ: 33 (20861 events)
I/O Ports: 0x3c0-0x3df (rw)
Module Alias: "pci:v000010DEd0000104Asv00003842sd00002615bc03sc00i00"
Driver Info #0:
Driver Status: nouveau is active
Driver Activation Cmd: "modprobe nouveau"
Config Status: cfg=new, avail=yes, need=no, active=unknown
Attached to: #30 (PCI bridge)
Primary display adapter: #16
enziosavio@Linux-Area51:~> grep -i ‘blacklist nouveau’ /etc/modprobe.d/*
/etc/modprobe.d/nvidia.conf:blacklist nouveau
/etc/modprobe.d/nvidia-installer-disable-nouveau.conf:blacklist nouveau
This 2 files will prevent nouveau from loading/working.
Delete them and reboot.
Would this have anything to do with the problem?
https://preview.redd.it/2ejyle86nyj51.png?width=1024&auto=webp&8a406808
The expected checksum of file /var/tmp/AP_0xp8UPnm/repodata/susedata.xml.gz
is fa47ae8244fe7640f149d28964e604bf7e756da989d7e520f67216a96bb6cfb2,
but the current checksum is 1f4dc420bfbf3d1b423cd7c4e1ca6aa25d7a13daff70c64e412f0ed449602406.
The file has been changed by accident or by an attacker since the repository creator signed it. Using it is a big risk for the integrity and security of your system.
Use it anyway?
Adding repository nVidia Graphics Drivers failed.
download.nvidia.com-tumbleweed_1: [download.nvidia.com-tumbleweed_1|https://download.nvidia.com/opensuse/tumbleweed] Valid metadata not found at specified URL
History:
- susedata.xml.gz has wrong checksum
- Can't provide ./repodata/susedata.xml.gz
Hi
Very likely since the repository doesn’t exist at the moment, I would say updates are going on…
https://download.nvidia.com/opensuse/tumbleweed/ == 404