I’m having a similar problem except not even the closed source drivers work which I use regularly. I’m using an RTX 2060. And I only have the tumbleweed nvidia repo. I can use my PC but i3 freezes it until a new tab is summoned, I have to use a 4:3 aspect ratio with a low resolution, only one monitor works, it’s not at 120hz like I set it and nvidia-setting gives me Error: NVIDIA driver is not loaded.
Transition to the new 590 (Beta) driver time… I would assume it’s going to be G07(?) since it will only support Turing or better… I install the hard way and the default which is the open driver…
System:
Kernel: 6.18.0-1-default arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
clocksource: tsc avail: hpet,acpi_pm
parameters: BOOT_IMAGE=/boot/vmlinuz-6.18.0-1-default
root=UUID=1ab675ed-aba2-4593-9ebc-1c893015b607 splash=silent quiet
security=apparmor intel_iommu=on iommu=pt fbdev=1
nvidia_drm.modeset=1 mitigations=auto
.....
Device-2: NVIDIA TU104GL [Quadro RTX 4000] vendor: Hewlett-Packard
driver: nvidia v: 590.44.01 alternate: nouveau,nvidia_drm
non-free: 550-580.xx+ status: current (as of 2025-11; EOL~2026-12-xx)
arch: Turing code: TUxxx process: TSMC 12nm FF built: 2018-2022 pcie:
gen: 1 speed: 2.5 GT/s lanes: 16 link-max: gen: 3 speed: 8 GT/s ports:
active: none empty: DP-4, DP-5, DP-6, Unknown-1 bus-ID: 07:00.0
chip-ID: 10de:1eb1 class-ID: 0300
1 Like
Show inxi -GSaz, zypper se -si nvidia and zypper lr -d
I have switched to closed drivers and nvidia drivers works again under 6.18 kernel.
I have installed nvidia-driver-G06-kmp-default and nvidia-driver-G06-kmp-meta.
I removed nvidia-open-driver-G06-signed-kmp-default and nvidia-open-driver-G06-signed-kmp-default-meta.
2 Likes
S | Name | Type | Version | Arch | Repository
---+--------------------------------------------------+---------+---------------------------+--------+------------------------
i | kernel-firmware-nvidia | package | 20251018-1.1 | noarch | openSUSE-Tumbleweed-Oss
i | libnvidia-egl-gbm1 | package | 1.1.2-7.17 | x86_64 | NVIDIA
i | libnvidia-egl-gbm1-32bit | package | 1.1.2-7.12 | x86_64 | NVIDIA
i | libnvidia-egl-wayland1 | package | 1.1.20-52.3 | x86_64 | NVIDIA
i | libnvidia-egl-wayland1-32bit | package | 1.1.20-52.3 | x86_64 | NVIDIA
i | libnvidia-egl-x111 | package | 1.0.3-21.5 | x86_64 | NVIDIA
i | libnvidia-egl-x111-32bit | package | 1.0.3-21.4 | x86_64 | NVIDIA
i | libnvidia-gpucomp | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | libnvidia-gpucomp-32bit | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-common-G06 | package | 580.105.08-44.1 | x86_64 | NVIDIA
i+ | nvidia-compute-G06 | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-compute-G06-32bit | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-compute-utils-G06 | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-gl-G06 | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-gl-G06-32bit | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-libXNVCtrl | package | 580.105.08-47.1 | x86_64 | NVIDIA
i | nvidia-modprobe | package | 580.105.08-20.1 | x86_64 | NVIDIA
i+ | nvidia-open-driver-G06-signed-cuda-default-devel | package | 580.105.08-2.4 | x86_64 | openSUSE-Tumbleweed-Oss
i+ | nvidia-open-driver-G06-signed-kmp-default | package | 580.105.08_k6.17.9_1-2.4 | x86_64 | openSUSE-Tumbleweed-Oss
i | nvidia-open-driver-G06-signed-kmp-longterm | package | 580.105.08_k6.12.59_1-2.4 | x86_64 | openSUSE-Tumbleweed-Oss
i | nvidia-open-driver-G06-signed-kmp-meta | package | 580.105.08-25.1 | x86_64 | NVIDIA
i | nvidia-persistenced | package | 580.105.08-2.1 | x86_64 | NVIDIA
i+ | nvidia-settings | package | 580.105.08-47.1 | x86_64 | NVIDIA
i | nvidia-userspace-meta-G06 | package | 580.105.08-24.1 | x86_64 | NVIDIA
i+ | nvidia-video-G06 | package | 580.105.08-44.1 | x86_64 | NVIDIA
i | nvidia-video-G06-32bit | package | 580.105.08-44.1 | x86_64 | NVIDIA
System:
Kernel: 6.12.60-1-longterm arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
clocksource: tsc avail: hpet,acpi_pm
parameters: BOOT_IMAGE=/boot/vmlinuz-6.12.60-1-longterm
root=UUID=aec7219b-fc2e-4ba2-8cff-64bcc514ae06 splash=silent
mitigations=auto quiet security=selinux selinux=1
rd.driver.blacklist=nouveau
Desktop: IceWM v: 3.9.0 wm: i3 v: 4.24 tools: avail: i3lock, swayidle,
swaylock, xscreensaver vt: 7 dm: 1: GDM v: 49.2 2: LightDM v: 1.32.0
note: stopped Distro: openSUSE Tumbleweed 20251204
Graphics:
Device-1: NVIDIA TU106 [GeForce RTX 2060 Rev. A] vendor: ASUSTeK
driver: nvidia v: 580.105.08 alternate: nouveau,nvidia_drm
non-free: 550-580.xx+ status: current (as of 2025-11; EOL~2026-12-xx)
arch: Turing code: TUxxx process: TSMC 12nm FF built: 2018-2022 pcie:
gen: 3 speed: 8 GT/s lanes: 16 ports: active: none off: DP-1,HDMI-A-1
empty: DVI-D-1,HDMI-A-2 bus-ID: 01:00.0 chip-ID: 10de:1f08 class-ID: 0300
Display: x11 server: X.org v: 1.21.1.21 with: Xwayland v: 24.1.8
compositor: Picom v: 12.5 driver: X: loaded: nvidia
unloaded: modesetting,vesa alternate: fbdev,nouveau,nv
gpu: nv_platform,nvidia,nvidia-nvswitch display-ID: :0 screens: 1
Screen-1: 0 s-res: 5120x1440 s-size: <missing: xdpyinfo>
Monitor-1: DP-1 mapped: DP-0 note: disabled pos: primary,left
model: Samsung Odyssey G61SD serial: <filter> built: 2245 res:
mode: 2560x1440 hz: 120 scale: 100% (1) dpi: 109 gamma: 1.2
size: 598x336mm (23.54x13.23") diag: 686mm (27") ratio: 16:9 modes:
max: 2560x1440 min: 640x480
Monitor-2: HDMI-A-1 mapped: HDMI-0 note: disabled pos: right
model: Samsung LC27G5xT serial: <filter> built: 2023 res: mode: 2560x1440
hz: 120 scale: 100% (1) dpi: 108 gamma: 1.2 size: 600x340mm (23.62x13.39")
diag: 685mm (27") ratio: 16:9 modes: max: 2560x1440 min: 640x480
API: OpenGL v: 4.6.0 vendor: nvidia v: 580.105.08 glx-v: 1.4
direct-render: yes renderer: NVIDIA GeForce RTX 2060/PCIe/SSE2
memory: 5.86 GiB
API: Vulkan v: 1.4.328 layers: 8 device: 0 type: discrete-gpu
name: NVIDIA GeForce RTX 2060 driver: nvidia v: 580.105.08
device-ID: 10de:1f08 surfaces: N/A device: 1 type: cpu name: llvmpipe
(LLVM 21.1.3 256 bits) driver: mesa llvmpipe v: 25.2.5 (LLVM 21.1.3)
device-ID: 10005:0000 surfaces: N/A
API: EGL Message: EGL data requires eglinfo. Check --recommends.
Info: Tools: api: glxinfo,vulkaninfo gpu: nvidia-settings,nvidia-smi
wl: kanshi, swaymsg, wlr-randr x11: xprop,xrandr
(the kernel is lts because I’m doing something so I cannot restart it and this one works correctly)
# | Alias | Name | Enabled | GPG Check | Refresh | Keep | Priority | Type | URI | Service
---+---------------------------------+------------------------------------------------------------+---------+-----------+---------+------+----------+--------+----------------------------------------------------------------------------------------------------+--------
1 | Google-Chrome | Google-Chrome | Yes | (r ) Yes | No | - | 50 | rpm-md | http://dl.google.com/linux/chrome/rpm/stable/x86_64 |
2 | KDE_KDE3 | KDE 3.5.10 and Qt3-related software (openSUSE_Tumbleweed) | Yes | (r ) Yes | No | - | 99 | rpm-md | https://download.opensuse.org/repositories/KDE:/KDE3/openSUSE_Tumbleweed/ |
3 | NVIDIA | NVIDIA | Yes | (r ) Yes | Yes | - | 50 | rpm-md | https://download.nvidia.com/opensuse/tumbleweed |
4 | X11_COSMIC_Next | Development branch of COSMIC (unstable) (openSUSE_Factory) | Yes | (r ) Yes | Yes | - | 99 | rpm-md | https://download.opensuse.org/repositories/X11:/COSMIC:/Next/openSUSE_Factory/ |
5 | adoptium | adoptium | Yes | (r ) Yes | Yes | - | 99 | rpm-md | https://packages.adoptium.net/artifactory/rpm/opensuse/15.3/x86_64 |
6 | code | Visual Studio Code | Yes | (r ) Yes | Yes | - | 50 | rpm-md | https://packages.microsoft.com/yumrepos/vscode |
7 | devel_languages_ruby_extensions | Ruby Extensions (openSUSE_Tumbleweed) | Yes | (r ) Yes | No | - | 99 | rpm-md | https://download.opensuse.org/repositories/devel:/languages:/ruby:/extensions/openSUSE_Tumbleweed/ |
8 | google-chrome | google-chrome | Yes | (r ) Yes | Yes | - | 50 | rpm-md | https://dl.google.com/linux/chrome/rpm/stable/x86_64 |
9 | hardware | Hardware tools (openSUSE_Tumbleweed) | Yes | (r ) Yes | No | - | 50 | rpm-md | https://download.opensuse.org/repositories/hardware/openSUSE_Tumbleweed/ |
10 | home_Dead_Mozay | home:Dead_Mozay (openSUSE_Tumbleweed) | Yes | (r ) Yes | No | - | 99 | rpm-md | https://download.opensuse.org/repositories/home:/Dead_Mozay/openSUSE_Tumbleweed/ |
11 | home_getchoo | Prism Launcher (openSUSE_Tumbleweed) | Yes | (r ) Yes | No | - | 50 | rpm-md | https://download.opensuse.org/repositories/home:/getchoo/openSUSE_Tumbleweed/ |
12 | openSUSE-20250414-0 | openSUSE-20250414-0 | No | ---- | ---- | - | 99 | rpm-md | hd:/?device=/dev/disk/by-id/usb-Generic_Flash_Disk_F0124CBA-0:0-part2 |
13 | packman | packman | Yes | (r ) Yes | Yes | - | 50 | rpm-md | https://ftp.gwdg.de/pub/linux/misc/packman/suse/openSUSE_Tumbleweed/ |
14 | repo-debug | openSUSE-Tumbleweed-Debug | No | ---- | ---- | - | 50 | N/A | http://download.opensuse.org/debug/tumbleweed/repo/oss/ |
15 | repo-non-oss | openSUSE-Tumbleweed-Non-Oss | Yes | (r ) Yes | Yes | - | 51 | rpm-md | http://download.opensuse.org/tumbleweed/repo/non-oss/ |
16 | repo-oss | openSUSE-Tumbleweed-Oss | Yes | (r ) Yes | Yes | - | 50 | rpm-md | http://download.opensuse.org/tumbleweed/repo/oss/ |
17 | repo-source | openSUSE-Tumbleweed-Source | No | ---- | ---- | - | 50 | N/A | http://download.opensuse.org/source/tumbleweed/repo/oss/ |
18 | repo-update | openSUSE-Tumbleweed-Update | Yes | (r ) Yes | Yes | - | 50 | rpm-md | http://download.opensuse.org/update/tumbleweed/ |
Judging from the reported versions of packages, you should have the following CUDA repo enabled: https://developer.download.nvidia.com/compute/cuda/repos/opensuse15/x86_64/
It is a developer repo that should not be normally enabled (unless you know what you are doing and you are prepared to deal with any inconsistencies with other packages).
I doubt that the above has anything to do with the installation of openSUSE-repos-Tumbleweed-NVIDIA.
Newb here, so please be gentle…
I had Tumbleweed working perfectly with the repo-based driver and CUDA in a command-shell doing ML training on the gpu.
Everything was running perfectly until the new kernel deployed.
After deployment nvidia-smi could not find the driver.
I tried a forced redeployment via zypper and the kernel module compilation stage generated a lot of errors that were not visible via Myrlyn or Yast2.
It looks like there is a kernel version mismatch somewhere:
uname -r gives 6.18.0-1-default
Meanwhile the rpm packages are showing:
kernel-default → 6.18.0-1.1
kernel-default-devel → 6.18.0-1.1
My understanding is that the kernel-default-devel contains the headers used by dkms to rebuild the kernel modules when the kernel gets updated.
But somewhere between Grub, uname, and the rpms there is a .1 discrepancy which would cause the dkms rebuild to fail. I know from my old Debian days, dkms rebuilds typically fail silently in these situations…
I’m still attempting to trace this, but am throwing it out there in case someone knows what is really going on here and can either correct me or use the info to zero in on the real situation.
1 Like
@woneill1701 Hi and welcome to the Forum 
So yesterday things changed for the CUDA repo by the looks…
See https://en.opensuse.org/SDB:NVIDIA_drivers#CUDA
Also what is your GPU? inxi -GSaz
From a clean installation of tumbleweed using online repos, after installing inxi, sway and chromium (the latter two so I can access the forum), inxi gives:
bill@localhost:~> inxi -GSaz
System:
Kernel: 6.18.0-1-default arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
clocksource: tsc avail: hpet,acpi_pm
parameters: BOOT_IMAGE=(hd0,gpt1)/opensuse-tumbleweed/6.18.0-1-default/linux-920be6329be2211fef6d2bda38b8289a290ea832
resume=/dev/disk/by-uuid/8c51000e-f9cf-4377-8d5d-0ffdbc60c8b3 quiet
security=selinux selinux=1 mitigations=auto
root=UUID=070acf67-d22b-463b-b36b-37254b80a28d
rootflags=subvol=@/.snapshots/1/snapshot
Desktop: Sway v: 1.11 with: waybar tools: swayidle avail: swaylock vt: 1
dm: N/A Distro: openSUSE Tumbleweed 20251204
Graphics:
Device-1: Intel Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics
vendor: Gigabyte driver: i915 v: kernel arch: Gen-7.5 process: Intel 22nm
built: 2013 ports: active: HDMI-A-3 empty: HDMI-A-1,HDMI-A-2,VGA-1
bus-ID: 00:02.0 chip-ID: 8086:0412 class-ID: 0300
Device-2: NVIDIA GB206 [GeForce RTX 5060 Ti] vendor: Micro-Star MSI
driver: nouveau v: kernel non-free: 550-580.xx+
status: current (as of 2025-11) arch: Lovelace code: AD1xx
process: TSMC n4 (5nm) built: 2022+ pcie: gen: 1 speed: 2.5 GT/s lanes: 8
link-max: gen: 5 speed: 32 GT/s lanes: 16 ports: active: none empty: DP-1,
DP-2, DP-3, HDMI-A-4 bus-ID: 01:00.0 chip-ID: 10de:2d04 class-ID: 0300
Display: wayland server: Xwayland v: 24.1.8 compositor: Sway v: 1.11
driver: gpu: i915 display-ID: 1
Monitor-1: HDMI-A-3 model: HDMI serial: <filter> built: 2018 res:
mode: 2560x1440 hz: 60 scale: 100% (1) dpi: 255 gamma: 1.2
size: 255x220mm (10.04x8.66") diag: 337mm (13.3") ratio: 5:4 modes:
max: 2560x1440 min: 720x400
API: OpenGL v: 4.6 vendor: intel mesa v: 25.3.0 glx-v: 1.4 es-v: 3.2
direct-render: yes renderer: Mesa Intel HD Graphics 4600 (HSW GT2)
device-ID: 8086:0412 memory: 1.46 GiB unified: yes
API: EGL Message: EGL data requires eglinfo. Check --recommends.
Info: Tools: api: glxinfo wl: kanshi,swaymsg
@woneill1701 that’s some old hardware… are you even getting PCIe 3.0 from the CPU? I have a similar setup, but it’s only running Pascal GPU’s but the CPU upgrade does allow PCIe 3.0…
Did you look at the link and run the respective commands to add the CUDA repo and open driver?
Yes it’s an old NAS server that I pressed into service to host the new GPU. I know I’m pushing the kernel to its limits - using an ancient mb with a card that requires the latest drivers - but the openSUSE kernel handles it like a champ!
Reading through the link now!
The install process described mirrors the process I used previously:
Installing the openSUSE tumbleweed repo and then installing manually via zypper. I then added the cuda repo for openSUSE 15…
Now the same process gives compilation errors.
From my recent experience with openSUSE distros, Nvidia drivers have two issues.
- Nvidia driver has a meta package that deals with multiple components of driver. All these packages are not updated at the same time, resulting in mismatched verisons of components, and driver failure. Arch, and Fedora (RPM fusion repo) drivers get updated package with all components at same verison.
- On tumbleweed, kernel updates are not recompiling the nvidia drivers somehow. This results in driver failure. On OpenSUSE Leap, at least the kernel is fixed, so this kernel update is not an issue.
I hope OpenSUSE developers look into these issues and fix them. Also having different repos for Leap, Tumbleweed, MicroOS, makes things unnecessarily complicated.
It’s just a driver. User is supposed to just install it, and forget about it. If use has to put this much effort into making a graphics card working, it is not very encouraging.
Both statements are quite incorrect. The meta packages have the purpose to pull in the matching package versions. If it is not working, file a precise bugreport.
Regarding the kernel modules, for the closed module, they are automatically rebuild on the user machine when updating the driver/kernel. The open module is precompiled and also matching.
These packages are specifically build for the named distributions. In your logic, it also makes no sense to have different repos for the different openSUSE distro flavors.
Btw the repos are automatically managed by the openSUSE-repos-xxxxxx-NVIDIA packages. xxxxxx stands for the distro flavor. If you have a look at the URL which is provided by these service packages, you will notice, that these 4 packages boil down to 2 different repos (TW, SR and MIcroOS use the same repo URL; Leap has its own repo URL).
1 Like
Another way to avoid this problem is using the .run installer, Then enable DKMS.
But generally open version is recommended for Linux.
- In my case, with Leap, the meta package, and persistenced didn’t get updated, but other components did. This kind of mismatch doesn’t happen with Fedora RPM fusion, and arch linux.
- On tumbleweed, i am using the open module, and it is not working with new kernel. With components, i notice only persistenced has 580.105.08.2 while others have 580.105.08.25. Not sure if this is an issue.
- I agree on different repos for different distros. I got confused on that.
I like Leap16 and use it as my workstation distro. And testing out tumbleweed. Nvidia driver is the only issue i faced. And the fact that this is the main topic on forums, says it.
Yeah,I have same issue
When I upgrade to Kernel 6.18 I will black when boot to stage 2
I am rolled back to 6.17.9 and wait OPEN MODULE to update for Kernel 6.18
openSUSE’s NVIDIA bin driver is easy to use, But Updates’ flexibility is worse than RH’s akmod and General DKMS They will autobuild when update
I am waiting RH & NOVIDEO to release their NOVA driver to workable,and into the mainline, NVIDIA = NOVIDEO maybe be a history
1 Like
@csrinivas the Nvidia run file defaults to the open driver, to install the proprietary version a command line switch is needed…
/sbin/modinfo nvidia | grep license
license: Dual MIT/GPL
inxi -Gz
Graphics:
Device-1: Intel DG2 [Arc A380] driver: i915 v: kernel
Device-2: NVIDIA TU104GL [Quadro RTX 4000] driver: nvidia v: 590.44.01
....
2 Likes
Also having this issue now after latest update using the open driver.