Do you suggest that I move all nvidia packages to G07 so it might solve this particular issue?
You can simply update this two packages without that the dependency solver is complaining. This shouldn’t be the case. So the dependency isn’t implemented properly.
It shouldn’t be possible to upgrade this two packages without complaining of zypper due to version missmatches.
I have completely lost on what you mean here.
This comment was more directed at OrsoBruno.
The mismatches were likely from the period when G07 was still experimental.
Unfortunately the OP upgraded from 550.144 packages, so that system didn’t see any upgrade for a long time.
At that time nvidia-compute-G06-550.144.03-30.1 didn’t include the dependency for nvidia-persistenced and nvidia-common-G06 didn’t even exist.
So that is likely why the OP saw all that list of conflicts etc.
After a manual upgrade to the right current packages there should be no need to add locks, which might actually cause trouble on future upgrades of G06 packages.
@moshpuke So I did a quick install here along with the nvidia container tool kit. my last few installs were on leap 16.0 and the run file…
I see nvidia-smi is needed, so that’s provided by the nvidia-compute-utils-G06 package which does pull in some others…
zypper se -si nvidia
Loading repository data...
Reading installed packages...
S | Name | Type | Version | Arch | Repository
---+-------------------------------+---------+--------------------------+--------+-------------------------
i | kernel-firmware-nvidia | package | 20260408-1.1 | noarch | repo-oss
i+ | libnvidia-container-tools | package | 1.19.0-1 | x86_64 | nvidia-container-toolkit
i+ | libnvidia-container1 | package | 1.19.0-1 | x86_64 | nvidia-container-toolkit
i | libnvidia-gpucomp-G06 | package | 580.159.03-52.1 | x86_64 | repo-non-free
i | nvidia-common-G06 | package | 580.159.03-52.1 | x86_64 | repo-non-free
i | nvidia-compute-G06 | package | 580.159.03-52.1 | x86_64 | repo-non-free
i+ | nvidia-compute-utils-G06 | package | 580.159.03-52.1 | x86_64 | repo-non-free
i+ | nvidia-container-toolkit | package | 1.19.0-1 | x86_64 | nvidia-container-toolkit
i+ | nvidia-container-toolkit-base | package | 1.19.0-1 | x86_64 | nvidia-container-toolkit
i+ | nvidia-driver-G06-kmp-default | package | 580.159.03_k7.0.5_1-52.1 | x86_64 | repo-non-free
i | nvidia-modprobe | package | 580.159.03-26.2 | x86_64 | repo-non-free
i | nvidia-persistenced | package | 580.159.03-2.2 | x86_64 | repo-non-free
i | openSUSE-repos-MicroOS-NVIDIA | package | 20260423.1a6a0f3-2.1 | x86_64 | repo-oss
kgp -n gpu-operator
NAME READY STATUS RESTARTS AGE
gpu-feature-discovery-5h7bt 1/1 Running 0 32m
gpu-operator-6c79d6dd65-qsktl 1/1 Running 2 (12m ago) 38m
nvidia-container-toolkit-daemonset-wflsp 1/1 Running 0 32m
nvidia-cuda-validator-wx2lj 0/1 Completed 0 12m
nvidia-dcgm-exporter-zsl8w 1/1 Running 0 32m
nvidia-device-plugin-daemonset-dsckz 1/1 Running 0 32m
nvidia-operator-validator-lrldn 1/1 Running 0 32m
nvidiagpu-node-feature-discovery-gc-79448dc8cb-pc6xd 1/1 Running 1 (12m ago) 38m
nvidiagpu-node-feature-discovery-master-945cb5d95-8wn2s 1/1 Running 2 (12m ago) 38m
nvidiagpu-node-feature-discovery-worker-579pf 1/1 Running 1 (12m ago) 38m
nvidia-smi
Thu May 14 19:41:21 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.159.03 Driver Version: 580.159.03 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 Tesla P4 On | 00000000:01:00.0 Off | 0 |
| N/A 40C P8 6W / 75W | 0MiB / 7680MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
You are saying utils-G06 was the culprit? I completely uninstalled- reinstalled nvidia again (3rd time) now there’s no complaint.
I wonder why I have two firmware-nvidia listed in my zypper se -is nvidia though compared to yours:
zypper se -si nvidia
Loading repository data...
Reading installed packages...
S | Name | Type | Version | Arch | Repository
---+--------------------------------+---------+----------------------------+--------+----------------------------
i | kernel-firmware-nvidia | package | 20260408-1.1 | noarch | openSUSE-Tumbleweed-Oss
i | kernel-firmware-nvidia | package | 20260408-1.1 | noarch | openSUSE-MicroOS-20260426-0
...
Are they the same package??
@moshpuke you have two repositories that point to the same URL, so yes they are the same package… the openSUSE-MicroOS-20260426-0 one was the install… you can remove.
I’m using the Agama installer, so it leaves some agama ones that are duplicated by the openSUSE repository service, so I remove…
You can see by using zypper lr -d
To me, nvidia-smi needs to be split out when using containers.
My install was;
transactional-update --non-interactive pkg install --auto-agree-with-licenses \
nvidia-driver-G06-kmp-default nvidia-compute-utils-G06
Yeah, you reminded me how I took an hour with chatgpt just to findout which package ‘nvidia-smi’ tool was hiding in on opensuse.
@moshpuke You can use zypper, since it’s a user binary and you have the repository present (in this case the nvidia one) zypper se --provides /usr/bin/nvidia-smi or if cnf (command-not-found) is installed cnf nvidia-smi.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.