I’m trying to make an update on nvidia proprietary drivers and theproblem persists for a long time that I can not do it with the following errors:
#### Myrlyn dependency conflicts list - generated 2026-09-25 08:59:21 ####
the to be installed nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 requires 'nvidia-common-G06 = 580.178.04', but this requirement cannot be provided
not installable providers: nvidia-common-G06-580.178.04-lp160.48.1.x86_64[NVIDIA:repo-non-free]
nvidia-common-G06-580.178.04-lp160.49.1.x86_64[NVIDIA:repo-non-free]
[ ] break nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 by ignoring some of its dependencies
[ ] Following actions will be done:
do not install nvidia-compute-G06-580.178.04-lp160.49.1.x86_64
deinstallation of nvidia-compute-G06-580.159.03-lp160.60.1.x86_64
[ ] Following actions will be done:
keep nvidia-common-G06-580.159.03-lp160.60.1.x86_64
keep nvidia-persistenced-580.159.03-lp160.2.1.x86_64
the installed nvidia-userspace-meta-G06-580.159.03-lp160.53.1.x86_64 requires 'nvidia-compute-utils-G06 = 580.159.03', but this requirement cannot be provided
not installable providers: nvidia-compute-utils-G06-580.159.03-lp160.48.1.x86_64[NVIDIA:repo-non-free]
nvidia-compute-utils-G06-580.159.03-lp160.49.1.x86_64[NVIDIA:repo-non-free]
nvidia-compute-utils-G06-580.159.03-lp160.52.1.x86_64[NVIDIA:repo-non-free]
nvidia-compute-utils-G06-580.159.03-lp160.54.1.x86_64[NVIDIA:repo-non-free]
nvidia-compute-utils-G06-580.159.03-lp160.56.1.x86_64[NVIDIA:repo-non-free]
[ ] keep nvidia-compute-utils-G06-580.159.03-lp160.60.1.x86_64
[ ] deinstallation of nvidia-userspace-meta-G06-580.159.03-lp160.53.1.x86_64
[ ] break nvidia-userspace-meta-G06-580.159.03-lp160.53.1.x86_64 by ignoring some of its dependencies
the to be installed nvidia-video-G06-580.178.04-lp160.49.1.x86_64 requires 'nvidia-common-G06 = 580.178.04', but this requirement cannot be provided
not installable providers: nvidia-common-G06-580.178.04-lp160.48.1.x86_64[NVIDIA:repo-non-free]
nvidia-common-G06-580.178.04-lp160.49.1.x86_64[NVIDIA:repo-non-free]
[ ] do not install nvidia-video-G06-580.178.04-lp160.49.1.x86_64
[ ] keep nvidia-modprobe-580.159.03-lp160.26.1.x86_64
[ ] break nvidia-video-G06-580.178.04-lp160.49.1.x86_64 by ignoring some of its dependencies
#### Myrlyn dependency conflicts list END ###
This has been a problem for at least a month is there anyone else with this problem ?
I removed the lock and it updated.
Should I set the lock again after the update ?
I tried to remove the lock before on myrlyn, but it did not work. Now removing the lock on zypper and making the upgrade worked.
There is no need for locks normally, unless there is another problem in your system that should be fixed or there is a bug that should be reported. But as already stated, no locks and no problems here.
I maintined the lock and tried make an ugrade on myrlyn …
Trouble yet again …
I started myrlyn and I get some options to upgrade nvidia-Persistence and nvidia-modprobe. And they have the errors with broken dependencies. And I showed results on attached.
#### Myrlyn dependency conflicts list - generated 2026-09-26 17:33:21 ####
the installed nvidia-common-G06-580.178.04-lp160.49.1.x86_64 requires 'nvidia-modprobe-580.178.04', but this requirement cannot be provided
not installable providers: nvidia-modprobe-580.178.04-lp160.28.1.x86_64[NVIDIA:repo-non-free]
[ ] do not install nvidia-modprobe-595.99.02-lp160.2.1.x86_64
[ ] Following actions will be done:
downgrade of nvidia-common-G06-580.178.04-lp160.49.1.x86_64 to nvidia-common-G06-570.172.08-lp160.37.1.x86_64
install nvidia-common-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-common-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-video-G06-580.178.04-lp160.49.1.x86_64 to nvidia-video-G06-570.172.08-lp160.37.1.x86_64
install nvidia-video-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-video-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-compute-utils-G06-580.178.04-lp160.49.1.x86_64 to nvidia-compute-utils-G06-570.172.08-lp160.37.1.x86_64
install nvidia-compute-utils-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-compute-utils-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-userspace-meta-G06-580.178.04-lp160.54.1.x86_64 to nvidia-userspace-meta-G06-570.172.08-lp160.10.1.noarch
install nvidia-userspace-meta-G06-570.172.08-lp160.10.1.noarch from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-userspace-meta-G06-580.178.04-lp160.54.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-driver-G06-kmp-meta-580.178.04-lp160.54.1.x86_64 to nvidia-driver-G06-kmp-meta-570.172.08-lp160.10.1.noarch
install nvidia-driver-G06-kmp-meta-570.172.08-lp160.10.1.noarch from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-driver-G06-kmp-meta-580.178.04-lp160.54.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-driver-G06-kmp-default-580.178.04_k6.12.0_160000.37-lp160.49.4.x86_64 to nvidia-driver-G06-kmp-default-570.172.08_k6.12.0_160000.19-lp160.37.1.x86_64
install nvidia-driver-G06-kmp-default-570.172.08_k6.12.0_160000.19-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-driver-G06-kmp-default-580.178.04_k6.12.0_160000.37-lp160.49.4.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-gl-G06-580.178.04-lp160.49.1.x86_64 to nvidia-gl-G06-570.172.08-lp160.37.1.x86_64
install nvidia-gl-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-gl-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
[ ] break nvidia-common-G06-580.178.04-lp160.49.1.x86_64 by ignoring some of its dependencies
the installed nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 requires 'nvidia-persistenced-580.178.04', but this requirement cannot be provided
[ ] Following actions will be done:
downgrade of nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 to nvidia-compute-G06-570.172.08-lp160.37.1.x86_64
install nvidia-compute-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-compute-utils-G06-580.178.04-lp160.49.1.x86_64 to nvidia-compute-utils-G06-570.172.08-lp160.37.1.x86_64
install nvidia-compute-utils-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-compute-utils-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-userspace-meta-G06-580.178.04-lp160.54.1.x86_64 to nvidia-userspace-meta-G06-570.172.08-lp160.10.1.noarch
install nvidia-userspace-meta-G06-570.172.08-lp160.10.1.noarch from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-userspace-meta-G06-580.178.04-lp160.54.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-driver-G06-kmp-meta-580.178.04-lp160.54.1.x86_64 to nvidia-driver-G06-kmp-meta-570.172.08-lp160.10.1.noarch
install nvidia-driver-G06-kmp-meta-570.172.08-lp160.10.1.noarch from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-driver-G06-kmp-meta-580.178.04-lp160.54.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-gl-G06-580.178.04-lp160.49.1.x86_64 to nvidia-gl-G06-570.172.08-lp160.37.1.x86_64
install nvidia-gl-G06-570.172.08-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-gl-G06-580.178.04-lp160.49.1.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
downgrade of nvidia-driver-G06-kmp-default-580.178.04_k6.12.0_160000.37-lp160.49.4.x86_64 to nvidia-driver-G06-kmp-default-570.172.08_k6.12.0_160000.19-lp160.37.1.x86_64
install nvidia-driver-G06-kmp-default-570.172.08_k6.12.0_160000.19-lp160.37.1.x86_64 from vendor SUSE LLC <https://www.suse.com/>
replacing nvidia-driver-G06-kmp-default-580.178.04_k6.12.0_160000.37-lp160.49.4.x86_64 from vendor obs://build.suse.de/Proprietary:X11:Drivers
[ ] break nvidia-compute-G06-580.178.04-lp160.49.1.x86_64 by ignoring some of its dependencies
[ ] do not install nvidia-persistenced-595.99.02-lp160.2.1.x86_64
#### Myrlyn dependency conflicts list END ###
So there is indeed something deeper I think I should reimpose the lock on those packages …
nvidia-modprobe-595.99.02-lp160.2.1.x86_64 is for systems using the G07 drivers, so on your system with G06 setup Myrlyn should recognize the conflict and automatically skip that upgrade… I should fire up a test Leap 16.0 to check…
lt-b:~ # zypper locks
# | Name | Type | Repository | Comment
--+---------------+---------+------------+--------
1 | libopenh264-8 | package | (any) |
lt-b:~ # zypper up
Refreshing service 'NVIDIA'.
Refreshing service 'openSUSE'.
Loading repository data...
Reading installed packages...
The following item is locked and will not be changed by any action:
Installed:
libopenh264-8
The following 4 package updates will NOT be installed:
libnvidia-egl-gbm1 libopenh264-8 nvidia-modprobe nvidia-persistenced
Nothing to do.
lt-b:~ #
and if I hit “Package update” in Myrlyn the two nvidia 595.xx packages are marked with “sad faces” and are not installed without any user interaction.
So check that your list of installed packages matches that posted by @malcolmlewis .
Exactly the same behavior …on myrlyn they are marked with the sad face if one chooses to update the packages … and they do not install without interaction.
The post I made initially was exactly the action of marking the packages for update manually/forcefully. they gave the initial error messages i posted.
I checked and have the same packages @malcolmlewis showed.
I have to say that the system works fine, I do not have crashes and the system is stale.
The GPU is also working fine . The objective of this particular PC was to build a Dual-GPU local AI system but I do not have yet the dual GPU’s so I used and old nvidia 1070. I was making this one ready for the migration … when I noticed this issue with the drivers update …
That is the “normal” behaviour that should be used to update.
That is not the way to update the system. If you try that (for debugging, bug reporting etc.) expect the system to spit out every sort of conflict, like you apparently already found out.
So apparently everything is working on your system, relax, only update via “Package update” in Myrlyn or zypper up and have a lot of fun with Leap.
As a side note, your current Pascal GPU needs the G06 driver, so everything OK.
You may need to upgrade to G07 when you switch to a new card.
And IMHO you have way too many repos enabled and all with the default priority at 99: that is likely to cause conflicts sooner or later, so please check what you really need.
If you cannot disable any, at least lower the priority (higher number, e.g. 101) of all “non standard” repos, so that new packages get installed from those non-standard repos only when you want, not because an apparently “higher” version is available on a random repo.
Will try to lower then priority on some of those repos. And also I deleted the install repo number 9 like @malcolmlewis suggested.
I will in the future stay simply with package update and that’s it. The system is now stable and working fine.
the only problem was the locked nvidia repos that stopped me from updating the drivers.
Other then that now everything works fine .
Thanks for the help.
The idea is to test the setup of dual GPU’s on this desktop. Hence the need for updated driver for the nvidia card, the current one and future ones.
But I think the only option now is to have a dedicated Server, one of those cheap and still available servers from some years ago. They are on sale and they can take multiple PCIe GPU’s …
The only way to get cheap RAM now now is to use ECC DDR3 RAM. Slower PCIe and RAM does actually make a big impact on the response of local LLM’s … not ideal but sufficient.
RAM is also plenty and still cheap and available on those servers.
So I think every single day that Desktop local AI is truly not economical at all … but a dedicated used server changes the entire scenario.