I have the same issue.
After I run sudo dnf upgrade today, I get error logs related to nvidia-powerd.service.
I got following by systemctl status nvidia-powerd.
× nvidia-powerd.service - nvidia-powerd service
Loaded: loaded (/usr/lib/systemd/system/nvidia-powerd.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Wed 2022-02-09 22:28:20 JST; 7min ago
Process: 1009 ExecStart=/usr/bin/nvidia-powerd (code=exited, status=1/FAILURE)
Main PID: 1009 (code=exited, status=1/FAILURE)
2月 09 22:28:20 PC-NAME systemd: Starting nvidia-powerd service...
2月 09 22:28:20 PC-NAME /usr/bin/nvidia-powerd: nvidia-powerd version:1.0(build 1)
2月 09 22:28:20 PC-NAME /usr/bin/nvidia-powerd: No matching GPU found
2月 09 22:28:20 PC-NAME /usr/bin/nvidia-powerd: Failed to initialize RM Client
2月 09 22:28:20 PC-NAME systemd: nvidia-powerd.service: Main process exited, code=exited, status=1/FAILURE
2月 09 22:28:20 PC-NAME systemd: nvidia-powerd.service: Failed with result 'exit-code'.
2月 09 22:28:20 PC-NAME systemd: Failed to start nvidia-powerd service.
with kernel version 5.16.5 and nvidia proprietary driver (from rpmfusion) version 510.47.03
Both reinstalling the kernel and booting an older kernel (5.15.17 with nvidia 510.47.03) did not solve the problem.
However, I think before I performed upgrade (including kernel upgrade) today there had been no issue, because at that time there was no notification of service failure in my Cockpit’s overview page.
I found this post on the nvidia forum: No matching GPU found with 510.47.03
According to the post, we can ignore this error.
Those who have AMD CPU or desktop GPU could be able to ignore this error(?).
(Sorry. I have desktop GPU, but others might not.)
Additionally, most of characters and icons in the GNOME shell are not rendered well (become like corrupted texture) when resuming from suspend, after the same upgrade. This symptom went away after the next upgrade (to kernel 5.16.7), performed yesterday. (The service failure still remains.) reference (nvidia forum)
Is this related?
If my post does not match the case of the original questioner @sojonny, I’m sorry, and I will make another topic.
@oyster I split the threads since even though your issue may be the same it is difficult to reply to 2 questions that are similar and keep track of who is being spoken to with the responses. The way you jumped in may be considered hijacking by some.
The person in that post whose answer was marked as the solution is correct in that it was developed for mobile users who often suspend when closing the lid on their laptop. The nvidia-powerd service does however do the same thing for desktop systems with an nvidia gpu.
Users with a desktop system who never suspend or hibernate should never have that installed anyway. If they do suspend or hibernate it is needed for a recovery of the GPU at the same state when suspended.
It was what I was concerned, but I didn’t know how much I might talk about my symptoms of the related/same issue on a topic which another user opened. (I didn’t know that the term hijacking, but I was worried that it could be something like that. However, I posed some question, and I should have opened new one.)
I’m sorry for my impoliteness and thank you very much for teaching me and splitting the issue.
The failure of nvidia-powerd is confirmed in the kernel 5.15.17, 5.16.5 and 5.16.7, and corrupted texture (perhaps because of corruption of graphic memory after suspend?) persists in kernel 5.16.7 (not checked in kernel 5.15.17).
To me it looks like the xorg-x11-drv-nvidia-power package may not have gotten properly updated for the newer nvidia driver. You are not the only one to report the problem with the nvidia-powerd.service not starting so I suggest either reporting this as a bug at ReportingBugs - RPM Fusion or just waiting until it gets caught up.
The problem with video recovery after suspend/hibernation will likely continue until that package gets updated properly and those 3 services are running properly again. The link previously noted tells what the package does and why it was created.
In simple terms, that package saves and restores the gpu memory content for suspend and restart.
@rabiny thank you. I tried the commands you wrote, but it produced balanced outputs (4 were disabled and then enabled) and after that the issue remains.
@computersavvy thank you, and I think there are no more thing I can do to workaround this issue, as you said. I am considering reporting at RPM Fusion Bugzilla, if I can gather some meaningful information on my computer about this issue.
I got the same message with an update & reboot a few days ago and after an update & reboot last night it is now gone.
I mis-spoke above.
The problem seemed to appear when my system updated akmod-nvidia from
akmod-nvidia x86_64 3:510.47.03-1.fc35 rpmfusion-nonfree-updates 28 k
akmod-nvidia x86_64 3:510.47.03-2.fc35 rpmfusion-nonfree-updates 28 k
and the remainder of the nvidia packages from rpmfusion remained at 3:510.47.03-1.fc35
This definitely appears to be a bug introduced by an rpmfusion incomplete update. On my laptop where I may suspend the error does not appear. On my desktop, where I won’t suspend, it does appear. Both systems have the same packages and versions installed.
I have never installed (intentionally) the xorg-x11-drv-nvidia-power package on my desktop but it appears that it may be installed when installing or updating the akmod-nvidia package so that seems to be why it was there on my desktop.
It appears to be a benign error so should not cause issues unless actually suspending or hibernating on the machine where it is reported.
I am prevented from filing a bug report at rpmfusion since they do not allow gmail accounts to connect.