Arch Linux

ThomasN

Member

Registered: 2016-02-03

Posts: 31

[SOLVED]Nvidia Persistence Daemon fails to start

After upgrading yesterday, using sudo pacman -Syu, Nvidia Persistence Daemon fails to start, this in turn means xorg fails to start and I can't get into my desktop environment. The error I get is:

Failed to start NVIDIA Persistence Daemon.
See "systemctl status nvidia-persistenced.service" and "journalctl -xe" for details.

Looking into these suggestions, I found they both say:

Failed to query NVIDIA devices. Please ensure that the NVIDIA device files (/dev/nvidia*) exist, and that user 143 has read and write permissions for those files.

I then tried checking this and saw that "/dev/nvidia" does not exist. So it's clear why Nvidia persistence daemon fails, but I'm not sure how to proceed to get it to work again.

The system is a desktop with a Nvdia GTX1050 GPU. It worked normally until this morning. At the moment I'm typing any error message I see on my screen, because I'm posting this on my laptop while looking at my desktop screen.

Edit: Solved thanks to V1del

Last edited by ThomasN (2020-10-12 08:54:50)

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 21,846

Re: [SOLVED]Nvidia Persistence Daemon fails to start

There's no reason for the persistence daemon to be enabled or running for a normal desktop system. You should generally be able to just disable it.

However the fact that it fails in the first place might point to a race condition, follow setting up the modules in initramfs: https://wiki.archlinux.org/index.php/NV … de_setting

Also these files might miss if your kernel module doesn't load, check dmesg, maybe also post the pacman log of the problematic transaction.

Kabir

Member

From: India

Registered: 2016-12-06

Posts: 59

Re: [SOLVED]Nvidia Persistence Daemon fails to start

It's probably because I see a version mismatch. On my system I see 455.23.04-4, which is the same version in the repository. However the version in the package list is different 455.28.4

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 21,846

Re: [SOLVED]Nvidia Persistence Daemon fails to start

@Kabir if you do actually see that your primary mirror is broken pick one with a low delay time and put it at the top of your list: https://www.archlinux.org/mirrors/status/#successful

Kabir

Member

From: India

Registered: 2016-12-06

Posts: 59

Re: [SOLVED]Nvidia Persistence Daemon fails to start

@V1del thanks a bunch, that fixed it. I also got a whole lot of updates that I didn't get before.
Thanks again!

ThomasN

Member

Registered: 2016-02-03

Posts: 31

Re: [SOLVED]Nvidia Persistence Daemon fails to start

Thank you for your help.

There's no reason for the persistence daemon to be enabled or running for a normal desktop system. You should generally be able to just disable it.

In that case, the failing Daemon is just another symptom of the issue that also makes me end up with a failed xorg, right? (rather than the cause of it)

However the fact that it fails in the first place might point to a race condition, follow setting up the modules in initramfs: https://wiki.archlinux.org/index.php/NV … de_setting

Following those instructions, I added the nvidia, nvidia_modeset, nvidia_uvm and nvidia_drm modules to Mkinitcpio. When I now run mkinitcpio -P I get the following error for each of these modules:

ERROR: module not found: `nvidia(_...)'

What am I doing wrong?

Also these files might miss if your kernel module doesn't load, check dmesg, maybe also post the pacman log of the problematic transaction.

I'm sorry, I'm not sure what I'm looking for in dmesg and the output is enormous. Any suggestions on what should/shouldn't be there? My pacman log shows the nvidia driver getting updated to 455.28 without error.

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 21,846

Re: [SOLVED]Nvidia Persistence Daemon fails to start

The kernel log/dmesg should show nvidia getting loaded. in doubt post them (the kernel log and the pacman log) here, but if mkinitcpio fails that's a good tell that there's some mismatch going on, could also be that you are loading the old kernel but have the newer nvidia module. (you should have 5.8.14-1 on linux and 455.28-4 on the nvidia package)

Also double check what I told Kabir as well.

Last edited by V1del (2020-10-12 08:30:45)

ThomasN

Member

Registered: 2016-02-03

Posts: 31

Re: [SOLVED]Nvidia Persistence Daemon fails to start

(...) could also be that you are loading the old kernel but have the newer nvidia module. (you should have 5.8.14-1 on linux and 455.28-4 on the nvidia package)

Also double check what I told Kabir as well.

This was it! I had Linux 5.8.13, but did have the newer nvidia module. After Kabir posted, I checked to see if I had the latest nvidia package (and I did), but I didn't check the Linux package version. Turns out I had a bad mirror in my mirrorlist as well. Fixing my mirrorlist, then running pacman -Syu and then rebooting solved it and gave me a working Xorg again.

Thanks!