SOLVED - Will not play .avi files

[Huecuva]

I have Jellyfin running in a headless Debian VM on Proxmox. My VM has a GTX1070 passed through to it for the purpose of transcoding. I put it in the server several years ago so I don't remember for sure, but I think I recall nvidia-smi resulting in the expected information about the GPU. The other night my girlfriend and I were watching movies and I noticed that Jellyfin would not play any .avi files. I don't recall having this many issues playing movies in Jellyfin in the past. I've been using it for three or four years at this point and I don't remember it having trouble playing so many movies. I checked the jellyfin logs and learned that jellyfin does not seem to detect a cuda device, however inxi -F, nvidia-detect, and lspci all list my GTX 1070 as present and that it is supported by all drivers, though there appears to be no driver loaded for it. When I attempt to use nvidia-smi now it says that it cannot communicate with the driver. Elsewhere, I have read that purging the nvidia driver and reinstalling should solve that issue. I have purged and reinstalled the driver several times and nvidia-smi still cannot communicate with it. 

To be honest, I'm not even 100% certain that the driver is the reason Jellyfin will not play .avi files, but since Jellyfin doesn't seem able to detect a cuda device it seems like the driver might be related. If anyone here can help me figure this out, it would be much appreciated.

[Fate]

post the jellyfin log and transcoding logs to pastebin.
post the output of "vainfo"

Which driver are you using?

[Huecuva]

After attempting to play a .avi file of Constantine, these logs appear:

Transcode log
Jellyfin log

The driver I have installed:

Code:

sudo apt install nvidia-driver
[sudo] password for <user>:
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
nvidia-driver is already the newest version (525.147.05-7~deb12u1~bpo11+2).
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

My VM is running Debian 11. I tried to upgrade it to Debian 12 but there are space constraints and that's not going to work. I will have to live with 11 for now. It has been suggested to me that with Debian 11 I should not have the 525 driver. However, my apt sources list has only Debian 11 Bullseye and Bullseye backport repositories. Nothing for Debian 12. Again, I've purged the driver and reinstalled several times and that's the driver that installs.

I should also mention that the VM is booting BIOS, not UEFI so SecureBoot should not be an issue. Also, during boot, the VM complains that it fails to start the nvidia persistence daemon.

Code:

$ sudo systemctl status nvidia-persistenced
● nvidia-persistenced.service - NVIDIA Persistence Daemon
    Loaded: loaded (/lib/systemd/system/nvidia-persistenced.service; enabled; vendor p>
    Active: failed (Result: exit-code) since Mon 2024-05-20 10:20:24 PDT; 16h ago
    Process: 823 ExecStart=/usr/bin/nvidia-persistenced --user nvpd (code=exited, statu>
    Process: 914 ExecStopPost=/bin/rm -rf /var/run/nvidia-persistenced (code=exited, st>
        CPU: 15ms

May 20 10:20:24 Ilithid systemd[1]: Starting NVIDIA Persistence Daemon...
May 20 10:20:24 Ilithid nvidia-persistenced[836]: Started (836)
May 20 10:20:24 Ilithid nvidia-persistenced[836]: Failed to query NVIDIA devices. Pleas>
May 20 10:20:24 Ilithid nvidia-persistenced[823]: nvidia-persistenced failed to initial>
May 20 10:20:24 Ilithid nvidia-persistenced[836]: Shutdown (836)
May 20 10:20:24 Ilithid systemd[1]: nvidia-persistenced.service: Control process exited>
May 20 10:20:24 Ilithid systemd[1]: nvidia-persistenced.service: Failed with result 'ex>
May 20 10:20:24 Ilithid systemd[1]: Failed to start NVIDIA Persistence Daemon.

[Fate]

there is a note in the arch wiki that driver 525 breaks va-api.
https://wiki.archlinux.org/title/Hardwar...celeration

Can you try 520?

Also btw. if you run proxmox. You can run jellyfin as a container instead of a VM, this saves a bit of overhead and also some storage.

[Huecuva]

Well, whenever I feel ambitious enough to try setting everything up again from scratch to upgrade to Debian 12, maybe I will try a container instead. Until then, a VM works just fine for me.

How do I go about installing a specific driver? As I mentioned, the 525 driver installs by default when I install the nvidia-driver package.

Also, I neglected to post the results of vainfo. That command does not appear to exist on my server.

I just tried purging and reinstalling the nvidia driver again and for some reason this time it appears to have installed the 470 driver. However, it says that dependency problems prevent configuration of the nvidia-driver package. When I attempt to sudo apt --fix-broken install it is unable to fix them:

Code:

$ sudo apt --fix-broken install
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
2 not fully installed or removed.
After this operation, 0 B of additional disk space will be used.
Setting up nvidia-kernel-dkms (470.223.02-1) ...
Removing old nvidia-current-470.223.02 DKMS files...

------------------------------
Deleting module version: 470.223.02
completely from the DKMS tree.
------------------------------
Done.
Loading new nvidia-current-470.223.02 DKMS files...
Building for 5.10.0-29-amd64
Building initial module for 5.10.0-29-amd64
Error! Bad return status for module build on kernel: 5.10.0-29-amd64 (x86_64)
Consult /var/lib/dkms/nvidia-current/470.223.02/build/make.log for more information.
dpkg: error processing package nvidia-kernel-dkms (--configure):
installed nvidia-kernel-dkms package post-installation script subprocess returned error exit status 10
dpkg: dependency problems prevent configuration of nvidia-driver:
nvidia-driver depends on nvidia-kernel-dkms (= 470.223.02-1) | nvidia-kernel-470.223.02; however:
  Package nvidia-kernel-dkms is not configured yet.
  Package nvidia-kernel-470.223.02 is not installed.
  Package nvidia-kernel-dkms which provides nvidia-kernel-470.223.02 is not configured yet.

dpkg: error processing package nvidia-driver (--configure):
dependency problems - leaving unconfigured
Errors were encountered while processing:
nvidia-kernel-dkms
nvidia-driver
E: Sub-process /usr/bin/dpkg returned an error code (1)

Maybe it would be easier to simply reinstall the OS (and upgrade to Debian 12) and set my server back up from scratch?

[TheDreadPirate]

You can specify driver versions. Nvidia-driver-520 or something like that.

[Huecuva]

Okay. I'm pretty sure I've tried that and it just said the package did not exist, but I will give it another try when I get a chance. I have to leave for work now.

[Huecuva]

Yeah, so I'm unable to install specific drivers for some reason:

Code:

$ sudo apt install nvidia-driver-520
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
E: Unable to locate package nvidia-driver-520

Code:

$ sudo apt install nvidia-driver-490
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
E: Unable to locate package nvidia-driver-490

Either way, now when I sudo apt install nvidia-driver I get the 470 driver instead of 525, but it still doesn't work so even if I could install a specific driver, I don't think it would solve anything.

[TheDreadPirate]

In your VM, what is the output of "sudo lshw -C display"?

[Huecuva]

Code:

$ sudo lshw -C display
sudo: lshw: command not found