nvidia hardware transcoding breaks after update

Posted by RyuikuX

Havent done much digging into the issue other than troubleshooting the symptom. Everytime i update Sparkbox, i need to reboot the host in order to regain hardware transcoding with my nvidia solution. i suspect it is something related to the containers being remade, but im not entirely sure. the symptom that tells me something is wrong is media will not load in jellyfin, it just hits a fatal error crash. i reboot the host and everything works as expected.

1 replies

Chris wrote:

Good news, this one's not on you. Restarting or updating SparkBox currently gets applied from inside a helper container that doesn't have your GPU attached to it, so the setting that turns Nvidia transcoding back on for Jellyfin doesn't always get reapplied afterward. A full reboot works because everything starts fresh, but there's a lighter fix: over SSH, run sudo sparkbox up media That reapplies your GPU setup for Jellyfin without a reboot. This is a known gap on our side that's already flagged to get fixed properly so you won't need to run that by hand. Try the command next time it happens and let me know if it clears the crash. — Chris