papaHoff

Yeah, that won't be hard, Jellyfin is the only active docker and VM at the moment.

No, never, it becomes completely impossible to use. It was just now i added the "-m 28G" to limit the amount of RAM it uses, since when it locks down, I also enabled SSH to be able to go in here and check while it happens. Looking at it right now, it is doing some background tasks and directly shoot up to It looks as if it is doing some form of library scan, it is using ffprobe a lot and simultaneously.

My unRaid 6.12.13 has been up for some months now and i rarely use it for anything else but these three dockers containers: - pi-hole - jellyfin (tried other containers) - homeassistant Even before having this issue i only had jellyfin and nothing else. So i've concluded that it must be here some issue is. So what's the issue then, well, when streaming any content that needs to be transcoded by ffmpeg to a target it seems to choke up at some point. It can be randomly at night, in the afternoon and sometimes even when it is being unused. So my first conclusion was that it must be some form of scheduled task that goes wrong... right? But then again, the tasks takes less than 2 minutes to complete and most of the time I've been awake for it to happen, the server seems usable. Since this randomly happen I suppose it could be a hardware issue as well as a software issue, but I would like to know more about the error. So when I come back after some time the system can be locked like this: And there's no other way to stop it from doing this than to simply do a hard reboot the whole server. tower-diagnostics-20241111-1852.zip

I want to add to this that on a new container, i also had to go to: http://<ip>:8123/config/integrations/dashboard And actually add the usb-device explicitly (perhaps obvious to anyone). And then once i added it by clicking on Configure and in my case choosing Zigbee - and the wizard was done, i could confirm it by going in to the uri: http://<ip>:8123/config/hardware. The above guide, from @PzrrL was therefor completely correct and there's no need to set the docker instance in Privileged mode (make everything accessible). You can also make sure it exists by entering the console on the docker-image and go: ls /dev There you should see the ttyUSB0 as well.

Well, the most common error i see is that the CPU gets overloaded and all cores go to 100% and stays there. Sometimes even unraid and all docker-images get unresponsive and the only solution is to kill the server by holding the powerbutton. Supposedly i believe that could damage the OS in the longrun. I suppose this could might as well be an issue with the docker for jellyfin doing something, but Playback on phone: Playback on chromecast: At this point, just as you pointed out, the gpu-statistics plugin notices that the GPU is being used, and something is clearly happening on the jellyfin-side, since the log prompts this: The attempt was made with several variations of all encoded movies using H.264, NVEnc (using HandBrake). I also waited for 10 minutes, but the process information never gets populated. Even when they presumably use the GPU, since the stats of the GPU raises. Right now, doing playback to the TV: But still the same from nvidia-smi. TMI; but makes me feel like this all the time, cause i gotta kill the server, which i suppose could be from it trying to do everything on the CPU instead: It happened again now, I'm doing nothing on the server at the moment, except stream a 576p bad dvd and a 1080p clip at the same time: And the GPU was doing nothing.

Is there a good reason as to why the hardware acceleration isn't working? The id is correct. tower-diagnostics-20241103-1106.zip When playing a h.264 1080p clip on my chromecast: