Server Version#: 1.25.4.5487
Player Version#: 1.40.1.2823-6f60f3a7
I recently upgraded to an i5-12600k for QuickSync hw transcoding. However, if I turn on hw acceleration (with or without decoding) my server will inevitably crash at some point after it starts transcoding videos. Everything seems fine until the crash eventually happens. The last two lines of the server log are always as follows:
Feb 01, 2022 14:56:57.227 [0x7f87e065bb38] DEBUG - [Transcode] Codecs: testing hevc (decoder) with hwdevice vaapi
Feb 01, 2022 14:56:57.228 [0x7f87e065bb38] DEBUG - [Transcode] Codecs: hardware transcoding: testing API vaapi
Is this a known issue? I’m having real trouble diagnosing this.
What kernel version are you using? I am able to do hardware transcoding with my Alder Lake using the 5.13.0-28-generic kernel. I haven’t had a lot of luck using and of the newer kernels, but they are supposed to have better/more support for Alder Lake CPU’s.
i have an alder lake as well and so far not many problem it’s pretty stable
don’t use rod for now because that will break hw
share your full specs please
This is not directly related to plex, but should help. I was having similar transcoding problems with ffmpeg raw ffmpeg -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format vaapi -i /tmp/sample.mkv -c:v:0 hevc_vaapi /tmp/sample-out.mkv - the latest build of jellyfin-ffmpeg fixed this issue for me /usr/lib/jellyfin-ffmpeg/ffmpeg -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format vaapi -i /tmp/sample.mkv -c:v:0 hevc_vaapi /tmp/sample-out.mkv .
Sorry - v silly of me. The default ffmpeg transcode using vaapi to hevc would produce total garbage - nothing viewable, I’ll try to go back and get a screenshot.
The parched ffmpeg produced proper output.
It wasn’t dissimilar to some of the issues I’ve seen on Plex transcoding on alder lake, so I thought it might have a similar root cause.