2024-01-12, 02:29 AM
I'm running Jellyfin server 10.8.10.0 on a Ubuntu server with an Intel CPU (Gen 11, integrated graphics). On some specific files, I get an ffmpeg memory allocation bug when trying to do hardware acceleration via QSV on HEVC encoded videos. The error message always looks like this:
This happens no matter where I start playback from, but it always occurs 2-5 minutes before the end of the video. It's strange because when I run
I don't think the files are corrupted because I can play them with no issues using VLC. Also, if I re-encode the video to use h264 then Jellyfin can handle it no problem. However, this only happens with certain HEVC files, and I can't tell what separates the ones that work from the ones that don't.
Has anyone seen an error like this before? I'd like to avoid re-encoding a bunch of files if possible.
[Parsed_scale_qsv_4 @ 0x56365ccf2e00] Can't allocate a surface.
[Parsed_scale_qsv_4 @ 0x56365ccf2e00] Failed to query an output frame.
Error while filtering: Cannot allocate memory
Failed to inject frame into filter network: Cannot allocate memory
Error while processing the decoded data for stream #0:0
Conversion failed!
This happens no matter where I start playback from, but it always occurs 2-5 minutes before the end of the video. It's strange because when I run
htop
to see memory usage, I still have plenty of RAM available, so I'm not sure why ffmpeg would be running out of it. I don't think the files are corrupted because I can play them with no issues using VLC. Also, if I re-encode the video to use h264 then Jellyfin can handle it no problem. However, this only happens with certain HEVC files, and I can't tell what separates the ones that work from the ones that don't.
Has anyone seen an error like this before? I'd like to avoid re-encoding a bunch of files if possible.