ffmpeg streams sometimes don't die (docker container 0.6.2)

[scumola]

I'm finding that sometimes, ffmpeg streams aren't dying properly. After a few days, I'll find ffmpeg still running in the container. I've got a workaround that just restarts the container once at 3am every morning, but it's kind of a hacky workaround. Could we add something to kill off ffmpeg processes that have been running for > 12 hours or something maybe?

Other than that, I'm a big fan of the project. Good work!

[deathbybandaid]

@tgorgdotcom I fixed this in fHDHR by adding except GeneratorExit:. This allows ffmpeg to safely shut down when a client disconnects. However I'll also note that I run the ffmpeg subprocess in a generator.

[scumola]

So this is fixed in the current (> 0.6.2) docker build then?

[deathbybandaid]

@scumola fHDHR is a project similar to l2p and I was just suggesting to tgorgdotcom what seems to have fixed it for that project.

[scumola]

I'm still seeing streams open for > 24 hours (see the ffmpeg stream that was open since Jan24)

root@docker01:~/locast2plex# ps auxw | grep ffmpeg
root     1716435  0.0  0.4  84696 37720 ?        S    Jan24   0:02 ffmpeg -i https://den.locastnet.org/proxy/den/variant/wBCLh5JpeYVNoiPQfzjoXY6Z9YZu6impsGhWTCeCDx2zglPu3GGZ1WihqUXFQrKWamm9OkjMVEYpuD5UbWnf5ikkslNjPf2LC67rLFM2ygp.m3u8 -c:v copy -c:a copy -f mpegts -nostats -hide_banner -loglevel warning pipe:1
root     1785025  0.4  0.4  85068 39032 ?        S    07:00   0:24 ffmpeg -i https://hls.locastnet.org/proxy/srte2b1/variant/TIRj39AB6tIYoXQPLubozOLG1m10NK9Bzy4dAV6HRi5KdZ5AhXJpAYa7eJ5VCfEHGY8oMn1hcNJnXXFItwpZmw.m3u8 -c:v copy -c:a copy -f mpegts -nostats -hide_banner -loglevel warning pipe:1