Jibri experienced a situation where the service filled up memory after following the recording in the background
You need to reserve more resources for it. At least 4 or 8GB of Ram and enough disk storage to accommodate the maximum time of recording you want before the recording is over and you transfer the recording to some storage.
If you are doing 1080(the default) recording that requires more resources compared to 720p of course.
May I ask you a question? Is this recording always in memory and not written to the disk until the recording is completed? Why can I only record for 10 minutes when running 32g of memory and release memory space after the meeting is over
No, its saved on disk, is there enough disk. What is the cpu usage? You can see in ffmpeg logs is it struggling to because of not enough cpu.
ffmpeg.0.txt (565.2 KB)
thanks ，i did not find any related issues in the log， Is there any other settings that I haven’t set for the jibri that I started using Docker? This issue will occur after recording for about ten minutes now
Your problem is insufficient resources. Your ffmpeg log proves that. How many cores do you have reserved for Jibri on your server?
How to determine the specific number of CPU cores occupied by jibri? I think the number of cores occupied should be reasonable
How to determine the specific number of CPU cores occupied by jibri? I think the number of cores occupied should be reasonable
You can search in internet how to do that, how to check it from sh inside container or from the host system. Also there are mentions there that for example if you are using it on Windows host it uses less than the host on and that needs adjustments … there are plenty of resources on the subject. Basically what you need is to make ffmpeg happy to run with speed of 1 and more … but to no allow it to be lower than 1 as it starts to use RAM and swap till it stops or crashes.
Okay, thank you. I’m researching it myself. If there are any other questions, I hope I can consult you again
Can you help me analyze it? The hand that is currently using the backend for recording will crash，The following is a complete log of the crash during the recording process
ffmpeg.0.txt (252.8 KB)
log.0.txt (58.7 KB)
The following is a complete log of the crash during the recording process
You can see how ffmpeg becomes slower and slower till it returns an error.