I’m using NextCloudPi on a RaspberryPi 4 for some time now. It’s working great and Nextcloud is very responsive. A few weeks ago I decided to install Talk App. I followed the instructions HowTo: Setup Nextcloud Talk with TURN server to install coTURN and configure Talk. Talk is now usable. Or let’s say bearable… As soon as two participants join the same conversation via Browser (Firefox, Chrome, Safari… doesn’t matter) things are getting slow. Every single action takes approx. 20s. Sending a message, starting a call, changing names, even leaving Talk takes up to 20s. Monitoring CPU load with “htop” while Talk is lagging shows: CPU is either in idle or has very little load. (0.12-0.25 average) So it’s not caused by lack of CPU power.
I’m now trying to find the root cause of this and improve speed. What I did so far:
- https://github.com/nextcloud/spreed/issues/2211 and https://techwombat.com/enable-http2-apache-ubuntu-16-04/ gave me some hints. I’ve checked the Apache, HTTP/2 and PHP settings. NextCloudPi seems to be configured correctly according to these links. But it’s still slow.
- Someone mentioned that he had to use mod_php and mpm_prefork instead of php-fpm and mpm_event. So I tried:
Talk is now as fast as it should be and not stalling anymore.
But there’s a problem. I don’t know if these changes have an impact on system stability, performance, security or maintainability. I’ll switch back to the old configuration. Better to have a slow system that was configured by an expert rather than a fast and unstable one configured by a noob
Nevertheless: It seems some people are running Talk with Apache2, mpm_event, PHP-FPM and HTTP\2 without problems. I would like to debug and find out what’s wrong with the NextCloudPi configuration. But I don’t know where to start.
- Does anybody know what could cause Talk to stall the system?
- What’s the best way to find the process or configuration that causes the problem?
- Where can I find the log-file of Talk?
Nextcloud: 18.0.4 (installation based on NCP image “NextCloudPi_RPi_03-28-20.tar.bz2”)
Hardware: Raspberry Pi 4B 4GB