Open ivnc opened 6 months ago
I'm experiencing this problem on both Ubuntu 22.04 and 24.04. Maybe @beqabeqa473 is thinking of something.
I'm experiencing this problem on both Ubuntu 22.04 and 24.04. Maybe @beqabeqa473 is thinking of something.
Also in docker or with a direct installation in the host?
Also in docker or with a direct installation in the host?
I'm using a direct installation with an isolated user on the host.
CC: @cyrmax
@ivnc, does this issue still occur? Is there a log for this?
I dont know because I have already downgraded the sdk and dont have an alternative machine to test without affecting my production environment. Nor the bot container nor the bot itself produced any relevant log messages.El 17 ago 2024, a las 15:29, Ozancan Karataş @.***> escribió: @ivnc, does this issue still occur? Is there a log for this?
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>
Since TeamTalk SDK was updated to 5.15, the bot causes a extreme & rising CPU load in the system, causing audio micro-cuts to the bot and performance problems for other services in the machine. I've tried to downgrade all components except ttsdk and the problem persists. It can only be resolved downgrading the TeamTalk SDK to 5.8.1. With 5.15 I have seen CPU usage values of up to 60 by the process running the bot, while with 5.8.1 it remains at 4/5 with eventual peaks, especially after startup, of 10/11. With 5.15 it starts at 10/11 as well, but keeps constantly rising. If there are various bot instances running in the same machine this situation can cause a heavy CPU load and serious performance issues to the whole machine and the services it hosts. I'm running Centos 7 (7.9.2009) with Docker 26.1.1, build 4cf5afa. I've also tested the new dockerfile in docker-behavior-improvements and the issue is still reproducible, the problem is apparently in the ttsdk.