We are currently on v.24. Things work well but Talk is sluggish and the use of it sometimes appears to drive the web server to a halt. Is there a big improvement migrating to v.25 based on other’s results?
It’s always good to not be too far behind in terms of updates/upgrades. Generally there usually is quite some work on performance done for each major version. But your issue sound like a infrastructure issue, can you check if the requirements are met Server system requirements - Nextcloud Talk API documentation and please share some details about your setup?
I cannot find any reason that the instance behaves this way. It is very intermittent that it is a problem. Could be a couple weeks without any issue, or it could be a couple times in the same day. I have setup a TURN server and I’m running through the HPB settings as well. I do not see particular log issues. Htop doesn’t report any system usage issues. No signs of network congestion or blocks. We are even willing to pay an expert that can look over our installation for recommendations of settings we may have missed. We would rather pay for time accumulated rather than an annual subscription.
As I said, please check what I have linked. There are configuration known to creating problems.
Yes, thank you. I have previously used this page and set our installation up according to it. Still having issues. In all honesty, I do not even know if it is Talk that is creating the issue. It just seems to be when we are using it, even just for messaging, we notice the problem happening. But also, sometimes we get by using it without problems, even video chats.
Then it would be good to get some system details and maybe someone has an idea
Details, yes. I’m running on a DELL R640. ESXi with and a VM with 8 CPUs, 128GB RAM, 5.75 TB (for file storage). Ubuntu 20.04 with Nextcloud 24.0.6 running on Apache. I’m authenticating with an OpenLDAP server (also running iRedMail) as another VM on the same DELL box. Collabora is also installed as a third VM on the same box. All of my Nextcloud apps are up to date. I only have about 20 users on my server so far but will be adding about 100 more eventually.
Are you running mod-php oder php-fpm? Are there enough worker configured for the latter?
Also, just to be sure, which mpm model are you using? Just from the specs this sounds powerful enough.
Any reverse proxy in front?
php-fpm. I’m on 7.4. Reverse proxy is configured. mpm_prefork is used.
Then we are back at square one: Server system requirements - Nextcloud Talk API documentation
Ahh! So if I am following correctly, I’m mismatched on what is recommended. If I want to keep the fpm, I need to switch from mpm_prefork to mpm_events, correct?
Next month is EOL if you use non enterprise version. The question of whether the upgrade improves performance does not arise at all. It is necessary anyway.
Whoops. I think I see what you are saying now. If I’m going to stick with fpm I need to switch to events instead of prefork
Thanks. I didn’t find an EOL list previously. Major upgrades always make me nervous, even with backups… I just dread any downtime. But I look forward to being on the other side of the move. More excited about v.26 I think.