Increase Nextcloud HPB limit?

nextcloud / all-in-one

📦 The official Nextcloud installation method. Provides easy deployment and maintenance with most features included in this one Nextcloud instance.

https://hub.docker.com/r/nextcloud/all-in-one

GNU Affero General Public License v3.0

5.61k stars 652 forks source link

Increase Nextcloud HPB limit? #400

Closed muellerm01 closed 2 years ago

muellerm01 commented 2 years ago

How to use GitHub

Please use the 👍 reaction to show that you are affected by the same issue.
Please don't comment if you have no relevant information to add. It's just extra noise for everyone subscribed to this issue.
Subscribe to receive notifications on status change and new comments.

Steps to reproduce

Software Install Nextcloud AIO 0.8 like documented

Hardware: Azure North Europe Standard D4as v4 (2 vcpus, 8 GiB Arbeitsspeicher) 8 Gbit download (measured) 1,8 Gbit upload (measured)

Steps:

start a meeting with allowed guestes
the 7th person will take a little bit longer to join
the 8th person is not able to join the meeting

Expected behaviour

Nextcloud talk participant limit should be releated to CPU, RAM and upload/download speed

Actual behaviour

CPU runs on 0,5 cores RAM 2,5 GB used Upload/Download ~20Mbit and participant limit is between 7 and 8

Nextcloud AIO version

AIO 0.8

Current channel

latest channel

Other valuable info

szaimen commented 2 years ago

Hi, thanks for your report. I fear though that there is not much we can do here because we install everything based on the docs.

Feel free to read the following docs though: https://github.com/nextcloud/spreed#how-to-have-the-maximum-number-of-participants-in-a-call

szaimen commented 2 years ago

Also it seems like your cpu is pretty limitied. Feel free to try to increase the cpu count

muellerm01 commented 2 years ago

OK, sorry I thought HPB was part of AIO because it is configured. I requested your sales team but think the minimum user count is around 100 to get enterpise support and search for a solution below that. Let´s see what i will get from them.

I also tried it with a larger CPU, but that doesn't seem to change anything.

I think in the range of 50 participants and an on premise installation. But I could not find a purchasable solution so far.

Hosting is not an option for me. Then I could just take teams.

szaimen commented 2 years ago

OK, sorry I thought HPB was part of AIO because it is configured.

yes, the HPB is part of AIO.

szaimen commented 2 years ago

It seems like it is just not scaling as good as promised...

muellerm01 commented 2 years ago

But than HPB with a limit of 7-8 thems to be confusing.

szaimen commented 2 years ago

But than HPB with a limit of 7-8 thems to be confusing.

yes, it is

muellerm01 commented 2 years ago

Is something configured incorrectly?

szaimen commented 2 years ago

Noth that I am aware of.

szaimen commented 2 years ago

You can have a look yourself: https://github.com/nextcloud/all-in-one/blob/main/Containers/talk/start.sh

szaimen commented 2 years ago

Maybe this is the problem for you? https://github.com/nextcloud/all-in-one/blob/8fedcb48109a62e794eb4ae6a5ae258ae43c4ef8/Containers/talk/start.sh#L55

szaimen commented 2 years ago

feel free to adjust that number in the container and try again.

muellerm01 commented 2 years ago

I looked, but I only know the documentation from others and not what is recommended.

Hm ok good point, I'll try that.

szaimen commented 2 years ago

Feel free to report back if adjusting this number makes it work.

szaimen commented 2 years ago

We can then adjust the number if it works for you.

szaimen commented 2 years ago

I'll reopen this until then :)

szaimen commented 2 years ago

Or lets go the other way around. I'll increase the value and you can test it with the next release.

muellerm01 commented 2 years ago

I tryed docker stop docker copy :/usr/bin/start.sh start.sh changed connectionsperhost = 40 docker cp start.sh :/usr/bin/start.sh docker restart

But nothing changed, the limit is still between 7 and 8

muellerm01 commented 2 years ago

the janus log looks like a crash loop but hard to read

szaimen commented 2 years ago

What is the output of cat /etc/signaling/server.conf after a restart?

szaimen commented 2 years ago

the janus log looks like a crash loop but hard to read

Interesting! Can you share the logs?

muellerm01 commented 2 years ago

connectionsperhost = 40

szaimen commented 2 years ago

connectionsperhost = 40

Hm...

muellerm01 commented 2 years ago

log-part-janus.log

muellerm01 commented 2 years ago

I also tryed to disconnect user 7 and 8 and than reconnect again 8 before 7 the last user is always the one with the problem

szaimen commented 2 years ago

I also tryed to disconnect user 7 and 8 and than reconnect again 8 before 7 the last user is always the one with the problem

So there seems to be a hard limit at 7-8 somehow...

muellerm01 commented 2 years ago

Hm I have a HPB running with these sources and docker-compose, it looks like this is not limited to 7-8

szaimen commented 2 years ago

Hm I have a HPB running with these sources and docker-compose, it looks like this is not limited to 7-8

So then it is probably a configuration issue, right.

muellerm01 commented 2 years ago

I used this guide for configuration https://markus-blog.de/index.php/2020/07/30/how-to-install-nextcloud-talk-high-performance-backend-with-stun-turnserver-on-ubuntu/

szaimen commented 2 years ago

Thanks! At first glance it looks comparable. I'll adjust our config to match this guide a bit better.

muellerm01 commented 2 years ago

thx but can´t say what is different hm everything is splitted in 4 container

szaimen commented 2 years ago

Hey, I just published a new talk container. Feel free to try that one out and please report back if it is now better than before. :)

muellerm01 commented 2 years ago

Hm, has not changed the behavior.

szaimen commented 2 years ago

Hm...