we have installed elgg 1.8 on two different servers with this specification:
1- 4cpu, 2.8 ghz, 4GB ram
2- 8cpu, 2.8 GHz, 16 GB ram
we used wapt pro 2.5 software to emulate users simultanious activities and the results are described below:
cpu usage is immediately increased, RAM usage reached full usage, and eventually the server crashed.
when we reached above 150 users, the server begins to respond "time out" warning.
what should we do about that? how can we resolve this problem?
info@elgg.org
Security issues should be reported to security@elgg.org!
©2014 the Elgg Foundation
Elgg is a registered trademark of Thematic Networks.
Cover image by RaĆ¼l Utrera is used under Creative Commons license.
Icons by Flaticon and FontAwesome.
- Cash@costelloc
Cash - 0 likes
- I would expect that cpu usage would increase when you start to load the server (wasn't sure what your point was there)
- RAM usage depends a lot on server configuration and also on number of plugins/what plugins. You will see different RAM usage patterns Apache vs nginx.
- Make sure you are running the latest version of Elgg
- Make sure you have all the caching turned on for Elgg
- Make sure you are using something like APC
- Search plugin hits the DB extremely hard
- DhrupDeScoop@Dhrup2000
DhrupDeScoop - 0 likes
You must log in to post replies.I hope you realize that hitting a server like that is not at all an accurate picture of how many users you can support. It is a worse case scenario - slashdot effect kind of analysis.
which server crashes ? #1 or 2 ? 100% ram usage ! does not sound right ;(
we've had 150 simultaneous users some 2 years ago.. with no problems except occasional 'warnings' or slower responses with similar server hardware.
one small piece of advice for this situation.. take from our experience
the fine-r tuning that we did not to get to that state was not achieved with short questions and quick answers -- a fair lot of apache, mysql, linux configuration kopw-how and.... a lot of patience over time to monitor performance and config changes' resultants was done before the server showed improvements. not a task for weak-hearts ;-) when the server behaves unexpectadly and requires very quick restores to config while scratching heas to deternine what went wrong..