New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Apache cpapacity - calculation
HI Guys ,
I really need help here, I know some system administrators have some formula to calculate to apache variables such as timeout, keepalive timeout ,maxclients and so on.
Anyone can suggest if my client needs a server which can handle 1000 or 10000 customers at a time, so how much RAM,processor needed for this server and how should we optimize apache or nginx accordingly, I have searched on google and ended up nowhere.,
If you are an apache tuner, suggest me some practical ways. Its purely related apache architecting.
Comments
Which application do you run?
If something PHP, switch to PHP-FPM and preferably use nginx, at least as a reverse proxy for static file serving.
Hi,
I need to know how do the calculation bro, how do we calculate such things
There is no "calculation". You look at monitoring and logs, figure out what takes long/what consumes most system resources and then research how it can be optimized.
Wait, you mean there's real work involved? Well that changes everything...
About fiddytwo.
You mean there's a difference between a php+mysql+merry-plugin-zoo site and a couple of static pages? Damn! You seem to be one of them guys who always make things complicicated when all he needs is some practical purely related apache or nginx architecting.
@bootstrap, hold em horses! Might actually be fiddythree. In case there are more than kinda many customers and you wanna go luxurious n' high end you might even turn it up to fiddyfour.
And don't you trust em guys who tell ya "42"! They're a buncha clueless liars. 42 ain't apache. 42 is sendmail tuning.
Nuh, just a bit of, you know, apache and customers and processor stuff calculations plus some architecting. No big thing. It's just that Google for some reason didn't tell him the system administrators formula.
Generally, Web server traffic or any application performance depends on server resources.
As we talking about apache config, we can use following standard values
Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
MinSpareServers 5
MaxSpareServers 10
StartServers 5
MaxClients 150
MaxRequestsPerChild 0
Also, Nginx use as a reverse proxy is a good option to handle server traffic.
Sorry but I think that for a totally unknown system and totally unknown usage and load the correct value is fiddytwo or max. fiddyfour.
42
I don't agree at all. You give some magic numbers (you know exactly nothing about the environment) that that are a "standard".
Some of these values make literally zero sense in pretty much ALL environments. Timeout 300. Why do you need 300 seconds? Default in latest 2.4 is 60 now, which is pretty good in most situations.
All my question is apache values depends on ulimit (open file limits) of a linux server, My system admin gives me a calculation based on that.
The answer still is fiddytwo.
An alternative answer would be: you will have to learn at least the basics. At the very minimum that would allow you to ask reasonable questions.
If you're still using Apache in 2018 then the only calculation you need to know is…
DON'T
She's the fat girl at the dance and has had ~20 years to sort her shit out so sorry love you're sitting this one out.
I read somewhere about calculation of MaxClient. Hard to describe so post your Apache usage and I will try my best to give you well calculated configuration.