Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I am trying to load test my apache server with jmeter (10 hosts, 10 threads) with various mpm config. I tried various settings from maxclient 150 to 3000, but there is negligible difference in the average response time.

The cpu usage reaches around 10% and the ram usage is much less around 100 mb or so in total. The box on which apache server runs is a dual core,16 GB box.

Any idea why the mpm config changes are not having any affect on the response time ?

edit : It is a standard apache which server which serves static content like images, js files, css files etc

Regarding 10 threads : I am really really sorry, I didn't explained the threads part fully. I did tried different values of thread from 6 to 30. for both the configs the degrade was similiar. For 150 maxclient settings it changed from ~45ms to ~430ms, for the 3000 maxclient settings it changed from ~45ms to ~420ms. I am not sure where I am making the mistake.

share|improve this question
    
what type of applications are you testing? –  ALex_hha Aug 30 '13 at 17:00
2  
10x10=100 < 150 :) –  GioMac Aug 30 '13 at 17:13

1 Answer 1

As GioMac pointed out, your math is flawed.

Apache MaxClients limits only have an effect if you hit the limit.
10 hosts with 10 threads each is 100 clients.
If MaxClients is set to 150 you will never turn away a client with those settings.

Either add more jmeter clients (another 10 hosts, or increase to 20 threads/host), or reduce MaxClients to a value below 100 so that you hit the limit. There should then be an observable difference in performance.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.