You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@httpd.apache.org by "Rose, John B" <jb...@utk.edu> on 2014/12/15 19:08:40 UTC

[users@httpd] httpd usage differences using "ab" and "httperf"

Red Hat 7, Apache 2.4.6, 48 GB RAM, 16 cores

We have been doing some testing us both "ab" and "httperf" sending what we think are the same parameters. The # of processes and RAM usage look pretty much the same, but cpu usage is vastly different. Using "ab" seems to use a total average of 71% of cpu resources. Using "httperf" looks to use 4.7% of total cpu resources. Also, the cpu% for each httpd process is pretty different.

Using "top" to analyze both ...

1. Command: ab -t 500 -n 300000 -c 300 ourserver.com

Selected "top" output

%Cpu(s): 71.0 us

 PID       USER      PR  NI    VIRT    RES        SHR S  %CPU %MEM     TIME+ COMMAND
10407 apache    20   0  415332  10064   3752 R  42.8  0.0   0:02.16 httpd

2.  Command: httperf --hog -server ourserver.com --num-conn 300000 --ra 300

Selected "top" output

%Cpu(s):  4.7 us,

 PID       USER      PR  NI    VIRT    RES        SHR S  %CPU %MEM     TIME+ COMMAND
10625 apache    20   0  415308  10060   3756 S  12.0  0.0   0:04.39 httpd


My questions are why are they so different, and which one most accurately reflects real traffic?

Thanks

Re: [users@httpd] httpd usage differences using "ab" and "httperf"

Posted by Eric Covener <co...@gmail.com>.
On Mon, Dec 15, 2014 at 1:08 PM, Rose, John B <jb...@utk.edu> wrote:
> My questions are why are they so different, and which one most accurately
> reflects real traffic?


Maybe httperf defaults to using keepalive?  ab doesn't (-k).


-- 
Eric Covener
covener@gmail.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org