>When i launch several HTTP requests on a simple page on a 100mbps LAN
>(for example 1 user with a loop of requests separated by 1000ms) i've
>not coherent response time:
If you are always requesting the _same_ resource, it will get 'pinned' in
memory, so that you will always be served from cache. Of course, this
is not a realistic test. There should be _some_ variation in results.
>A classic response time for my request should be around 60ms but it
>takes sometime 40ms and sometime more than 200ms; i don't understand
>how it could be that; that gives a very strange graph with big variations
>in time...
>
>does somebody have such performance problem?
This is easy to do ... overload your server (have you looked at your server
logs ?). For instance, if your server is set up for 10 connections, hit it
with 20 simultaneous requests.
If you can cause your connections (Apache processes, for example) to
crash, this will make for even more widely distributed response times.
Some will be almost instantaneous, others will time-out. Apache is
almost infinitely configurable, so problems such as these can sometimes
be fixed by a re-configuration.
Regards
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
