jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Deepak Shetty <shet...@gmail.com>
Subject Re: Test plan for (970 page requests every 5 min)
Date Fri, 17 Sep 2010 17:53:39 GMT
not enough to go through  - An easy way to verify is when you have generated
load , access your site using a different machine and see the amount of time
the page takes on clean browser , it should be close to average time

>Non HTTP response message: Connection timed out: connect
Its what it says it is. Your client(Jmeter) tried to connect but timed out.
Indicates server is under load and cant accept all the connections to it.
Again are you testing it against a dev system?



On Fri, Sep 17, 2010 at 10:49 AM, Prostak <strodion@gmail.com> wrote:

>
> >b. From firebug try out the home page when your server is loaded under
> >different loads. Estimate the proportions between the cumulative time and
> >observed times for static files
> >e.g. if you have 60 resources and if you sequentially added up the times
> it
> >takes 4 seconds , but firebug shows that the browser loads all 60
> resources
> >in 1 seconds. Repeat this with a browser and with cache cleared but with
> >your server under a different load. You should be able to come up with a
> >range that given a cumulative time for static resources you can estimate
> the
> >time it would have taken for the browser(sum of default.aspx + time for
> >static + time for render). Then use JMeter to figure out the cumulative
> time
> >for the static resources on your page (simple addition) and use the
> >proportion range you've calculated above to get the browser time. You can
> do
> >this for if its a returning user (where the cache might be hit , but the
> >browser may still make requests and get a 304 or something)
>
> This is what I've done. Since all my files are static, I've created one
> sample with "retrieve" check marked. I've run it with single thread (no
> load). Then I run Firebug (to get base times). Then I generated the load
> (always 5 second ramp-up). Here is my data:
>
> http://tinypic.com?ref=29veydy  http://i53.tinypic.com/29veydy.png
>
> So now I have a proportion in 59%. Then I generate a load 40 threads per 5
> second and I get Jmeter data 38 seconds. I use 59% and get 22 seconds for
> browser time. Is my calculation correct?
>
>
>
> Also, I have a question about unrelated test results. Do you know what
> these
> errors are? When do such errors usually occur?
>
> <httpSample t="21117" lt="0" ts="1284679206421" s="false" lb="Error:
> http://www.********.com/en/********/VLG2.jpg" rc="Non HTTP response code:
> java.net.ConnectException" rm="Non HTTP response message: Connection timed
> out: connect" tn="Home 1-14" dt="text" by="1773"/>
>
> <httpSample t="64541" lt="814" ts="1284679160822" s="false" lb="Error:
> http://********.com/App_Themes/********/Grid.Web20.css" rc="Non HTTP
> response code: java.net.SocketException" rm="Non HTTP response message:
> Connection reset" tn="Home 1-55" dt="text" by="1521"/>
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Test-plan-for-970-page-requests-every-5-min-tp2826174p2844087.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message