>Are you saying that everytime to measure the page loading, I need to generate a load with JMeter, and check the page loading on another >computer? No Im saying in the specific case that you really need accurate response times and the architecture is like yours , you might need to do this. Again there are so many factors that influence a page load time as a browser sees that it is not usually worth doing any exhaustive testing for this. (And page load time is just one part of a family of load testing)
>Servers also have cache function? yes. Plus your OS might do so for file read's >Do you mean server's CPU, or client's CPU? Servers >Again, it was uniformly slow for all files (for images and aspx). In some of your original screens the times for the same static file varied from a few ms to a few seconds. Thats not uniformly slow. If your Server was loaded , then it should be simple to find this out using perfmon >No response codes. No URL. Yes but you get a java stack trace that may have some more information or lets you look into JMeter source and see what it was trying to do. On Mon, Sep 20, 2010 at 9:47 AM, Prostak <[email protected]> wrote: > > >If your browser accessing a page with the same background load gives you > >similar results to your jmeter plus estimates - its a match and your > >calculations are reasonably accurate > > Are you saying that everytime to measure the page loading, I need to > generate a load with JMeter, and check the page loading on another > computer? > > >Static files shouldnt need the server to do any calculation it just needs > to > >stream (usually cached) bytes . > > Servers also have cache function? > > >if the CPU was loaded all requests should be > >uniformly slow. > > Do you mean server's CPU, or client's CPU? > > Again, it was uniformly slow for all files (for images and aspx). All files > increased to the 7s level respectively. > > >no Im referring to a jmeter.log file which gets created in > Jmeter_Home/bin. > >You might also ned to check your webserver logs to see why it closed the > >connection > > I found it. All information got deleted anyways. But even if the > information > was there, it contains even less information. No response codes. No URL. > Something like that: > 2010/09/20 09:36:31 INFO - jmeter.threads.JMeterThread: Thread finished: > HomePage 1-1 > So I don't see how it could help with finding out the reason for "Non HTTP > response message: Connection reset". > > >Whats the load? > I am starting with 10 threads and increasing by 10. > -- > View this message in context: > http://jmeter.512774.n5.nabble.com/Test-plan-for-970-page-requests-every-5-min-tp2826174p2846842.html > Sent from the JMeter - User mailing list archive at Nabble.com. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
