>Well when I use apache bench, I can easily get over 5K requests per second 10 Threads making 100 requests each within 1 second != 100 Threads making 10 requests each within 1 second. (even if the server still sees 1k requests per second)
What you have to also consider is fidelity - how closely does your test resemble unique machines making the same number of requests. >But how can I tweak the xml file before it performs the post? some of the pre/post processors can do it , ideally you want to create everything you want before you start the test , but if thats not possible - see something like http://theworkaholic.blogspot.com/2011/02/dynamic-values-within-files.html regards deepak On Tue, May 15, 2012 at 4:41 PM, S Ahmed <[email protected]> wrote: > Well when I use apache bench, I can easily get over 5K requests per second > with about 20-30 threads so I guess this shouldn't be an issue. > > But how can I tweak the xml file before it performs the post? > > > > On Tue, May 15, 2012 at 4:35 PM, Deepak Shetty <[email protected]> wrote: > > > unlikely to get this with some degree of accuracy with one machine - but > > possible with multiple machines running jmeter . > > My rule of thumb is no more than 100 threads per machine (and usually > much > > lesser), so if each request takes on an average 0.5 second => 5 machines > , > > but you always have to check if your jmeter client machine is able to > > generate that lo1d without significant reduction in throughput. > > > > On Tue, May 15, 2012 at 8:00 AM, S Ahmed <[email protected]> wrote: > > > > > I have an xml file that I want to http post to a url. > > > Before posting it, I want to modify an element in the xml (its a > > > credentials xml node value). > > > > > > I also need to send 1K connections per second, and do this via the > > command > > > line. > > > > > > Is this possible with jmeter? Is 1K per second too high? > > > > > >
