If I understand correctly, it seems to be you are doing a jmeter recording in which case it should be enough to set port of the HTTP Proxy Server to 9090 and then start it. That will not work properly if you are running fiddler or other monitoring tools that also act as proxies, so make sure that jmeter alone is the proxy.
However, if it is the case that you are doing the recording from scratch, I would strongly advice you to look at the Scalability Tools that will assist in creating load tests for you. Here is the link:
Thanks for the prompt reply. I only running jmeter and QV server in the same machine. I find that when I use 127.0.0.1 or localhost, Jmeter cannot record the action . I have read some thread and some of them mentioned using hostname for access AccessPoint.
When I use hostname, e.g http://hostname/qlikview, Jmeter can record action but the AccessPoint is not stable. It have error when open reports and there are error in AccessPoint dashboard icons.
I also read the link you provided but I want to test the performance on specific report instead of hardware brenchmark. So, is there a easy way to test on performance on report?
Thanks a lot!
That link should take you to the scalability tools that you can use to test the document you specify, and create your own scenario on how the virtual users should perform selections. There is also a hardware benchmarking package but that is something separate and as you said not what you are after.
This is of course valid if you want to test qlikview versions 10 or 11. If it is Qlikview 9, then you need to do a jmeter recording and modify it afterwards. The tutorial for QV9 scripting is here: Tutorial on load tesing Qlikview 9 with JMeter