Rational Performance Tester

[Kristof Stevens]

Hi all,

Rational Performance Tester is a performance testing tool for validating the scalability and reliability of complex e-business applications before deployment. It is one of a set of IBM Rational Software products for automated software testing.

We prepared many scripts on Rational Performance Tester to see if OTM can handle an extremely high load with 500 users running bulkplans, tender, split shipments, integration, unassign orders, running reports etc...

Does people here have experience with Rational because we are facing some issues with our recorded scripts in OTM...It seems OTM is not always reacting on the scripts...
For example we recorded a script to widthdraw a tender but when we try to run a performance test we see the script is logging on, but when he tries to get into one shipment to widthdraw, it doesnt do anything..;

Has this something to do with the number generated by OTM in the URL?

otmlb.group.upm-kymmene.com/GC3/glog.webserver.util.QueryResponseServlet/1189164787232?state=tender&finder_set_gid=UPM.UPM_ BUY_SHIPMENT

Thx,

Kristof Stevens

[chrisplough]

Re: Rational Performance Tester

Kristof,

I've seen a variety of tools used to test OTM, including Rational, Mercury and JMeter. JMeter was often the most successful, but also the hardest to configure.

This usually occurs because the Web Tool isn't handling the OTM sessions correctly. In addition to handling the session cookie that OTM's Tomcat instance hands out, it needs to handle the unique identifier that you're seeing generated in the URL. This is used by OTM to facilitate more granular session control and may (though I could be wrong) be used by the newer JavaScript and JSP-based UI enhancements.

I don't know the facilities for managing this in Rational - but you need to ensure this number is captured dynamically and included within the associated URLs for that particular web session -- or at the very least that Rational isn't ignoring them and replacing them with the numbers generated at the time it was recorded.

Thanks,
Chris