Here's what I've done:
1. Recorded a script that: enters an SSN into a CGI field and opens that person, browses through 2 different pages inside that person's record, adds a note to the person's record, then logs out.
2. Found and parameterized the entry of that SSN with a datapool with thousands of SSNs.
Here's my problem:
When I run the test with xx number of VUs, I get xx number of notes added to that first person's record (the one I recorded it with).
Now, I know that every application and implementation is different, but it's clear that the Session ID and SSN are not the only things that need to be different between the VUs. Where I'm at right now is recording two identical scripts with different SSNs and I'm diff'ing the C++ to find all the datapoints that are different. Unfortunately, the CGI field that I enter the SSN into does not have a unique name. It looks like there's probably another unique ID that is returned from the DB that I will have to extract and re-enter into successive page requests. ...But I don't know what the devs are using or sending back, so I'm at the diff'ing stage here.
Does anybody have any better suggestions than recording two identical scripts and diff'ing them every single time?