I am facing problem while trying to use unique users for my testing. I have my website which verifies users using the LDAP however, I can not use unique data as we are not authorized to create one. However, one of my friend had done this saying that he had executed the same script 3-4 times and after that server had recognised that and then he had performed the load test for more than 25 users using single username and password.
I had asked him but his answers are not satisfactory can anyone please help me out with this.
Not sure what you are asking about. But I think even you can use just a single username/password combination to do the load testing with multiple users. It all depends upon the application you are testing. If the application is designed to have multiple logins with the same username and password then I think this approach is fine. But, if the AUT doesn't allow multiple logins then your application would anyway be throwing error messages. Another scenario is that the application is desired to be for one user-name, one login but has no functionality built in to identify/stop multiple logins thru one username. Your script (with one username) would still be able to push the desired load, but then you would not be simulating real-life usage of the application under load.
Your second scenario does identifies my problem. however, let me clarify my issue. I have an application which identifies unique users each time for login. Once the user logs in, it gets verified in LDAP. Now the problem which I am facing is that I need unique users for login however, right now I have been provided with only 1 username and password and I have to test somewhere around unique 200VU. Is there anyway I can load test this or is there any work around.
Even if you can log in multiple times with 1 user your test may not be realistic as locking issues can occur due to the fact 1 user account is used / will not occur due to the fact 1 user account is used.
If you really know your applcaition well then you can make a judgement on wether this is a feasible approach.
If you're doing load testing then the output has to be realistic or is there any point doing it at all?