I am planning to test an application. I have two sets of users that perform completely separate transactions and are connected to different parts of the system under test. However they all use the same url. I've set up 2 server definitions with different variables for the user ids. The scripts have the separate server definitions coded in them. However when I run Check User path or a test, from time to time the wrong set of users are picked and the test will fail. (I track the signed on user from the actual responses.) How to fix?
The users are following 2 separate user paths.
They sit in separate populations.
Don't think you've got what's happening.
The authentication method used for the application requires NeoLoad to have the user credentials (id/password) in the server definition.
So there's two server definitions, say A and B. The host definition is the same just the credentials are different.
There are 2 separate user paths, one for each set of users. The user paths work fine for the users they are meant to do. In the user paths, each call has the server definition included.
What happens is intermittently when running a check user path or a test, the scripts pick up an entry from the wrong datapool of users. This causes a complete meltdown.
I need to control the number of users running each user path so cannot just detect which user gets signed on and executing suitable transactions.
This seems to be a bug in NeoLoad or is there another way to address getting two different sets of users to point at the same host URL?.
Your logic sounds fine and as the problem is intermittent, it sound like you implemented it correctly. I can’t explain why it would jump to the wrong dataset but if you have a valid support contract, you could raise it with support. As a plan b, you could have an extra column in the data, with values set_a and set_b and have a check in the test so if userpath A uses a set_b record it can perform some recovery steps?
