I have a set of 200 unit tests which do run ok when executed on a single thread with py.test.
I do have a fixture which is configuring a webserver so I can run the tests against it.
Still, I need to be sure that this is run only once before all the other tests are executed, if this fails the entire process must be stopped. If this succeeds, it can perform the testing on as many threads it wants.
How can I achieve that?