As far as I can remember there have never been any such restrictions. As Trudy says, this is by design.
The lock in the database is created for the Run. But another user can start another Run with another ID, so the test instance is still available for execution.
If you really think this is something that needs to be prevented, you should be able to do it from the workflow. For example, by actually creating a lock on the test instance when the run is starting, and removing that lock in Run_AfterPost.
The lock in the database is created for the Run. But another user can start another Run with another ID, so the test instance is still available for execution.
If you really think this is something that needs to be prevented, you should be able to do it from the workflow. For example, by actually creating a lock on the test instance when the run is starting, and removing that lock in Run_AfterPost.