Create DVs and VMs sequentially instead of using Python multiprocessing. 

This will still allow us to have concurrent imports, but will reduce the automation issues associated with multithreading.