If this is common knowledge, pardon me; but I just discovered this fact this morning, so I figured I'd share.

If you're using an app_config file to limit the number of WUs that are crunched simultaneously on a project, and your other projects run out of work, you will NOT receive WUs from your backup project.

As an example, I'm running 1 Atlas WU/computer. I'm also trying to crunch Rioja and IGEM for goals. However, both of those projects ran out of work, leaving my machines dry. I had set up NFS as a backup, but that did not kick in, since I had enough Atlas WUs downloaded to cover all my threads. [NFS has been a great backup on machines where I've not been running Atlas.]