When we make essentially a skim of a skim of a skim (a chain of skims), it doesn't always find back the original, non-empty skim to run over. We think that what's happening is that if the secondary files are also empty skims, then there's a problem. We think we need to fix the code here:
https://github.com/OSU-CMS/OSUT3Analysis/blob/master/Configuration/python/processingUtilities.py#L261-L279
in order to make sure the secondary files are indeed non-empty skims (and if not, then go back one step further in the chain)