Don't know if you rebooted in between the time it wouldn't work and the time that it did just now or what. Unless you are sharing it, then I have no other ideas. If you are sharing it, then it's a possibility that someone closed it in the meantime. If not sharing, then I don't know, but would suggest that you implement some sort of periodic (daily?) backup of the back end after splitting it. You are running the risk of losing ALL of your data in an unsplit db, especially one that raised such an issue at the outset.