Thanks for your suggestion. Yes, modifying the WHERE clause of the UPDATE to exclude b=0 will prevent the overflow and the error, and was my next step.

I guess I'm just very curious as to why the app would suddenly choke on an overflow situation it has handled smoothly for nearly three years. I'd also like to make sure nothing else is amiss, as I was unable to reproduce the error in my testing environment, only on the server after waiting through an hour and a half of processing.