Comment 4 for bug 1494359

well... or maybe if there's some crazy info in the on-disk hashes that doesn't match up with the current schema in swift.conf for the storage policy?

Have you changed EC settings at all? Maybe disks didn't get fully wiped out between the change?