On 17 Nov, Steve Young went to index 2 Timothy in 30384 only to find that previously indexed pages were off by 3-4 pages. That is, the mapping between the manuscript images and the indexing was off.
On 05 Dec, he found a similar problem with 31149--now the indexing and images were off by 4-5 pages.
I don't know what kind of logs are available for debugging, but this seems like a problem we don't want to have!