Good news to report and an embarrassing confession. First the confession, I was missing a critical piece of YAF logic in that each board creates its own user. So while my report was correct (user daveburke in Board 1 had a rankID of 10, which was associated with Board 2--causing the runtime on yaf_user_getalbumsdata null data), I found no bad RankID assignments from any Board 2 processes.
SO FORGET THE BAD RANKID ISSUE, in other words. How certain of my key user RankIDs got screwed up, I don't know, but I'm not worrying about it. Mek, we don't have to budget for Excedrins anymore.
My apologies for the noise. I do have some thoughts on the yaf_user_getalbumsdata logic which I'll post on a separate thread.