Description
From parent, saw unreferenced rows in a run yesterday against tip of branch-2. Saw similar in a run from a week or so ago.
Enabling DEBUG and rerunning to see if I can get to root of dataloss. See https://docs.google.com/document/d/14Tvu5yWYNBDFkh8xCqLkU9tlyNWhJv3GjDGOkqZU1eE/edit# for old debugging trickery.