Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> writes:
> Tom> One thing that seems odd is that the xids are kinda small. Did
> Tom> the system just recently have a wraparound event?
> The system was recently dump/restored from a different box. The
> failing rows are all new inserts since the restore.
So the table has been insert-only so far? I was just thinking that HOT
bugs seemed like a probable explanation, but that idea goes out the
window if there have been no UPDATEs.
regards, tom lane