Romuald Brunet <romuald(dot)brunet(at)gmail(dot)com> writes:
> After some investigation, we found out that the PostgreSQL server
> (8.1) changed the execution plan (I'm assuming because the number of
> rows increased).
The problem seems to be that the estimate for the number of rows fetched
from dc_post changed drastically --- it was in the right ballpark, and
now it's off by a factor of 30. Why is that? Maybe you had had a
special statistics target setting, and it got dropped?
> The statistics are at the default value everywhere (10)
Almost certainly not enough.
regards, tom lane