On Tue, 26 Aug 2003, Pedro Alves wrote:
>
> First of all tkx for the answer, Dennis.
>
> The vacuum analyze is run on daily basis, so that cannot be the point.
Have you tried raising the statistics targets on ra_datacolh? Is there a
significant correlation between ra_datacolh and isactive values (which I
think is problematic right now since cross column stats aren't generated)?
Is isactive effectively a boolean (just 0 and 1 values?), if so maybe once
some of the other issue are done, you might want a partial index where
isactive=0 depending on how often you do queries where that's important.
As for settings, lowering random_page cost a bit may help for this case
although to balance it seems problematic so it'll probably just move
the switch over point. effective_cache_size was mentioned already I
believe.