On Wed, 12 Nov 2003, Reece Hart wrote:
> On Wed, 2003-11-12 at 09:04, jake johnson wrote:
>
> > I also posted about the performance increase of 7.4, but I think that
> > much of the difference you're seeing (because it's such a large
> > difference) is probably due to the cleanliness of a newly restored
> > database from backup.
>
>
> I agree that this seems likely, except that the 7.3.4 database is
> vacuumed nightly, and analyzed periodically. And about a week ago I
> reclustered on the index intended to most facilitate this select.
> Furthermore, merely hardcoding the subselect result achieves a
> tremendous improvement (which was the workaround I used). So, I'm pretty
> sure that it's not a vacuum, index use, or cleanliness issue.
Do you vacuum full every so often? If not, and if you've been overflowing
your fsm, then your tables will just grow without shrinking.
Also, index growth could be a problem.
The real test is to dump the database and reload it to give 7.3.4 a fair
shake.