Re: Slow count(*) again...

On Sat, Oct 9, 2010 at 7:44 PM, Mladen Gogala <mladen(dot)gogala(at)vmsinfo(dot)com>wrote:

> The architects of Postgres database would be well advised to operate under
> the assumption that every production database has a competent DBA keeping an
> eye on the database.
>

I'd actually go so far as to say that they have already made this
assumption. The out of the box config needs modification for all but the
most low-volume applications and postgres really benefits from having some
attention paid to performance. Not only does tuning the db provide enormous
gains, but it is often possible to dramatically improve query responsiveness
by simply restructuring a query (assuming an aggregating query over a fairly
large table with a few joins thrown in). My team does not have a competent
DBA (though I've got 15+ years of experience developing on top of various
dbs and certainly don't make overly naive assumptions about how things work)
and the gains that we made, when I finally just sat down and read everything
I could get my hands on about postgres and started reading this list, were
really quite impressive. I intend to take some of the courses offered by
some of the companies that are active on this list when my schedule allows
in order to expand my knowledge even farther, as a DBA is a luxury we cannot
really afford at the moment.