Re: Moving postgresql.conf tunables into 2003... - Mailing list pgsql-performance

From scott.marlowe
Subject Re: Moving postgresql.conf tunables into 2003...
Date
Msg-id Pine.LNX.4.33.0307091336020.8918-100000@css120.ihs.com
Whole thread Raw
In response to Re: Moving postgresql.conf tunables into 2003...  (Kaarel <kaarel@future.ee>)
Responses Re: Moving postgresql.conf tunables into 2003...  (Martin Foster <martin@ethereal-realms.org>)
List pgsql-performance
On Wed, 9 Jul 2003, Kaarel wrote:

> >>Are you willing to say that the PostgreSQL database system should only be
> >>used by DBAs?  I believe that Postgres is such a good and useful tool that
> >>anyone should be able to start using it with little or no barrier to entry.
> >
> >
> > I quite agree.  But there is a difference between saying "you should get
> > decent performance with no effort" and "you should get optimal
> > performance with no effort".  I think we can get to the first with
> > relatively little trouble (like boosting the default shared_buffers to
> > 1000), but the second is an impractical goal.
>
>
> Just wanted to repeat some of the thoughts already been expressed.
>
> There are no reasons why shouldn't PostgreSQL be reasonably well
> configured for a particular platform out of the box. Not for maximum
> performance but for good enough performance. The many complaints by new
> users about PostgreSQL being suprisingly slow and the all the so
> standard answers (vacuum, pump up memory settings) imho prove that the
> default installatio can be improved. Already mentioned in the mail
> lists: using multiple standard conf files, quering system info and
> dynamically generating all or some parts of the conf file, automating
> the vacuum process...

It would be nice to have a program that could run on any OS postgresql
runs on and could report on the current limits of the kernel, and make
recommendations for changes the admin might want to make.

One could probably make a good stab at effective cache size during
install.  Anything reasonably close would probably help.

Report what % of said resources could be consumed by postgresql under
various circumstances...


pgsql-performance by date:

Previous
From: Kaarel
Date:
Subject: Re: Moving postgresql.conf tunables into 2003...
Next
From: Martin Foster
Date:
Subject: Re: Moving postgresql.conf tunables into 2003...