Re: Volunteer to build a configuration tool

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Greg Smith <gsmith@gregsmith.com>
Cc: pgsql-performance@postgresql.org
Date: 2007-06-20T05:43:50Z
Lists: pgsql-performance
Greg Smith <gsmith@gregsmith.com> writes:
> On Wed, 20 Jun 2007, Mike Benoit wrote:
>> I don't think something like this would be very difficult at all to
>> write

> Here I just smile and say that proves you've never tried to write one :) 

I'm with Greg on this.  It's not that easy to optimize in a
multi-parameter space even if all conditions are favorable, and they
never are.

I think what would be much more useful in the long run is some
serious study of the parameters themselves.  For instance,
random_page_cost is a self-admitted oversimplification of reality.
We know that good settings for it depend critically on how large
your DB is relative to your RAM; which means there are at least two
parameters there, but no one's done any serious thinking about how
to disentangle 'em.

			regards, tom lane