Re: Default cost variables in postgresql.conf - Mailing list pgsql-performance

From Andrew Sullivan
Subject Re: Default cost variables in postgresql.conf
Date
Msg-id 20021021101621.A19970@mail.libertyrms.com
Whole thread Raw
In response to Default cost variables in postgresql.conf  (Ludwig Lim <lud_nowhere_man@yahoo.com>)
List pgsql-performance
On Mon, Oct 21, 2002 at 12:04:58AM -0700, Ludwig Lim wrote:
> Hi:
>
>   Are the "cost" variables (e.g.
> random_page_cost,cpu_tuple_cost,cpu_index_tuple_cost)
> in postgresql.conf optimal for a particular set of
> platform / hardware requirements? (i.e. the configs

Not exactly.  They're best guesses.  If you check the admin guide,
you'll see that there's a note about these which says that there is
not a well-defined method for calculating these things, so you are
encouraged to experiment and share your findings.  They _are_ known
to be conservative defaults, like everything else in the system.

A

--
----
Andrew Sullivan                         204-4141 Yonge Street
Liberty RMS                           Toronto, Ontario Canada
<andrew@libertyrms.info>                              M2P 2A8
                                         +1 416 646 3304 x110


pgsql-performance by date:

Previous
From: Andrew Sullivan
Date:
Subject: Re: Self-generating statistics?
Next
From: Ludwig Lim
Date:
Subject: Selective usage of index in planner/optimizer (Too conservative?)