Re: Recommended optimisations slows down PostgreSQL 8.4 - Mailing list pgsql-performance

From Waldo Nell
Subject Re: Recommended optimisations slows down PostgreSQL 8.4
Date
Msg-id FE19EF2B-56BF-4DB1-9459-4DEA8D56B8BF@telkomsa.net
Whole thread Raw
In response to Re: Recommended optimisations slows down PostgreSQL 8.4  ("ktm@rice.edu" <ktm@rice.edu>)
Responses Re: Recommended optimisations slows down PostgreSQL 8.4  (Merlin Moncure <mmoncure@gmail.com>)
List pgsql-performance
On 2011-08-11, at 17:18 , ktm@rice.edu wrote:

> One guess is that you are using the defaults for other costing parameters and they
> do not accurately reflect your system. This means that it will be a crap shoot as
> to whether a plan is faster or slower and what will affect the timing.

Ok, but I thought the way to best optimise PostgreSQL is to start with the parameters having the biggest impact and
workfrom there.  To adjust multiple parameters would not give a clear indication as to the benefit of each, as they may
canceleach other out. 

To test your theory, what other parameters should I be looking at?  Here are some more with their current values:

random_page_cost = 4.0
effective_cache_size = 128MB

Remember this runs on SATA so random seeks are not as fast as say SSD.

pgsql-performance by date:

Previous
From: "ktm@rice.edu"
Date:
Subject: Re: Recommended optimisations slows down PostgreSQL 8.4
Next
From: Greg Smith
Date:
Subject: Re: Recommended optimisations slows down PostgreSQL 8.4