Subject: Re: 
From: Ed Tyrrill <tyrrill_ed@emc.com>
To: Stephen Frost <sfrost@snowman.net>
Cc: pgsql-performance@postgresql.org
In-Reply-To: <20070625215654.GF7531@tamriel.snowman.net>
References: <A23190A408F7094FAF446C1538222F76041E1C28@avaexch01.avamar.com>
	<20070625215654.GF7531@tamriel.snowman.net>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: Mon, 25 Jun 2007 16:39:02 -0700
Message-Id: <1182814742.6477.5.camel@nickel.avamar.com>
Mime-Version: 1.0

On Mon, 2007-06-25 at 17:56 -0400, Stephen Frost wrote:
> Was the 'work_mem' set to the same thing on all these runs?  Also, you
> might try increasing the 'work_mem' under 8.2.4, at least for this query
> (you can set it by just doing: set work_mem = '2GB'; or similar in psql,
> or you can change the default in postgresql.conf).
> 
> The big thing of note, it seems, is that you've got enough memory and
> it's coming out faster when doing a hash-join vs. a sort + merge-join.
> Could likely be because it doesn't think there's enough work memory
> available for the hash, which might change based on the values it gets
> from the statistics on how frequently something shows up, etc.
> 
> 	Enjoy,
> 
> 		Stephen

Yes, work_mem was set to 128MB for all runs.  All settings were the same
except for the change to default_statistics_target.  I'm certainly
memory constrained, but giving 2GB to one one session doesn't allow
other sessions to do anything.  Possibly when we upgrade to 16GB. :-)