Looks like that worked. I set work_mem to 256MB, and it looks like my
standard sql came back in just a couple of seconds.
Thanks!
-----Original Message-----
From: pgsql-performance-owner@postgresql.org
[mailto:pgsql-performance-owner@postgresql.org] On Behalf Of Tom Lane
Sent: Wednesday, October 01, 2008 07:30
To: David logan
Cc: pgsql-performance@postgresql.org
Subject: Re: [PERFORM] Mystefied at poor performance of a standard query
"David logan" <djlogan2@comcast.net> writes:
> (The question is why this simple select takes me 20 minutes to run...)
What have you got work_mem set to? The hash join is not going to be
real fast if it has to split the join into multiple batches, so you
want work_mem large enough to hold the whole inner relation. That would
be at least 20MB in this example, probably quite a bit more after
allowing for per-row overhead in the table.
regards, tom lane
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
