Re: SELECT with MANY tables

From: Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com>
To: Javier Carlos <fjcarlos(at)correo(dot)insp(dot)mx>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: SELECT with MANY tables
Date: 2003-11-24 22:44:39
Message-ID: 20031124144237.V91719@megazone.bigpanda.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs


On Mon, 24 Nov 2003, Javier Carlos wrote:

> Quoting Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com>:
>
> >
> > On Mon, 24 Nov 2003, Javier Carlos wrote:
> >
> > >
> > > Please enter a FULL description of your problem:
> > > -------------------------------------------------
> > > When I make a SELECT with many tables (more than 12), postgresql eats
> > all my
> > > %CPU and I've waited more than 1 hour and stays the same. The weird thing
> > is
> > > that with 10 tables the same select with the same joins only takes about 5
> > > seconds. First I thought that It was a problem related with one specific
> > table,
> > > but I've changed in the SELECT the tables and while the number of tables
> > remains
> > > less than 12 all is ok.
> >
> > As a question, does the query work if you SET geqo=off; or SET
> > geqo_threshold=<some higher number, say 20>?
> >
>
> I tried with setting geqo=off and It didn't work. Then I set
> geqo_threshold=20, and neither. In both the result was:
>
> server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> The connection to the server was lost. Attempting reset: Failed.

What does your postmaster log say? There should be something in there.
I also forgot to ask for EXPLAIN output under normal conditions and with
geqo off (although it's possible that it will fail as well for the latter)

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Marcin Kapski 2003-11-24 23:44:01 Performance postgresql
Previous Message Javier Carlos 2003-11-24 21:43:20 Re: SELECT with MANY tables