Received: from localhost (unknown [200.46.204.182]) by postgresql.org (Postfix) with ESMTP id B96D82E3635 for ; Fri, 30 Nov 2007 01:50:06 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.182]) (amavisd-maia, port 10024) with ESMTP id 16390-02-4 for ; Fri, 30 Nov 2007 01:50:01 -0400 (AST) Received: from davinci.ethosmedia.com (server227.ethosmedia.com [209.128.84.227]) by postgresql.org (Postfix) with ESMTP id DA30F2E3629 for ; Fri, 30 Nov 2007 01:50:01 -0400 (AST) X-EthosMedia-Virus-Scanned: no infections found Received: from [63.195.55.98] (account josh@agliodbs.com HELO [192.168.2.3]) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.1.8) with ESMTP id 14125295; Thu, 29 Nov 2007 21:54:39 -0800 From: Josh Berkus Organization: PostgreSQL @ Sun To: pgsql-performance@postgresql.org Subject: Re: Configuring a Large RAM PostgreSQL Server Date: Thu, 29 Nov 2007 21:50:14 -0800 User-Agent: KMail/1.8.2 Cc: Alex Hochberger References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200711292150.14545.josh@agliodbs.com> X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200711/390 X-Sequence-Number: 27805 Alex, > The new machine will have 48 GB of RAM, so figuring out starting =A0 > points for the Shared Buffers and Work_mem/Maintenance_work_mem is =A0 > going to be a crap shoot, since the defaults still seem to be based =A0 > upon 256MB of RAM or less. Why a crap shoot? Set shared_buffers to 12GB. Set work_mem to 20GB / # of concurrent active= =20 connections (check your logs). Set Maint_mem to 2GB (I don't think we can= =20 actually use more). Then tune from there. Also, use 8.2 or later, and you'd better compile 64-bit. =2D-=20 Josh Berkus PostgreSQL @ Sun San Francisco