X-Original-To: pgsql-performance-postgresql.org@postgresql.org Received: from localhost (wm.hub.org [200.46.204.128]) by postgresql.org (Postfix) with ESMTP id B45DD9FB365 for ; Tue, 5 Sep 2006 15:13:11 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.128]) (amavisd-new, port 10024) with ESMTP id 76288-04 for ; Tue, 5 Sep 2006 18:13:02 +0000 (UTC) X-Greylist: from auto-whitelisted by SQLgrey- Received: from net2.micro-automation.com (net2.micro-automation.com [64.7.141.29]) by postgresql.org (Postfix) with SMTP id 62B799FB361 for ; Tue, 5 Sep 2006 15:12:55 -0300 (ADT) Received: (qmail 27035 invoked from network); 5 Sep 2006 18:12:54 -0000 Received: from dcdsl.ebox.com (HELO ?192.168.1.2?) (davec@64.7.143.116) by net2.micro-automation.com with SMTP; 5 Sep 2006 18:12:54 -0000 In-Reply-To: References: <87veo7a5dn.fsf@meuh.mnc.lan> <44F8CF32.9040606@commandprompt.com> Mime-Version: 1.0 (Apple Message framework v752.2) Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: Cc: "Joshua D. Drake" , "Guillaume Cottenceau" , pgsql-performance@postgresql.org Content-Transfer-Encoding: 7bit From: Dave Cramer Subject: Re: increasing shared buffers: how much should be removed Date: Tue, 5 Sep 2006 14:12:46 -0400 To: Merlin Moncure X-Mailer: Apple Mail (2.752.2) X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200609/30 X-Sequence-Number: 20716 On 5-Sep-06, at 9:31 AM, Merlin Moncure wrote: > On 9/1/06, Joshua D. Drake wrote: >> >> >> >> I think 'shared buffers' is one of the most overrated settings >> from a >> >> performance standpoint. however you must ensure there is >> enough for >> >> things the server does besides caching. It used to be a bigger >> deal >> >> than it is in modern versionf of postgresql modern operating >> systems. >> >> Previous to 8.1 I would agree with you, but as of 8.1 it is >> probably the >> most underrated. > > really? what are the relative advantages of raising shared buffers? I > was thinking maybe there might be less context switches in high load > environments...I'm really curious what you have to say here. Have you tried it ? The results are quite dramatic. So if shared buffers aren't the first tool you reach for, what is ? > > merlin > > ---------------------------(end of > broadcast)--------------------------- > TIP 2: Don't 'kill -9' the postmaster >