X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [64.117.224.130]) by svr1.postgresql.org (Postfix) with ESMTP id C5F82D1B8D9 for ; Sat, 30 Aug 2003 15:37:12 +0000 (GMT) Received: from svr1.postgresql.org ([64.117.224.193]) by localhost (neptune.hub.org [64.117.224.130]) (amavisd-new, port 10024) with ESMTP id 41294-04 for ; Sat, 30 Aug 2003 12:36:42 -0300 (ADT) Received: from wight.ymogen.net (unknown [217.27.240.153]) by svr1.postgresql.org (Postfix) with SMTP id F2BF2D1B8B4 for ; Sat, 30 Aug 2003 12:36:40 -0300 (ADT) Received: (qmail 18512 invoked from network); 30 Aug 2003 15:36:11 -0000 Received: from host213-123-217-162.in-addr.btopenworld.com (HELO solent) (213.123.217.162) by wight.ymogen.net with SMTP; 30 Aug 2003 15:36:11 -0000 From: "Matt Clark" To: "Ron Johnson" , "PgSQL Performance ML" Subject: Re: Hardware recommendations to scale to silly load Date: Sat, 30 Aug 2003 16:36:20 +0100 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <1062220877.669.406.camel@haggis> Importance: Normal X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200308/535 X-Sequence-Number: 3351 > SELECT > IF > BEGIN > INSERT > or > UPDATE > COMMIT; > > He says his current h/w peaks at 1/10th that rate. > > My question is: is that current peak rate ("300 inserts/updates > *or* 2500 selects") based upon 1 connection, or many connections? > With 4 CPUs, and a 4 disk RAID10, I wouldn't be surprised if 4 con- > current connections gives the optimum speed. Well it's more like each user interaction looks like: SELECT SELECT SELECT SELECT SELECT SELECT INSERT SELECT SELECT SELECT SELECT INSERT SELECT SELECT SELECT UPDATE SELECT SELECT UPDATE SELECT And concurrency is very high, because it's a web app, and each httpd has one connection to PG, and there can be hundreds of active httpd processes. Some kind of connection pooling scheme might be in order when there are that many active clients. Any views?