X-Original-To: pgsql-performance@postgresql.org Received: from spampd.localdomain (postgresql.org [64.49.215.8]) by postgresql.org (Postfix) with ESMTP id B6C07475458 for ; Wed, 9 Apr 2003 19:47:46 -0400 (EDT) Received: from flake.decibel.org (flake.decibel.org [66.93.216.19]) by postgresql.org (Postfix) with SMTP id 83F9F474E42 for ; Wed, 9 Apr 2003 19:47:41 -0400 (EDT) Received: (qmail 57261 invoked by uid 1001); 9 Apr 2003 23:47:44 -0000 Date: Wed, 9 Apr 2003 18:47:44 -0500 From: "Jim C. Nasby" To: "scott.marlowe" Cc: Matthew Nuzum , 'Josh Berkus' , 'Pgsql-Performance' Subject: Re: choosing the right platform Message-ID: <20030409184744.W31861@flake.decibel.org> Reply-To: jim@nasby.net References: <00a001c2febb$631141d0$6900a8c0@mattspc> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: ; from scott.marlowe@ihs.com on Wed, Apr 09, 2003 at 11:55:56AM -0600 X-Operating-System: FreeBSD 4.7-RELEASE-p3 i386 X-Distributed: Join the Effort! http://www.distributed.net X-Spam-Status: No, hits=-39.4 required=5.0 tests=BAYES_01,EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT, REFERENCES,REPLY_WITH_QUOTES,SIGNATURE_SHORT_SPARSE, USER_AGENT_MUTT autolearn=ham version=2.50 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) X-Archive-Number: 200304/95 X-Sequence-Number: 1601 On Wed, Apr 09, 2003 at 11:55:56AM -0600, scott.marlowe wrote: > A couple more suggestions. One is to never allocate more than 50% of your > memory to a database's shared buffers, i.e. let the OS buffer the disks en > masse, while the database should have a smaller buffer for the most recent > accesses. This is because kernel caching is usually faster and more > efficient than the database doing it, and this becomes more an issue with > large chunks of memory, which both Linux and BSD are quite good at > caching, and postgresql, not so good. That seems odd... shouldn't pgsql be able to cache information better since it would be cached in whatever format is best for it, rather than the raw page format (or maybe that is the best format). There's also the issue of having to go through more layers of software if you're relying on the OS caching. All the tuning info I've seen for every other database I've worked with specifically recommends giving the database as much memory as you possibly can, the theory being that it will do a much better job of caching than the OS will. -- Jim C. Nasby (aka Decibel!) jim@nasby.net Member: Triangle Fraternity, Sports Car Club of America Give your computer some brain candy! www.distributed.net Team #1828 Windows: "Where do you want to go today?" Linux: "Where do you want to go tomorrow?" FreeBSD: "Are you guys coming, or what?"