X-Original-To: pgsql-performance@postgresql.org Received: from spampd.localdomain (postgresql.org [64.49.215.8]) by postgresql.org (Postfix) with ESMTP id 0B818474E44 for ; Wed, 9 Apr 2003 14:04:00 -0400 (EDT) Received: from davinci.ethosmedia.com (unknown [209.10.40.251]) by postgresql.org (Postfix) with ESMTP id 655DB474E42 for ; Wed, 9 Apr 2003 14:03:59 -0400 (EDT) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 2953260; Wed, 09 Apr 2003 11:03:53 -0700 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: "Matthew Nuzum" , "'Pgsql-Performance'" Subject: Re: choosing the right platform Date: Wed, 9 Apr 2003 11:03:48 -0700 User-Agent: KMail/1.4.3 References: <00a001c2febb$631141d0$6900a8c0@mattspc> In-Reply-To: <00a001c2febb$631141d0$6900a8c0@mattspc> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200304091103.48151.josh@agliodbs.com> X-Spam-Status: No, hits=-21.9 required=5.0 tests=BAYES_10,IN_REP_TO,REFERENCES,USER_AGENT_KMAIL autolearn=ham version=2.50 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) X-Archive-Number: 200304/93 X-Sequence-Number: 1599 Matthew, > I read through the entire archive at > http://archives.postgresql.org/pgsql-performance/ and didn't see much talk > on the subject. It only goes back 8 months though, so I don't know if th= ere > is another archive that is more comprehensive... Really? There was a long-running Mac OS X vs. Solaris thread that touched = on=20 most major platforms, about 2-3 months ago. > I don't suspect that individual tables will be 2GB, but that the size of = all > tables combined will be. If there is a limitation on the largest chunk of > RAM allocated to a program, will I have problems? No. Since PostgreSQL is a multi-process architecture, not a multi-threaded= ,=20 you only need enough RAM per process to load the current largest query. Plus, in my experience, Disk I/O issues are vastly more important than RAM = in=20 database performance. You're better off spending money on really fast dis= ks=20 in Linux RAID or really good hardware RAID 1+0 .... --=20 -Josh Berkus Aglio Database Solutions San Francisco