X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C99FBD1B8CB for ; Wed, 3 Mar 2004 02:06:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 28250-10 for ; Tue, 2 Mar 2004 22:06:29 -0400 (AST) Received: from jefftrout.com (h00a0cc4084e5.ne.client2.attbi.com [24.128.241.68]) by svr1.postgresql.org (Postfix) with SMTP id CBF42D1C9C3 for ; Tue, 2 Mar 2004 22:06:24 -0400 (AST) Received: (qmail 56958 invoked from network); 3 Mar 2004 02:06:35 -0000 Received: from unknown (HELO ?192.168.0.8?) (10.10.10.177) by 10.10.10.10 with SMTP; 3 Mar 2004 02:06:35 -0000 In-Reply-To: References: Mime-Version: 1.0 (Apple Message framework v609) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <7AACCD10-6CB7-11D8-AD74-000393D1F76E@torgo.978.org> Content-Transfer-Encoding: 7bit Cc: , Anjan Dave , , William Yu , Chris Ruprecht From: Jeff Subject: Re: Scaling further up Date: Tue, 2 Mar 2004 21:07:10 -0500 To: "scott.marlowe" X-Mailer: Apple Mail (2.609) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200403/36 X-Sequence-Number: 5892 On Mar 2, 2004, at 5:36 PM, scott.marlowe wrote: > Some folks on the list have experience with Postgresql on Solaris, and > they generally say they use Solaris not for performance reasons, but > for > reliability reasons. I.e. the bigger Sun hardware is fault tolerant. > Solaris isn't nearly as bad for PG as it used to be. But as you say - the #1 reason to use sun is reliability. (In my case, it was because we had a giant sun laying around :) I'm trying to remember exactly what happens.. but I know on sun if it had a severe memory error it kills off processes with data on that dimm (Since it has no idea if it is bad or not. Thanks to ECC this is very rare, but it can happen.). I want to say if a CPU dies any processes running on it at that moment are also killed. but the more I think about that th emore I don't think that is the case. As for x86.. if ram or a cpu goes bad you're SOL. Although opterons are sexy you need to remember they really are brand new cpus - I'm sure AMD has done tons of testing but sun ultrasparc's have been in every situation conceivable in production. If you are going to really have thousands of users you probably want to bet the farm on something proven. lots and lots of spindles lots and lots of ram You may also want to look into a replication solution as a hot backup. -- Jeff Trout http://www.jefftrout.com/ http://www.stuarthamm.net/