Received: from localhost (mx1.hub.org [200.46.208.251]) by postgresql.org (Postfix) with ESMTP id 50F809FA5DC for ; Wed, 11 Oct 2006 11:33:48 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.208.251]) (amavisd-new, port 10024) with ESMTP id 33376-05 for ; Wed, 11 Oct 2006 11:33:44 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey- Received: from main2.mycybernet.net (main2.mycybernet.net [209.222.63.140]) by postgresql.org (Postfix) with ESMTP id 8F97C9FB251 for ; Wed, 11 Oct 2006 11:33:43 -0300 (ADT) Received: from 227-54-222-209.mycybernet.net ([209.222.54.227] helo=phlogiston.dydns.org) by main2.mycybernet.net with esmtp (Exim 4.62) (envelope-from ) id 1GXfSD-000783-6b for pgsql-admin@postgresql.org; Wed, 11 Oct 2006 10:53:05 -0400 Received: by phlogiston.dydns.org (Postfix, from userid 1000) id 557B73FAF; Wed, 11 Oct 2006 10:33:27 -0400 (EDT) Date: Wed, 11 Oct 2006 10:33:27 -0400 From: Andrew Sullivan To: pgsql-admin@postgresql.org Subject: Re: postgres in HA constellation Message-ID: <20061011143327.GB25181@phlogiston.dyndns.org> References: <20061006063425.49239C01C@ogo.rapideye.de> <20061010151108.GE72517@nasby.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20061010151108.GE72517@nasby.net> User-Agent: Mutt/1.5.9i X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.195 tagged_above=0 required=5 tests=AWL, FORGED_RCVD_HELO X-Spam-Level: X-Archive-Number: 200610/81 X-Sequence-Number: 23384 On Tue, Oct 10, 2006 at 10:11:08AM -0500, Jim C. Nasby wrote: > couldn't setup HA on OpenBSD. The key is just to make sure that you > never bring up two servers on the same data directory. I think this highlights exactly what I'm trying to emphasise: in actual, shared-nothing systems like this, there's no possible guarantee of "never". There are possible guarantees of "very rarely". The problem is, you're already trying to address a teeny portion of the likely events on your machines. So you have to assume that more than one thing might break at the same time, and have a recovery plan for it. I submit that a recovery plan of "restore from pg_dump" is usually not going to be enough if it was worth the cost and hassle of setting up shared disk failover. YMMV, of course. A -- Andrew Sullivan | ajs@crankycanuck.ca The fact that technology doesn't work is no bar to success in the marketplace. --Philip Greenspun