X-Original-To: pgsql-admin-postgresql.org@localhost.postgresql.org Received: from localhost (av.hub.org [200.46.204.144]) by postgresql.org (Postfix) with ESMTP id AB92F11F65D7 for ; Wed, 29 Mar 2006 12:54:31 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 58692-09 for ; Wed, 29 Mar 2006 12:54:29 -0400 (AST) X-Greylist: from auto-whitelisted by SQLgrey- Received: from pat.uio.no (pat.uio.no [129.240.10.6]) by postgresql.org (Postfix) with ESMTP id 4CE8C11F65D5 for ; Wed, 29 Mar 2006 12:54:28 -0400 (AST) Received: from mail-mx7.uio.no ([129.240.10.52]) by pat.uio.no with esmtp (Exim 4.43) id 1FOdwA-00068q-3G; Wed, 29 Mar 2006 18:54:26 +0200 Received: from ti200710a080-3004.bb.online.no ([85.164.139.188] helo=[10.1.1.101]) by mail-mx7.uio.no with esmtpsa (SSLv3:RC4-MD5:128) (Exim 4.43) id 1FOdw8-0006Rw-5x; Wed, 29 Mar 2006 18:54:24 +0200 Subject: Re: Bloated pg_shdepend_depender_index From: Rafael Martinez To: andy.shellam@mailnetwork.co.uk Cc: "'Jim C. Nasby'" , pgsql-admin@postgresql.org In-Reply-To: References: Content-Type: text/plain Organization: University of Oslo Date: Wed, 29 Mar 2006 18:56:56 +0200 Message-Id: <1143651416.18329.66.camel@linux.site> Mime-Version: 1.0 X-Mailer: Evolution 2.4.0 Content-Transfer-Encoding: 7bit X-UiO-Spam-info: not spam, SpamAssassin (score=-5, required 12, autolearn=disabled, UIO_MAIL_IS_INTERNAL -5.00) X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, score=0.12 required=5 tests=[AWL=0.120] X-Spam-Score: 0.12 X-Spam-Level: X-Archive-Number: 200603/343 X-Sequence-Number: 21100 On Wed, 2006-03-29 at 13:16 +0100, Andy Shellam wrote: > This is true, but you could run the new version on the same server/different > port to the current version, and do a simultaneous dump/restore without > having to use extra disk space, or taking your production database offline - > then you can schedule a time to kill the old one, re-assign the port on the > new one and bring it up - would take about 30 seconds, and this could be > done at a time when traffic was at its lowest. Running the new version on another port in the samme server does not help much. You have to initialize the new version to use new data partitions (so you need the double of the disk space you have in production. Another minus is that this works only in a 'read only' database, all the updates from you start the dump/restore process until you kill/re-assign/start again won't be in the new version (and this can be a lot of data in a busy/large system) -- Rafael Martinez, Center for Information Technology Services University of Oslo, Norway PGP Public Key: http://folk.uio.no/rafael/