Received: from localhost (unknown [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id C6B0B9FC191 for ; Tue, 23 Oct 2007 15:26:22 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 29318-02-8 for ; Tue, 23 Oct 2007 15:26:05 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from lists.commandprompt.com (host-254.commandprompt.net [207.173.203.254]) by postgresql.org (Postfix) with ESMTP id 8EED39FC6A7 for ; Tue, 23 Oct 2007 15:23:20 -0300 (ADT) Received: from perhan.alvh.no-ip.org (201-221-215-196.bk12-dsl.surnet.cl [201.221.215.196]) (authenticated bits=0) by lists.commandprompt.com (8.13.7/8.13.6) with ESMTP id l9NINKwo000747; Tue, 23 Oct 2007 11:23:20 -0700 Received: by perhan.alvh.no-ip.org (Postfix, from userid 1000) id 92C7147C79; Tue, 23 Oct 2007 15:23:18 -0300 (CLST) Date: Tue, 23 Oct 2007 15:23:18 -0300 From: Alvaro Herrera To: Ron St-Pierre Cc: Gregory Stark , pgsql-performance@postgresql.org Subject: Re: 12 hour table vacuums Message-ID: <20071023182318.GN18013@alvh.no-ip.org> References: <471E18ED.3090704@shaw.ca> <87y7dtokfu.fsf@oxford.xeocode.com> <471E26E4.7040804@shaw.ca> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <471E26E4.7040804@shaw.ca> User-Agent: Mutt/1.5.16 (2007-06-11) X-Virus-Scanned: ClamAV 0.91.1/4572/Tue Oct 23 00:50:50 2007 on projects.commandprompt.com X-Virus-Status: Clean X-Greylist: Sender succeded SMTP AUTH authentication, not delayed by milter-greylist-1.6 (lists.commandprompt.com [192.168.2.159]); Tue, 23 Oct 2007 11:23:21 -0700 (PDT) X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200710/264 X-Sequence-Number: 27286 Ron St-Pierre wrote: > Gregory Stark wrote: >>> We were running autovacuum but it interfered with the updates to we >>> shut it off. >> >> Was it just the I/O bandwidth? I'm surprised as your >> vacuum_cost_delay is quite high. Manual vacuum doesn't do anything >> differently from autovacuum, neither should interfere directly with >> updates except by taking away I/O bandwidth. >> > I don't know what the problem was. I tried to exclude certain tables > from autovacuuming, but it autovacuumed anyway. Probably because of Xid wraparound issues. Now that you're vacuuming weekly it shouldn't be a problem. (It's also much less of a problem in 8.2). -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.