Received: from localhost (maia-2.hub.org [200.46.204.187]) by postgresql.org (Postfix) with ESMTP id 872029FB8B1 for ; Wed, 16 May 2007 13:35:33 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.187]) (amavisd-maia, port 10024) with ESMTP id 36436-03 for ; Wed, 16 May 2007 13:35:27 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.4 Received: from lists.commandprompt.com (host-254.commandprompt.net [207.173.203.254]) by postgresql.org (Postfix) with ESMTP id 01C749FB87E for ; Wed, 16 May 2007 13:35:27 -0300 (ADT) Received: from perhan.alvh.no-ip.org (201-221-217-243.bk12-dsl.surnet.cl [201.221.217.243]) (authenticated bits=0) by lists.commandprompt.com (8.13.7/8.13.6) with ESMTP id l4GGZNAT020405; Wed, 16 May 2007 09:35:23 -0700 Received: by perhan.alvh.no-ip.org (Postfix, from userid 1000) id 0DCB132A13; Wed, 16 May 2007 12:35:20 -0400 (CLT) Date: Wed, 16 May 2007 12:35:20 -0400 From: Alvaro Herrera To: Keaton Adams Cc: pgsql-admin@postgresql.org Subject: Re: WAL file utilization question Message-ID: <20070516163520.GZ4582@alvh.no-ip.org> References: <1179332909.22514.3.camel@MXLRMT-208.corp.mxlogic.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1179332909.22514.3.camel@MXLRMT-208.corp.mxlogic.com> User-Agent: Mutt/1.5.13 (2006-08-11) X-Virus-Scanned: ClamAV version 0.88.5, clamav-milter version 0.88.5 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]); Wed, 16 May 2007 09:35:23 -0700 (PDT) X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200705/143 X-Sequence-Number: 25463 Keaton Adams wrote: > Even with full_page_writes set to false and checkpoints taking place > every 60 seconds or so, the amount of WAL log data generated per minute > seems to be significant. Stop right there. Having checkpoints every 60 seconds is insane. Did you look into increasing checkpoint_segments a lot? > So my question is this: If I kick off a transaction that loads records > with a size of 100 bytes, does the insert for that record take 100 bytes > in the WAL file, or is the data archived in the WAL log in page size > (8k) portions? Vicinity of 100 bytes, except right after checkpoints as you noted. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.