Received: from localhost (unknown [200.46.204.187]) by postgresql.org (Postfix) with ESMTP id 9A4559F9611 for ; Mon, 22 Oct 2007 15:34:36 -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 95215-04 for ; Mon, 22 Oct 2007 15:34:21 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from mail.autorevenue.com (ip-208.178.167.106.gblx.net [208.178.167.106]) by postgresql.org (Postfix) with ESMTP id 557179FB301 for ; Mon, 22 Oct 2007 15:32:38 -0300 (ADT) Received: (qmail 14589 invoked from network); 22 Oct 2007 14:35:19 -0400 Received: from unknown (HELO D1RPQZB1) (msteben@192.168.1.181) by mail.autorevenue.com with SMTP; 22 Oct 2007 14:35:19 -0400 From: "Mark Steben" To: "'Tom Lane'" Cc: Subject: Re: bgwriter update question Date: Mon, 22 Oct 2007 14:30:35 -0400 Message-ID: <001101c814d9$a2fce2c0$b501a8c0@D1RPQZB1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Office Outlook 11 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.3790.2992 thread-index: AcgU2DG83VZELKJ5T+eJxmpc00X/0wAAJhzg In-Reply-To: <19802.1193077052@sss.pgh.pa.us> X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.128 tagged_above=0 required=5 tests=AWL=0.027, BAYES_50=0.001, RDNS_DYNAMIC=0.1 X-Spam-Level: X-Archive-Number: 200710/213 X-Sequence-Number: 27127 Thanks Tom, every piece of new information adds to my education. I'm going to put to bed this effort of trying to verify BGWRITER, knowing that it works in 8.2 and move on. Regards, Mark Steben -----Original Message----- From: Tom Lane [mailto:tgl@sss.pgh.pa.us] Sent: Monday, October 22, 2007 2:18 PM To: Mark Steben Cc: pgsql-admin@postgresql.org Subject: Re: [ADMIN] bgwriter update question "Mark Steben" writes: > I'm running on 7.4.5 (pre bgwriter) and moving to 8.2 in 3 weeks. I turn > fsync off on my testing server. When I update a table I expect the file in > the base directory to remain the same size until the next syncpoint when it > would then update. But it increases immediately. I thought updates remain in > memory until syncpoint in pre-8.0. Am I missing somethng? The updates don't go to disk immediately (they didn't in 7.4 either). But we always allocate disk space immediately when needed, to make sure it's available. Otherwise we could face the embarrassing situation of having noplace to put rows inserted by an already-reported-committed transaction. regards, tom lane