Received: from localhost (maia-3.hub.org [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id 8E4039FA222 for ; Mon, 18 Dec 2006 00:59:51 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-new, port 10024) with ESMTP id 24925-02 for ; Mon, 18 Dec 2006 00:59:09 -0400 (AST) X-Greylist: from auto-whitelisted by SQLgrey-1.7.4 Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130]) by postgresql.org (Postfix) with ESMTP id 5746E9F9C35 for ; Mon, 18 Dec 2006 00:59:15 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.13.6/8.13.6) with ESMTP id kBI4xCgV021941; Sun, 17 Dec 2006 23:59:12 -0500 (EST) To: tsuraan cc: "Greg Smith" , pgsql-performance@postgresql.org Subject: Re: Scaling concerns In-reply-to: <84fb38e30612172036s74183c5ey930eb4f608578f81@mail.gmail.com> References: <84fb38e30612160926l609ec55cs352b2495289bae0c@mail.gmail.com> <20061217032527.GL29195@andi-lap.la.revver.com> <84fb38e30612171359r6288de69y2553be256019ba4b@mail.gmail.com> <84fb38e30612172036s74183c5ey930eb4f608578f81@mail.gmail.com> Comments: In-reply-to tsuraan message dated "Sun, 17 Dec 2006 22:36:50 -0600" Date: Sun, 17 Dec 2006 23:59:12 -0500 Message-ID: <21940.1166417952@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200612/456 X-Sequence-Number: 22512 tsuraan writes: > Is the WAL at the same location as the xlog (transaction log?)? Same thing. > The checkpoint_segments doc says increasing that value is really only > useful if the xlog is separate from the data, Dunno where you read that, but it's utter bilge. If you've got a write-intensive workload, you want to crank checkpoint_segments as high as you can stand. With the default settings on a modern machine it's not hard at all to push it into checkpointing every dozen or seconds, which will completely kill performance. (Disk space for pg_xlog/ and potential delay during crash restart are the only negatives here. If you are willing to push the average inter-checkpoint interval past five minutes then you need to increase checkpoint_timeout too.) regards, tom lane