Received: from localhost (unknown [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id D0EF12E308B for ; Sat, 15 Mar 2008 01:07:58 -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 41971-05 for ; Sat, 15 Mar 2008 01:07:50 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from mailout.abaca.com (mailout.abaca.com [75.59.196.2]) by postgresql.org (Postfix) with ESMTP id B9B7C2E00BB for ; Sat, 15 Mar 2008 01:07:50 -0300 (ADT) Received: from v-abaca-epg.lab.abaca.com (localhost.localdomain [127.0.0.1]) by v-abaca-epg.lab.abaca.com (Postfix-out) with ESMTP id 1A3A6110097; Fri, 14 Mar 2008 21:11:13 -0700 (PDT) X-Propel-Return-Path: Received: from mailout.abaca.com ([172.16.83.19]) by [127.0.0.1] ([127.0.0.1]) (port 7027) (Abaca EPG outproxy filter 3.1.0.8749 $Rev: 8725 $) id 5uNZr83f4bc0; Fri, 14 Mar 2008 21:11:13 -0700 Received: from mail.abaca.com (unknown [172.16.83.4]) by v-abaca-epg.lab.abaca.com (Postfix-out) with ESMTP id 84F82110086; Fri, 14 Mar 2008 21:11:12 -0700 (PDT) Received: from [172.16.83.73] (172.16.83.73) by mail.abaca.com (172.16.83.4) with Microsoft SMTP Server id 8.0.744.0; Fri, 14 Mar 2008 21:08:39 -0700 Subject: Re: best way to run maintenance script From: Vinubalaji Gopal To: "Joshua D. Drake" CC: "Pgsql-Performance ((E-mail))" In-Reply-To: <20080314191722.13e19440@commandprompt.com> References: <1205542821.12661.16.camel@vinu-ubuntu.abaca> <20080314173235.58f854e6@commandprompt.com> <1205545912.12661.22.camel@vinu-ubuntu.abaca> <20080314191722.13e19440@commandprompt.com> Content-Type: text/plain Date: Fri, 14 Mar 2008 21:07:45 -0800 Message-ID: <1205557665.12661.51.camel@vinu-ubuntu.abaca> MIME-Version: 1.0 X-Mailer: Evolution 2.12.1 Content-Transfer-Encoding: 7bit X-Propel-ID: 5uNZr83f4bc0 X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests=none X-Spam-Level: X-Archive-Number: 200803/327 X-Sequence-Number: 29281 > > I think you will find if you do it the right way, which is to say the > way that it is meant to be done with the configurable options, your > life will be a great deal more pleasant than some one off hack. > yeah I agree. The pg_maintanence script which calls vacuum and analyze is the one of the thing that is causing more problems. I am trying out various vacuum options (vacuum_cost_limit, vacuum_cost_delay) and finding it hard to understand the implications of the variables. What are the optimal values for the vacuum_* parameters - for a really active database (writes at the rate of ~ 50 rows/seconds). I started with vacuum_cost_delay = 200 vacuum_cost_limit = 400 and that did not help much. -- Vinu