X-Original-To: pgsql-admin-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id 133C15349F for ; Tue, 5 Apr 2005 18:56:14 +0100 (BST) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 28735-05 for ; Tue, 5 Apr 2005 17:56:05 +0000 (GMT) Received: from mailhost.intellivid.com (mailhost.intellivid.com [64.32.200.11]) by svr1.postgresql.org (Postfix) with ESMTP id 1879F52A8B for ; Tue, 5 Apr 2005 18:56:04 +0100 (BST) Received: from [192.168.2.68] (spectre.intellivid.com [192.168.2.68]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by newmail.intellivid.com (Postfix) with ESMTP id 65745F18174; Tue, 5 Apr 2005 13:56:05 -0400 (EDT) Subject: Re: VACUUM and read-mostly tables From: Ian Westmacott To: "Jim C. Nasby" Cc: pgsql-admin@postgresql.org In-Reply-To: <20050405153904.GP93835@decibel.org> References: <000201c53993$b07bb440$76d81e42@hsd1.ma.comcast.net> <20050405044128.GL93835@decibel.org> <1112713986.8115.81.camel@spectre.intellivid.com> <20050405153904.GP93835@decibel.org> Content-Type: text/plain Message-Id: <1112723765.8115.119.camel@spectre.intellivid.com> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 Date: Tue, 05 Apr 2005 13:56:05 -0400 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests= X-Spam-Level: X-Archive-Number: 200504/65 X-Sequence-Number: 17078 On Tue, 2005-04-05 at 11:39, Jim C. Nasby wrote: > On Tue, Apr 05, 2005 at 11:13:06AM -0400, Ian Westmacott wrote: > > On Tue, 2005-04-05 at 00:41, Jim C. Nasby wrote: > > > We'll only answer if you do a write-up on your database. :P > > > > > > Seriously, those are some seriously big numbers. What else is the > > > database doing? What hardware is it running on? > > > > > > We run on a dual 3.2GHz P4 with 2GB RAM, but are still > > finalizing the storage hardware. We've tried various > > flavors of RAID, filesystems and volume management (and > > are anxious to try out tablespaces in 8). We've found > > fragmentation to be our largest limiting factor. XFS > > helps with that, and seems to provide the highest > > sustained throughput on raw tables, but its not the end > > of the story since fragmentation is still high. > > What else is the database doing besides the inserts? A proportionally small number of updates, no deletes, and a set of moderately complex queries. > And if UFS is available for linux you should might try it. Would UFS help the fragmentation issue? We have seen ext3 allocating blocks in 2-4 pages, while XFS manages 8-16 pages. Thanks, --Ian