X-Original-To: pgsql-admin-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 030BDD1B4DA for ; Tue, 10 Feb 2004 16:24:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 58008-06 for ; Tue, 10 Feb 2004 12:24:12 -0400 (AST) Received: from rambutan.pingpong.net (81.milagro.bahnhof.net [195.178.168.81]) by svr1.postgresql.org (Postfix) with ESMTP id 60E59D1D8AE for ; Tue, 10 Feb 2004 12:24:11 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by rambutan.pingpong.net (8.12.9p2/8.12.9) with ESMTP id i1AGO8eq097331; Tue, 10 Feb 2004 17:24:09 +0100 (CET) (envelope-from girgen@pingpong.net) Date: Tue, 10 Feb 2004 17:24:08 +0100 From: Palle Girgensohn To: Robert Treat Cc: pgsql-admin@postgresql.org Subject: Re: reduce downtime when upgrading 7.3 -> 7.4 Message-ID: <51450000.1076430248@rambutan.pingpong.net> In-Reply-To: <1076423601.17917.8.camel@camel> References: <331690000.1076207705@palle.girgensohn.se> <1076423601.17917.8.camel@camel> X-Mailer: Mulberry/3.1.0 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200402/142 X-Sequence-Number: 12316 --On Tuesday, February 10, 2004 09:33:20 -0500 Robert Treat wrote: > On Sat, 2004-02-07 at 21:35, Palle Girgensohn wrote: >> Hi, >> >> We use postgresql for rather large databases. For a typical >> installation, a pg_restore takes a couple of hours, at least (the >> dumpfiles are usually 2-4 gigabytes or so, including BLOBs). The >> machines are expected to be up 24/7, so this dump/restore procedure >> makes upgrading unpopular. Is there any (safe) way to speed this >> process up? >> >> If pg_upgrade is not a good idea, how can I speed up pg_restore? Best >> way to set things like fsync etc in postgresql.conf? Will it make a big >> difference? >> > > yes, setting fsync off should make a significant difference. I usually > recommend it cause if there is a machine failure during restore I will > want to start the process again anyway. Other items you should probably > change are cranking up sort_mem significantly (if your restore is the > only process running, let it take up most of the ram on the box) and you > can increase check_point segments as well. HTH Thanks, I'll try this. Regards, Palle