X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8E7AED1D362 for ; Thu, 12 Feb 2004 20:27:30 +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 91251-02 for ; Thu, 12 Feb 2004 16:27:25 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id D69FFD1BB9E for ; Thu, 12 Feb 2004 16:27:23 -0400 (AST) Received: from [63.195.55.98] (HELO 192.168.1.29) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 4417083; Thu, 12 Feb 2004 12:28:27 -0800 From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Leon Out , pgsql-performance@postgresql.org Subject: Re: Disappointing performance in db migrated from MS SQL Server Date: Thu, 12 Feb 2004 12:26:33 -0800 User-Agent: KMail/1.5.4 References: <70CC4F08-5D89-11D8-AB22-0030658FB514@comvision.com> In-Reply-To: <70CC4F08-5D89-11D8-AB22-0030658FB514@comvision.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200402121226.33823.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200402/151 X-Sequence-Number: 5675 Leon, > Hello all. I am in the midst of porting a large web application from a > MS SQL Server backend to PostgreSQL. The migration work is basically > complete, and we're at the testing and optimization phase of the > project. The results so far have been disappointing, with Postgres > performing queries in about the same time as SQL Server even though > Postgres is running on a dedicated box with about 4 times the clock > speed of the SQL Server box. For a chart of my results, please see > http://leonout.com/pggraph.pdf for a graph of some test results. Your settings look ok to start, but we'll probably want to tune them further. Can you post some details of the tests? Include: 1) the query 2) the EXPLAIN ANALYZE results of the query 3) Whether you ran the test as the only connection, or whether you tested multi-user load. The last is fairly important for a SQL Server vs. PostgreSQL test; SQL Server is basically a single-user-database, so like MySQL it appears very fast until you get a bunch o' users on it. Finally, for most queries the disk I/O and the RAM are more important than the CPU clock speed. From the looks of it, you upgraded the CPU + RAM, but did downgraded the disk array as far as database writes are concered; not a terrible effective way to gain performance on your hardware. -- -Josh Berkus Aglio Database Solutions San Francisco