Received: from localhost (unknown [200.46.204.191]) by postgresql.org (Postfix) with ESMTP id 598AA2E33F6 for ; Thu, 10 Jan 2008 10:35:02 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.191]) (amavisd-maia, port 10024) with ESMTP id 20903-02 for ; Thu, 10 Jan 2008 10:34:56 -0400 (AST) Received: from puck.nether.net (puck.nether.net [204.42.254.5]) by postgresql.org (Postfix) with ESMTP id 1A3B42E33E3 for ; Thu, 10 Jan 2008 10:34:55 -0400 (AST) Received: from puck.nether.net (puck.nether.net [204.42.254.5]) by puck.nether.net (8.14.1/8.12.9) with ESMTP id m0AEYfkV007222 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 10 Jan 2008 09:34:41 -0500 (EST) (envelope-from jared@puck.nether.net) Received: (from jared@localhost) by puck.nether.net (8.14.1/8.12.9/Submit) id m0AEYeSF007219; Thu, 10 Jan 2008 09:34:40 -0500 (EST) (envelope-from jared) Date: Thu, 10 Jan 2008 09:34:40 -0500 From: Jared Mauch To: Stephane Bailliez Cc: Jared Mauch , pgsql-performance@postgresql.org Subject: Re: big database performance Message-ID: <20080110143440.GB79941@puck.nether.net> References: <4784835A.8070604@careerjunction.co.za> <47848575.8000007@commandprompt.com> <20080109174243.GA8915@puck.nether.net> <4785FCB7.3090403@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4785FCB7.3090403@gmail.com> User-Agent: Mutt/1.5.17 (2007-11-01) X-Greylist: Sender is SPF-compliant, not delayed by milter-greylist-4.0 (puck.nether.net [204.42.254.5]); Thu, 10 Jan 2008 09:34:41 -0500 (EST) X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200801/55 X-Sequence-Number: 28294 On Thu, Jan 10, 2008 at 12:08:39PM +0100, Stephane Bailliez wrote: > Jared Mauch wrote: >> I do large databases in Pg, like 300GB/day of new data. > > That's impressive. Would it be possible to have details on your hardware, > schema and configuration and type of usage ? > > I'm sure there's something to learn in there for a lot of people (or at > least for me) http://archives.postgresql.org/pgsql-performance/2007-12/msg00372.php http://archives.postgresql.org/pgsql-performance/2006-05/msg00444.php The hardware specs are kinda boring since it's not I/O bound, so you could get the same disk performance out of some EIDE 7200 rpm disks (which I have done for testing). The current setup is a 4xOpteron 8218 (dual core) w/ 16G ram. I have roughly 12TB usable disk space on the sysem connected via some SATA <-> FC thing our systems folks got us. Problem I have is the linear cpu speed isn't enough and there would be challenges splitting the workload across multiple cpus. All my major reporting is done via pg_dump and I'm pondering what would happen if I just removed Pg from the equation for the major reporting tasks entirely. I may see much better performance without the database [in my way]. I've not done that as some types of data access would need to be significantly redone and I don't want to spend the time on that... - Jared -- Jared Mauch | pgp key available via finger from jared@puck.nether.net clue++; | http://puck.nether.net/~jared/ My statements are only mine.