X-Original-To: pgsql-performance@postgresql.org Received: from localhost (postgresql.org [64.49.215.8]) by postgresql.org (Postfix) with ESMTP id DB835474E61 for ; Fri, 7 Feb 2003 12:17:43 -0500 (EST) Received: from davinci.ethosmedia.com (unknown [209.10.40.251]) by postgresql.org (Postfix) with ESMTP id 41DBE4758C9 for ; Fri, 7 Feb 2003 12:17:33 -0500 (EST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 2829171; Fri, 07 Feb 2003 09:17:03 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Andreas Pflug , pgsql-performance@postgresql.org Subject: Re: how to configure my new server Date: Fri, 7 Feb 2003 09:16:27 -0800 User-Agent: KMail/1.4.3 References: <3E4326B6.5030100@web.de> In-Reply-To: <3E4326B6.5030100@web.de> MIME-Version: 1.0 Message-Id: <200302070916.27541.josh@agliodbs.com> Content-Transfer-Encoding: quoted-printable X-Virus-Scanned: by AMaViS new-20020517 X-Archive-Number: 200302/40 X-Sequence-Number: 1096 Andreas, > Josh wrote: > >> With a large database you may even think about > >> shifting individual tables or indexes to seperate disks. > > OK, I admit it was a bit provoking. It was my intention to stir things up= a > little bit ;-) IMHO, thinking about locating data on dedicated files is a > waste of time on small servers. Let the hardware do the job for you! It is > "good enough". Aha, by "large databases" I mean "several million records". In the odd ca= se=20 where you have a database which has one or two tables which are larger than= =20 the rest of the database combined, you can get a performance boost by putti= ng=20 those tables, and/or their indexes, on a seperate spindle. Frankly, for small servers, a pair of mirrored IDE drives is adequate. And= ,=20 of course, if you have a RAID 1+0 controller, that's better than trying to= =20 directly allocate different files to different disks ... except the WAL log. > I don't think solid state disks are a way out (unless you don't know where > to bury your money :-). Maybe the gurus can tell more about PostgreSQL's > caching, but for my opinion if enough RAM is available after some time all > of the DB should be in cache eliminating the need to access the disks for > read access. For writing, which is typically less than 10 % of total load, > an optimizing caching disk controller should be sufficient. Depends on the database. I've worked on DBs where writes were 40% of all= =20 queries ... and 90% of the system resource load. For those databases,=20 moving the WAL log to a RAMdisk might mean a big boost. Also, I'm currently working to figure out a solution for some 1U machines= =20 which don't have any *room* for an extra drive for WAL. A PCI ramdisk woul= d=20 be just perfect ... --=20 Josh Berkus Aglio Database Solutions San Francisco