X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id E0F105445B for ; Tue, 10 May 2005 14:46:33 -0300 (ADT) 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 25181-08 for ; Tue, 10 May 2005 17:46:24 +0000 (GMT) Received: from davinci.ethosmedia.com (server227.ethosmedia.com [209.128.84.227]) by svr1.postgresql.org (Postfix) with ESMTP id 10E5454326 for ; Tue, 10 May 2005 14:46:23 -0300 (ADT) X-EthosMedia-Virus-Scanned: no infections found Received: from [63.195.55.98] (account josh@agliodbs.com HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.1.8) with ESMTP id 7345177 for pgsql-performance@postgresql.org; Tue, 10 May 2005 10:48:24 -0700 From: Josh Berkus Organization: Aglio Database Solutions Subject: Re: Partitioning / Clustering Date: Tue, 10 May 2005 10:46:24 -0700 User-Agent: KMail/1.8 Cc: pgsql-performance@postgresql.org References: <5F94B6EB-3925-4CD3-A67F-4310890B438A@advfn.com> <4280C801.80603@arbash-meinel.com> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline To: Undisclosed.Recipients: ; Message-Id: <200505101046.25108.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.013 tagged_above=0 required=5 tests=AWL X-Spam-Level: X-Archive-Number: 200505/154 X-Sequence-Number: 12318 Alex, > This is why I mention partitioning. It solves this issue by storing =A0 > different data sets on different machines under the same schema. =A0 That's clustering, actually. Partitioning is simply dividing up a table in= to=20 chunks and using the chunks intelligently. Putting those chunks on sepera= te=20 machines is another thing entirely. =20 We're working on partitioning through the Bizgres sub-project: www.bizgres.org / http://pgfoundry.org/projects/bizgres/ =2E.. and will be pushing it to the main PostgreSQL when we have something. I invite you to join the mailing list. > These seperate chunks of the table can then be replicated as well for =A0 > data redundancy and so on. MySQL are working on these things,=20 Don't hold your breath. MySQL, to judge by their first "clustering"=20 implementation, has a *long* way to go before they have anything usable. I= n=20 fact, at OSCON their engineers were asking Jan Wieck for advice. If you have $$$ to shell out, my employer (GreenPlum) has a multi-machine=20 distributed version of PostgreSQL. It's proprietary, though. =20 www.greenplum.com. If you have more time than money, I understand that Stanford is working on= =20 this problem: http://www-db.stanford.edu/~bawa/ But, overall, some people on this list are very mistaken in thinking it's a= n=20 easy problem. GP has devoted something like 5 engineers for 3 years to=20 develop their system. Oracle spent over $100 million to develop RAC. =20 > but PG =A0=20 > just has a bunch of third party extensions, I wonder why these are =A0 > not being integrated into the main trunk :/=20 Because it represents a host of complex functionality which is not applicab= le=20 to most users? Because there are 4 types of replication and 3 kinds of=20 clusering and not all users want the same kind? =2D-=20 Josh Berkus Aglio Database Solutions San Francisco