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 2D3235372B; Tue, 5 Apr 2005 04:35:09 +0100 (BST) 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 86173-08; Tue, 5 Apr 2005 03:35:08 +0000 (GMT) Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130]) by svr1.postgresql.org (Postfix) with ESMTP id D3C0C5373D; Tue, 5 Apr 2005 04:35:07 +0100 (BST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.13.1/8.13.1) with ESMTP id j353Z07O006491; Mon, 4 Apr 2005 23:35:01 -0400 (EDT) To: Christopher Petrilli Cc: jd@commandprompt.com, "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: Sustained inserts per sec ... ? In-reply-to: <59d991c4050404195443fba84c@mail.gmail.com> References: <20050401140421.P18194@ganymede.hub.org> <59d991c4050401124638c4ab5d@mail.gmail.com> <1112388818.3487.34.camel@jd2.commandprompt.com> <59d991c4050401125956d6069d@mail.gmail.com> <59d991c40504040648457808e9@mail.gmail.com> <26454.1112629979@sss.pgh.pa.us> <59d991c405040409094e7bbd17@mail.gmail.com> <26669.1112631807@sss.pgh.pa.us> <59d991c40504041846bc39a6@mail.gmail.com> <5217.1112668589@sss.pgh.pa.us> <59d991c4050404195443fba84c@mail.gmail.com> Comments: In-reply-to Christopher Petrilli message dated "Mon, 04 Apr 2005 22:54:57 -0400" Date: Mon, 04 Apr 2005 23:35:00 -0400 Message-ID: <6490.1112672100@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.006 tagged_above=0 required=5 tests=AWL X-Spam-Level: X-Archive-Number: 200504/78 X-Sequence-Number: 11475 Christopher Petrilli writes: > On Apr 4, 2005 10:36 PM, Tom Lane wrote: >> The indicated fix of course is to increase shared_buffers. > Any idea where it should be set? Not really. An upper bound would be the total size of the finished indexes for one 10M-row table, but one would suppose that that's overkill. The leaf pages shouldn't have to stay in RAM to have reasonable behavior --- the killer case is when upper-level tree pages drop out. Or that's what I'd expect anyway. You could probably drop the inter-insert sleep for testing purposes, if you want to experiment with several shared_buffers values quickly. regards, tom lane