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 13A3CD1B1D9 for ; Mon, 12 Jul 2004 15:01:45 -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 72577-01 for ; Mon, 12 Jul 2004 18:01:35 +0000 (GMT) Received: from smtp105.mail.sc5.yahoo.com (smtp105.mail.sc5.yahoo.com [66.163.169.225]) by svr1.postgresql.org (Postfix) with SMTP id B23D3D1B1CD for ; Mon, 12 Jul 2004 15:01:32 -0300 (ADT) Received: from unknown (HELO jupiter.black-lion.info) (janwieck@68.80.245.191 with login) by smtp105.mail.sc5.yahoo.com with SMTP; 12 Jul 2004 18:01:32 -0000 Received: from Yahoo.com (vgateway.libertyrms.info [207.219.45.62] (may be forged)) (authenticated bits=0) by jupiter.black-lion.info (8.12.10/8.12.9) with ESMTP id i6CI1N3v025173; Mon, 12 Jul 2004 14:01:24 -0400 (EDT) (envelope-from JanWieck@Yahoo.com) Message-ID: <40F2D1E6.10504@Yahoo.com> Date: Mon, 12 Jul 2004 14:01:10 -0400 From: Jan Wieck User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Josh Berkus Cc: Rod Taylor , Christopher Browne , Postgresql Performance Subject: Re: Working on huge RAM based datasets References: <6EE64EF3AB31D5448D0007DD34EEB34101AEC2@Herge.rcsinc.local> <1089643087.42256.7.camel@jester> <200407120938.01764.josh@agliodbs.com> In-Reply-To: <200407120938.01764.josh@agliodbs.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=2.6 tagged_above=0.0 required=5.0 tests=RCVD_IN_DYNABLOCK, RCVD_IN_SORBS X-Spam-Level: ** X-Archive-Number: 200407/65 X-Sequence-Number: 7456 On 7/12/2004 12:38 PM, Josh Berkus wrote: > Rond, Chris, > >> > What would be most interesting to see is whether this makes it wise to >> > increase shared buffer size. It may be more effective to bump down >> > the cache a little, and bump up sort memory; hard to tell. >> >> How do we go about scheduling tests with the OSDL folks? If they could >> do 10 runs with buffers between 1k and 500k it would help us get a broad >> view of the situation. > > Yes. We'll need to. However, I'd like to wait until we're officially in > Beta. I'll be seeing the OSDL folks in person (PostgreSQL+OSDL BOF at Linux > World Expo!!) in a couple of weeks. > Don't forget to add that ARC needs some time actually to let the algorithm adjust the queue sizes and populate the cache according to the access pattern. You can't start a virgin postmaster and then slam on the accellerator of your test application by launching 500 concurrent clients out of the blue and expect that it starts off airborne. Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #