X-Original-To: pgsql-performance@postgresql.org Received: from spampd.localdomain (postgresql.org [64.49.215.8]) by postgresql.org (Postfix) with ESMTP id 2623D47592C for ; Thu, 10 Apr 2003 14:59:29 -0400 (EDT) Received: from beamish.nsd.ca (beamish.nsd.ca [205.150.156.194]) by postgresql.org (Postfix) with ESMTP id B4EBA474E42 for ; Thu, 10 Apr 2003 14:59:27 -0400 (EDT) Received: (from smap@localhost) by beamish.nsd.ca (8.9.3/8.9.3) id OAA16964; Thu, 10 Apr 2003 14:58:20 -0400 X-Authentication-Warning: beamish.nsd.ca: smap set sender to using -f Received: from reddog.nsd.ca(192.168.101.30) by beamish.nsd.ca via smap (V2.1/2.1+anti-relay+anti-spam) id xma016962; Thu, 10 Apr 03 14:58:16 -0400 Received: from nsd.ca (jllachan-linux.nsd.ca [192.168.101.148]) by reddog.nsd.ca (8.8.7/8.8.7) with ESMTP id OAA03220; Thu, 10 Apr 2003 14:52:59 -0400 Message-ID: <3E95BF2B.B93FD0EF@nsd.ca> Date: Thu, 10 Apr 2003 14:59:55 -0400 From: Jean-Luc Lachance X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.18-24.7.x i686) X-Accept-Language: en MIME-Version: 1.0 To: Tom Lane Cc: jim@nasby.net, "scott.marlowe" , Matthew Nuzum , "'Josh Berkus'" , "'Pgsql-Performance'" Subject: Re: Caching (was Re: choosing the right platform) References: <00a001c2febb$631141d0$6900a8c0@mattspc> <20030409184744.W31861@flake.decibel.org> <11732.1049934037@sss.pgh.pa.us> <3E957F44.A1106F6A@nsd.ca> <4072.1049985615@sss.pgh.pa.us> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Spam-Status: No, hits=-30.7 required=5.0 tests=BAYES_01,EMAIL_ATTRIBUTION,QUOTED_EMAIL_TEXT,REFERENCES, REPLY_WITH_QUOTES,X_AUTH_WARNING autolearn=ham version=2.50 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) X-Archive-Number: 200304/118 X-Sequence-Number: 1624 How can we solve the problem of cache trashing when scanning large tables? Tom Lane wrote: > > Jean-Luc Lachance writes: > > Shouldn't there be a way to lock some tables in PG cache? > > In my opinion, no. I do not think a manual locking feature could > possibly be used effectively. It could very easily be abused to > decrease net performance, though :-( > > It does seem that a smarter buffer management algorithm would be a good > idea, but past experiments have failed to show any measurable benefit. > Perhaps those experiments were testing the wrong conditions. I'd still > be happy to see LRU(k) or some such method put in, if someone can prove > that it actually does anything useful for us. (As best I recall, I only > tested LRU-2 with pgbench. Perhaps Josh's benchmarking project will > offer a wider variety of interesting scenarios.) > > regards, tom lane > > ---------------------------(end of broadcast)--------------------------- > TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org