X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id C4982534AC for ; Fri, 1 Apr 2005 02:51:52 +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 51400-02 for ; Fri, 1 Apr 2005 01:51:43 +0000 (GMT) Received: from floppy.pyrenet.fr (floppy.pyrenet.fr [194.116.145.2]) by svr1.postgresql.org (Postfix) with ESMTP id ACD97537FF for ; Fri, 1 Apr 2005 02:51:42 +0100 (BST) Received: by floppy.pyrenet.fr (Postfix, from userid 106) id 3CB2330952; Fri, 1 Apr 2005 03:51:43 +0200 (MET DST) From: "Qingqing Zhou" X-Newsgroups: pgsql.hackers Subject: Re: [ADMIN] Excessive growth of pg_attribute and other system tables Date: Fri, 1 Apr 2005 09:50:49 +0800 Organization: Hub.Org Networking Services Lines: 21 Message-ID: References: <200503171515.38214.scrawford@pinpointresearch.com> <423F17D7.7040904@zeut.net> <1219.1111434030@sss.pgh.pa.us> <200503310944.50647.scrawford@pinpointresearch.com> <7739.1112295848@sss.pgh.pa.us> X-Complaints-To: usenet@news.hub.org X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1437 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1441 To: pgsql-hackers@postgresql.org X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=1.21 tagged_above=0 required=5 tests=AWL, PRIORITY_NO_NAME X-Spam-Level: * X-Archive-Number: 200504/5 X-Sequence-Number: 65874 "Tom Lane" writes > Steve Crawford writes: > > On Monday 21 March 2005 11:40 am, Tom Lane wrote: > >> However, given that there are 9334 tuples in 82282 pages, I'd say > >> that autovacuum has already failed Steve rather badly :-(. There > >> shouldn't be more than a couple hundred pages given that number of > >> rows. Perhaps the FSM settings are too small? > Seems this is another question pointing to the inproper setting of "can-be-avoided" shared memory parameters. Maybe we should eliminate GUC parameters related to the FSM. Can we follow Alvaro's idea like spilling some data of FSM into disk while keeping the indices and maybe part of data in the memory? So no free page would be discarded due to no space to record them in FSM? Also, in this handling, efficiency should not be a problem. Regards, Qingqing