X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F3D02D1C4E0 for ; Wed, 21 Apr 2004 01:39:03 -0300 (ADT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75150-05 for ; Wed, 21 Apr 2004 01:39:03 -0300 (ADT) Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130]) by svr1.postgresql.org (Postfix) with ESMTP id 1EA49D1B4E1 for ; Wed, 21 Apr 2004 01:39:02 -0300 (ADT) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.11/8.12.11) with ESMTP id i3L4crj1026128; Wed, 21 Apr 2004 00:38:53 -0400 (EDT) To: "Matthew T. O'Connor" Cc: Christopher Kings-Lynne , Bruce Momjian , Jeff Boes , PostgreSQL-development Subject: Re: pg_autovacuum crashes when query fails for temp tables In-reply-to: <4085EE81.9080309@zeut.net> References: <200404201525.i3KFPp809934@candle.pha.pa.us> <4085D554.9010802@familyhealth.com.au> <4085EE81.9080309@zeut.net> Comments: In-reply-to "Matthew T. O'Connor" message dated "Tue, 20 Apr 2004 23:46:09 -0400" Date: Wed, 21 Apr 2004 00:38:52 -0400 Message-ID: <26127.1082522332@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200404/609 X-Sequence-Number: 52531 "Matthew T. O'Connor" writes: > System Tables: pg_autovacuum treats non-shared system tables just like > any other table. It monitors the activity and vacuums when it deems it > appropriate. As for shared system tables: In user databases they are > only analyzed by pg_autovacuum, while connected to template1, > pg_autovacuum will treat the shared tables as normal tables and vacuum > when appropriate. As long as you hit template1 reasonably often, this will work. But I'm a bit concerned about the possibility that some maverick will decide he doesn't need template1. (It's at least theoretically possible to run without it.) Plan B would be to ignore the sharedness issue and vacuum/analyze shared catalogs the same as anything else. While this would certainly result in more vacuums than really necessary, these tables are probably small enough that it'd hardly matter ... Comments? regards, tom lane