Received: from maia.hub.org (maia-5.hub.org [200.46.204.29])
	by mail.postgresql.org (Postfix) with ESMTP id 9C0C51337BD5
	for <pgsql-admin-postgresql.org@mail.postgresql.org>;
	Thu,  3 Mar 2011 04:17:30 -0400 (AST)
Received: from mail.postgresql.org ([200.46.204.86])
	by maia.hub.org (mx1.hub.org [200.46.204.29]) (amavisd-maia, port 10024)
	with ESMTP id 03921-08; Thu,  3 Mar 2011 08:17:23 +0000 (UTC)
X-Greylist: domain auto-whitelisted by SQLgrey-1.7.6
X-Greylist: domain auto-whitelisted by SQLgrey-1.7.6
Received: from exprod7og114.obsmtp.com (exprod7og114.obsmtp.com [64.18.2.215])
	by mail.postgresql.org (Postfix) with ESMTP id E93081337B68;
	Thu,  3 Mar 2011 04:17:21 -0400 (AST)
Received: from source ([209.85.215.42]) (using TLSv1) by
	exprod7ob114.postini.com ([64.18.6.12]) with SMTP
	ID DSNKTW9OkUNt6mtU+Z3OQm7IMGmxaOvV+KUT@postini.com;
	Thu, 03 Mar 2011 00:17:22 PST
Received: by ewy1 with SMTP id 1so363715ewy.15
	for <multiple recipients>; Thu, 03 Mar 2011 00:17:19 -0800 (PST)
Received: by 10.213.16.78 with SMTP id n14mr815827eba.99.1299140209244;
	Thu, 03 Mar 2011 00:16:49 -0800 (PST)
Received: from [192.168.1.183] (dsl-hkibrasgw2-ff7ac300-240.dhcp.inet.fi
	[88.195.122.240])
	by mx.google.com with ESMTPS id t50sm717840eeh.12.2011.03.03.00.16.46
	(version=TLSv1/SSLv3 cipher=OTHER);
	Thu, 03 Mar 2011 00:16:47 -0800 (PST)
Message-ID: <4D6F4E5D.6080507@enterprisedb.com>
Date: Thu, 03 Mar 2011 10:16:29 +0200
From: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; fi-FI;
	rv:1.9.1.16) Gecko/20101226 Icedove/3.0.11
MIME-Version: 1.0
To: daveg <daveg@sonic.net>
CC: Robert Haas <robertmhaas@gmail.com>, Greg Stark <gsstark@mit.edu>,
	David Christensen <david@endpoint.com>,
	Maxim Boguk <maxim.boguk@gmail.com>, Fujii Masao <masao.fujii@gmail.com>,
	pgsql-admin@postgresql.org,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Re: [HACKERS] Re: PD_ALL_VISIBLE flag was incorrectly set happend
	during repeatable vacuum
References: <AANLkTino+Gc2vvOVdzCPRRXW-N7R5=ikYMnNKT5W=mk1@mail.gmail.com>
	<AANLkTikw==USr3+4Mt+w97P3ymux-Bhn0G7Ad0ZJHbqG@mail.gmail.com>
	<4D2D6A56.9060808@enterprisedb.com>
	<20110228212832.GH26397@sonic.net>
	<1B3CCBDD-4AC6-404A-9E26-DB4CA2554228@endpoint.com>
	<AANLkTi=xq4SV6jwA7txvrbEWocyeiU6XiTG=-zVy+7Dq@mail.gmail.com>
	<AANLkTi=Ye7AoF7_uVQfPVn-ok2VOp65mi0xU3H7kCke3@mail.gmail.com>
	<20110303071208.GA4053@sonic.net>
In-Reply-To: <20110303071208.GA4053@sonic.net>
Content-Type: multipart/mixed; boundary="------------050707090009000309090802"
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Spam-Status: No, hits=-4.2 tagged_above=-10 required=5 tests=BAYES_00=-1.9,
	RCVD_IN_DNSWL_MED=-2.3
X-Spam-Level: 
X-Archive-Number: 201103/28
X-Sequence-Number: 36900

This is a multi-part message in MIME format.
--------------050707090009000309090802
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 03.03.2011 09:12, daveg wrote:
> Question: what would be the consequence of simply patching out the setting
> of this flag? Assuming that the incorrect PD_ALL_VISIBLE flag is the only
> problem (big assumption perhaps) then simply never setting it would at least
> avoid the possibility of returning wrong answers, presumably at some
> performance cost. We possibly could live with that until we get a handle
> on the real cause and fix.

Yes. With that assumption.

If you really want to do that, I would suggest the attached patch 
instead. This just disables the optimization in seqscans to trust it, so 
an incorrectly set flag won't affect correctness of query results,  but 
the flag is still set as usual and you still get the warnings so that we 
can continue to debug the issue.

-- 
   Heikki Linnakangas
   EnterpriseDB   http://www.enterprisedb.com

--------------050707090009000309090802
Content-Type: text/x-diff;
 name="dont-trust-all-visible-flag.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="dont-trust-all-visible-flag.patch"

diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
index 7dcc601..d53aede 100644
--- a/src/backend/access/heap/heapam.c
+++ b/src/backend/access/heap/heapam.c
@@ -255,6 +255,11 @@ heapgetpage(HeapScanDesc scan, BlockNumber page)
 	 * transaction in the standby.
 	 */
 	all_visible = PageIsAllVisible(dp) && !snapshot->takenDuringRecovery;
+	/*
+	 * XXX: there seems to be something wrong with the way the flag is set,
+	 * so don't trust it. Remove this when the cause is found.
+	 */
+	all_visible = false;
 
 	for (lineoff = FirstOffsetNumber, lpp = PageGetItemId(dp, lineoff);
 		 lineoff <= lines;

--------------050707090009000309090802--