DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com;
	h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:References:Message-ID:Date:From:Reply-To:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=ZQB3dZogvmV/RIpno5XRrpW9i5F2urXkw4Udn+TlLZjv5NHLxGoLLcTD/Iidz5RzuHYXIGcYa/FaUxuBm6HIGkNj9lxeZYrHDMAR2n1uBU1j6nbCS+hSmIgeC9OytV03ww1LIc3qfW7l9TmRuXtztDDJfSx6Wj1cAkoLL6iNw58=;
References: 
 <CBAC86BE623FDB4E8B6225471691724291E1BC19@EXCHMBX-ADL6-01.staff.internode.com.au>
Message-ID: <1353368441.76758.YahooMailNeo@web125903.mail.ne1.yahoo.com>
Date: Mon, 19 Nov 2012 15:40:41 -0800 (PST)
From: Greg Williamson <gwilliamson39@yahoo.com>
Reply-To: Greg Williamson <gwilliamson39@yahoo.com>
Subject: Re: Resolving Index Bloat
To: "pgsql-admin@postgresql.org" <pgsql-admin@postgresql.org>
In-Reply-To: 
 <CBAC86BE623FDB4E8B6225471691724291E1BC19@EXCHMBX-ADL6-01.staff.internode.com.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Samuel --=0A=0A=0A=0A>________________________________=0A> From: Samuel Ste=
arns <SStearns@internode.com.au>=0A>To: "pgsql-admin@postgresql.org" <pgsql=
-admin@postgresql.org> =0A>Sent: Monday, November 19, 2012 1:59 PM=0A>Subje=
ct: [ADMIN] Resolving Index Bloat=0A> =0A>=0A><...>=0A>=0A>Any ideas on how=
 to resolve?=0A>=A0=0A=0A=0AI have custody of one database that is fairly s=
mall but subjected to lots of updated, inserts and deletes, and that databa=
se shows bloat that we have to knock down hourly or performance tanks.=0A=
=0AThis is on postgres 9.1 so I am not sure how much translates to earlier =
versions.=0A=0AWe ran ""REINDEX" hourly for a while; we currently use a han=
d rolled script that reindexes each regular index and does a dance to reind=
ex primary keys and constraints. I could email it to you but as I said, not=
 sure how much would work.=0A=0AThe straight REINDEX did work 99% of the ti=
me, with very occasional reports of failures, never repeated. The hand roll=
ed version is a bit safer and a little less intrusive.=0A=0AHTH,=0A=0AGreg =
Williamson