Message-ID: <492D97D6.7080506@gmx.de>
Date: Wed, 26 Nov 2008 19:39:18 +0100
From: Jan-Peter Seifert <Jan-Peter.Seifert@gmx.de>
User-Agent: Thunderbird 2.0.0.18 (Windows/20081105)
MIME-Version: 1.0
To: Rafael Domiciano <rafael.domiciano@gmail.com>
CC: "pgsql-admin@postgresql.org" <pgsql-admin@postgresql.org>
Subject: Re: Vacuum Problems
References: <3a0028490811260543x74d02719w775a457999cfc74@mail.gmail.com>	
	<492D5C87.5090505@usit.uio.no>
	<3a0028490811260700w6e8d9d86o24effbe0bea2bddb@mail.gmail.com>
In-Reply-To: <3a0028490811260700w6e8d9d86o24effbe0bea2bddb@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Rafael Domiciano wrote:

> The database has around 40 Gb.
> 
> If I not use vacuum full everyday the database get very slow.
> 
> There is no deadlock on the database.
> The vacuum does a clean in the table and in every index of the table
> "clifatura". And in the end of the vacuum, seems that vacuum is working
> hard on the table (Vacuum is using CPU and memory), but doesn't end.

Well - maybe an index on the table is broken? I did once run a script on
a database to insert a rather big amount of data. During reindexing the
db ran out of diskspace, but obviously it was freed after that so I
missed the really important error message. I just wondered why the
server had restarted in the meantime. A vacuum on the table in question
took forever. I'm not sure whether I just did a reindex or a dump and
restore to fix this. Just a suggestion. Good luck.

Peter