Message-ID: <4E790E13.8060308@catalyst.net.nz>
Date: Wed, 21 Sep 2011 10:05:07 +1200
From: Mark Kirkwood <mark.kirkwood@catalyst.net.nz>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US;
	rv:1.9.2.21) Gecko/20110831 Lightning/1.0b2 Thunderbird/3.1.13
MIME-Version: 1.0
To: Josh Berkus <josh@agliodbs.com>
CC: pgsql-performance@postgresql.org
Subject: Re: : Performance Improvement Strategy
References: 
 <CAFrxt0j5UysgEU8hQxZ_mGbdex8Ev-bGdGsdpOGYYSO-jO24Fw@mail.gmail.com>
	<4E78D6EC.30205@agliodbs.com>
In-Reply-To: <4E78D6EC.30205@agliodbs.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 21/09/11 06:09, Josh Berkus wrote:
> Venkat,
>
>> I see lot of free spaces or free pages in Tables and Indexes. But, I need to
>> give an exact calculation on how much space will be reclaimed after VACUUM
>> FULL and RE-INDEXING.
> At present, there is no way to calculate this precisely.  You can only
> estimate, and estimates have significant error factors.  The query which
> Marcin linked for you, for example, can be as much as 500% off (although
> usually only 50% off).
>

If you have autovacuum on (which should be typical thee days), then 
using the freespacemap contrib module should give very accurate results:

SELECT oid::regclass,
                pg_relation_size(oid)/(1024*1024) AS mb,
                sum(free)/(1024*1024) AS free_mb
         FROM
             (SELECT oid, (pg_freespace(oid)).avail AS free
              FROM pg_class) AS a
         GROUP BY a.oid ORDER BY free_mb DESC;


regards

Mark