Message-ID: <4FDA002E.2080605@elitecore.com>
Date: Thu, 14 Jun 2012 20:45:58 +0530
From: Siddharth Shah <siddharth.shah@elitecore.com>
User-Agent: Mozilla/5.0 (X11; Linux i686;
	rv:11.0) Gecko/20120327 Thunderbird/11.0.1
MIME-Version: 1.0
To: pgsql-performance@postgresql.org
Subject: High CPU Usage
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hello,

I have around 1000 schema in database, Each schema having similar data 
structure with different data
Each schema has few tables which never updates (Read only table) and 
other tables rewrites almost everyday so I prefer to TRUNCATE those 
tables and restores with new data

Now facing issue on high CPU & IO on database primarily of Stats 
Collector & Vacuuming, size of statfile is almost 28MB and when I 
manually vacuum analyze complete database it takes almost 90 minutes 
though auto vacuum is configured

Restoring dump on each schema may minor data variations
Executing SQL statements on schema are few , Affecting less than 50 
touple / day

My Questions :

Increasing Maintainace_Work_Mem improves auto / manual vacuum 
performance ? If it improves will it require more IO / CPU resource ?
If I stops Stats Collector process & auto vaccuming & Execute manual 
vaccum based on schema restoration with major change what performance 
parameter I need to consider ? (Restoring data has vary few changes)
Is Vacuuming & Stats required here for Metadata for improving 
performance ? (Table structures remain same)

Any more on this which can help to reduce IO without affecting major 
performance

regards,
Siddharth