Date: Thu, 05 Jun 2008 14:47:16 -0500
From: Robert Lor <Robert.Lor@Sun.COM>
Subject: Re: Overhauling GUCS
In-reply-to: <7473.1212625397@sss.pgh.pa.us>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Andrew Dunstan <andrew@dunslane.net>, Greg Smith <gsmith@gregsmith.com>,
	Andreas Pflug <pgadmin@pse-consulting.de>,
	Decibel! <decibel@decibel.org>, Peter Eisentraut <peter_e@gmx.net>,
	pgsql-hackers@postgresql.org, Josh Berkus <josh@agliodbs.com>
Message-id: <484842C4.4070801@sun.com>
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=ISO-8859-1
Content-transfer-encoding: 7BIT
References: <48409D1E.3070208@agliodbs.com>
	<200806010032.53337.peter_e@gmx.net>
	<Pine.GSO.4.64.0806020452220.26912@westnet.com>
	<17932.1212415926@sss.pgh.pa.us> <20080604131810.GH65238@decibel.org>
	<4846A483.8000001@pse-consulting.de>
	<Pine.GSO.4.64.0806041045520.2258@westnet.com>
	<29525.1212613052@sss.pgh.pa.us> <48472E47.7030701@dunslane.net>
	<7473.1212625397@sss.pgh.pa.us>
User-Agent: Thunderbird 2.0.0.6 (X11/20071119)

Tom Lane wrote:
> If those aren't enough questions, what else must we ask?  Or maybe they
> aren't the right questions at all --- maybe we should ask "is this a
> dedicated machine or not" and try to extrapolate everything else from
> what we (hopefully) can find out about the hardware.
>   
I think probably a combination of  high and low-level questions and make 
the low-level (more specific) questions optional since some users may 
not be able to provide low-level info.  Here's a rough idea of how I 
envision this tool should work.

$ pg_config_wizard

Is this machine dedicated to Postgres? (y/n) n   (now tool 
auto-discovers available HW resources)
Your system has 32GB memory. What percentage do you want to allocate to 
Postgres? (1=50%, 2=33%, 3=25%, etc) 1
What type of workload? (OLTP, DSS, etc)
...

At the end, the tool runs for a while checking to see if certain 
thresholds are reached to determine which parameters need to be 
increased.  The tool would change the parameters causing the 
bottlenecks, rerun Postgres/workload, and iterate a few times until the 
results are satfisfactory. Write the recommended settings to 
postgresql.conf.recommend and let the user update postgresql.conf 
himself or whatever.

I update my postgresql.conf, rerun the app and see 100% increased in 
throughput :-)

-Robert