thread-index: AcjLj8Zsf+1P6N1rQ6C+OJoJQQP+9w==
Content-Class: urn:content-classes:message
Importance: normal
Message-ID: <484F7613.5070202@enterprisedb.com>
Date: Wed, 11 Jun 2008 09:52:03 +0300
From: "Heikki Linnakangas" <heikki@enterprisedb.com>
Organization: EnterpriseDB
User-Agent: Mozilla-Thunderbird 2.0.0.14 (X11/20080509)
MIME-Version: 1.0
To: "Tom Lane" <tgl@sss.pgh.pa.us>
Cc: "Josh Berkus" <josh@agliodbs.com>, <pgsql-hackers@postgresql.org>,
	"Robert Treat" <xzilla@users.sourceforge.net>,
	"Peter Eisentraut" <peter_e@gmx.net>, "Greg Smith" <gsmith@gregsmith.com>,
	"Andreas Pflug" <pgadmin@pse-consulting.de>,
	"Decibel!" <decibel@decibel.org>
Subject: Re: Overhauling GUCS
References: <48409D1E.3070208@agliodbs.com>
	<200806061624.29105.xzilla@users.sourceforge.net>
	<200806100937.38124.josh@agliodbs.com>
	<200806100949.17918.josh@agliodbs.com>
	<21875.1213122184@sss.pgh.pa.us>
In-Reply-To: <21875.1213122184@sss.pgh.pa.us>
Content-Type: text/plain;
	format=flowed;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit

Tom Lane wrote:
> Josh Berkus <josh@agliodbs.com> writes:
>> Oh, and wal_buffers, the default for which we should just change if it
>> weren't for SHMMAX.
> 
> Uh, why?  On a workload of mostly small transactions, what value is
> there in lots of wal_buffers?

None. But there's also little to no harm in having a higher setting; at 
worst you waste a few megabytes of memory. Besides, most databases are 
initialized from some outside source in the beginning, and data loading 
does benefit from a higher wal_buffers setting.

Ideally, of course, there would be no wal_buffers setting, and WAL 
buffers would be allocated from shared_buffers pool on demand...

-- 
   Heikki Linnakangas
   EnterpriseDB   http://www.enterprisedb.com