Subject: Re: Long Running Commits - Not Checkpoints
From: Brad Nicholson <bnichols@ca.afilias.info>
To: Greg Smith <gsmith@gregsmith.com>
Cc: pgsql-performance@postgresql.org
In-Reply-To: <1189700385.8624.32.camel@bnicholson-desktop>
References: <1189692915.8624.22.camel@bnicholson-desktop>
	<Pine.GSO.4.64.0709131149390.20144@westnet.com>
	<1189700385.8624.32.camel@bnicholson-desktop>
Content-Type: text/plain
Date: Thu, 13 Sep 2007 12:52:21 -0400
Message-Id: <1189702341.8624.38.camel@bnicholson-desktop>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit

On Thu, 2007-09-13 at 12:19 -0400, Brad Nicholson wrote:
> On Thu, 2007-09-13 at 12:12 -0400, Greg Smith wrote:
> > On Thu, 13 Sep 2007, Brad Nicholson wrote:
> 
> > I'd be curious to see how you've got your background writer configured to 
> > see if it matches situations like this I've seen in the past.  The 
> > parameters controlling the all scan are the ones you'd might consider 
> > turning down, definately the percentage and possibly the maxpages as well.
> 
> 
> bgwriter_delay = 50                     # 10-10000 milliseconds between
> rounds
> bgwriter_lru_percent = 20.0             # 0-100% of LRU buffers
> scanned/round
> bgwriter_lru_maxpages = 300             # 0-1000 buffers max
> written/round
> bgwriter_all_percent = 20               # 0-100% of all buffers
> scanned/round
> bgwriter_all_maxpages = 600             # 0-1000 buffers max
> written/round

I should add, there are 6 back ends running on this disk array
(different servers and different data partitions) with these bgwriter
settings. 

-- 
Brad Nicholson  416-673-4106
Database Administrator, Afilias Canada Corp.