Message-ID: <4CC07291.8020601@2ndquadrant.com>
Date: Thu, 21 Oct 2010 13:04:17 -0400
From: Greg Smith <greg@2ndquadrant.com>
User-Agent: Thunderbird 2.0.0.24 (X11/20100317)
MIME-Version: 1.0
To: Bruce Momjian <bruce@momjian.us>
CC: Scott Marlowe <scott.marlowe@gmail.com>, jd@commandprompt.com,
	Ben Chobot <bench@silentmedia.com>,
	Steve Crawford <scrawford@pinpointresearch.com>,
	pgsql-performance@postgresql.org
Subject: Re: BBU Cache vs. spindles
References: <201010211651.o9LGp9I28474@momjian.us>
In-Reply-To: <201010211651.o9LGp9I28474@momjian.us>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Bruce Momjian wrote:
> With a BBU you can turn off full_page_writes, which should decrease the
> WAL traffic.
>
> However, I don't see this mentioned in our documentation.  Should I add
> it?
>   

What I would like to do is beef up the documentation with some concrete 
examples of how to figure out if your cache and associated write path 
are working reliably or not.  It should be possible to include "does 
this handle full page writes correctly?" in that test suite.  Until we 
have something like that, I'm concerned that bugs in filesystem or 
controller handling may make full_page_writes unsafe even with a BBU, 
and we'd have no way for people to tell if that's true or not.

-- 
Greg Smith, 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD
PostgreSQL Training, Services and Support  www.2ndQuadrant.us