Message-ID: <4CC078DD.3010107@2ndquadrant.com>
Date: Thu, 21 Oct 2010 13:31:09 -0400
From: Greg Smith <greg@2ndquadrant.com>
User-Agent: Thunderbird 2.0.0.24 (X11/20100317)
MIME-Version: 1.0
To: Kevin Grittner <Kevin.Grittner@wicourts.gov>
CC: Bruce Momjian <bruce@momjian.us>, jd@commandprompt.com,
	Scott Marlowe <scott.marlowe@gmail.com>,
	Steve Crawford <scrawford@pinpointresearch.com>,
	pgsql-performance@postgresql.org, Ben Chobot <bench@silentmedia.com>
Subject: Re: BBU Cache vs. spindles
References: <4CC02A520200002500036C14@gw.wicourts.gov>
	<201010211659.o9LGxkV01781@momjian.us>
	<4CC02ED30200002500036C24@gw.wicourts.gov>
In-Reply-To: <4CC02ED30200002500036C24@gw.wicourts.gov>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Kevin Grittner wrote:
> Bruce Momjian <bruce@momjian.us> wrote:
>  
>   
>> full_page_writes is designed to guard against a partial write to a
>> device.  I don't think the raid cache can be partially written to
>>     
>  
> So you're confident that an 8kB write to the controller will not be
> done as a series of smaller atomic writes by the OS file system?

Sure, that happens.  But if the BBU has gotten an fsync call after the 
8K write, it shouldn't return success until after all 8K are in its 
cache.  That's why full_page_writes should be safe on a system with BBU 
as Bruce is suggesting.  But I'd like to see some independent proof of 
that fact, that includes some targeted tests users can run, before we 
start recommending that practice.

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