Message-ID: <4C4E13A7.2080100@2ndquadrant.com>
Date: Mon, 26 Jul 2010 19:00:55 -0400
From: Greg Smith <greg@2ndquadrant.com>
User-Agent: Thunderbird 2.0.0.24 (X11/20100317)
MIME-Version: 1.0
To: Greg Spiegelberg <gspiegelberg@gmail.com>
CC: Yeb Havinga <yebhavinga@gmail.com>, pgsql-performance@postgresql.org
Subject: Re: Testing Sandforce SSD
References: <4C4A9452.9070100@gmail.com>	<4C4B1F84.5030306@2ndquadrant.com>
	<4C4C83DF.3010703@gmail.com>	<4C4C8CCC.7080406@gmail.com>
	<alpine.DEB.2.00.1007261045470.20471@aragorn.flymine.org>
	<4C4D6396.5020707@gmail.com>	<4C4DAAE8.3060501@2ndquadrant.com>
	<4C4DE50B.1060102@gmail.com>	<4C4DE5C9.8090004@2ndquadrant.com>
	<AANLkTimwAqzvWvSfscqDKJDG3Yod=6KQ6e0Xo+c+k6s=@mail.gmail.com>
In-Reply-To: <AANLkTimwAqzvWvSfscqDKJDG3Yod=6KQ6e0Xo+c+k6s=@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Greg Spiegelberg wrote:
> I know I'm talking development now but is there a case for a pg_xlog 
> block device to remove the file system overhead and guaranteeing your 
> data is written sequentially every time?

It's possible to set the PostgreSQL wal_sync_method parameter in the 
database to open_datasync or open_sync, and if you have an operating 
system that supports direct writes it will use those and bypass things 
like the OS write cache.  That's close to what you're suggesting, 
supposedly portable, and it does show some significant benefit when it's 
properly supported.  Problem has been, the synchronous writing code on 
Linux in particular hasn't ever worked right against ext3, and the 
PostgreSQL code doesn't make the right call at all on Solaris.  So 
there's two popular platforms that it just plain doesn't work on, even 
though it should.

We've gotten reports that there are bleeding edge Linux kernel and 
library versions available now that finally fix that issue, and that 
PostgreSQL automatically takes advantage of them when it's compiled on 
one of them.  But I'm not aware of any distribution that makes this easy 
to try out that's available yet, paint is still wet on the code I think.

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