Date: Thu, 14 Jul 2005 11:47:51 -0400
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Dan Harris <fbsd@drivefaster.net>
Cc: pgsql-performance@postgresql.org
Subject: Re: Quad Opteron stuck in the mud
Message-ID: <20050714154750.GC19232@alvh.no-ip.org>
References: <E5E21B7B-FFE0-4D24-8CDD-62179AEBBB92@drivefaster.net>
	<87oe95ykz5.fsf@stark.xeocode.com>
	<F755CCAD-55D9-4E8F-BBB4-2FE881B0CD3B@drivefaster.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <F755CCAD-55D9-4E8F-BBB4-2FE881B0CD3B@drivefaster.net>
User-Agent: Mutt/1.5.9i

On Thu, Jul 14, 2005 at 12:28:05AM -0600, Dan Harris wrote:

> Ok, so I remounted this drive as ext2 shortly before sending my first  
> email today.  It wasn't enough time for me to notice the ABSOLUTELY  
> HUGE difference in performance change.  Ext3 must really be crappy  
> for postgres, or at least is on this box.  Now that it's ext2, this  
> thing is flying like never before.   My CPU utilization has  
> skyrocketed, telling me that the disk IO was constraining it immensely.

Were you using the default journal settings for ext3?

An interesting experiment would be to use the other journal options
(particularly data=writeback).  From the mount manpage:

       data=journal / data=ordered / data=writeback
              Specifies  the  journalling  mode  for  file  data.  Metadata is
              always journaled.  To use modes other than ordered on  the  root
              file system, pass the mode to the kernel as boot parameter, e.g.
              rootflags=data=journal.

              journal
                     All data is committed into the  journal  prior  to  being
                     written into the main file system.

              ordered
                     This  is  the  default mode.  All data is forced directly
                     out to the main file system prior to its  metadata  being
                     committed to the journal.

              writeback
                     Data ordering is not preserved - data may be written into
                     the main file system after its metadata has been  commit-
                     ted  to the journal.  This is rumoured to be the highest-
                     throughput option.  It guarantees  internal  file  system
                     integrity,  however  it  can  allow old data to appear in
                     files after a crash and journal recovery.


-- 
Alvaro Herrera (<alvherre[a]alvh.no-ip.org>)
Officer Krupke, what are we to do?
Gee, officer Krupke, Krup you! (West Side Story, "Gee, Officer Krupke")