Message-ID: <4DDD696C.5040306@2ndQuadrant.com>
Date: Wed, 25 May 2011 16:41:16 -0400
From: Greg Smith <greg@2ndQuadrant.com>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US;
	rv:1.9.1.16) Gecko/20110505 Icedove/3.0.11
MIME-Version: 1.0
To: pgsql-performance@postgresql.org
Subject: Re: "error with invalid page header" while vacuuming pgbench
 data
References: <20110523221603.GA31823@renesys.com>
In-Reply-To: <20110523221603.GA31823@renesys.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 05/23/2011 06:16 PM, John Rouillard wrote:
> OS: centos 5.5
> Filesystem: data - ext4 (note 4 not 3); 6.6T formatted
>              wal  - ext4; 1.5T formatted
> Raid: data - level 10, 8 disk wd2003; controller LSI MegaRAID SAS 9260-4i
>        wal  - level 1,  2 disk wd2003; controller LSI MegaRAID SAS 9260-4i
>
> Could it be an ext4 issue? It seems that ext4 may still be at the
> bleeding edge for postgres use.
>    

I would not trust ext4 on CentOS 5.5 at all.  ext4 support in 5.5 is 
labeled by RedHat as being in "Technology Preview" state.  I believe 
that if you had a real RedHat system instead of CentOS kernel, you'd 
discover it's hard to even get it installed--you need to basically say 
"yes, I know it's not for production, I want it anyway" to get preview 
packages.  It's not really intended for production use.

What I'm hearing from people is that they run into the occasional ext4 
bug with PostgreSQL, but the serious ones aren't happening very often 
now, on systems running RHEL6 or Debian Squeeze.  Those kernels are way, 
way ahead of the ext4 backport in RHEL5 based systems, and they're just 
barely stable.

-- 
Greg Smith   2ndQuadrant US    greg@2ndQuadrant.com   Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support  www.2ndQuadrant.us
"PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books