Message-ID: <4B692A43.1070007@2ndquadrant.com>
Date: Wed, 03 Feb 2010 02:48:19 -0500
From: Greg Smith <greg@2ndquadrant.com>
User-Agent: Thunderbird 2.0.0.23 (X11/20090817)
MIME-Version: 1.0
To: Andy Colson <andy@squeakycode.net>
CC: Rob <rclemley@yahoo.com>, pgsql-performance@postgresql.org
Subject: Re: System overload / context switching / oom, 8.3
References: <4B6878DF.3000004@yahoo.com> <4B688F76.9010002@squeakycode.net>
In-Reply-To: <4B688F76.9010002@squeakycode.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Andy Colson wrote:
> The oom-killer is kicking in, at some point, so somebody is using too 
> much ram.  There should be messages or logs or something, right?  
> (I've never enabled the oom stuff so dont know much about it).  But 
> the log messages might be helpful.

They probably won't be.  The information logged about what the OOM 
killer decided to kill is rarely sufficient to tell anything interesting 
about the true cause in a PostgreSQL context--only really helpful if 
you've got some memory hog process it decided to kill.  In this case, 
seems to be a simple situation:  way too many connections for the 
work_mem setting used for a 8GB server to support.  I'd take a look at 
the system using "top -c" as well, in good times and bad if possible, 
just to see if any weird memory use is showing up somewhere, perhaps 
even outside the database.

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