Skip site navigation (1) Skip section navigation (2)

Re: oom_killer

From: Claudio Freire <klaussfreire(at)gmail(dot)com>
To: Cédric Villemain <cedric(dot)villemain(dot)debian(at)gmail(dot)com>
Cc: Tory M Blue <tmblue(at)gmail(dot)com>, Merlin Moncure <mmoncure(at)gmail(dot)com>, pgsql-performance(at)postgresql(dot)org
Subject: Re: oom_killer
Date: 2011-04-22 23:19:19
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgsql-performance
On Fri, Apr 22, 2011 at 6:45 PM, Cédric Villemain
<cedric(dot)villemain(dot)debian(at)gmail(dot)com> wrote:
> Are you sure it is a PAE kernel ? You look limited to 4GB.

If my memory/knowledge serves me right, PAE doesn't remove that limit.
PAE allows more processes, and they can use more memory together, but
one process alone has to live within an addressable range, and that is
still 4GB, mandated by the 32-bit addressable space when operating in
linear addressing mode.

But linux kernels usually reserve 1GB for kernel stuff (buffers and
that kind of stuff), so the addressable portion for processes is 3GB.

Take away 2.5GB of shared buffers, and you only leave 0.5G for general
data and code.

Really, lowering shared_buffers will probably be a solution. Moving to
64 bits would be a better one.

In response to

pgsql-performance by date

Next:From: Greg SmithDate: 2011-04-23 02:48:04
Subject: Re: OT (slightly) testing for data loss on an SSD drive due to power failure
Previous:From: Marinos YannikosDate: 2011-04-22 21:43:58
Subject: Re: postgresql random io test with 2 SSD Kingston V+100 500GB in (software) Raid1

Privacy Policy | About PostgreSQL
Copyright © 1996-2017 The PostgreSQL Global Development Group