Received: from localhost (unknown [200.46.204.182]) by postgresql.org (Postfix) with ESMTP id D87DA9FA5DF for ; Fri, 7 Sep 2007 07:13:06 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.182]) (amavisd-maia, port 10024) with ESMTP id 84443-06-5 for ; Fri, 7 Sep 2007 07:12:43 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from vms046pub.verizon.net (vms046pub.verizon.net [206.46.252.46]) by postgresql.org (Postfix) with ESMTP id A9BB29FA6A9 for ; Fri, 7 Sep 2007 07:11:45 -0300 (ADT) Received: from osgiliath.mathom.us ([96.231.24.225]) by vms046.mailsrvcs.net (Sun Java System Messaging Server 6.2-6.01 (built Apr 3 2006)) with ESMTPA id <0JNZ00JWDTMZPHFC@vms046.mailsrvcs.net> for pgsql-performance@postgresql.org; Fri, 07 Sep 2007 05:11:23 -0500 (CDT) Received: from localhost (localhost [127.0.0.1]) by osgiliath.mathom.us (Postfix) with ESMTP id 8C18B1F816 for ; Fri, 07 Sep 2007 06:11:22 -0400 (EDT) Received: from osgiliath.mathom.us ([127.0.0.1]) by localhost (osgiliath.home.mathom.us [127.0.0.1]) (amavisd-new, port 10024) with LMTP id BHwlhHVHw7Wm for ; Fri, 07 Sep 2007 06:11:21 -0400 (EDT) Received: by osgiliath.mathom.us (Postfix, from userid 1000) id 6A3AC1F80D; Fri, 07 Sep 2007 06:11:21 -0400 (EDT) Date: Fri, 07 Sep 2007 06:11:21 -0400 From: Michael Stone Subject: Re: postgres memory management issues? In-reply-to: To: pgsql-performance@postgresql.org Mail-followup-to: pgsql-performance@postgresql.org Message-id: <20070907101120.GF1795@mathom.us> MIME-version: 1.0 Content-type: text/plain; charset=us-ascii; format=flowed Content-disposition: inline X-Virus-Scanned: Debian amavisd-new at mathom.us X-Pgp-Fingerprint: 53 FF 38 00 E7 DD 0A 9C 84 52 84 C5 EE DF 7C 88 References: User-Agent: Mutt/1.5.13 (2006-08-11) X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200709/121 X-Sequence-Number: 26666 On Thu, Sep 06, 2007 at 09:06:53AM -0700, Richard Yen wrote: >My understanding is that if any one postgres process's memory usage, >plus the shared memory, exceeds the kernel limit of 4GB, On a 32 bit system the per-process memory limit is a lot lower than 4G. If you want to use 16G effectively it's going to be a lot easier to simply use a 64bit system. That said, it's more likely that you've got a number of processes using an aggregate of more than 16G than that you're exceeding the limit per process. (Hitting the per-process limit should result in a memory allocation failure rather than an out of memory condition.) Mike Stone