From owner-freebsd-questions@FreeBSD.ORG Mon Apr 3 13:36:45 2006 Return-Path: X-Original-To: freebsd-questions@freebsd.org Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4290916A425 for ; Mon, 3 Apr 2006 13:36:45 +0000 (UTC) (envelope-from freebsd@meijome.net) Received: from sigma.octantis.com.au (ns2.octantis.com.au [207.44.189.124]) by mx1.FreeBSD.org (Postfix) with ESMTP id C0D6E43D5F for ; Mon, 3 Apr 2006 13:36:44 +0000 (GMT) (envelope-from freebsd@meijome.net) Received: (qmail 2178 invoked from network); 3 Apr 2006 23:36:44 +1000 Received: from 210-84-32-140.dyn.iinet.net.au (HELO localhost) (210.84.32.140) by creativesilk.com.au with (DHE-RSA-AES256-SHA encrypted) SMTP; 3 Apr 2006 23:36:43 +1000 Date: Mon, 3 Apr 2006 23:36:39 +1000 From: Norberto Meijome To: usleepless@gmail.com Message-ID: <20060403233639.21a65930@localhost> In-Reply-To: References: <442B2FC6.9040001@123.com.sv> <20060403225821.45fcc26e@localhost> X-Mailer: Sylpheed-Claws 2.0.0 (GTK+ 2.8.16; i386-portbld-freebsd6.1) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: Miguel , freebsd-questions@freebsd.org Subject: Re: terrible performance in 6.1beta4 X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 03 Apr 2006 13:36:45 -0000 On Mon, 3 Apr 2006 15:18:02 +0200 usleepless@gmail.com wrote: > Beto, > > i believe he has given up. > :( without knowing WHAT is the problem? > i would say that on a dedicated database machine 50-75% > dedicated(shared) memory is not overkill. of course not. I am not sure I read his vmstat screenshot right, but it seemed to me that the box was using a lot of disk, but had lots of free memory - i.e., it was probably just reading the data and the disks maxed out? Miguel, what does gstat show ? (it'll show the % busy of each device in GEOM , which also includes non-RAID devices, so it's quite useful) > > wasn't there some other switch which determined wether to page > shared-memory out? i dont actually recall - my DB project had to be put on hold about 75% ready for prod, trying to get back onto it soon.