From owner-freebsd-current@FreeBSD.ORG Fri Feb 27 05:58:38 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4E8B916A4CE for ; Fri, 27 Feb 2004 05:58:38 -0800 (PST) Received: from clueful.shagged.org (clueful.shagged.org [212.13.201.101]) by mx1.FreeBSD.org (Postfix) with ESMTP id F36DA43D2F for ; Fri, 27 Feb 2004 05:58:37 -0800 (PST) (envelope-from chris@clueful.shagged.org) Received: from chris by clueful.shagged.org with local (Exim 4.30; FreeBSD) id 1AwiVh-000Dhv-0k; Fri, 27 Feb 2004 13:58:37 +0000 Date: Fri, 27 Feb 2004 13:58:37 +0000 From: Chris Elsworth To: Dag-Erling Sm?rgrav Message-ID: <20040227135836.GA52534@shagged.org> References: <20040227101410.GA45369@shagged.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.6i Sender: Chris Elsworth X-SA-Exim-Mail-From: chris@clueful.shagged.org X-SA-Exim-Scanned: No; SAEximRunCond expanded to false cc: freebsd-current@freebsd.org Subject: Re: mdconfig unable to allocate memory X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 27 Feb 2004 13:58:38 -0000 On Fri, Feb 27, 2004 at 02:31:37PM +0100, Dag-Erling Sm?rgrav wrote: > Chris Elsworth writes: > > this fails: > > cluestick# mdconfig -a -t malloc -o reserve -s 1g > > mdconfig: ioctl(/dev/mdctl): Cannot allocate memory > > it tries to allocate 1 GB of non-pageable kernel memory. Use -t swap > instead of -t malloc. Hello, Thanks for the reply. Is creating 1GB of non-pageable memory not possible then? Even if I explicitly give the kernel far more than that to play with? I've recently been trying to fiddle with KVA_PAGES etc on 5.2.1 but not really getting very far. It seems to be a bit of a black art. An out of the box 5.2.1-RELEASE on a machine with 4GB of memory is giving kmem_map panic errors from "vinum start" on a clean boot. Some Googling suggests raising KVA_PAGES and VM_KMEM_SIZE_MAX. So I set those to 512 and (1024*1024*1024) respectively, and that's killed the machine; it hasn't come back from a remote reboot, but I suspect init is dying and forcing a reboot loop. I'll be able to check that in a few hours when I get home. So, in short, is there any rule of thumb to follow when tweaking these settings? It's a bit worrying that an out of the box 5.2.1 managed to crash at all. -- Chris