From owner-freebsd-arch Mon Jul 24 15:12:11 2000 Delivered-To: freebsd-arch@freebsd.org Received: from mail.integratus.com (miami.integratus.com [63.209.2.83]) by hub.freebsd.org (Postfix) with SMTP id 49E5337BD04 for ; Mon, 24 Jul 2000 15:12:06 -0700 (PDT) (envelope-from jar@integratus.com) Received: (qmail 25695 invoked from network); 24 Jul 2000 22:12:01 -0000 Received: from kungfu.integratus.com (HELO integratus.com) (172.20.5.168) by tortuga1.integratus.com with SMTP; 24 Jul 2000 22:12:01 -0000 Message-ID: <397CBF31.BABAB0F8@integratus.com> Date: Mon, 24 Jul 2000 15:12:01 -0700 From: Jack Rusher Organization: Integratus X-Mailer: Mozilla 4.73 [en] (X11; I; Linux 2.2.12 i386) X-Accept-Language: en MIME-Version: 1.0 To: mjacob@feral.com Cc: freebsd-arch@FreeBSD.ORG Subject: Re: SANs, disks, & devfs References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-arch@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Matthew Jacob wrote: > > It wasn't meant to be exclusionary- just additive. And the WWN and VPD > serial number are alternate methods, both possibly qualified by lun. This seems perfectly sensible. I am interested to see if there will be some method to give a common access method (excuse the pin) that solves both sets of requirements (abstract and direct). This is one of the areas in which FreeBSD can really step ahead of the other server OSs. > physical, while /dev names would be logical. I argued that /dev/sd[0..n] was > perfectly adequate in this model (rather than trying to encode position > information into the name as the /dev/[r]dsk names do). Furthermore, naming The c0d0t0s0 naming convention is a funny idea, I agree. It is sort of worst of both worlds. :-) > disks (at least) by volume name (the "Fred" volume) is even more convenient. This is a nice feature. Do we want all three things (physical, logical, and name) as potential disk ids? If so, how do you tell the tools which name space you are working in? Some sort of accesor character before the "name"? Search order? What about duplicates (i.e. someone names the disk "sd0")? > I'd much rather see some framework in FreeBSD (and NetBSD and OpenBSD.... > Linux has a devfs going...) to accomodate this. But that isn't there yet and > won't be for a while. I want to solve this for 5.0 for sure. Some of the implementation details of the JINI model are pretty hairy. It would be pretty wasteful if that stuff had to end up in every driver. Something else that I would like to see done with devfs and procfs in the future of FreeBSD is the ability to unify these namespaces over several machines. There was some really good work done at SunLabs for a thing called Solaris MC that allowed single system image over a set of machines. They used a networked version of the ddi/devfs facility and added vprocs to the proc structure to allow one machine to address the processes and devices on another. This leads to some very nice parallel computation, load balancing, and HA implications. -- Jack Rusher, Senior Engineer | mailto:jar@integratus.com Integratus, Inc. | http://www.integratus.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message