From owner-freebsd-scsi Mon Sep 7 08:04:22 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA11705 for freebsd-scsi-outgoing; Mon, 7 Sep 1998 08:04:22 -0700 (PDT) (envelope-from owner-freebsd-scsi@FreeBSD.ORG) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA11641 for ; Mon, 7 Sep 1998 08:03:55 -0700 (PDT) (envelope-from gibbs@narnia.plutotech.com) Received: (from gibbs@localhost) by narnia.plutotech.com (8.8.8/8.7.3) id IAA29563; Mon, 7 Sep 1998 08:57:39 -0600 (MDT) Date: Mon, 7 Sep 1998 08:57:39 -0600 (MDT) From: "Justin T. Gibbs" Message-Id: <199809071457.IAA29563@narnia.plutotech.com> To: Chris Csanady cc: scsi@FreeBSD.ORG Subject: Re: CAM device naming.. Newsgroups: pluto.freebsd.scsi In-Reply-To: <199809071416.JAA06605@friley-185-114.res.iastate.edu> User-Agent: tin/pre-1.4-971204 (UNIX) (FreeBSD/3.0-CURRENT (i386)) Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In article <199809071416.JAA06605@friley-185-114.res.iastate.edu> you wrote: > >> >> >>On Sun, 6 Sep 1998, Chris Csanady wrote: >> >>> As the device naming is different in CAM anyways, would there be any >>> possibility of taking it one step further? I have found it to be >>> quite nice in other OS's that actually use subdirectories for disk >>> names, terminals, etc. It makes /dev a much cleaner place imho. >>> >>> Also, I would love for the names to actually be based upon the >>> scsi controller, id, lun, partition etc.. although, I suppose there >>> is no way in hell this will happen. :P It sucks when a disk goes >>> away, and then every last entry in your fstab is incorrect though. >> >>In a multipathing environment this is not right. It has no meaning >>on fibre fabrics. > > I'm sorry, but I don't follow here. I was under the impression that > the cam layer was intended for scsi.. SCSI is one of the supported protocols across Fibre-channel. CAM can handle other protocols than SCSI although the current implementation only supports SCSI. >>I really strongly suggest that leaving the names the way they >>are right now is fine until a suitable volume binder exists. > > The thing I don't like about the current scheme is that a particular > device name, in no way corresponds to a particular device. The names can be bound to particular devices by hardwiring devices in your kernel config file. > > Chris -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message