From owner-freebsd-current Thu Mar 4 3: 8:44 1999 Delivered-To: freebsd-current@freebsd.org Received: from ns.oeno.com (ns.oeno.com [194.100.99.145]) by hub.freebsd.org (Postfix) with SMTP id 8F47514E85 for ; Thu, 4 Mar 1999 03:08:34 -0800 (PST) (envelope-from will@ns.oeno.com) Received: (qmail 23583 invoked by uid 1001); 4 Mar 1999 11:08:16 -0000 To: Sheldon Hearn Cc: current@freebsd.org Subject: Re: ATAPI and ATAPI_STATIC with the new ATA* driver? References: <79192.920464790@axl.noc.iafrica.com.newsgate.clinet.fi> From: Ville-Pertti Keinonen Date: 04 Mar 1999 13:05:39 +0200 In-Reply-To: Sheldon Hearn's message of "3 Mar 1999 15:12:04 +0200" Message-ID: <86iuchms58.fsf@not.oeno.com> Lines: 38 X-Mailer: Gnus v5.5/XEmacs 20.4 - "Emerald" Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Sheldon Hearn writes: > I'm not sure I understand what real-world frustrations people are having > here. Is this thread the product of reactionary criticism, or are there > real examples of situations in which there are serious disadvantages to > the way Soren has things working? So far, you seem to have received replies arguing for "convenience", which isn't the only valid reason for static device numbering. When a boot fails with something like /dev/rwd2s1g: Device not configured, how long will it take you to figure out that it was actually wd1 that didn't probe and what is now visible as wd1 is what you used to know as wd2? If the disks are identical, you have to look carefully at the boot messages. If that isn't a POLA violation, what is? What about if your /etc/fstab only checks/mounts partitions on wd0/wd1, and /etc/rc doesn't fail because of the missing wd2? Instead, you have partitions in unexpected places... Oh and of course if you aren't pedantic about partitioning conventions and have a non-'b' swap partition or a 'b'-partition used as a filesystem, you might be swapping onto a filesystem...even if the fsck fails and drops you into single user mode. Disks can fail to probe, even when they aren't permanently broken, e.g. because of failure to spin up. Fixed numbering isn't merely a convenience for those who add and remove devices routinely. Changing the default behavior from a safer alternative to a more dangerous one might not be a good thing. Of course dangerous numbering is the default for SCSI devices, as well... Perhaps it's reasonable to expect people who don't know what they're doing to only have one or two disks. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message