From owner-freebsd-scsi Mon Jun 22 16:30:54 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA21403 for freebsd-scsi-outgoing; Mon, 22 Jun 1998 16:30:54 -0700 (PDT) (envelope-from owner-freebsd-scsi@FreeBSD.ORG) Received: from nomis.simon-shapiro.org ([209.86.126.163]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id QAA21387 for ; Mon, 22 Jun 1998 16:30:46 -0700 (PDT) (envelope-from shimon@nomis.Simon-Shapiro.ORG) Received: (qmail 3855 invoked by uid 1000); 22 Jun 1998 23:31:29 -0000 Message-ID: X-Mailer: XFMail 1.3 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: Date: Mon, 22 Jun 1998 19:31:29 -0400 (EDT) Reply-To: shimon@simon-shapiro.org Organization: The Simon Shapiro Foundation From: Simon Shapiro To: Chris Parry Subject: RE: DPT support binaries - How to Setup Cc: freebsd-SCSI@FreeBSD.ORG, alex@nac.net, Tom Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 22-Jun-98 Chris Parry wrote: >> > It is when the array is in degraded mode, and you try to boot. >> >> No. Done that here several times. >> >> However, you can't boot if the array is dead (RAID5 with two failed >> drives) but that goes without saying. I've seen an interesting scenario >> where a rebuild caused a marginal drive to fail, so you always want to >> try >> at least two simulated failures before bringing a new server online. > > It was my understanding that Alex was talking about a RAID-0 array > (originally), which is the concatenated array (no redunancy). I > definitely plan to trash this machine and see how it responds before > bringing online for production work :) Ah. RAID-0 is definitely ``all or nothing'' proposition. BTW, RAID-0 is NOT concatenated. It is striped; Every N kilobytes the data shifts to another drive. This is why, in case of failure of a single drive you can get almost nothing useful out. Were it to be concatenated, only a disk-size single hole would have existed. AS it is, RAID-0 failure leaves your data looking like a shotgun target. > Anyway I'd like to bounce that architechure off of the group. First my > goal is to have a RAID-1 mirror cvs depsitory for my company and want > this > machine to be rock stable. As I understand it, I can't really take two > disks, install dos on one of them and then mirror them. So I was going > to > take two identical IDE drives, install DOS on a small partition, and then > the FreeBSD OS on the rest of it. On the second IDE controller (so they > are both masters on thier respective channels) have this disk be the > target of a daily lowlevel dd of the the first disk. Then have the DPT > controller be doing real RAID-1 on the cvs repository. IDE is not SCSI :-) What you want is doable, but you will not have automatic switchover, and you will be copying a lot of ``empty'' junk. The copy will not be efficient, as all the data will travel from the disk to userspace and back to disk. Also, note that changes occuring during the update will probably NOT end up in the mirror disk the way you like/think them to be. This can mean ``corrupted'' mirror. > Is there another way you all would recommend? Perhaps just doing > everything from floppy and then RAID-1 the scsi drives (and just forget > IDE), then install DOS and Freebsd? Yup. If you use a mirroring/RAID controller (they ARE available for IDE), then the O/S, and even the BIOS are not aware of the mirror, and the copying is done automatically. > Also, when a drive fails (or I unplug it), will it only log to > /var/log/messages? Anyone have an automated response mechanism installed > to notify of disk failure? I do, but it is not released yet. ---------- Sincerely Yours, Simon Shapiro Shimon@Simon-Shapiro.ORG Voice: 503.799.2313 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message