Date: Mon, 10 Apr 2006 14:23:07 -0400 (EDT) From: Jerry McAllister <jerrymc@clunix.cl.msu.edu> To: petermatulis@yahoo.ca (Peter) Cc: freebsd-questions <freebsd-questions@freebsd.org> Subject: Re: seeking help on "adding a disk" Message-ID: <200604101823.k3AIN7EP016528@clunix.cl.msu.edu> In-Reply-To: <20060410153524.25115.qmail@web60024.mail.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>
>
> --- Jerry McAllister <jerrymc@clunix.cl.msu.edu> wrote:
>
> > > Anyway, when I try to create a filesystem with newfs I get an
> > error:
> > >
> > > # newfs /dev/ad2s1a
> > > ..., 450493504, 450869856, 451246208,newfs: wtfs: 65536 bytes at
> > sector
> > > 451622560: Input/output error
> > >
> > > Now please let me know if my method is sound. I feel the drive
> > size is
> > > not being properly recognized and that
> > > the last command is trying to write past the edge of the disk. I
> > would
> > > also like to not be using s1a but s1e instead.
> > >
> > > Furthermore, the docs [1] for this drive say that an 80-wire cable
> > is
> > > required. I didn't have one handy so I had to
> > > use a 40-wire cable. Could this be causing the trouble?
> > >
> > > Thanks for any insights (I have a hell of a time working with disks
> > on
> > > FreeBSD).
> >
> > It looks like things worked up to the bsdlabel which didn't do
> > anything.
> > Namely, after doing the fdisk, even though you saw some errors,
> > you seem to have gotten a single slice with everything in it just
> > as you wanted.
> >
> > > The data for partition 1 is:
> > > sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
> > > start 63, size 586114641 (286188 Meg), flag 80 (active)
> > > beg: cyl 0/ head 1/ sector 1;
> > > end: cyl 854/ head 15/ sector 63
> > > The data for partition 2 is:
> >
> > I usually do fdisk -BI -v -b /boot/mbr ad2
> > but, I believe the -b /boot/mbr is default with you use -B
> > and I think it will take /dev/ad2 as well as just ad2. Actually I
> > am usually working on SCSI drives so it is da2, etc, but that
> > shouldn't
> > be any different.
> >
> > To answer your question on that, It looks like the data in the
> > bsdlabel read you did is probably correct. It shows values for size
> and
> > offset although the offset seems weird. I would expect it to be 0.
> Maybe
> > that is because you didn't do the first bsdlabel step.
>
>
> Which first step?
The first of the two bsdlabel steps - eg:
bsdlabel -w
>
> > For the disklabel you need two steps - noting that you were writing
> > the label to make it bootable, you need:
> >
> > bsdlabel -w -B ad2s1 auto
> > and then
> > bsdlabel -e ad2s1
> >
> > The first step (which you appear to have skipped) makes the base
> > label on the slice and the second edits it to be the way you want.
>
>
> I cannot find the 'auto' option in the bsdlabel man page.
Yes, looks like it has disappeared now. It used to be there in disklabel.
So, don't worry about the auto, but do use the -w
>
> Also, I do not need the drive to be bootable. I redid the procedure
> like this:
>
> # dd if=/dev/zero of=/dev/ad2 bs=1k count=64
> # fdisk -I ad2 { same "error" msg as before }
> # bsdlabel -w ad2s1
> # bsdlabel -e ad2s1
OK. If you don't want it to be bootable, then leave out those parameters.
You had used them in your earlier post so I continued them.
fdisk complains about geometry, but then it does things the way it
wants so I generally just ignore it. Some people get all worked
up, but on modern disks, disk geometry from the external point of
view is virtual (fiction).
>
> ------------------------
> # /dev/ad2s1:
> 8 partitions:
> # size offset fstype [fsize bsize bps/cpg]
> e: 586114625 16 unused 2048 16384 89
> c: 586114641 0 unused 0 0 # "raw" part,
> don't edit
> ------------------------
>
>
> # newfs -U /dev/ad2s1e
Newfs used to require the raw device name, as in
newfs -U /dev/rad2s1e
but I see the man page doesn't show that in its example now, so
maybe it no long does. Try it once and see. It can't hurt as long
as it is a new disk and you haven't used it for anything yet.
>
> > > e: * * 4.2BSD 2048 16384 89
> >
> > You may actually want bigger fsize, bsize and some different bps/cpg
> > I sometimes set bps/cpg to 64 on larger disks and 16 on smaller, but
> > I usually let it do whatever bps/cpg that it wants, but often specify
> > the fsize and bsize. bsize should be 8X fsize.
>
> How do you explain this on another (5.4) system of mine:
Explain what? It looks reasonably normal.
Unless you mean the 0-s in fsize, bsize and bps/cpg. I think that may
mean it is doing its own thing - using defaults.
But, I don't know. Those fields have something in them on the machines I
have handy to look at. But, then, I probably put something there when I
bsdlabeled them.
> # bsdlabel ad0s3
>
> # /dev/ad0s3:
> 8 partitions:
> # size offset fstype [fsize bsize bps/cpg]
> a: 524288 0 4.2BSD 0 0 0
> b: 4142736 524288 swap
> c: 24900750 0 unused 0 0 # "raw" part, don't edit
> d: 524288 4667024 4.2BSD 0 0 0
> e: 524288 5191312 4.2BSD 0 0 0
> f: 19185150 5715600 4.2BSD 0 0 0
>
> > The newfs should then work after the bsdlabel is fixed up.
>
> Nope. Same error. Retries of newfs causes crashes at random sectors.
> I am guessing that the 40-wire cable is causing poor signaling. I also
> tried changing the offset of 'e' to 0 and modifying the 'size' so that
> it matches 'c'.
Could be. If using rad2s1e doesn't help, then maybe that is your problem.
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200604101823.k3AIN7EP016528>