Subject: Re: [Q] DS3100 installation? (trying something new)
To: None <bickford@jimmy.harvard.edu>
From: None <Havard.Eidnes@runit.sintef.no>
List: port-pmax
Date: 10/17/1997 11:40:31
> I assumed that I should try a command to the effect of:
>
> 	dd if=3D32M-diskimage of=3D/dev/sd1c
>
> However, when I try that (or "/dev/rsd1c"), I get the following error:
> 	"device not configured"

Probably due to missing label, as you probably guessed.

> So, I re-labeled the disk so that the "a" partition spans the entire disk,
> removed all the other partitions, and lastly "newfs /dev/sd1a".

newfs should not be required.

> Then, I:
> 	dd if=3D32M-diskimage of=3D/dev/sd1a
>
> and get no errors.

I'd use the raw partition instead: /dev/rsd1a in this case.  That
way you won't have the buffer cache interposed between your
writes and your disk.

> >* Boot it from prom typing
> >  boot -f rz(0,X,0)netbsd
> >  where X is the SCSI ID of your disk.
>
> boot -f rz(0,0,0)netbsd
>
> which returns the error:
> ?417 btblk fmt
> ?470 bt err: rz(0,0,0)netbsd
>
> What now?

Compare what's on the first parts of the raw disk with what's in
the image file, if possible.

Reconnect the disk (with ID 1 I presume) to the NetBSD/i386
machine, and boot up.  If the NetBSD/i386 machine says the label
on sd1 is bogus (as there now ought to be a pmax disklabel on the
disk, which probably won't be recognized as valid by the i386),
NetBSD on the i386 will probably make up a "dummy" in-core disk
label where partition 'd' covers the entire disk.  You can check
this by doing "disklabel sd1" on the i386 (note: no '-r' option),
and if I'm right you can get the data from the first part of the
disk off the platters by using a suitable combination of dd and
/dev/rsd1d.  If it doesn't compare as equal, try dd'ing to
/dev/rsd1d as mentioned above, and try again.

Regards,

- H=E5vard