I remember seeing something about this a while ago - I think it had to
do with the geometry of the disk for the installation (when putting on a
disklabel) - I think it was something like that the geometry that was
detected was used - but if the information was altered in disklabel
incorrectly the machine would not boot afterwards and would require a
re-install. 

I am using a DEC 5000/133 with OpenBSD so I don't know how this might
relate to a Sparc machine or even if this is relevant to NetBSD - anyway
you might try the mailing list archives at www.openbsd.org for some more
info.

> -----Original Message-----
> From:	woods@most.weird.com [SMTP:woods@most.weird.com]
> Sent:	Tuesday,22 September 1998 1:10
> To:	Paul F. Wells
> Cc:	port-sparc@netbsd.org
> Subject:	Re: Is it really the disk drive or something else?
> 
> [ On , September 21, 1998 at 21:53:05 (-0400), Paul F. Wells wrote: ]
> > Subject: Is it really the disk drive or something else?
> >
> > I have installed NetBSD 1.3.2 on an SS2 with 64 meg with SUN2.1G
> > hard drive.  The problem I am experiencing is that if I power off
> > the machine and restart, the disk drive acts as though it has
> > media errors...there are CRC errors, soft read errors, etc.  The
> > system sometimes will not boot (short read) or panic immedaitely
> > after loading the kernel.
> 
> This sounds like an interface problem, not a drive problem.  It could
> be
> cables, termination, connectors, line driver ICs on the drive itself,
> line driver ICs on the sparcstation, etc.  It could even be a power
> supply problem, or a power connector or cable problem.
> 
> How *exactly* do you have the drive connected?  What, if anything, is
> connected to the other SCSI connectors on the system, and *exactly*
> how
> are any such things connected.
> 
> Does the drive exhibit any similar errors when connected to the i386
> machines?
> 
> Formatting is not likely to fix/change anything here, esp. if it
> didn't
> work after the first attempt....
> 
> Do you have any other SCSI drives you could use to experiment with on
> the sparcstation?
> 
> In general the SS2's SCSI interface is quite robust and reliable,
> though
> it's very finicky about termination and cable length.
> 
> -- 
> 							Greg A. Woods
> 
> +1 416 218-0098      VE3TCP      <gwoods@acm.org>
> <robohack!woods>
> Planix, Inc. <woods@planix.com>; Secrets of the Weird
> <woods@weird.com>