Subject: Re: ERR_NOOS in stand/mbr/mbr.S
To: Brian A. Seklecki <lavalamp@spiritual-machines.org>
From: Steven M. Bellovin <smb@cs.columbia.edu>
List: port-i386
Date: 04/28/2006 10:23:57
On Fri, 28 Apr 2006 10:01:10 -0400 (EDT), "Brian A. Seklecki"
<lavalamp@spiritual-machines.org> wrote:

> On Thu, 27 Apr 2006, Steven M. Bellovin wrote:
> 
> > On Thu, 27 Apr 2006 17:06:03 -0400 (EDT), "Brian A. Seklecki"
> > <lavalamp@spiritual-machines.org> wrote:
> >
> >>
> >>
> >> The catch here is that this what it looks like as a vnd(4).  When I write
> >> image it looks different on the destination media.  vnd(4) also shows a
> >> different virtual disk geoemtry than wd(4):
> >>
> >>
> > I had similar problems trying to build a CF image for my Soekris, as
> > opposed to doing an over-the-net installation.  I eventually gave up.  My
> > tentative conclusion is that it was a geometry issue.
> 
> What we need is that nice OpenBSD fdisk(8) functionality that automaticlly 
> generates a proper MBR using that "reload" command.
> 
> I'm gonna try what Gray Watson did with the interactive fdisk(8) stuff.
> 
> What I really want to know is what am I doing differently than sysinst?! 
> >:}
> 
> 1) dd(1) /dev/zero
> 2) fdisk -i
> 3) disklabel
> 4) installboot

It may be a hardware issue, in terms of how your CF adapter sees the
geometry.  Note the "geomspec" operand to vnconfig; if you get that right,
maybe it will work.

*NOTE WELL* -- I'm guessing; I didn't actually try it.

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb