On Fri, 30 Jul 1999, Daniel R. Killoran,Ph.D. wrote:

> It may indeed be benign; I asked about it because I have not been able to
> boot, & thought it might be a clue.
> 
> I should perhaps have explained that this is a newly created system, from
> the bare hard drive, and I have not yet been able to get it to boot. I get
> the message:
> 
> Could not open kernel "(6, ?) netbsd"
> 
> If the "6" means the SCSI number, that is correct. I don't know how to fill
> in where it has the "?", so I have been looking around for anything that
> might be suspicious.

If you can cpout the kernel, you can rule out the file system by
booting from a MacOS file using the booter option, and running fsck.
You could also do that with a gzipped kernel, but not with a tarred
gzipped kernel. (That's why you should try to cpout.)

There is a known problem where the Booter gets confused if the kernel
is not using a low numbered inode in the first cylinder. If that's the
problem, and if you can boot from a MacOS file kernel, you might be
able to fix this with

	fsck -f
	mount -a
	cp -p netbsd netbsd~
	rm netbsd; mv netbsd~ netbsd

or some such thing. The idea is to bring down the kernels inode number
as viewed with "ls -l". While you're there, you could set rc.conf to
boot multi-user, as explained in the installation instructions. With
any luck, you could "shutdown -r now", set the Booter back to boot
from a NetBSD partition, and it will boot.

> Since I had trouble with too many partitions last time I created a system,
> this time I went to a lot of trouble to eliminate unneeded ones. The
> partitions are:

Scott Reynolds re-wrote the disklabel code to fix this for 1.4. Any
more than 6 or 7 should should simply be ignored.