First, you should be specifying `sd1*' rather than `sd1b' if
you're booting the miniroot.  (Don't ask me why....)


	   On Tue, 31 Oct 2000, matthew green wrote:
	   % what this is is sysinst finding an "old" ffs around and using fsck's upgrade
	   % ability to upgrade it to a "newer" ffs.  "fsck_ffs -c 2 /dev/..." will upgrade
	   % to level 2.
	   
If it found a Solaris 8 ufs partition I'm not surprised it
failed.  If it's a NetBSD partition it should not be the
"old" ffs since the machine did not exist when "old" ffs
partitions were generated by default.  In either case I
think you should consider destroying the start of the disk
to get rid of whatever is confusing fsck.

	     But that command will not be successfully executed because it
	   complains that / is full.  I turned off logging in sysinst but that
	   didn't solve the problem.  It still dies and will not complete the
	   fsck because it says / is full.  What is it trying to write?  I don't
	   have the NetBSD source around but I suppose I can go take a peak
	   myself if noone knows where this is coming from.

If the old ffs that existed before was larger than the
size of the current partition it is possible that fsck
is falling off the end of the partition.  If it's a
Solaris 8 filesystem with logging, etc. who knows what
fsck_ffs is trying to do to it.

Eduardo