And yet another note from that Centris 610 owner...

I think I've found part of my difficulties, now I need
help figuring out how to fix it.

I was having trouble booting into multi-user--it would panic
just before checking the SCSI bus. If I booted into single-user
and mounted /dev/sd0g onto /usr, then went multi-user, it would work.
As it went into multi-user, it would complain that /dev/sd0a and
/dev/sd0g weren't clean and that I should run fsck. So I did.
Next time I tried mounting /usr, it complained of a bad superblock.
Hmph. So I remade the partitions, using mkfs (MacOS side). This
time I selected the optimize for time, instead of the default
optimize for space--is this not a good idea, or does it matter?

As a test, I installed everything on /dev/sd0a,and left /dev/sd0g
empty. Voila, now I could boot directly to multi-user.

I ran mkfs again on /dev/sd0a and /dev/sd0g, just to be anal.
I then used the minishell on the installer to build devices
and to mount /dev/sd0g on /usr (well, first I had to mkdir /usr
with the minishell, and then mount). Then I installed base12, etc12,
man12. I then used the minishell of the installer to do a fstab force,
and did a cat /etc/fstab to make sure the /dev/sd0g entry was there.
So far so good. 

I booted into single-user, checked things out (are there standard walk-through
sanity checks?) and went into multiuser. Once again it died, which
I expected, because in single-user, a df showed that /dev/sd0g was
not mounted on /usr.

I tried booting straight to multi-user, and again it died very quickly.
I tried the single-user/mount usr/multi-user combo, and again that worked
fine.

So... when everything is on one partition, all is bliss. When I
have two partitions, I have to manually mount the non-root partition.
My question is, have I done something wrong +/or is there further
configuration to do to get it to boot straight into multi-user
w/out manually mounting partitions??

I still get segmentation faults and core dumps left and right. When
things get gnarly and it dies on me, the panic or message is usually
complaining about the MMU. Any thoughts?

And finally, I am thinking about using the -current snapshot Steve
Allen mentioned today, but what is SBC SCSI? I am starting to wonder
which of all these kernels to settle on (netbsd.quadra, netbsd.quadra_dual,
netbsd.scampbel, netbsd.current) and what the differences really are
between them.

My next step is to install the source code and start cranking out my own kernels.
Where is the source/patches for these various kernel builds located?

OK, I'm finished now. If you reached this point, you deserve a soggy
milky way.

Mike Hitchcock
mike@unival.com
(still temporarily vitality@teleport.com)