Hello,

I posted this a few days ago to netbsd-help. But appearantly I had
choosen the wrong list. So please bear with me posting it gain.

I want to install NetBSD 2.0.x on a i368 server over the network. As I
have only access to a linux rescue system I plan the following.

Boot rescue, fdisk harddisk into 2 partitions. First will be swap,
second / (root).
I will format the second partion with ext2. Then I mount it, untar all
sets onto it, copy the kernel, install grub, configure the basics part
(rc.conf, network, ...) and reboot.
I mad a special kernel for this, as I need to hardcode the root device,
it contains: root on wd0f dumps on none
This works pretty well under vmware, but I am a little unsure about
a few things.

A) I do not put a disklabel on the hd and rely on the fact, that my root
partition will always be recognized as wd0f (swap wd0e). Can I take this
granted w/o a disklabel?

B) I hardcoded "dumps on none" in my kernel config. Am I right that the
dump device is only needed if my machine crashes or I would create a
dump with ddb? The dump device is a little mysterious to me. Coult I use
'dumps on wd0e'?

C) Will it cause problems running my root partition with ext2fs? I guess
no, but I am not really sure.

TIA,
        Ralf