Re: Difficulty with netbooting - le0: lost carrier when bringing down thenetbsd kernel

> But I've long lost my boot drives of yesteryear, and I can't find anything S$

Where are you geographically?  It's a long shot, but, if you're near
me, well, I still have numerous SCSI disks and wouldn't mind donating a
couple to the cause.  (I tried checking www.hell.org in case it said
something useful, but all I got was a redirect to HTTPS, which I would
have to find a work machine to use.)

> I see that you say that the kernel is delivered via nfs.  Does that mean the$

It depends.

Depending on the hardware - or more precisely the firmware - there may
or may not be.  Suns, to cite an example I know better, use reverse ARP
to get an address and then TFTP a file named based on that address;
that file's code controls what further happens.  Assuming the text

> The bootstrap sequense is:

> - the firmware uses rbootd protocol to load a bootloader from a server
> - the bootloader uses BOOTP/DHCP to get IP addresses of server and client
> - the bootloader also tries to mount servers directory using NFS
> - the bootloader loads the specified kernel from the NFS server

is accurate (I don't know - if I ever netbooted an hp300, it was long
enough ago I don't recall details, and even if I did they would have
been for decades-ago NetBSD), then, once rbootd finishes, whatever else
happens is up to the bootloader.

One thing I've found very useful when diagnosing netboot issues is to
just snoop the network for packets to/from the subject host's MAC
address.  When it gets stuck, typically, the last few packets will
point you towards what you need to fix or setup next.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B