Hi, I've been using my Sun 3/60 dual-headed (mono w. salvaged cgfour
retro-fitted) and booting from Linux using Xkernel. I'm very interested in
NetBSD, and so I was delighted to see that a NetBSD Xkernel has been assembled.
Luckily I was familiar with the procedure, as there's no instructions! I've got
(I believe) to the same point as Wil, but unfortunately I don't understand your
solution. I'd _really_ appreciate some hand-holding here!

netboot loads via tftp, netbsd loads and boots. At this stage I have messages in
/var/log/messages telling me that the root directory has been mounted by the
Sun, but the Sun says (repeatedly)
RPC timeout for server 0x81d7461a
which is indeed my server's IP. So, following from Wil's lead I try tcpdump, as
follows: (The Sun 3/60 is dunbar.ucs.ed.ac.uk, the Linux box is lammermuir)

$ tcpdump host dunbar
eth0: Setting promiscuous mode
eth0: Setting promiscuous mode
tcpdump: listening on eth0
09:15:45.450000 dunbar.1022 > lammermuir.ucs.ed.ac.uk.637: udp 80

This last line simply repeats with the  time updated, every 5 seconds.

I'm stuck. All advice appreciated,
TIA,
al.

-- 
/* Allan.Kelly@ed.ac.uk ..   . .     . .           http://edina.ed.ac.uk
 * Software Engineer       *     . .     .    . .     Edina Data Library
 *                                *       . .     .    . .   
 * Doughnut by aMeringue                   *       . .     .  ..
 * http://www.ameringue.freeserve.co.uk             *       *    * . . .
 */

On Fri, 12 Feb 1999, W. Reilly Cooley wrote:

>
>After fiddling around with my 3/60 for too many hours, I have finally
>discovered the source of my problem.  I'm netbooting off of a BSDI 3.1
>machine, which configures pretty much like NetBSD.  (There was no supplied
>bootparamd, but I used the one from a free UNIX, I forget which right now.  
>[I started this several months ago.])
>
>At any rate, I got it to go all the through booting, to loading the
>kernel, but it would hang when it went to mount the NFS root.  I noticed
>with tcpdump that it was broadcasting.  I thought the problem was that it
>was broadcasting, rather than directly contacting the machine it had
>called bootparamd from previously.  (Someone explain why it calls
>bootparamd twice, and why it broadcasts the second time.)
>
>*Finally* I figured out that it was broadcasting as if it were on a /24
>subnet, but we have a /23 subnet.  This caused it to use the wrong
>broadcast address. Changing the netmask on the host machine changed the
>broadcast address, which allowed the system to boot.  Who-ee!
>
>But now I've got a machine with the wrong netmask.  Can I configure the
>Sun to use a different broadcast or netmask?  Will it even understand a
>/23 subnet, being from the days before CIDR?
>
>Also, I'm having a bit of a problem with messages about the getty
>respawning too quickly (from syslog):
>
>init: getty repeating too quickly on port /dev/ttya, sleeping 
>init: getty repeating too quickly on port /dev/kd, sleeping
>
>Before I dig into this, does anyone have a quick explaination?  Sorry to
>ask before even looking at it... This brings up request: searchable mail
>archives.  Good idea eh?  (Or maybe even a link to another site that has a
>searchable archive.)
>
>Finally, one last question.  I notice archives of Sept-Nov 1998 are
>missing.  Is this due to the disk crash that happened several months ago?
>(Or maybe I'm mis-remembering.)
>
>
>Wil
>--
>W. Reilly Cooley
>
>   Internet Meta-Resources: http://nakedape.navi.net/meta-res/
>            "All the Net you need to be a geek"
>
>