Subject: Re: diskless problem
To: None <port-hp300@netbsd.org>
From: Andre Schulze <as8@rcs.urz.tu-dresden.de>
List: port-hp300
Date: 05/19/2000 19:21:17
Am Don den 18 Mai 2000 um 11:10:08 +0100 schrieb David Brownlee:
> On Thu, 18 May 2000, Andre Schulze wrote:
> 
/* can't mount root from diskless apollo */
> 
> 	Option a)
> 	    If its a recent redhat or similar, there should be a way to
> 	    disable its bogus nfs v3 mountd responses. I think the netbsd
> 	    netboot howto should have the details.
> 
> 	Option b)
> 	    Switch operating systems on the server :)
> 
it's getting worse: i can't mount root from solaris either. tcpdump 
says something like this while the apollo tries to mount:
18:54:33.654276 rarp reply 8:0:9:16:f9:3b at rapollo
18:54:33.720351 arp who-has rapollo tell rapollo
18:54:33.722694 rapollo.1023 > 141.30.255.255.sunrpc: udp 96
18:54:34.719880 rapollo.1023 > 141.30.255.255.sunrpc: udp 96
repeated serveral times.

on the console:
mountroot: trying nfs...                                                        
nfs_boot: trying RARP (and RPC/bootparam)                                       
nfs_boot: client_addr=141.30.67.234 (RARP from 141.30.67.231)                   
nfs_boot: timeout...                                                            
nfs_boot: timeout...                                                            
nfs_boot: timeout...                                                            
nfs_boot: bootparam whoami, error=60

the strange thing is the reply to the rarp request comes before the
apollo asked. might be an error in tcpdump but there seems to be indeed 
a rarp problem, as this "bootparam whoami" on the console suggests.

	andre

ps.: this box already ran diskless - so it is possible to get those hp425e
(no framebuffer) working