I use a netboot procedure for bootstrapping netbsd on sparc's.
I have a sun4c that I've booted many times without problem.

I've been trying to boot a classic and it keeps getting:


SPARCclassic, No Keyboard
ROM Rev. 2.12, 32 MB memory installed, Serial #3380091.
Ethernet address 8:0:20:1f:25:7c, Host ID: 8033937b.


Rebooting with command: net netbsd -s                                 
Boot device: /iommu/sbus/ledma@4,8400010/le@4,8c00000   File and args: netbsd -s
>> NetBSD/sparc Secondary Boot, Revision 1.8
>> (pk@flambard, Tue Oct 27 14:09:00 MET 1998)
boot: client IP address: 192.148.164.20
bootparamd: 'whoami' call failed
Can't open network device `/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000'
open: netbsd: Unknown error: code 60
device[/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000] ("halt" to halt): halt

That's with a Solaris system as boot server (what I normally use).
I get much the same result with a NetBSD 1.4 machine (sun4c) as the boot
server though:

ok boot net netbsd -s
Resetting ... 
m
SPARCclassic, No Keyboard
ROM Rev. 2.12, 32 MB memory installed, Serial #3380091.
Ethernet address 8:0:20:1f:25:7c, Host ID: 8033937b.


Rebooting with command: net netbsd -s                                 
Boot device: /iommu/sbus/ledma@4,8400010/le@4,8c00000   File and args: netbsd -s
>> NetBSD/sparc Secondary Boot, Revision 1.8
>> (pk@flambard, Sat May  8 21:54:06 MEST 1999)
boot: client IP address: 192.148.164.20
bootparamd: 'whoami' call failed
Can't open NFS network connection on `/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000'
open: netbsd: Unknown error: code 60
device[/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000] ("halt" to halt): 

The NetBSD bootparamd(8) says that it will return the
name in /etc/bootparams rather than the cannonical name in response to
the whoami request, but the client still says that 'whoami' call failed.

In both cases everything runs fine until the client does:

12:30:29.469152 rarp who-is 8:0:20:1f:25:7c tell 8:0:20:1f:25:7c
12:30:29.470387 rarp reply 8:0:20:1f:25:7c at 192.148.164.20
12:30:29.516933 192.148.164.20.1023 > 255.255.255.255.111: udp 96
12:30:31.513737 192.148.164.20.1023 > 255.255.255.255.111: udp 96
12:30:35.514650 192.148.164.20.1023 > 255.255.255.255.111: udp 96
12:30:43.514490 192.148.164.20.1023 > 255.255.255.255.111: udp 96

at which point we get the error above.

Why is the client not trying to talk to portmap on the boot server?
Neither Solaris nor NetBSD (1.4) pay any attention to
255.255.255.255.111 (I'm running portmap -l and nothing is logged).

[please cc me on any replies]

--sjg