--- "James K. Lowden" <jklowden@schemamania.org> wrote:
 
> To make sure your hardware is functioning (cards/cables/hub), you can
> run
> tcpdump(1) on the netbsd box and ping it from somewhere else.  It's
> not
> obvious right away, but tcpdump doesn't even require a functioning ip
> setup; it will listen on any ethernet interface, whether or not
> ifconfig(8) has been used.  
> 
> Even if ping reports a problem, tcpdump should show some activity. 
> If it
> doesn't, you have an OSI Layer 1 problem.  If you see something you
> can't
> interpret on an otherwise quiet network, post it here and let other
> eyes
> look.  

I now think there is a hardware problem with the NIC. With one machine
pinging away continuously, tcpdump -i ep0 on the problem machine showed
nothing (a third, functioning host running tcpdump showed plenty). I
let this go for some time, during which I sat and observed the activity
lights on the hub. There was no activity on the problem machine's port
save for a few occasionally blips. When I later checked tcpdump on the
problem machine, there were a number of the following:

	21:19:29.xxxxxx arp who-has jeffoh1 tell netbsd386

jeffoh1 is the gateway, same machine that was trying to ping the
problem machine. Netbsd386 is the problem host.

I've tried three cables from various ports on the hub to the NIC in the
problem machine, to no effect. This NIC did work last week with a brief
install of Linux. I'm not sure if I ever used this box/NIC combination
with Windows.

In case it's useful, I'll include the information about the NIC from
the 3c35x9cfg DOS utility:

Ethernet Address: 00A024EC81F9
NIC type: Combo
Date of Manufacture: 8/27/96
Division Code: 6
Product Code: BZ
ASIC Revision: 2
Software compatibility failure level: 0
Software compatibility warning level: 0
I/O Base address: 340H
Interrupt request level: 5
Boot PROM size: disabled
Transceiver type: auto select
Network driver optimization: Windows or OS/2 client
Maximum interrupt disable time: 500 microseconds
Plug and Play capability: disabled
Full duplex: disabled

> 
> BTW, there's no /etc/ipf.conf or /etc/ipnat.conf, is there?  
 
No, neither of these.

Lastly in case it is still a softare/config issue, I compared the
running processes on the problem machine to the processes on an install
of 1.4.1/mac68k and observed some differences. This might not be
unusual but I'm too green with 1.5.x to know. For what it's worth, here
is the output from ps -ax on the problem machine:

PID TT STAT    TIME COMMAND
  0 ?? DLs  0:00.00 (swapper)
  1 ?? Is   0:00.02 init 
  2 ?? DL   0:00.00 (pagedaemon)
  3 ?? DL   0:00.02 (reaper)
  4 ?? DL   0:00.01 (ioflush)
109 ?? Is   0:00.10 /usr/sbin/syslogd -s 
190 ?? Is   0:00.01 /usr/sbin/inetd -l 
193 ?? Ss   0:00.01 /usr/sbin/cron 
195 E0 Ss   0:00.18 -csh 
213 E0 R+   0:00.00 ps ax 
196 E1 Is+  0:00.02 /usr/libexec/getty Pc ttyE1 
197 E2 Is+  0:00.02 /usr/libexec/getty Pc ttyE2 
198 E3 Is+  0:00.02 /usr/libexec/getty Pc ttyE3 


Thanks very much,

Jeff
08.21.2002


__________________________________________________
Do You Yahoo!?
HotJobs - Search Thousands of New Jobs
http://www.hotjobs.com