Cannot ping C3000 workstation

[Bruce Baillie]

I cannot ping, ftp or telnet this system. Sometimes ping will work a few times then stop.
The network is OK going out with FTP, telnet and Netscape. Lanscan shows lan0 is up.
C3000, HP-UX 11.00, onboard nic, 100BT, network. Swapped cat 5 cable, tried different drop, Changed IP address. Anything else to check before I place a hardware call?

[Biswajit Tripathy]

Might want to try using "linkloop" to see
if the MAC address connectivity is okay.

- Biswajit

[Sanjay_6]

Hi,

what is your lan's netmask.

Is it set to 255.255.255.0 or something else.

Hope this helps.

Regds

[Ron Kinner]

HP allows you to set the MAC address for a machine. Could be somebody cloned this workstation and the MAC address came with it. I think it's in a file with a name like hpbtlanconf down in /etc/rc.config.d. If you have faster than 10/100 the file may have a different name but it will start with hp and end in conf.

If you see a line like:

HP_BTLAN_STATION_ADDRESS[0]=0x080009C4686E

then remove it and reboot.

Also if we are not local to the machine and the default gateway does not respond to pings the connection will work for about 3 minutes after a reboot and then fail. This is dead gateway detection and is turned off by ndd option ip_ire_gw_probe which should be set to 0 to stop it from complaining.

FTP and Telnet like to use a DNS lookup to see who is calling. Test it with nslookup.

Use
lanadmin
lan
display
to look for Ethernet errors.

You don't say what the IP address is but sometimes if your IP address is in the first or last subnet of a divided class address HPUX will refuse to use it. There is an ndd option ip_check_subnet_addr which if set to 0 will stop HPUX from objecting to the address.

If you need to make an ndd change also make it in /etc/rc.config.d/nddconf

Ron

[Steven E. Protter]

Another possible cause is another device on the network with the same ip address. Doesn't have to be a Unix box.

To diganose, i pull the C3000 off the network and then ping the address. If I get an answer, I know what the issue is and resolve it with the networking folk.

It is however equally possible that the hardware is bad. You might try isolating the box on its own hub and see if the situation improves. If yes, its environmental, if not, lets check the hardware.

xstm (requires X) cstm mstm can let you run the excercize function on the NIC card. If it has link it should execute and suceed. If it returns an error code call hardware support with it.

SEP

[TwoProc]

with one of the terminals that you have still connected - see if "inetd" is still running. You're situation sounds like inetd died. HPUX has been plagued with the very infrequent bug in which this deamon just ups and quits.

To check - "ps -ef | grep inetd"
if you don't get anything - re-run the following...

To restart - you can a) /sbin/init.d/inetd start, or b) just type "inetd".

If it's happening often, open a support call.
In the meantime - you can write a small shell script to watch for it being up, and if it's every down, just restart it...

something like...

while /bin/true
do
if [ `ps -ef | grep inetd| grep -v grep | wc -l ` -eq 0 ]
then
/sbin/inetd start
fi
sleep 30
done

(Check out the script for errors - I just typed that in off the top of my head).

Anyways, if it is inetd - the above will tide you over until you get a real resolution...