HPUX 11.0: Host Unreachable

[LG Porter]

I have a N4000 server running HPUX 11.0 inside the DMZ with HP Secure Shell installed. Several days ago the server was rebooted and as result, the connectivity is unreachable. I have checked to assure the at the gateway detection was turned off using the "ndd" command as well as confirmed that the parameters in the /etc/rc.config.d/netconf is right. I verified the present routes using the "netstat -rn" command. The sshd daemon is running and I can ssh from other systems on that same network. Outside performing a nettl trace, what other methods or commands should I use to determine the problem? Thanks

[john kingsley]

Are you sure there isn't a problem with your router? Can you ping your default route?

[Steven E. Protter · Accepted solution]

I'd check DNS first.

nslookup servername (fully qualified is better)

If you don't get an answer or it doesn't match the netconf file you have a way to proceed.

Also, check with the firewall admin and see if they've closed some ports recently. That gets me every time, and they almost NEVER tell the Unix admin.

SEP

[rick jones]

IIRC, "Host Unreachable" messages are triggered by ICMP Destination Unreachable messages with a subcode of host (or some such terminology). I suspect if you were to tcpdump trace (preferable to nettl, and tcpdump, along with ethereal are available from the HP-UX Internet Express bits on software.hp.com) your attempt to ssh out from the system you would see those ICMP messages coming back.

If nothing else the network trace wold be interesting, and you can filter it by adding a "port " to the end of the tcpdump command line.

Perhaps your firewall has decided it does not like your N4000 and does not want to let it communicate to the outside world.

Whether or not the sshd is running should have no (IIRC) effect on running ssh from that system to another.