cannot telnet to hpux 11.11 host

[ecorban]

Hi
I rebooted my host 11.11 host this morning and can no longer remsh to the host or telnet with telnet refused. I get the following on the console and cant figure how to restart the udp services.

Clean up old log files .............................................. OK
Start system message logging daemon ................................. May 26 14:45:08 cp092043 syslogd: syslog/udp: unknown service
OK


how can i fix it ?

[Patrick Wallek]

It sounds like networking did not start at all.

Check your /etc/rc.log file and look for an errors.

You could also try a

/sbin/init.d/net start

and see what happens.

[ecorban]

i cannot remsh out of the host. i can ping the host but not remsh/rlogin out to the host.


#
# remsh cp092059
rlogin: login/tcp: Unknown service
# ping 10.32.92.59
PING 10.32.92.59: 64 byte packets
64 bytes from 10.32.92.59: icmp_seq=0. time=0. ms

----10.32.92.59 PING Statistics----
1 packets transmitted, 1 packets received, 0% packet loss
round-trip (ms) min/avg/max = 0/0/0
# remsh 10.32.92.59
rlogin: login/tcp: Unknown service
#

[Mel Burslan]

I was just going to say something in that manner. Usually connection refused message to telnet request means that networking is up but the service is not working or blocked someway.

And your messages are telling me that your inetenet services have problems.

If you can attach your computer via a serial cable to the console port of your hpux server and by capturing the terminal screen contents in to a log file, you can put the contents of /etc/services and /etc/inetd.conf files here, maybe it can help understanding the problem.

[Mel Burslan]

also a dump of /etc/rc.log would be useful.

[ecorban]

i also extend fs the /usr partition, maybe that may have screwed something else up that effects udp services.

[Patrick Wallek]

Since you can't telnet or remsh to the machine, I think ALL networking is down, not just UDP. Forget about UDP itself and look at all networking.

It is possible that something was messed up when you extended /usr.

Things to check:

Is /usr mounted? Are all directories under /usr visible like /usr/bin, /usr/sbin, /usr/lib, etc. If not, you've really got a problem.

Also, please check /etc/rc.log and tell us what errors occurred there.

[ecorban]

problem was a correct /etc/services file.

go figure ? thanks anyway.

[Bill Hassell]

What happened is that your /etc/services file was ovewrwritten or truncated. When the networking code is started some (not all) network services need to know their name and /etc/services is like /etc/hosts--it put a name to a particular service. The key in all the errors was that the service name was unknown, and the place where the name is defined is /etc/services.

NOTE: ping is unique in that it is a low level service and does not require a daemon to work. The ping command will use DNS or /etc/hosts to get an IP, but once it has an IP addr, it just sends the ICMP echo request.

[Suraj Singh_1]

Strange! How did /etc/services corrupt?
What could have gone wrong?

[Bill Hassell]

There is nothing in the opsystem that would cause this. /etc/services is a read-only file as far as the OS is concerned. Someone may have run vi on it and accidently deleted all the entries, or perhaps did a mv or cp command that overlaid the file. Look in root's $HOME/.sh_history:

grep services $HOME/.sh_history

Hopefully you do not have duplicate root accounts:

logins -d

but if so, look in those accounts too for evidence of editing or manipulating the services file.

[ecorban]

I had no /etc/services so it never got loaded. how i lost it i really dont know.