Hi All,
Our server OS info:
HPUX-UNIX>uname -a
HP-UX hps B.11.11 U 9000/800 2983362429 unlimited-user license
We access our server from Windows m/c through telnet sessions. But for unknown reason any telnet sessions are getting hung after about 15 minutes idle time. Then we close it and open another new session which works fine.
Any help on how solve this session hung problem will be appreciated.
Thanks in advance
Hey -
There could be a few things going on:
1) If you are using HP-UX outta the box 11.11 telnetd(aemon) via inetd, then check to make sure you don't have a $TMOUT variable is set upon login. The standard telnetd with HP-UX will look at this var. IE: echo $TMOUT If there is a value, you can set to =0 OR find out you .profile, .login, etc and change. Then you can update the 'skel' login files which should live in /etc/skel.
2) If 1) is not the case, then talk to your network admin, they might have certain rules invoked that even though you have an active TCP connection over telnet, if there isn't any generating traffic, the inline router, (if routing) TCP session timeout configured at switch level, or even an active state-full firewall that will drop connections after X time. You can also sorta trick 'schtuff' like this by setting a TCP keep-alive ACK via ndd in HP-UX. I say, "sorta" because depending on the level of experience your network admin has and the equipment, you can also set known keep-alive watchdogs that will do deep packet inspection and see that it's nothing but a keep-alive and still close the connection for security reasons; once again, depending on your environment.
To modify your HP-UX host to send keep-alives for telnet try this:
1) cp -p /etc/rc.config.d/nddconf /etc/rc.config.d/nddconf.prior_keep_alive
2) vi /etc/rc.config.d/nddconf
Look for "NDD_NAME" (minus quotes)
Uncomment (remove #) if needed and put:
NDD_NAME=[0]=tcp_keepalive_interval
THEN set the interval:
NDD_VALUE[0]=300000
This will send a keep-alive data packet every 5 minutes. Each minute is 60000, so set for how often you would like.
Save and exit and run: ndd -c which will change the parameter without rebooting.
Once again, could be some timeout on the host, or network idle timeout. Good luck!
Hi, many thanks for your reply with suggestions. Case 1 is not applicable as value of TMOUT is 0, it also not found in .profile, .login etc. /etc/skel is empty.
Now about case 2, I am now going through nddconf file and shall discuss with our Network admin before making any change.
Hope it will solve our problem.