Logged on into CLUSTER message when not logged on??

[johnslayton1]

I am running the IBM Personal Comm telnet client to connect to a OpenVMS 7.1-2 Alpha host from my WIndows XP workstation and once in a while, the IBM PCOMM will just close due to some application failure in the telnet client itself and then it just closes out my VMS session too without me logging off properly from the session like I always do to avoid issues like these below:

Then, when I try logging backon to the VMS session, it gives me this error message below EVEN when I reboot the PC....And then I have to wait like 20 minutes and then it will allow me to log back on again.
**********************************************
MYUSERNAME IS ALREADY LOGGED ON TO THE CLUSTER ACCESS IS DENIED
***********************************************

Like said, this always happens after when an application failure from the IBM PCOMM.

And I AM NOT logged on since I always restart the PC in order for this message to clear itself..

How can I resolve this issue on the server side?

thanks.

[Hein van den Heuvel]

John,

This message is NOT coming from OpenVMS itself.

It is generated by some tool/script/command file which the local cluster management has put in place. You would have to check with them for clarification.

We can but speculate how they would have implement the (lame!) mechanism to dissallow a second login and how it failed to detect you lost the session. OpenVMS itself would be perfectly happy to create more than one session for you, ot the same system or on multipe nodes in the cluster.

Cheers,
Hein.

[Volker Halle]

John,

when talking to your system manager about this problem, please mention the 'virtual terminal' feature.

Please see the following thread:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1048081

Volker.

[Robert Gezelter]

John,

I concur with Hein and Volker. A mechanism has been implemented beyond the base mechanisms provided by OpenVMS to prevent multiple logins. That mechanism appears to have some shortcomings.

It is certainly possible to implement such a mechanism within a cluster without such problems. Personally, I would suggest that you bring the problem to the attention of the system manager(s).

The virtual terminal facility mentioned in a previous response is also an excellent idea, but if it is used, there is a need to reconnect to the SAME node that you disconnected from (indeed, that could be the problem that you are experiencing).

- Bob Gezelter, http://www.rlgsc.com

[Paul Jerrom]

Gents, I think the issue isn't the ability of the users to log in more than once, but the fact that the sessions persist after a PC crash.
We had similar where user sessions were persisting after PC crashes, and their processes had devices allocated (and which therefore weren't deallocated after the crash).
It's due to the keepalive time - I think the default settings are that the server will check whether the client is there after 150 seconds of inactivity, and if inactive will poll up to 8 times at 150 second intervals (= 20 minutes).
I think there is a way of changing this for Telnet only (tcpip set service telnet/prot=tcp/probe_timer=desired seconds or something). I guess too small a setting will cause a large amount of traffic just checking that clients exist, so you'll probably need to play until you get the best setting for your environment.
Have fun!

[Hoff]

I'd work to identify the source of the instability -- the PC client, the network link, the PC itself -- and remediate it.

There are good free and non-free telnet clients available for Windows systems. VTstar on the OpenVMS Freeware and PuTTY are around, and HP offers the PATHWORKS32 client package.

This OpenVMS system is also in the range that had a version of TCP/IP Services that had some telnet-related issues. V5.0 had a looping telnet server bug that could be nasty. Upgrades to more recent versions -- these versions of OpenVMS and of TCP/IP Services date back seven years or so -- are to be encouraged.