HP OpenVMS Systems

Failure to deallocate TCP/IP TN (telnet) device?

» close window

The Question is:

 
	Oh, mighty Oracle, I have a conundrum which tests my ASE and 25 years''
 experience:
 
	I have a hospital system running IDX. IDX provides what appears to be a
 rehashed version of VT320 emulator from older versions of Pathworks which the
 client medical technicians and nurses use.
 
	The IDX database and software runs on an Alpha 4100 running VMS V7.1
 (all patched up properly), and Digital Services for TCP/IP V5.0A (similarly
 patched up by yours truly).
 
	What folows is a description of the problem:
 
	If the user's Windows systems crashes (imagine that), the user's BG
 device goes away, but the TNAxxx device remains attached to the process and
 thinks it still knows the client's remote port number. These processes have to
 be removed via a tedious IDX-su
pplied FORCEX procedure.
 
	The question is, how come the TNAxx device persists even though the
 associated port and BG device have disappeared?
 
	Thank you, oh wise one, for your help.
 
	W.
 

The Answer is :

 
  It would appear that the application is not responding to errors reported
  from the TN device, and thus not deallocating the device as would normally
  be appropriate -- if a $forcex call clears this, the most obvious behaviour
  of such a tool would be to cause the application to exit and thus deassign
  its channel(s), and thus the TN device would be vaporized.
 
  Check for channels assigned to the TN device, and -- if you find any -- then
  check with the maintainer of the application to determine what (if anything;
  anything other than the FORCEX tool) can be done to resolve this behaviour
  and to cause the process to (cleanly) exit.
 

  

     

     
answer written or last revised on ( 15-JUN-2001 )
     
» close window