You'd code a REXX exit and ask for execution via a DTCPARMS file.
Create a SYSTEM DTCPARMS file on TCPMAINT 198.
Here's my entry for SMTP, with an EXIT
:nick.smtp    :type.class
              :name.SMTP daemon
              :command.SMTP
              :runtime.Pascal
              :diskwarn.YES
              :exit.LDIMMCMS
In the REXX exit exec -that is called at various events- you'd code
something like
  parse upper arg keyw .
  if keyw='SETUP' then do
    'EXECLOAD ....
    .....
 end

Changing the code without stopping the TCPIP stack:
- yes if TCPIP could execute CMS commands while it runs
  (and I don't think it provides a mechanism)
- if the code is not EXECLOADed, you could store the exec in an SFS directory
  accessed by TCPIP and then you could change it dynamically.  But, when the
  SFS pool is stopped, TCPIP will no longer find the exec.

You could write a REXX EXEC, compile it into a MODULE, load that module
as an IMMEDIATE CMS command in TCPIP and then you could do everything.
I wrote such an exec years ago to cleanup MONWRITE's A-disk, and have
re-used it to cleanup SMTP's A-disk.  But:
 - EXECs loaded as immediate commands should not perform screen IO
 - maybe not execute some special CMS commands
I solved the first problem by executing the CMS command via a PIPE
and show the response via CP MSG.  I never had problems with MONWRITE
nor SMTP where it was used daily, never tested this with TCPIP.

If you're still interested, I can send you the code:
 IMMERASE and IMMCMS
 both the source and the MODULE
 you need at least the REXX compiler library.

2009/1/29 Hughes, Jim <[email protected]>:
> Can the active Telnet Session Connect Exit(SCEXIT EXEC)  be changed without
> stopping and restarting the tcpip stack?   If so, how??
>
>
>
> The TCPIP programmers guide mentions you should EXECLOAD this rexx exec for
> performance.  Where would be the appropriate place to add the EXECLOAD
> statement?
>
>
>
> I've looked through the manuals and I cannot put my fingers on an example of
> where the EXECLOAD/EXECDROP statements would go.
>
>
>
> ____________________
>
> Jim Hughes
>
> 603-271-5586
>
> "It is fun to do the impossible."
>
>



-- 
Kris Buelens,
IBM Belgium, VM customer support

Reply via email to