On Wednesday, 08/29/2007 at 05:48 EDT, Aria Bamdad 
<[EMAIL PROTECTED]> wrote:
> This is a problem not just for what I need done but in general for
> any exit processing.  If an exit processing routine is provided and
> documented for use, then the stack needs to allow the server some
> known time window to do it's job.  As it is now, the time window changes
> each time.  It could be a few seconds, to tens of seconds.  I am not
> sure what it is dependent on but I know it is not consistent.
> 
> It would be helpful if the server could be aware of exit processing
> so that it can allow for the exit to finish properly.

I understand what you are saying.  The exit capability is provided so you 
can make custom configuration changes, not get tied up running some other 
program.  That stack has no idea whether the server is still coming up or 
has stalled.  If you don't want the stack to monitor the port, put 
NOAUTOLOG on the PORT statement for the server.  Then the server can take 
as long as it like to come up.  Of course, if it goes down, the stack 
won't restart it automatically.

Alan Altmark
z/VM Development
IBM Endicott

Reply via email to