Re: Call program(s) on subsystem start

So a NEP is running in a user job in the subsystem, such that if someone issues an ENDJOB and the RAGE transpires, the subsystem has not really ended but the NEP falsely concludes and reports the subsystem has ended.?

While I would generally accept that implementation for my own use as well [i.e. I find it perfectly acceptable, but I would also enable the job as self-submitted to a single-threaded job queue, behind itself], I do not see how that implementation is any less of a "work around" than some other suggestions that have been made, because the termination processing would not be running in the subsystem monitor job itself; i.e. the RAGE would need to occur as the effect of the subsystem ending, not as the effect of the user NEP job ending, in order to conclude that the subsystem is ending.

Regards, Chuck

On 2/2/11 12:28 PM, dieter.bender@xxxxxxxxxxxx wrote:

... the program simple has to be a NEP (never ending program)

CRPence on Wednesday, February 02, 2011 9:08 PM wrote:

On 2/2/11 11:57 AM, dieter.bender@xxxxxxxxxxxx wrote:

.... why using work arounds? as I mentioned before, there is an
API CEE4RAGE to register a procedure, called at end of activation
group. Just call a program in your startup in a named activation
group CONTROL and register CEE4RAGE for this Programm. At
Subsystem end, the activation group CONTROL will end an the exit
handler is called. You will find this technic in many of my open
Source stuff at my german website bender-dv.de, google translate
will assist you to find it there.

A good approach for a user job, but how does a user get such a
program to run in the OS subsystem monitor job? The intercept of
the end of the AG in a user job that runs "in" the subsystem will
detect when that job\activation group ends, not when the subsystem
[monitor job] has ended.