RE: ILE questions: activation groups, rpg-cgi, classic web server.

Seth, Nathan and Bob,

thank you all for the replies.

So, what you're saying is, as long as I can guarantee that my programs will
run properly (leftover data in variables, left-open files, etc), after
multiple calls without 'closing' the activation group, AND that I don't
have an enormous number of programs that could be called from one request
processor, then I should 'name' my activation group.  Right?

or is that even the way web server request processor jobs work?  the job
itself (DEFAULT QTMHHTTP in subsystem QHTTPSVR) runs forever (until the
subsystem ends or the server ends or I cancel the job).  do the activation
groups work the same way in this job?  ie eventually I'll get a copy of all
my CGI programs running in each RP job?

as the application grows, will this hinder or help performance (both of the
web ap and any other batch or interactive jobs on the system)?

If I name my activation groups, how should I name them?  all first level
programs with the same activation group?  different names?

Thanks again for the help.

rick

---original messages---
>>
>> What do you mean by "run out of memory"?  Wouldn't all resources
allocated
>> by the program be released, and the *new activation group be deleted,
when
>> the CGI program ended?
>>
>> Nathan M. Andelin
>> www.relational-data.com
>>
>
>I looked it up and apparently I'm wrong.  It was so long ago that I can't
be
>sure of the reason for using named (as opposed to *new) activation groups
>but I thought it was that the system wasn't releasing resources.

If program A calls program B where B is compiled with activation group
*new,
then when program B ends, the activation group B was running in ends.  If
program B is in a named activation group, when program B ends the
activation
group remains open until it is ended by one of three ways: 1) the job ends
2)You reclaim the activation group using RCLACTGRP or 3) use CEETREC API to
reclaim activation group.

The benefit of using *new is that activation groups are cleaned up
automatically when programs end.  The advantage of using named activation
groups is the activation group remains open while you job is running along
with all the resources within it.  This can be a real advantage in
performance if the program gets called multiple times within the job.
There are advantages to using names activation groups such as sharing
open data paths.

Seth Shields
Julius Blum Inc.

-----and Bob's message--------

The biggest issue here, is that *NEW causes the activation group to be
created each time the program or service program is used in a job. Not
each time a procedure is called within a service program, but each time
the program that uses the service program is created.

A named activation group is somewhat "static". And is reused each time
the program is called.

Bob Cozzi
cozzi@rpgiv.com
Visit the on-line  Midrange  Developer  forum at: http://www.rpgiv.com