Using JDE as a base example, my comments weren’t about making the application unavailable. They were about correctly bringing down servers that may be in use by an application.

If the *DATABASE host server has been serving connections into the JDE database, using ENDSBS to end host server connections can (and eventually probably will) cause damage to the database. Most often it can be corrected with RCLSTG *DBXREF because only the system catalog synchronization is disturbed.

One difficulty can be that you won’t know that it all must be resynchronized until days or weeks later. In the meantime, odd inexplicable errors get signaled that seem to go away simply by logging off and back on or by rerunning jobs or other weird actions.

But every once in a while, actual object damage is caused to tables in the database. Recovery from that can be painful.

So, instead of ENDSBS used to force jobs to shutdown, the proper sequence depends on what’s going on in the subsystem. For QUSRWRK, for example, it should include ENDHOSTSVR *DATABASE ENDACTCNN(*DATABASE) to cause connections to be ended before the server jobs are ended.

Then, before ENDSBS QUSRWRK is done, the jobs in the subsystem should be tested to see if they’re gone. If they aren’t, then ENDSBS should not be run yet.

The *DATABASE server is only one of many server functions. And QUSRWRK is only one example subsystem. And each LPAR may be configured to break work into different subsystems or to have different kinds of work.

But all of that is just warning that there are meaningful risks. It’s separate from your question.

What kind of “script” are you looking for? And What programming background will you be working from? (I.e., what languages have you worked with?) If ‘scripts’ are provided, is there a language that’d you’d be more comfortable with?

Because this involves system control functions, CL is most appropriate. If you understand that CL is a compiled language rather than a ‘script’, it won’t be a problem.

Also, if you’re already going through steps manually, those steps might easily be converted into the exact ‘script’ that you want. If you simply listed the steps, we could probably turn them around quickly.

Tom

I’m looking to automate the recycle of this application. I will using STRPDM.

This isn’t as simple as it might seem. Bringing subsystems down shouldn’t be done in a production environment without preparation. Ending the subsystems is easy, but handling all of the possible active work going on within a subsystem can be difficult.

Asking a question for an example ‘script’ indicates some uncertainty about system management software. A multi-LPAR system with a JDE environment, ROBOT products and others indicates a reasonably sophisticated production environment that shouldn’t be shut down without care. Simply coordinating management of multiple system images brings challenges of its own without consideration for applications that might be active.

What programming background will you be working from?

Tom