Re: executing jobs in sequence

Hi, Michael:

I think the biggest disadvantage to maintaining a single large CL program 
is what to do if the program fails (not just goes into message waiting 
status). If an untrapped/unhandled error occurs when the CL has executed 
some but not all of its calls, then there has to be a mechanism to recover 
and resume at the correct point. This mechanism may be programmatic or may 
require operator intervention. Proponents of this approach will suggest 
that the solution is to write the CL correctly, and in theory they will be 
absolutely right, but factors external to the program can sometimes gum up 
the works. At least, they do in OUR shop [grin].

Generally speaking, if a large CL submits jobs exclusively (nothing 
"inline") and the submits are to one single-threaded job queue, then 
recovery from an unhandled error can be relatively straightforward from an 
operations standpoint.

The second disadvantage in my opinion is that you reach a certain critical 
mass of code at which efficient maintenance becomes an issue. The better 
your shop is at code maintenance in general, the less this should be a 
concern.

We have a number of these omnibus CLs. As opportunity arises I replace 
them with discrete programs. This is in part because discrete programs are 
much more easily managed with Robot/SCHEDULE (and take advantage of more 
of Robot's capabilities). However, the same principles would apply to 
other environments -- in many cases I would guess even more so.

A sequence of discrete programs has its own weaknesses, but that isn't the 
question you asked.

Darrell

Darrell A. Martin  -  630-754-2187
Manager, Computer Operations
dmartin@xxxxxxxxxxxxx

midrange-l-bounces@xxxxxxxxxxxx wrote on 01/08/2007 08:04:57 AM:

"Jones, John (US)" <John.Jones@xxxxxxxxxx> wrote in message 
news:mailman.42.1168262354.12056.midrange-l@xxxxxxxxxxxxxxx

A second option which would be more complex to develop but potentially
nicer to use (depends on your job stream) would be to have a single CL
program drive all of the batch work (either inline or via SBMJOBs) and
have that CL program monitor for errors and enable whatever corrective
measures you require.

Thanks for the suggestion. On the VSE system, all of these "jobs" are 
contained in a single library "book". There are a lot of them and it is 
easier to maintain this way. Are there any disadvantages to maintaining 
single large CL program for the entire run?

-- 
Regards,

Michael Rosinger

This e-mail, including attachments, may contain information that is 
confidential and/or proprietary, and may only be used by the person to 
whom this email is addressed. If the recipient of this e-mail is not the 
intended recipient or an authorized agent, the reader is hereby notified 
that any dissemination, distribution, or copying of this e-mail is 
prohibited. If this e-mail has been delivered to you in error, please 
notify the sender by replying to this message and deleting this e-mail 
immediately.