×
The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.
In the past, I worked with a lot of code that was migrated from a S/36.
In that world, they'd often use a single-threaded job queue... they'd
submit many procedures so that each one would run after the other was
finished.
It was more efficient than making it all one single job, becuase it
would have less in memory at one time. It was also more like the
punched card environments (many S/36 software was migrated from a S/3,
possibly via the S/34). In this case, you could put multiple jobs in the
card stack and they'd run one after the other, etc. (At least, I think
so... my mind is fuzzy, it was too long ago.)
Anyway.. this type of coding is commonplace in old solutions. And one
of the unfortunate things about this particular platform is that people
strongly resist change -- it worked 40 years ago, why not still do it
today?? Haha.
Another thing is that back then multitasking wasn't as good as it is
today, and systems had far less memory, so running one task at a time
was useful.
Anyway, I can't tell you whether it'll break anything in your
environment. You need to analyze that and figure it out. But, the
"right" way forward is to determine which jobs should be sitting on a
single threaded queue to wait for each other, and which ones should run
simultaneously with other jobs, get that all straightened out, and
create separate queues for each situation, then change the code to use
them properly.
On 7/15/2013 10:24 AM, Hoteltravelfundotcom wrote:
What do you think of a shop which has only 1 jobs que? SHouldn't there be a
test system
for a programmer? This has happened twice now. I have a long running
RPG program, working over the weekend, it didn't sink in the first time.
so then its still running and causes some jobs to not run. Consequently
subsystems
did not come up. I don't recall in many years either a system that shuts
down every night and comes back up.
It is frustrating. Am I wrong in thinking that they should have more than 1
jobq, that is, that other jobs can run at the same time that there should
be a test system? Is it possible for a newcomer to not make these mistakes?
I suppose I am to blame that it happened again. BUt I really
have trouble acclimating to this type of environment, and not sure what can
be done.
As an Amazon Associate we earn from qualifying purchases.