Performance issues regarding merging libraries
Home > IT Answers > AS/400 > Performance issues regarding merging librarie...
Leah Rosin
480 pts.
0
Q:
Performance issues regarding merging libraries
multithreading, AS/400 performance, library merging
I received an email yesterday from a reader who has a big task. She is working with a hospital to merge three different libraries into one using AS/400.

"One Hospital is going for data merger of three different libraries to one single library including some data mapping and conversion.
There are lots of programs written for conversion of data to single library. But we are facing performance issue as converting billions of records takes time and CPU usage goes high, resulting in slowness of application.
We can implement the concept of multithreading, but monitoring each job for its conversion may become a big task.
What should we do?"
ASKED: Apr 23 2008  5:05 PM GMT
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
RELATED QUESTIONS
0
DanTheMan 2
395 pts.
0
A:
 RATE THIS ANSWER
0
Click to Vote:
  •   0
  •  0
  • AddThis Social Bookmark Button
It look like this project is pretty huge so in that case, I would probably ask to create a special subsystem with a specific shared pool insulated from the one used by QBATCH so each of them would have their individual memory and CPU specs so they wont disturb the other one while running Monster jobs.

If you do so, dont forget to start your sub system at IPL time in your QSTRUP pgm.

Have Fun !

__________________
DanTheMan
Last Answered: Apr 28 2008  7:51 PM GMT by DanTheMan 2   395 pts.
Latest Contributors: Lovemyi   1470 pts.
  •   
0
0
RSS - Discussions Help
Discuss This Answer:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _



_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Karank   55 pts.  |   Apr 25 2008  7:43AM GMT

yes , time is factor here also……….
we have to prioritize the job…we have to assigne less than 10 priority for them…….if we set the priority to 70 then it will get delayed or will take longer time……..

May be we can have another solution other than multithreading which is our last left option >

 
0