cancel
Showing results for 
Search instead for 
Did you mean: 

performance problem with UPC_BUNDLE_EXECUTE_STEP

Former Member
0 Kudos

I am trying to use UPC_BUNDLE_EXECUTE_STEP program to execute global planning sequence to copy data from actual's cube to plan cube. I followed the steps in OSS note 546464. For the partition characteristics, I give the rebate number and from & to range.

When I execute the program in the back ground, it takes close to an hour in our development box. If I run the planning function in BPS0, it completes in 5 minutes. In our quality, if I split the range in the BPS0, I am able to get the data loaded in 2 hours for all rebates. Using UPC_BUNDLE_EXECUTE_STEP , ran for 2 days and did not complete.

Is there anything that I needed to do the speed up the processing.

Thanks,

VJ Sudharsan

Accepted Solutions (0)

Answers (5)

Answers (5)

Former Member
0 Kudos

Hi Chris,

The answer is yes for all the points.

Former Member
0 Kudos

Thanks everyone for the reply.

Even if I give a small range, it takes considerably more time to run the global function using UPC_BUNDLE_EXECUTE_STEP than if I run the function independently. This is at the time when I am the only one who is in the system.

UPC_BUNDLE_EXECUTE completes much quicker but I can not use as I get a short dump due to memory overflow in the internal table.

Thanks,

Former Member
0 Kudos

Just to check the simple things:

Is your chosen 'partition' characteristic existing in every level contained in your global planning sequence?

Is the partition characteristic checked as available for selection in the package in every level used?

Is your partition characteristic set up with master data (check in trx RSD1)?

Former Member
0 Kudos

Generate export data source and load. Easy, fast and works fine in background.

Ravi Thotahdri

Former Member
0 Kudos

Hi,

If you want to execute a global planning sequence, why do not you use UPC_BUNDLE_EXECUTE directly ?

Regards

Laurent

Former Member
0 Kudos

1. UPC_BUNDLE_EXECUTE_STEP runs for each value in the master data for the characteristics used to split, irrespective of whether there is any transaction data in the cube or not. Try to restrict the range or create a job with multiple steps, each step using a value for the characteristic that has transaction data.

2. Background jobs typically run with a lower priority (usually C) whereas foreground jobs (BPS0) run with higher priority (A). Try running the background job with higher priority.

Regards - Ravi