cancel
Showing results for 
Search instead for 
Did you mean: 

Data Load (Info package and DTP) performance

Former Member
0 Kudos

Hi,

We are planning to perform initialisation expecting 30 million records.

To improve the data load performance i am planning to increase data package size from default value to double the default value

Will that improve the performance

Accepted Solutions (0)

Answers (4)

Answers (4)

0 Kudos

Hello,

If the issue still could not be resolved, please have a look at th recently released guided answer for DTP perforamance:
Guided Answers: Data Transfer Process (DTP) performance/memory issues

Best Regards,
Jiandong

0 Kudos

Hello,

If the issue still could not be resolved, please have a look at th recently released guided answer for DTP perforamance:
Guided Answers: Data Transfer Process (DTP) performance/memory issues

Best Regards,
Jiandong

RamanKorrapati
Active Contributor
0 Kudos

Hi,

As per the data volume, you better to work on this activity on weekends or where there is no business or no burden on source/bw servers.

1. You can play data packets sizes,parallel processes and run in background.

2. While running your loads, ask basis team to keep on monitoring at source side T code - SM58

3. Check about existing application servers to process your requests.if its needed, ask basis team to increase more

4. Better to run full loads with selections in multiple info packs.

5. later you can do init without data transfer

6.later you can set delta settings.

7. for LO loads you need to delete and refill setup tables.

8.Even you need to lock source system other wise leads to duplication.

Thanks

Former Member
0 Kudos

Hi,

What is the source your extraction? is it LO extractors or others?

Please let us know.

Thanks

Amit

Former Member
0 Kudos

Source is LO extractor

Former Member
0 Kudos

Adjusting the info package size would be helpful...

But don't go with INCREASING the numbers, It's all depends on the your extract structure size, source system memory settings.

I would suggest to do some trail in DEV/QUALITY systems with enough data.

I remember playing in production itself for CRM - 0BPARTNER extractor - we actually reduced the number of records to be extracted per package to improve the performance - we did see the impacts.

First calculate how much is your current package size by number of records extracted per package now with your extract structure size..and then play around..sure it will help unless you DON'T have any excessive ABAP code user exits.

Let us know if you need further help on this