topic Re: Performance difference in AUFM table. in Application Development and Automation Discussions

Performance difference in AUFM table.

jogeswararao_kavala — Fri, 05 Jul 2013 02:38:32 GMT

Dear ABAP Experts,

Background

After recent client copy from PRD 400 client our QAS 400 client is having identical database upto 31st January 2013.

Issue:

Present issue is about the drastic performance difference of a report pragram in the QAS and PRD servers.

This program takes few milliseconds in QAS and about 5minutes in PRD, with the same set of selection-parameter values, both returning identical values in the ALV.

Investigation:

While investigating, I have zeroed down to a the main table used in the program ie., AUFM and through SE16 with the same selection-set referred above, I have tried. The same experience.

So now the focus shifted from the report program to the AUFM table.

The key fields used in selection-set. (Both in program as well as querying through SE16)

MJAHR BUDAT BWART MATNR (Obvious that the BUDAT date range of selection-set is before 31stJan2013)

Expectation from forum:

Want to know the reason behind this.

The fact that at any point of time, load on AUFM table in PRD is much more than that of QAS, should make such a difference!

What can be done to rectify this by an ABAPer?

What can be done to rectify this by our BASIS?

Thank you for going through these many line of query.

Jogeswara Rao K

Re: Performance difference in AUFM table.

Arun_Prabhu_K — Fri, 05 Jul 2013 02:45:47 GMT

Hi Jogeswara Rao Kavala,

One reason may be the secondary index created for the table AUFM might not have been transported to PRD from QAS. Kindly check that.

Regards.

Re: Performance difference in AUFM table.

Former Member — Fri, 05 Jul 2013 03:18:52 GMT

Hi Jogeswara,

Assuming that you have run SQL trace (ST05) on your report and anlaysed that AUFM is the culprit, Can you please post the select on AUFM from your report. It might be the case that we may not be hitting key or secondary index. If this does not suffice and if we have number ranges used we can try leveraging the same.

BR.

Re: Performance difference in AUFM table.

raviahuja — Fri, 05 Jul 2013 03:41:16 GMT

Hi Jogeswara,

You said that same experience you had with SE16 as well, this means report or query you have provided has nothing to do with this.

Then the to-do points can be:

1. Secondary Indexes (if any) not created in PRD or got deleted. Compare the indexes of table in QAS to PRD.

2. Compare the technical settings of table in both systems.

3. Do ST05 - SQL trace on table for both systems and identify the pain point.

Regards,

Ravi

Re: Performance difference in AUFM table.

former_member209120 — Fri, 05 Jul 2013 04:38:56 GMT

Hi Jogeswara Rao,

We too faced this problem,SAP recommended this notes, this may help you

0000038855 Info on settlement: Performance improvements

0000099508 CO88: Settlement of order groups

0000308513 Variances on completed orders

0000372197 SETTLEMENT: Performance problem with cancellation in COPA

0000397799 INFORMATION: CO-OM-OPA (Order & Project Accounting)

0000400506 Termination in settlement of product cost collector KD557

0000418021 SETTLEMENT: Performance problem with reversal in COPA II

0000545932 Performance probl. in CO-PC-OBJ period-end closing: DB hints

0000740626 Cost Object Controlling: Performance Objectselection

0000874671 Service SAP BPPO for period-end closing

0001054121 The SAP Ecosystem in a Nutshell

0001058528 Performance during billing with KO8G or CO88

0001079126 Performance probl. in CO-PC-OBJ per-end closing: Oracle Hint

0001457783 Settlement: Update termination SAPSQL_ARRAY_INSERT_DUPREC

0001481867 Poor performance when checking settlement rule

0001791724 Long runtime of order settlement

Regards,

Ramesh.T