SAP Central Finance & managing performance issues

[nitin_gupta10]

Execution part of SAP Central Finance project is always a issue when it comes to system performance. There are several Initial Loads to be executed and all of them cause lot of performance issues specially on source system. In this blog I will discuss some options to overcome those issues but of course on the relevant projects and depending on the situation and size of the customer system you may need more things than listed in this blog.

List of performance intensive topics

1. AUFK Load via SLT - In this the Order master data is loaded from ECC to S/4. Although you can limit the data volume via SLT filters (by controlling area, company code, order type etc) but still you will see issue as the process works in 2 steps
   
   
   
   1. First the AUFK read starts when the action is triggered from SLT. In this system will read entire table and filters doesn’t come into picture. This action will take time based on data volume but it puts heavy load on source systems
   
   
   2. When the read is complete the second action comes for loading the data in target system. Now the load of source system is released and the performance issue is shifted to CFIN system where almost all work processes will be consumed as AIF will be flooded with the number of messages (1 message per order)
   
   
   
   

During this process ensure that the reading type is always 5 in SLT and also the number of records are mentioned in the below option.

For example if there is a table with 108 million records and you specify a value of 20 million for the table, the system will generate 6 access plans which can run in parallel. The first 5 access plans will handle 20 million records each, while the last access plan will handle the remaining 8 million records. To define the value for parameter Number of Records, you can divide the total number of records in the table by the number of access plan that you want to run in parallel.

2. AVL Data load via SLT- This applies to all AVL load as well which includes 4 tables from source and data is filled in delta tables in CFIN. Filters applied can be Purchase Organization, Sales Organization, Creation date etc

Read more here about AVL - CLICK HERE

 

Document Type	Table
Customer Invoice	VBRK
Sales Order	VBAK
Supplier Invoice	RBKP
Purchase Order	EKKO

3. FI Extract via RFC – In this process the data is extracted from source system and is filled in staging tables- CFIN_ACCHD, CFIN_ACCIT etc. This is done by Company code by standard SAP Initial load process. SLT is not in place in this process. Number of jobs should be carefully given when this is triggered via CFINIMG. Always good to starts this process during end of week and avoid month end window. To have this process smooth it is recommended to
   
   
   
   1. BSIS, BSAD tables should be indexed
   
   
   2. BKPF, BSEG, BSIS, BSAD, BKIS, and BSAK table should be reorganized
   
   
   
   

4. FI Load via RFC – This process starts posting in Central Finance. All load is focused in CFIN so relevant resources should be available to ensure system is not going down.

5. Co Preparation via RFC– this is the Marathon step where Data is read from COBK and COEP and once read the table CFIN_COPA is updated or rather exploded with data. This process can result in billions of entries in CFIN_COPA. Preparation job does not have much parameters to be given – only number of jobs and package size. Read this blog for more on CO postings – CLICK HERE

 

6. CO Load via SLT– Data prepared in previous step is loaded to CFIN system in ACDOCA table in this step. Messages get flooded in AIF so need to ensure that sufficient system capacity is available in CFIN system when this process is running

Over all – Source system sizing should be reviewed, work processes should be increased, and preferred is to have additional application server to prevent issues in business as usual. Also it is important to check the custom programs in ECC which may result in performance issues in normal system behavior.

It is not recommended to do load of derivations and enhancements in ECC as when the data is being prepared for Initial load there is already a heavy volume running from tables to tables and if any more custom solution comes it may make situation bad.

Also, it is proven that standard SAP solution works for executing Initial load. I have myself done this will millions and billions of line items and from multiple heavily loaded systems. Its just about doing right thing and following the recommendations of the experts

Many SAP notes are available to manage performance issues – Refer 2741509, 2718730, 2716000, 2715002, 2585653, 2611949, 2754709, 2749549 and many more are available

 

CFIN Master note - 2184567

Enjoy the load..