cancel
Showing results for 
Search instead for 
Did you mean: 

CIF_IMAX buffer invaldations casuing performance issues in ECC

dennis_crampton
Explorer
0 Kudos

In our ECC 6.0 system when we run the CIF process to move either new accounts or new materials to one of our two EWN ( Extended Warehouse Management) systems we see performance issues in ECC on the table CIF_IMAX while the table is invalidated in the app servers. CIF_IMAX is delivered fully buffered and it is optimized and compressed by SAP to improve performance and most of the time it performs well.

It appears when the jobs run to process the data to our EWM systems (2), reports RIMODAC2 and RCPTRAN4, the table CIF_IMAX is invalidated from the Application server buffer there is a performance hit to order entry and other processes while the application server buffers are generated.

If appears that even if there is no change in the CIF models or in the table CIF_IMOD the buffered table is invalidated. It appears this is done via function module. it is explicitly being invalidated in the FM CIF_IM_MAXVECTOR_BUILD. See lines 213 to 235.

We found some relief from OSS note 563359 - Performance optimization for tables with LOB columns". We buffered CIF_IMAX in the Oracle database. We are still seeing users via /nSM66 waiting on table CIF_IMAX while the buffers are being invalidated and rebuilt.

I don't think the increase in buffer parameters will stop the invalidations of the table CIF_IMAX because the invalidation isn't happening due to space shortage but via the Function Module invalidating the buffered version on every run.

We are running the CIF consistency check and all jobs defined in the best practices document as prescribed. We have a very busy system and we have reduced the runs of the CIF process during the day to the minimum runs to allow us to meet our service level goals.

I wonder if anyone else has experienced this issue and if so what was done to resolve this situation?

We have consider un-buffing the table CIF_IMAX and allowing the program to read for the database directly for each call but that appears to be a performance issue, OSS note 1376176.

Any other recommendations?  We have an open OSS message but seem to be going back and forth.

Thanks,

DC

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Hello Dennis,

Did you run any Data extraction program from your ERP?

Do you see the Object type T_IMAX in CIF_IMAX Table?  If not, please run the RCIFIMAX program , copy all the Object types which you current use and Check the "generation" Indicator.

Run this in the background, once you run this you can see an entry in CIF_IMAX as T_IMAX without any Logical system.

This will solve your performance issue.

It is important to check this T_IMAX if any data extraction program is deleting this especially when the FM CIF_IM_MAXVECTOR_BUILD being used.

hope it helps.

former_member906762
Participant
0 Kudos

Hello Dennis ,

              We are having a similar issue , do you have any resolution for this issue .

Regards,

Nick