There a variation in daily SNP optimizer job's finish time

[samir_ali]

Hi Experts,

Issue :- There is a variation in daily SNP optimizer job's finish time.

We have observed that our daily SNP optimizer job is finishing at different time. ( usually it complete at 7:15am but some time it is finishing at 6:30, sometime 6:45 & sometime at 7:10am)

I understand max time given is 80 min. to find optimum solution.

Analysed optimizer's log:-all three step's ( 1step reading master data ,2nd step execution time at optimizer server and 3rd step result to APO.)

Found that step1 :- Reading master has lot of variation some time it is taking 10 min, some time 20 and some time 25 min.

Need your input to understand why such variation is happening . Is there some system performance issue ?

What all other check I should do to avoid variation, this will help to schedule other dependant jobs and business activities.

Regards

Samir

Edited by: Samir Ali on May 19, 2009 11:49 PM

[Former Member]

Hi Samir,

1) Look into SAP note 587407 which states that in step 1 of optimisation run, the old orders of selected location products are deleted in livecache database. Followed by it carries out model consistency check and calculates solution.

2) Check SAP note 601121 for improving performance of step1

Please confirm whether this works for you,

Regards

R. Senthil Mareeswaran.

[samir_ali]

Hi Senthil

I forgotten to mention which SCM version. We are having SCM 5.0, support pack 13

So the above two notes I think will not be applicable

Regards

Samir

Edited by: Samir Ali on May 20, 2009 5:55 PM

Edited by: Samir Ali on May 21, 2009 11:08 AM

[Former Member]

Hi Samir Ali,

Looking at your scenario, there is some problem associated with reading master data. I would suggest carry out both model consistency check and livecache consistency check and look for any abnormalities.

Please let me know your observations,

Regards

R. Senthil Mareeswaran.

[Former Member]

Hi Samir,

The job runtime depends on various factors like

1) Queues in the system

2) Livecache performance at that point of time during job runs

3) Volume of data

4) Master data changes in system, (pls confirm r u running with same master data on all days, there will be delta change i believe)

5) Inconsistency in master data(like resource, ppm etc.,)

6) Nature of demand, if demand pattern changes (ie on higher side), then runtime will be more

7) Lotsize pattern

Regards

R. Senthil Mareeswaran.

[Former Member]

Hi

As the above reply says, there can be various factors that affect the read times, the main one being the Master Data queues for Livecache, there are level 1 and Level 2 queues that have to be processed before any planning run follwoing changes in Master Data. These are necessary to ensure that Livecache is consistent before it satrts planning. If you have a lot of updates to Master Data then these queues can take longer to process each time.

If you search the SAP Service Marketplace for "/sapapo/tsq*" you will see a lot of SAP Notes relating to performance of Data Queues, it may be worth looking through ones that are relevant for your SCM system and Patch Level and import ones relating to system performance improvements.

Regards

Ian

[samir_ali]

Thanks Senthil

From past two weeks I don't see much changes to masterdata.

I believe step 2 of optimizer run happens on optmz server and we have set 80 min max time. The only variation I can see in step 1.

Regards

Samir