Incomplete planning bu SNP Heuristics

[Former Member]

Hi Friends,

I have a situation here that SNP heuristics is not completely planning all location products in background run. When these location products are selected interactively and run, heuristics generates receipt elements. Pegging area, low level codes, planning book, selection profile consistency checks are not showing any abnormalities. Just to have better usability control, a separate background planning book is used for all backgrounbd runs, which is copy of foreground planning book. There is not any errors in planning book as interactive running using background planning book is also found giving desired results. Overnight background SNP run is found incomplete with respect to location products to be seleted. One more observation is the location products ommitted by SNP are not frequently demanded items. Demand element for these are found to be from reduction in safety stock and not from forecast or sales orders. We are running 'net change planning' mode, with required global and version setting. LC and TS consistency checks are fine. One more observation is LC catalog cache hit rate here is consistently around 80% (yellow)..

Am I missing something here? Any pointers are most welcome.

Cheers!

Raghav

Edited by: Raghav_Udupa on Jun 22, 2011 6:53 AM

[Former Member · Accepted solution]

Not sure it matters, but is the setting for Heur: Order update under Global SNP settings in complete mode or delta mode?

[Former Member · Accepted solution]

Hi Sajeev,

Thank you for your response. The order update has been enabled for 'delta mode'.

Cheers!

Raghav

[aparna_ranganathan]

Raghav

My suspicion is that it could be a performance issue. Do you have parallel processing profile for the heuristic run. We had a similar issue with forecast release job. After releasing the forecast for 70% of the items the job used to end abruptly with success status- without showing any errors in the job log . We realized later that it was because of the parallel processing threads and the load on the system during that time frame.

That is if you are running other performance intensive jobs during the same time , your server / database may not be able to handle the load . So i would start by looking at the load on the system during that time frame .

Thanks

Aparna

[Former Member]

Hi Aparna,

Looks like this could be the reason. I was thinking in similar lines as we have gone live with additional business region for APO. Parallelly we started running SNP heuristics for two business regions and started experiencing this problem. How and where can we check the load on the system? Is it the live cache load and performance checks? Whos is the best to monitor this - basis?

Cheers!

Raghav

[aparna_ranganathan]

You can use /n/sapapo/perfmon. And yes basis team will be able to monitor the load on the system

Thanks

Aparna