Hi there,
we are facing an increase of the MRP runtime of about 20-25% in last last few weeks. The runtime in the past has been very constant. We are wondering what may be the reason for this. There have been no changes in BOM structure, no changes in the performance in general (runtime of other jobs did not change) and, to my opinion, no significant increase of orders.

Does anybody know which data (tables) may be responsible for a longer MRP runtime?

Any help is highly appreciated.

Martin

Hello Martin,

we had the same problem. We changed the parameter regenerative to net change and startet MPS manually, after that run the time was the original one. Baan support couldn't find a solution.

A second solution is to run the MPS in littel pieces, but this helps only sometimes.

Regards Torsten

Ran any traces ?

tables involved in MRP:
mrp010 (which is dropped in a regenerative run)
mrp020 and mrp021 (also dropped)
mrp030 and mrp031 (....dropped)
If there aren't more orders than usual, this won't be the problem..

Maybe the capacity planning is killing performance
(crp020 or was it crp200...I forgot....)

On the other hand, itm001, and inv150 are also updated....might be wise to run some traces on those tables during the MRP (maybe inv150 has increased significantly lately, or lots of 'locks' on one of those tables during MRP ?)

Hi just_fro,
you are correct. I have a significant growth of the mrp tables (fortunately we are monitoring the table size :D ). The growth matches with the longer runtime of the MRP. This is a great step forward - thank you. Currently I have no idea what might cause this growth, but I think this can easily be achieved with some additional monitoring on these tables.

Thanks again and best regards,

Martin

The more daily buckets that you have will cause your runtime to increase as it creates a record for every Plan Item for Every Daily Bucket. You need to determine How many days that you need to actually look at in daily buckets and then look at using some weekly buckets and then monthly buckets.
At one client that was having long MRP/MPS Run Times we changed the setting to 28 daily buckets (4 Weeks) then 9 weekly buckets and and 9 monthly buckets to give them 1 year. If you have a test system this is something you should test.

Smaller order intervals could cause the problem.

Did you fill any data into MPS Demand forecast, we had an issue of significant time increase of MRP generation runtime, when we filled Demand forecast into MPS. We are currently testing Performance boosters for MRP batch job