Question

MRP Times in E9

  • 16 December 2020
  • 3 replies
  • 58 views
daniel-white
Userlevel 2

We have had issues for some time with MRP in that over the space of 2 weeks the time increases from around 5 to 6 hours up to 13/14/15 hours to run. We have to restart our server every 2 weeks to correct the issue.

I am thinking if i run a Trace on the locks when MRP is running it will identify any locks we may have when this is running which may help identify the issue however is there a list of tables anywhere that MRP uses when running as it may be that there are indexes that are all over the place on those tables.

 

Any Help would be appreciative.

Daniel White ERP Manager Surface Technology International Limited Hook 01256740820 ------------- EPICOR - 9.05.700C (Production) Kinetic 2022.1.9 (Test Bed)

3 replies

T
Userlevel 3

Hi Daniel

I believe from another recent post that you are running E9?

If so this is an issue I’ve experienced in multiple businesses (around 9.05.400-700). Indexes on the databases can certainly help ease the problem but I don’t believe the issue will go away.

With one of my previous businesses I spent more than 80 hours investigating the problem, trying to isolate, like you, which tables were locking or using poorly constructed indexes. To help with the investigation we also brought in our Epicor partner at the time. 

In the end we both came to the same conclusion, that the indexes and a memory leak in MRP was what was killing the process (over the years I found several other processes with similar symptoms).

The simplest, most cost effective (both time and money) solution was to schedule a weekly reboot of the Epicor environment (App servers, Database servers, service connect server). We also had a problem where by service connect would increase from 1-2 minutes to load a sales order to 10-15 minutes over the period of a week.

As the database and volume of transactions grew, the scheduled reboots went from once per week to 3 times per week as we looked to optimise the downtime with the performance gains of the reboots.

Hope this helps. Send me a msg if you have any specific questions and i’ll see if I can dig out any notes.

Tim 

G
Userlevel 3

Daniel,

We moved off of E905.700c this past January 2020.  Prior to that we had major issues with MRP running anywhere from 5-6 hours upto 20 hours.  We decided to run MRP in NetChange mode for Monday to Thursday and a Full Regeneration on either Friday evening or on Saturday to help us minimize the impact to our user community.  Even our NetChange process ran into the 14 hour range near the end of our E9 usage.

Are you seeing the increase in processing time only on the Regen process or on the Net Change process?

I can tell you that with 10.2.400.9 we are running Net Change in less than 60 minutes most nights and even our Full Regeneration process is only taking on average less than 2 hours.  It’s amazing the performance you can get from Epicor by just removing the Progress overhead!!

Glenn

Glenn Owers Sr. Business Analyst Safran Electronics & Defense Canada Peterborough, ON Canada Version: 2021.2..23
daniel-white
Userlevel 2

We have the issue on both although we only run ReGen on a Sunday Afternoon/evening, we run Net Change Monday Night to Thursday Night and even with a day of no changes that would cause MRP changes the time increases.

Daniel White ERP Manager Surface Technology International Limited Hook 01256740820 ------------- EPICOR - 9.05.700C (Production) Kinetic 2022.1.9 (Test Bed)

Reply


Terms & Conditions