100,000 Customers in RMS

Hi Butch,

We have a client who has 450,000 customers in their database and here's what we've done.

1) They are not using a standard HQ situation but instead have a customization that allows Great Plains as HQ so I'm not sure how this would affect HQ syncs - I know HQ does A LOT of queries even when there are only a few new customers in HQ

2) To improve performance we have a scheduled task to re-index the database daily - this helped very much.

3) Make sure you have plenty of RAM in the database server and the registers. We never send ANY machine out the door with less than 512 megs of ram and are now using dual core processors exclusively. We've seen a dramatic performance improvement on RMS database servers with the dual core processors versus same speed hyper threaded processors.

4) They are currently using the customer find window instead of the standard lookup but it's very easy for them to accidentally search for all records and RMS doesn't have much for fine tuning searches in such a large database so we are customizing the customer lookup to make it more fool proof (validating and requiring some customer data and also not allowing them to perform a search that will return all records) and allow them to do things like search in the first and last name fields at the same time. In our testing searches will be preformed much more quickly and will also be easier for cashiers.

Hope that helps!

Rick Feuling Retail Information Technology Enterprises (RITE)

"Butch" wrote:

As far as getting the customers loaded, you could break the batch up into smaller sets - maybe by zip code or alphabetically by name. You would need to let each batch flow through the system - at least two 401s per store before starting the next. Break the batch by using SQL to set only selected customers to Global at a time, or if you are loading them into HQ directly, just split it into smaller files. It sounds like you already know how to do that, so I'll leave it there...

As far as operating with that many customers... Ouch! You may have to add RAM at least to the servers. SQL Server can easily handle the data, but RMS is not good at displaying huge datasets like that - you may have to change some things operationally. Set all of your stores to use Find for customers - the customer list is going to be useless.

Glenn Adams Tiber Creek C> We are looking at bringing our biggest store onto RMS. They have just over

Thanks to both of you, that is the information I was looking for. I like the idea of setting the customers to local as a way to batch them to the stores. I didn't think of that.

Involving GP is not an option for us as we have JD Edwards/Peoplesoft/Oracle Entrprise One as our ERP/Accounting software. I have already customized an integration to that system. RMS is strictly for our retail stores.

I loaded the customers straight into the Store DB to see how the lookup would work. I ran into the problems you guys mentioned with the customer lookup vs customer find. Find was still pretty slow on my test machine as it only has 256MB RAM. the 512 Machine was much faster, but not blazing.

Thanks again for the replies.

Hi Butch,

We have a client who has 450,000 customers in their database and here's what we've done.

1) They are not using a standard HQ situation but instead have a customization that allows Great Plains as HQ so I'm not sure how this would affect HQ syncs - I know HQ does A LOT of queries even when there are only a few new customers in HQ 2) To improve performance we have a scheduled task to re-index the database daily - this helped very much. 3) Make sure you have plenty of RAM in the database server and the registers. We never send ANY machine out the door with less than 512 megs of ram and are now using dual core processors exclusively. We've seen a dramatic performance improvement on RMS database servers with the dual core processors versus same speed hyper threaded processors. 4) They are currently using the customer find window instead of the standard lookup but it's very easy for them to accidentally search for all records and RMS doesn't have much for fine tuning searches in such a large database so we are customizing the customer lookup to make it more fool proof (validating and requiring some customer data and also not allowing them to perform a search that will return all records) and allow them to do things like search in the first and last name fields at the same time. In our testing searches will be preformed much more quickly and will also be easier for cashiers.

Hope that helps!

Rick Feuling Retail Information Technology Enterprises (RITE)

"Butch" wrote:

Rick, Could you please share on how you schedule a reindex? I've been trying to figure out how to do this and haven't got very far. Thanks Craig "Rick Feuling [RITE]" wrote in message news: snipped-for-privacy@microsoft.com...