on 05-03-2006 4:29 PM
Hi
We have a demo repository that we are running on my laptop.
We have created a batch that add 10000 records to the repository using the API.
On my laptop 1.7GHZ machine with 2MB MEM its working nice.
we have loaded it to a server 3.2G Xenon dual CPU with 3GB MEM. and now the MDM server is working very slow.
I can see that when i am running my batch code the CPU is on 1-5% only but on my laptop 100%.
Some one know what maybe the problem holding the MDM server to run 100% of the CPUs?
Thanks
Noam
Hi,
Based on my experience, usually the oracle database will help you achive a better performance when managing big volumens of information and transaction. One of the reasons is that in Oracle you can configure a big set of parameters and in sql you have to stick to what sql natural optimizer can do.
It is also interesting to mention that because of the mdm nature (that is to sinch a lot of information from a lot of systems, being sinchronize often the same time), you need to asure that readers dont block writers and viceversa. In sql server there are block locking issues and something called locks scalation that kill performance. Oracle concurrency model is very good handling this so that readers dont block writers and viceversa.
Hope it helps.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
hi,
As far as my knowledge goes there is no separate
repository for IT scenarios, GDS and catalog publishing. GDS stores the trade item data in MDM repository iteself.
Though GDS is an application built upon MDM platform and uses hence its repository this repository can not be used for other purposes than GDS. In other words, this is a private GDS repository and if you want to leverage product data for other purposes you have to build another MDM repository.
In regards to IT scenarios it actually depends on your business scenarios. It may be one repository used for harmonization and central data management or two different. Anyway it's one repository per object family (Customers, Products, Employees and so on
hope it helps. Assign points if its useful.
thanks
Varun CN
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
deleted
Message was edited by: Marina Sinichenko
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
How many qualified tables are there in the repository?
Are you adding the qualified table records thu the batch process ?
What database you are using?
These are my observations from my previous projects.
1.More No of qualified tables import will decrease the performance.
2.Oracle is slow when compared to SQL Server. Even more slow when Database,MDM Server and Import Batch Process are in different machines.
Network connection also might be one of the root cause.
Thanks and Regards
Subbu
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
88 | |
10 | |
10 | |
9 | |
7 | |
7 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.