Running multiple copies of an App on one NetWeaver server

[Former Member]

Hi there, (V12 with ERP/ECC transactions)

We are setting up our landscapes/environments for development, beta testing, final QA and production. I am wondering if we can combine development and beta testing on the same system.

The obvious concern is that they will have copies of the "same" transactions. If we don't do anything special, the data source name is hard coded in the transaction and as we are on the same MII server, they can't point to different data sources unless we edit the transactions. Is there a way to have a transaction use a configuration of some sort to dynamically specify the actual data source?

Are there any other issues attempting to do this?

Thanks,

--Amy Smith

--Haworth

[jcgood25 · Accepted solution]

Assuming the server is up to the task, just go ahead and setup 2 NW instances on the same machine. They will each run on their own port number and have their own unique SID DB connection. You can mirror projects and configuration (export from current system and import into newly installed instance), then repoint your data sources and SAP server aliases to their corresponding mirrors in your SAP landscape (if available).

By having this mirror you really don't have to do anything in your application content. Promoting from one instance to the other is as simple as export and import.

Regards,

Jeremy

[Former Member · Accepted solution]

Thanks Jeremy,

Thanks for the suggestion Jeremy.

Our server is not capable of another instance. It is not easy to get our SAP Basis people to install and configure a new system and we are not allowed to do it ourselves. I was trying to find a way to use our existing instance of NetWeaver/MII to run multiple copies of the application. In J2EE apps I have worked on, it is as simple as copying the software, changing the URL, changing the application config files for the data source and deploying the new duplicate app under a different name.

Seems that in MII the transaction "hard" links the data source name to the data source definition and that link is not dynamically configurable. Would that be a fair statement?

Thanks again,

--Amy Smith

--Haworth

[Former Member · Accepted solution]

While the data source configuration is not dynamic, the name of the data source you would like your objects to use can be passed dynamically at runtime.

[Former Member · Accepted solution]

Hi Christian,

Cool! So how does this work, or where is a reference that discusses this technique.

Thanks,

--Amy

[Former Member · Accepted solution]

This is a perfect use case for VMWare...

[Former Member]

Thanks to all.

--Amy