Adding an additional Enterprise Server (with a twist)

gregzin

Active Member
Right now we have one Enterprise Server, (Unix HP9000) that also holds our database (Oracle) We are planning on purchasing another HP9000 box and wish to use it for a Stand By Database, essentially a mirror of our production database that will be about 30 minutes behind in updates. We would like to also use this extra server to process long batch jobs, having the processing occur on the new server and the database requests happen on the new server as well.

My thinking is this:

Install the Prod path code to this new server as well. In the server map for the new server, point the data sources to the connect string of the stand by database. Use OCM mappings to configure the UBE's that we wish to run on the new server.

Does this make sense and would it work? Is there any other way to configure where UBE's will process? Could we setup a queue that will direct anything submitted to this queue to the new server?

Thanks for your help,
 
Greg,

will these batch jobs update data or just be reports with no updates?
 
Greg,

I think you might be confusing the issue. I would not try to run the ube's against the standby database. The reason being that you may find it more complex than you think.

In order to access a standby database, you need to take it out of standby mode and open it as read only. If the ube's are long running, you will fall behind in applying your logs. This may or may not be a problem. Also when you log on via OneWorld, there are some updates which occur to tables, so you would need a complex environment mapping where some data sources (or tables) are configured to look at your production environment in order to make this work.

A simpler solution would be to treat the server as 2 separate functions. ie a server which simply hosts your standby database (with OneWorld environments configured to enable failover in a disaster). Then install OneWorld as a separate application server. You could still run ubes on this application server with the data being retrieved from the production database, relieving the pressure on your existing application server.

Personally, I think this would be a better approach.

Regards
Marty
 
Greg,

will these batch jobs update data or just be reports with no updates?

Hi Larry, I should have specified that my intention was to run no update reports. Sounds like this is more complicated than I thought?
 
Back
Top