Deployment/Storage Strategy for Cube using Multiple Dispatcher Environment

[javelina93]

We have a cube which is built on a daily basis, call it daily.mdc. We need to deploy the cube to a C10 BI environment which has 3 servers where each server is running a separate but clustered dispatcher.

Does each dispatcher need it's own copy of the cube? In other words, do we need to deploy and activate the cube on each server running a dispatcher? Or can we create a shared folder that is visible to all 3 servers and deploy and activate once from one of the servers?

Are there other options?

Thanks in advance.

[tjohnson3050]

You can create a shared folder that all three dispatchers can see and use the unc path (\\servername\sharename\file.mdc) to the mdc file in the datasource connection. 

When you test the datasource connection, you have the option of testing it from all the dispatchers in the environment, just to be sure each one can connect.

[bdouglas]

I am wondering how this solution worked?  I have always been concerned about file locks on the cubes if there are multiple servers accessing the same mdc.

[Rahul Ganguli]

Dispatchers are only going to read data from cubes hence a concurent access to same cube from different dispatchers does not causes any issue.