I've been trying to set up the Coherence processing pattern from the incubator. However I'm struggling with the configuration. Ideally I want to create a small subset of processing nodes that exist within the cluster. These node will not be storage nodes, but will have access to the storage nodes. I do not want my storage nodes to take part in the processing.
This seems to be difficult to setup. Unless I make my storage nodes eligible for processing I get an error when creating a session claiming that there are no storage nodes. I have tried both GRID and SINGLE configuration. The documentation seems to hint that I want to set up several SINGLE nodes (thus forming a cluster of workhorses).
Does anyone have any experience with this and perhaps could offer some kind of guidance?