Hi, please see mongopi.wordpress.com for background and progress to date.

I am trying to build a scale able, ultra low-cost BI solution (eg load up some data via a recurring ETL process (Pentaho), create some data structures for analysis (aggregation) and produce some nice reports!). This is my MSC project a the University of Dundee - my tutor is doing something similar with Cassandra.

I have an array of 5 Raspberry Pis (512MB), 32GB SD cards. I have successfully installed RickP's fork of MongoDB (2.2.1) via GitHub.

Additionally, I have a Dell Dimension 5100 (4GB RAM) which I intended to use as the 'master/control' machine. This is running the latest version of MongoDB for Linux.

I am now struggling with 2 things
1) 'Network config' ie Setting ports/static IPs on each device (so that they can all 'talk to/see' each other). I assume I need to add some lines into/create a config file, somewhere? (The R Pis are running Raspbian, the Dell Debian 6.0)
2) What 'roles & config' I need to create/the topography of the cluster - eg config servers? Can I create some 'roles' via virtual machines too
3) How to kick everything off and get it running/working.

Thanks for any help.
KR
Stuart