[one-users] Multiple OpenNebula Servers

[Lars Kellogg-Stedman]

> Does anyone have any experience with this type of configuration or is it
> even possible with OpenNebula?

We were just discussing this model for our use earlier today as a way
of working around the lack of a granular authorization model.  Our use
case is that we want to have some networks only available to our IT
staff, while making other networks available for course and research
use.  At first glance, there's nothing to stop you from deploying
multiple OpenNebula servers pointing at the same backend hosts, but
you're going to find it difficult to control resource (primarily CPU
and memory) utilization since the two instances won't be aware of each
other.  In theory you could adjust for this by limiting the total
resources available to each server, but I'm not sure off the top of my
head if this is actually possible.

We may ultimately go with shared accounts (e.g., everyone in IT
authenticates as the "IT" user, and everyone in CS50 authenticates as
the "CS50" user) -- unless the forthcoming release addresses some of
these issues.

-- 
Lars Kellogg-Stedman <lars at seas.harvard.edu>
Senior Technologist
Harvard University SEAS
Academic and Research Computing (ARC)