This document represents the state of things, not any best practices. In particular, management of nodes and our network bridging setup probably need work
1. Network Bridging
Some information: https://bugzilla.redhat.com/show_bug.cgi?id=512206
We are currently using a single bridge br0 that works atop the primary ethernet. Each kvm is then given a sub-device named br0:$hostname. ClintonEbadi suspects there is something wrong with this. If any network wizards know better, please mail hcoop-sysadmin and let us know.
2. Node Management
Account clinton_admin is in the libvirt group on fritz
We only have one virtualization node now, and soon will have but two. Therefore adding admins who are permitted to manage libvirt to the groups file on each node isn't the worst thing. We probably should use a config package, domtool, or centralized groups database to make it cleaner.