It makes virtualization possible
Part of nova must be running on every node in the cluster
We have 7 different nova component, here are some:
1. nova-api
interface to nova
maintain openstack compute DB
compute DB holds existence info of a vm, where vm is running, owner, IP, storage assigned to vm
2. nova-compute
start & stop VMs on hosts (because of that must be running on every virtualization node)
3. nova-scheduler
responsible for provisioning (decide where new vm are launched)
know available resources and load of virtualization nodes
it’s implemented centralized
4. nova-novncproxy
HTTP-capable proxy server that make vm console available through the dashboard
vnc proxy server typically runs on an api node or several nodes behind LB
5. nova-conductor
it is DB abstraction layer
used because of 1st. security reason and access control 2nd. easier for live upgrade
permit all compute instances to function without direct access to the central DB
CLI
openstack server list | create | stop | delete
Config Files
/etc/nova/api-paste.ini /etc/nova/nova.conf #containing all config options for all nova component