Compute services in OpenStack are provided by a project that goes by the name Nova. Nova is an API-driven system that manages physical and virtual compute resources in an OpenStack cloud, providing Infrastructure as a Service (IaaS). OpenStack operators, administrators, and users can leverage the Nova API to manage the life cycle of compute resources.

Nova is primarily responsible for managing two resource types: instances, which are the running virtual machines, application containers, or even full bare-metal machines a user has requested, and hosts, that provide the hardware resources required by instances. In most circumstances, instances are synonymous to virtual machines.

It is important to make a distinction between two main features of Nova: the Nova API service (and associated services such as the nova-scheduler, nova-conductor, and nova-placement), which runs on our cluster of three controller nodes, and the nova-compute service, which runs on each compute host in our environment.

In this chapter, we will first cover the administration of the physical host and Nova services, then the remainder of the chapter will take a task-centric approach to working with Nova. This means that the depth and detail on Nova features and functionality are beyond the scope of this book.