Development Quickstart

This page describes how to setup and use a working Python development environment that can be used in developing nova on Ubuntu, Fedora or Mac OS X. These instructions assume you’re already familiar with git.

Following these instructions will allow you to build the documentation and run the nova unit tests. If you want to be able to run nova (i.e., launch VM instances), you will also need to — either manually or by letting DevStack do it for you — install libvirt and at least one of the supported hypervisors. Running nova is currently only supported on Linux, although you can run the unit tests on Mac OS X.


For how to contribute to Nova, see HowToContribute. Nova uses the Gerrit code review system, GerritWorkflow.


There are two ways to create a development environment: using DevStack, or explicitly installing and cloning just what you need.

Using DevStack

See Devstack Documentation. If you would like to use Vagrant, there is a Vagrant for DevStack.

Explicit Install/Clone

DevStack installs a complete OpenStack environment. Alternatively, you can explicitly install and clone just what you need for Nova development.

The first step of this process is to install the system (not Python) packages that are required. Following are instructions on how to do this on Linux and on the Mac.

Linux Systems


This section is tested for Nova on Ubuntu (14.04-64) and Fedora-based (RHEL 6.1) distributions. Feel free to add notes and change according to your experiences or operating system.

Install the prerequisite packages.

On Ubuntu:

sudo apt-get install python-dev libssl-dev python-pip git-core libxml2-dev libxslt-dev pkg-config libffi-dev libpq-dev libmysqlclient-dev graphviz libsqlite3-dev python-tox python3-dev python3 gettext

On Fedora-based distributions (e.g., Fedora/RHEL/CentOS/Scientific Linux):

sudo yum install python-devel openssl-devel python-pip git gcc libxslt-devel mysql-devel postgresql-devel libffi-devel libvirt-devel graphviz sqlite-devel python3-devel python3 gettext
sudo pip-python install tox

On openSUSE-based distributions (SLES 12, openSUSE 13.1, Factory or Tumbleweed):

sudo zypper in gcc git libffi-devel libmysqlclient-devel libvirt-devel libxslt-devel postgresql-devel python-devel python-pip python-tox python-virtualenv python3-devel python3 gettext-runtime

Mac OS X Systems

Install virtualenv:

sudo easy_install virtualenv

Check the version of OpenSSL you have installed:

openssl version

The stock version of OpenSSL that ships with Mac OS X 10.6 (OpenSSL 0.9.8l) or Mac OS X 10.7 (OpenSSL 0.9.8r) or Mac OS X 10.10.3 (OpenSSL 0.9.8zc) works fine with nova. OpenSSL versions from brew like OpenSSL 1.0.1k work fine as well.

Getting the code

Once you have the prerequisite system packages installed, the next step is to clone the code.

Grab the code from git:

git clone
cd nova

Building the Documentation

Install the prerequisite packages: graphviz

To do a full documentation build, issue the following command while the nova directory is current.

tox -edocs

That will create a Python virtual environment, install the needed Python prerequisites in that environment, and build all the documentation in that environment.

Running unit tests

See Running Python Unit Tests.

Note that some unit and functional tests use a database. See the file tools/ on how the databases are set up in the OpenStack CI environment and replicate it in your test environment.

Using a remote debugger

Some modern IDE such as pycharm (commercial) or Eclipse (open source) support remote debugging. In order to run nova with remote debugging, start the nova process with the following parameters –remote_debug-host <host IP where the debugger is running> –remote_debug-port <port it is listening on>

Before you start your nova process, start the remote debugger using the instructions for that debugger. For pycharm - For Eclipse -

More detailed instructions are located here -

Using fake computes for tests

The number of instances supported by fake computes is not limited by physical constraints. It allows you to perform stress tests on a deployment with few resources (typically a laptop). But you must avoid using scheduler filters limiting the number of instances per compute (like RamFilter, DiskFilter, AggregateCoreFilter), otherwise they will limit the number of instances per compute.

Fake computes can also be used in multi hypervisor-type deployments in order to take advantage of fake and “real” computes during tests:

  • create many fake instances for stress tests
  • create some “real” instances for functional tests

Fake computes can be used for testing Nova itself but also applications on top of it.