OpenStack command line tool installation using package manager tools
This article describes how to install the OpenStack command line tools on Ubuntu, Red Hat and OS X based systems to efficiently manage various features of OpenStack. The installation instructions are based on Python's pip package. Once pip is in place, the steps are the same for all systems.
If you do not have root/administrator access to the system on which you want to run the command line clients, see the instructions about virtual environments.
Installing in Windows is also possible, but it is beyond the scope of this guide. Rackspace maintains a guide for installing the python-novaclient on Windows.
We recommend first familiarizing yourself with the Pouta web GUI and the main concepts.
Overview of the OpenStack command line tools
There is a generic command line tool called "openstack" that can be used to manage most things in OpenStack. There are individual tools for managing only some specific functionalities of OpenStack such as "neutron" for managing networks and "glance" for managing virtual machine images. In most cases, the common openstack tool should be used. The other tools are deprecated by the OpenStack project and are generally only necessary if you have scripts that depend on them or you need to run certain administrative commands.
|Tool name||Package name||Used for|
|openstack||python-openstackclient||Managing the entire OpenStack|
|nova||python-novaclient||Managing virtual machines and some limited functionality for managing e.g. volumes and images.|
|neutron||python-neutronclient||Managing virtual networks and routers.|
|glance||python-glanceclient||Managing virtual machine images.|
|cinder||python-cinderclient||Managing volumes that can be attached to virtual machines.|
|swift||python-swiftclient||Managing objects with the Swift API|
To prepare for the installation of the actual command line tools, we will install the pip and Python development packages.
We assume here that you already have Python installed. This is most likely the case if you are running any of the operating systems for which we have instructions here.
Find out if Python is installed and which version it is:
You will need Python 3 or newer in order to install the command line client tools. If you are running even a relatively recent version of your operating system, this should not be an issue. However, if you have a Red Hat based system that is older than version 7, you will not be able to install Python 3 the normal way. See the note for RHEL/CentOS 6 users under the corresponding subheading below.
If for some reason you do not have Python installed, please install it first.
Preparation: Ubuntu-based systems
If running Ubuntu 16.04 or newer:
sudo apt install python3-pip python3-dev
For versions older than 16.04:
sudo apt-get install python3-pip python3-dev
Preparation: Red Hat based systems
If running version 7 or newer:
sudo yum install python3-pip python3-devel
If you are running RHEL/CentOS 6, the latest version of Python available by default is 2.6. This is too old to run recent versions of the OpenStack client tools. It is possible to install Python 3 for these operating systems as well, but that is out of the scope of this guide. You can find information on how to install Python 3 at SoftwareCollections.org. Once you have done that, you should be able to follow the instructions for newer Red Hat based systems above.
Preparation: OS X systems
Download the latest Python setuptools"
Install the tools and pip:
sudo python ez_setup.py sudo /usr/local/bin/easy_install pip
Installing the openstack tool (all operating systems)
From this point onwards, the installation instructions are the same for all operating systems.
Optional: installation in a Python virtual environment
If you also run other Python software besides the OpenStack command line tools on your computer, we recommend using Python's virtual environments. Python's virtual environments are a way to create an isolated set of Python packages that are only available in that virtual environment. They are very convenient for keeping your Python installation clean and manageable. You will also need to use them if you wish to install the command line client tools on a computer where you do not have root access. In that case, your administrator will need to install Python, pip and the virtualenv Python package for you. For more information about virtual environments, see The Hitchhiker's Guide to Python.
Here we give basic instructions on setting up a simple virtual environment for installing the client tools. If you do not want to use virtual environments, you can skip these steps and move directly to the installation instructions using pip.
First install the python3-venv package. On Ubuntu:
sudo apt install python3-venv
Note that on Red Hat based systems there is no need of installing a dedicated package.
Create a directory for your virtual environments. We give it a specific name but you can call it something else as well:
Go to the directory and create a new virtual environment:
cd python_virtualenvs python3 -m venv osclient
This creates a new directory called "osclient" in the "python_virtualenvs" directory. This directory will contain everything you install within the virtual environment. It also contains a script for activating the environment that you need to run next:
After activation, further Python or pip commands will run in the context of the virtual environment. If you now proceed with the installation instructions below, all packages will be installed in your newly created virtual environment. You will know that the "osclient" virtual environment is activated when you see the text "(osclient)" before your prompt in the shell you are using. If you install the command line clients inside a virtual environment, they will only be available within that virtual environment.
After installation, if you want to deactivate the virtual environment, you can simply type:
Note that doing this will also mean that any of the command line tools you install will not be available for use until you reactivate the virtual environment.
Installing the client tools using pip
OpenStack provides a set of Python tools for managing various aspects of its operation. Each subcomponent of OpenStack has its own tool. It also provides the common shared tool python-openstackclient. The easiest way to install most of the command line tools is to install python-openstackclient. It should pull several other clients as dependencies. If a tool you need is missing even after installing python-openstackclient, please refer to the table at the beginning of this page for a list of packages to install.
If you followed the instructions for setting up a virtual environment, remove the "sudo" from the beginning of the commands below. As you are installing in your own virtual environment, you do not want to run the commands as root.
Install python-openstackclient with pip:
sudo pip install python-openstackclient
If you also want to install all the service-specific tools listed at the beginning of this page:
sudo pip install python-keystoneclient python-novaclient python-glanceclient python-neutronclient
Depending on your computer's setup, some dependencies might be
missing. The error message
ImportError: No module named
<module> can usually be fixed by installing the missing module
with pip and then repeating the previous, failed command. For some
errors during installation, updating setuptools or pip might help.
Upgrade a package:
sudo pip install -U python_module_to_be_upgraded
Configure your terminal environment for OpenStack
The environment variables need to be set before you can use the client. You can set them using a script provided in the OpenStack web interface. You can download this script at the following URL after logging in:
To download this in the web interface, navigate to the API Access section, and press Download OpenStack RC File v3.
Once you have the openrc script from the web UI, you can add the environment variables:
You will be asked to type in a password. Please use the password of your CSC account because the command line interface does not support yet the usage of credentials from linked accounts, such as Haka or Virtu. After doing this, the current terminal session will have the proper environment variables for using the command line tools. You need to do this again if you open a new terminal.