Chapter 6. Running PackStack

There are several different methods of running packstack and using it to deploy OpenStack. If you just want to deploy an OpenStack environment as quickly as possible for evaluation purposes then refer to Section 6.1, “Quick Start Deployment using PackStack”.
These methods of running packstack are provided for more complex deployments:
When run interactively packstack provides prompts for entry of each configuration value required to complete deployment. Alternatively the user may accept the provided default value.
Refer to Section 6.2, “Running PackStack Interactively” for more information on running packstack interactively.
When run non-interactively packstack expects an "answer" file to be provided as a command line option. This file contains the desired settings for all configuration values that are required to complete deployment.
Refer to Section 6.3, “Running PackStack Non-interactively” for more information on generating an answer file and using it to run packstack non-interactively.


To deploy OpenStack using packstack each machine targeted for deployment must be configured to allow access using the account of the root user over SSH on port 22.


By default packstack will configure a volume group named cinder-volumes on the system targeted for volume storage (Cinder) deployment if one does not already exist. This volume group will be backed by a loopback device and is not appropriate for production use.
If you intend to use physical storage for the cinder-volumes volume group then you must create the volume group in advance on the system to be used for Cinder.


It is strongly recommended that each compute node has two network interfaces available. One for access to the public network and one for the internal Nova network. While it is possible to use a single interface for both purposes, this approach may result in virtual machine instances obtaining addresses from the wrong DHCP server.

6.1. Quick Start Deployment using PackStack

The quickest way to deploy an OpenStack environment using packstack is to provide a host, or list of hosts, on the command line. The first host listed will be deployed as a compute controller node, subsequent hosts will be deployed as compute nodes.
When using this deployment method packstack will use default values for all other deployment options unless they are overridden on the command line.
For a list of available command line options refer to Table 6.1, “PackStack Configuration Keys”.

Procedure 6.1. Quick Start Deployment using PackStack

  1. Run packstack with the --install-hosts parameter. The parameter expects a comma separated list of IP addresses.
    $ packstack --install-hosts=CONTROLLER_ADDRESS,NODE_ADDRESSES
    Replace CONTROLLER_ADDRESS with the IP address of the system that you intend to use as a compute controller. Replace NODE_ADDRESSES with IP addresses of the systems that you intend to use as compute nodes.
    Alternatively provide only a single IP address to deploy an "all in one" OpenStack installation, even on the system from which you are running packstack.

    Example 6.1. Single Node Deployment

    In this example packstack is instructed to deploy an "all in one" installation to the system with IP address
    $ packstack --install-hosts=

    Example 6.2. Multiple Node Deployment

    In this example packstack is instructed to deploy a controller node on the system with IP address
    Additional compute nodes are deployed on the systems with IP addresses and
    $ packstack --install-hosts=,,
  2. The packstack utility will prompt you to enter the password of the root user for each system in the deployment. This is required to connect to the system and install Puppet which is the tool used to facilitate the rest of the deployment.
    root@'s password:
  3. The Puppet manifests used to deploy each component will be run on each of the target systems. The amount of time this takes to complete varies based on the hardware and existing workload of each system. It can be significant.
    When the deployment has successfully completed this message is displayed:
    **** Installation completed successfully ******
You have successfully deployed an OpenStack environment using packstack. Please note that:
  • An answer file containing all chosen configuration options is saved to disk on the system from which you ran packstack. This file can be used to automate future deployments.
     * A new answerfile was created in: /root/packstack-answers-20130306-051043.txt
  • A file containing the authentication details of the OpenStack admin user is saved to disk on the system to which the OpenStack client tools were deployed. You will need these details to manage the OpenStack environment.
     * To use the command line tools you need to source the file /root/keystonerc_admin created on
Refer to Part V, “Using OpenStack” to begin using your OpenStack environment.