Vagrant Openstack Cloud Provider

This is a Vagrant 1.4+ plugin that adds an Openstack Cloud provider to Vagrant, allowing Vagrant to control and provision machines within Openstack cloud.

Note: This plugin was originally forked from mitchellh/vagrant-rackspace

Features

Create and boot Openstack instances
Halt and reboot instances
Suspend and resume instances
SSH into the instances
Automatic SSH key generation and Nova public key provisioning
Automatic floating IP allocation and association
Provision the instances with any built-in Vagrant provisioner
Boot instance from volume
Attach Cinder volumes to the instances
Create and delete Heat Orchestration stacks
Support Openstack regions
Minimal synced folder support via rsync
Custom sub-commands within Vagrant CLI to query Openstack objects

Usage

Install using standard Vagrant 1.1+ plugin installation methods. After installing, vagrant up and specify the openstack provider. An example is shown below.

$ vagrant plugin install vagrant-openstack-provider
...
$ vagrant up --provider=openstack
...

Of course prior to doing this, you'll need to obtain an Openstack-compatible box file for Vagrant.

Quick Start

After installing the plugin (instructions above), the quickest way to get started is to specify all the details manually within a config.vm.provider block in the Vagrantfile

Create a Vagrantfile that looks like the following, filling in your information where necessary.

This Vagrantfile shows the minimal needed configuration.

require 'vagrant-openstack-provider'

Vagrant.configure('2') do |config|

  config.vm.box       = 'openstack'
  config.ssh.username = 'stack'

  config.vm.provider :openstack do |os|
    os.openstack_auth_url = 'http://keystone-server.net/v2.0/tokens'
    os.username           = 'openstackUser'
    os.password           = 'openstackPassword'
    os.tenant_name        = 'myTenant'
    os.flavor             = 'm1.small'
    os.image              = 'ubuntu'
    os.floating_ip_pool   = 'publicNetwork'
  end
end

And then run vagrant up --provider=openstack.

Note that normally a lot of this boilerplate is encoded within the box file, but the box file used for the quick start, the "dummy" box, has no preconfigured defaults.

Configuration

This provider exposes quite a few provider-specific configuration options:

Credentials

username - The username with which to access Openstack.
password - The API key for accessing Openstack.
tenant_name - The Openstack project name to work on
region - The Openstack region to work on
openstack_auth_url - The endpoint to authentication against. By default, vagrant will use the global openstack authentication endpoint for all regions with the exception of :lon. IF :lon region is specified vagrant will authenticate against the UK authentication endpoint.
openstack_compute_url - The compute service URL to hit. This is good for custom endpoints. If not provided, vagrant will try to get it from catalog endpoint.
openstack_network_url - The network service URL to hit. This is good for custom endpoints. If not provided, vagrant will try to get it from catalog endpoint.
openstack_volume_url - The block storage URL to hit. This is good for custom endpoints. If not provided, vagrant will try to get it from catalog endpoint.
openstack_image_url - The image URL to hit. This is good for custom endpoints. If not provided, vagrant will try to get it from catalog endpoint.

VM Configuration

server_name - The name of the server within Openstack Cloud. This defaults to the name of the Vagrant machine (via config.vm.define), but can be overridden with this.
flavor - The name of the flavor to use for the VM
image - The name of the image to use for the VM
availability_zone - Nova Availability zone used when creating VM
security_groups - List of strings representing the security groups to apply. e.g. ['ssh', 'http']
user_data - String of User data to be sent to the newly created OpenStack instance. Use this e.g. to inject a script at boot time.
metadata - A Hash of metadata that will be sent to the instance for configuration e.g. os.metadata = { 'key' => 'value' }
scheduler_hints - Pass hints to the OpenStack scheduler, e.g. { "cell": "some cell name" }

Floating IPs

floating_ip - The floating IP to associate with the VM. This IP must be formerly allocated.
floating_ip_pool - The floating IP Pool or a list of floating IP pool from which a floating IP will be allocated to be associated with the VM. alternative to the floating_ip option.

floating_ip_pool attribute can be either a string or an array. In case of an array, if an IP can't be allocated from the first pool for any reason, it will try with the second one and so on. Finally, if it does not manage to allocate a floating IP from any pools of the list, it will fail.

config.vm.provider :openstack do |os|
  ...
  os.floating_ip_pool = ['External-Network-01', 'External-Network-02']
  ...
end

floating_ip_pool_always_allocate - if set to true, vagrant will always allocate floating ip instead of trying to reuse unassigned ones

Networks

networks - Network list the server must be connected on. Can be omitted if only one private network exists in the Openstack project

Networking features in the form of config.vm.network are not supported with vagrant-openstack, currently. If any of these are specified, Vagrant will emit a warning, but will otherwise boot the Openstack server.

You can provide network id or name. However, in Openstack a network name is not unique, thus if there are two networks with the same name in your project the plugin will fail. If so, you have to use only ids. Optionally, you can specify the IP address that will be assigned to the instance if you need a static address or if DHCP is not enable for this network.

Here's an example which connect the instance to six Networks :

config.vm.provider :openstack do |os|
  ...
  os.networks = [
    'net-name-01',
    '287132f0-57e6-4c31-a1ee-4823e9786ff2',
    {
      name: 'net-name-03',
      address: '192.168.22.43'
    },
    {
      id: '7dfdcf01-5177-4774-9473-2ae92a6447d4',
      address: '192.168.43.76'
    },
    {
      name: 'net-name-05'
    },
    {
      id: '01e0950f-c668-4efe-821b-93ff6e427562'
    }
  ]
  ...
end

Volumes

volumes - Volume list that have to be attached to the server. You can provide volume id or name. However, in Openstack a volume name is not unique, thus if there are two volumes with the same name in your project the plugin will fail. If so, you have to use only ids. Optionally, you can specify the device that will be assigned to the volume.

Here comes an example that show six volumes attached to a server :

config.vm.provider :openstack do |os|
  ...
  os.volumes = [
    '619e027c-f4a9-493d-8c15-c89de81cb949',
    'vol-name-02',
    {
      id: '410096ff-ef71-4ca4-8006-e5bd9e99239a',
      device: '/dev/vdc'
    },
    {
      name: 'vol-name-04',
      device: '/dev/vde'
    },
    {
      name: 'vol-name-05'
    },
    {
      id: '9e419e91-8f66-4803-bc45-4600182cfd8d'
    }
  ]
  ...
end

volume_boot - Volume to boot the VM from. When booting from an existing volume, image is not necessary and must not be provided.

Orchestration Stacks

stacks - Heat Stacks that will be automatically created when running vagrant up, and deleted when running vagrant destroy

Here comes an example that show two stacks :

config.vm.provider :openstack do |os|
 ...
os.stacks = [
  {
    name: 'mystack1',
    template: 'heat_template.yml'
  }, {
    name: 'mystack2',
    template: '/path/to//my/heat_template.yml'
  }]
end

SSH authentication

keypair_name - The name of the key pair register in nova to associate with the VM. The public key should be the matching pair for the private key configured with config.ssh.private_key_path on Vagrant.
public_key_path - if keypair_name is not provided, the path to the public key will be used by vagrant to generate a keypair on the OpenStack cloud. The keypair will be destroyed when the VM is destroyed.

If neither keypair_name nor public_key_path are set, vagrant will generate a new ssh key and automatically import it in Openstack.

ssh_disabled - if set to true, all ssh actions managed by the provider will be disabled during the vagrant up. We recommend to use this option only to create private VMs that won't be accessed directly from vagrant. By contrast, others commands like vagrant ssh or vagrant provision is run normally but it is likely to fail. Default value is false.

Synced folders

sync_method - Specify the synchronization method for shared folder between the host and the remote VM. Currently, it can be "rsync" or "none". The default value is "rsync". If your Openstack image does not include rsync, you must set this parameter to "none".
rsync_includes - If sync_method is set to "rsync", this parameter give the list of local folders to sync on the remote VM.
rsync_ignore_files - Is used for the rsync prameter "--exclude-from". Set rsync_ignore_files to a list of files that contain patterns to exclude from the rsync to /vagrant on a provisioned instance. ".gitignore or ".hgignore" for example.

There is minimal support for synced folders. Upon vagrant up, vagrant reload, and vagrant provision, the Openstack provider will use rsync (if available) to uni-directionally sync the folder to the remote machine over SSH.

This is good enough for all built-in Vagrant provisioners (shell, chef, and puppet) to work!

Vagrant standard configuration

There are some standard configuration options that this provider takes into account when creating and connecting to OpenStack machines

config.vm.box - A box is not mandatory for this provider. However, if you are running Vagrant before version 1.6, vagrant will not start if this property is not set. In this case you can assign any value to it. See section "Box Format" to know more about boxes.
config.vm.box_url - URL of the box when it is necessary
ssh.username - Username used by vagrant for SSH login
ssh.port - Default SSH port is 22. If set, this option will override the default for SSH login
ssh.private_key_path - If set, vagrant will use this private key path to SSH on the machine. If you set this option, the public_key_path option of the provider should be set.

Box Format

Every provider in Vagrant must introduce a custom box format. This provider introduces openstack boxes. You can view an example box in the example_box/ directory. That directory also contains instructions on how to build a box.

The box format is basically just the required metadata.json file along with a Vagrantfile that does default settings for the provider-specific configuration for this provider.

Custom commands

Custom commands are provided for Openstack. Type vagrant openstack to show available commands.

$ vagrant openstack

Usage: vagrant openstack command

Available subcommands:
     image-list           List available images
     flavor-list          List available flavors
     network-list         List private networks in project
     subnet-list          List subnets for available networks
     floatingip-list      List floating IP and floating IP pools
     volume-list          List existing volumes
     reset                Reset Vagrant OpenStack provider to a clear state

For instance vagrant openstack image-list lists images available in Glance.

$ vagrant openstack image-list

+--------------------------------------+---------------------+
| Id                                   | Name                |
+--------------------------------------+---------------------+
| 594f1287-9de3-4f3e-b82a-6ad223943ab2 | ubuntu-12.04_x86_64 |
| 3e5aca4a-bf12-4721-87df-7bc8fd1fc36c | debian7_x86_64      |
| 3e561121-d8d0-4328-b319-7076bfb3b18a | ubuntu-14.04_x86_64 |
| 5c576643-7ea3-49db-b1c0-9b245d955ee0 | rhel65_x86_64       |
| d3145dd5-654a-4936-b421-9333f02ae66c | centos6_x86_64      |
+--------------------------------------+---------------------+

Contribute

Development

To work on the vagrant-openstack plugin, clone this repository out, and use Bundler to get the dependencies:

Note: Vagrant 1.6 requires bundler version < 1.7. We recommend using last 1.6 version.

$ gem install bundler -v 1.6.6

Install the plugin dependencies

$ bundle install

Once you have the dependencies, verify the unit tests pass with rake:

$ bundle exec rake

If those pass, you're ready to start developing the plugin. You can test the plugin without installing it into your Vagrant environment by just creating a Vagrantfile in the top level of this directory (it is gitignored) that uses it, and uses bundler to execute Vagrant:

$ bundle exec vagrant up --provider=openstack

Troubleshooting

Logging

To enable all Vagrant logs set environment variable VAGRANT_LOG to the desire log level (for instance VAGRANT_LOG=debug). If you want only Openstack provider logs use the variable VAGRANT_OPENSTACK_LOG. if both variables are set, VAGRANT_LOG takes precedence.

CentOS/RHEL/Fedora (sudo: sorry, you must have a tty to run sudo)

The default configuration of the RHEL family of Linux distributions requires a tty in order to run sudo. Vagrant does not connect with a tty by default, so you may experience the error:

sudo: sorry, you must have a tty to run sudo

The best way to take deal with this error is to upgrade to Vagrant 1.4 or later, and enable:

config.ssh.pty = true

Sponsoring

We thanks Numergy for giving us access to free compute resources on their OpenStack cloud that enabled us to test our provider on a real OpenStack installation.

If you are also powering an OpenStack cloud, we'd like to hear from you. Test the plugin and report us issues or features you'd like to see.

nathairtras/vagrant-openstack-provider