Rackspace is a major public cloud platform which may be configured using either the openstack driver.
my-rackspace-config: # Set the location of the salt-master # minion: master: saltmaster.example.com # Configure Rackspace using the OpenStack plugin # identity_url: 'https://identity.api.rackspacecloud.com/v2.0/tokens' compute_name: cloudServersOpenStack protocol: ipv4 # Set the compute region: # compute_region: DFW # Configure Rackspace authentication credentials # user: myname tenant: 123456 apikey: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx driver: openstack
Changed in version 2015.8.0.
provider parameter in cloud provider definitions was renamed to
change was made to avoid confusion with the
provider parameter that is used in cloud profile
definitions. Cloud provider definitions now use
driver to refer to the Salt cloud module that
provides the underlying functionality to connect to a cloud host, while cloud profiles continue
provider to refer to provider configurations that you define.
Rackspace currently has six compute regions which may be used:
DFW -> Dallas/Forth Worth ORD -> Chicago SYD -> Sydney LON -> London IAD -> Northern Virginia HKG -> Hong Kong
Note: Currently the LON region is only available with a UK account, and UK accounts cannot access other regions
user is the same user as is used to log into the Rackspace Control
apikey can be found in the API Keys area of the
Control Panel. The
apikey will be labeled as API Key (and may need to be
tenant will be labeled as Cloud Account Number.
An initial profile can be configured in
openstack_512: provider: my-rackspace-config size: 512 MB Standard image: Ubuntu 12.04 LTS (Precise Pangolin)
To instantiate a machine based on this profile:
# salt-cloud -p openstack_512 myinstance
This will create a virtual machine at Rackspace with the name
This operation may take several minutes to complete, depending on the current
load at the Rackspace data center.
Once the instance has been created with salt-minion installed, connectivity to it can be verified with Salt:
# salt myinstance test.ping
Rackspace offers a hybrid hosting configuration option called RackConnect that allows you to use a physical firewall appliance with your cloud servers. When this service is in use the public_ip assigned by nova will be replaced by a NAT ip on the firewall. For salt-cloud to work properly it must use the newly assigned "access ip" instead of the Nova assigned public ip. You can enable that capability by adding this to your profiles:
openstack_512: provider: my-openstack-config size: 512 MB Standard image: Ubuntu 12.04 LTS (Precise Pangolin) rackconnect: True
Rackspace offers a managed service level of hosting. As part of the managed service level you have the ability to choose from base of lamp installations on cloud server images. The post build process for both the base and the lamp installations used Chef to install things such as the cloud monitoring agent and the cloud backup agent. It also takes care of installing the lamp stack if selected. In order to prevent the post installation process from stomping over the bootstrapping you can add the below to your profiles.
openstack_512: provider: my-rackspace-config size: 512 MB Standard image: Ubuntu 12.04 LTS (Precise Pangolin) managedcloud: True
Rackspace provides two sets of virtual machine images, first, and next
generation. As of
0.8.9 salt-cloud will default to using the next
generation images. To force the use of first generation images, on the profile
configuration please add:
FreeBSD-9.0-512: provider: my-rackspace-config size: 512 MB Standard image: FreeBSD 9.0 force_first_gen: True
By default salt-cloud will not add Rackspace private networks to new servers. To enable
a private network to a server instantiated by salt cloud, add the following section
to the provider file (typically
networks: - fixed: # This is the private network - private-network-id # This is Rackspace's "PublicNet" - 00000000-0000-0000-0000-000000000000 # This is Rackspace's "ServiceNet" - 11111111-1111-1111-1111-111111111111
To get the Rackspace private network ID, go to Networking, Networks and hover over the private network name.
The order of the networks in the above code block does not map to the order of the ethernet devices on newly created servers. Public IP will always be first ( eth0 ) followed by servicenet ( eth1 ) and then private networks.
Enabling the private network per above gives the option of using the private subnet for all master-minion communication, including the bootstrap install of salt-minion. To enable the minion to use the private subnet, update the master: line in the minion: section of the providers file. To configure the master to only listen on the private subnet IP, update the interface: line in the /etc/salt/master file to be the private subnet IP of the salt master.