Rackspace Neutron-based Networking API FAQ
Neutron is an OpenStack project that will be the basis for all upcoming networking services in the Rackspace Public Cloud. If you deploy Rackspace Cloud Servers, you can use Rackspace Networking API based on OpenStack Neutron in order to create and manage Cloud Networks and perform other Networking operations.
Where are the docs?
- Security Groups concepts and introduction
- Cloud Networks Developer Guide
- Cloud Networks Getting Started Guide
Why are you introducing Rackspace Networking based on the OpenStack Neutron API?
At Rackspace we have been using the /os-networksv2 Cloud Servers extension to provision networking capabilities in the public cloud. This extension was based on nova-network API and could be used primarily to take advantage of Rackspace Cloud Networks. Since we introduced the existing API, the Neutron OpenStack project has become the primary Networking service in OpenStack. Moving forward, Rackspace Networking API will be based on OpenStack Neutron API that is more granular than the nova-network API and offers richer functionality and flexibility to consume networking services.
What new capabilities are being introduced with the new API?
The API provides three primary top level resources (networks , ports and subnets). Users can now:
- Create and manage Cloud Networks using the new API
- Assign routes to Cloud Servers at boot-time (Host routes)
- Configure allocation pools for subnets (CIDRs) on Cloud Networks to control the dynamic IP address assignment on your Cloud Servers
- Provision an IP address of your choice on Isolated Networks ports
- Dual stack your Isolated networks so that you have IPv4 and IPv6 addresses on the same port
Will I still be able to use the old Networking API based on nova-network?
The existing Networking API (/osnetworksv2 servers extension) will still function. However, we encourage you to move to the new Networking API to take advantage of some of the new capabilities.
Can I attach and detach networks from servers using the new API?
Currently, you still have to use the Cloud Networks virtual interfaces extension to attach and detach networks from a Cloud Server.
Is this API available to RackConnect customers?
RackConnect v2.0 customers can use this API. However, RackConnect v3.0 users will not be able to take advantage of this API yet. We are planning to address this shortly.
Can I perform the API functions using the Neutron client?
Neutron client is not currently supported. We plan to make the client available very soon.
Where is my XML?
The Networking API returns JSON exclusively, there is no option to receive responses in XML.
How do I create Networks with the new API?
You need to create a network and add a subnet to that network . This is different from the older way of creating networks which was a one step process. Please see the Getting Started guide for examples.
How many subnets can be provisioned on a network
We have a limit of one IPV4 subnet and one IPv6 subnet per network.
Can I create, update or delete subnets on Public and ServiceNet?
No, These networks are Rackspace Provider networks and cannot be modified. You cannot control subnetting and IP addressing on these networks.
Can I modify allocation pools on the subnet?
You can specify allocation pools for the subnet when you create a subnet but you cannot modify it.
When I change attributes on subnets will they take effect on ports that are using it ?
No, the changes to the attributes will only take effect on any new ports and servers that are created
Can I create ports on Public and ServiceNet ?
No, ports can be created only on Isolated Networks (cloud networks).
Can I have both an IPv4 and an IPv6 address on the same Isolated network port?
Yes, you can provision both an IPv4 address and an IPv6 address on the same port using the fixed_ips attribute.
Continue the conversation in the Rackspace Community.
©2017 Rackspace US, Inc.
Except where otherwise noted, content on this site is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License