IPv6¶

IPv6 addressing¶

The ipv6_address_mode attribute is used to control how addressing is handled by OpenStack. There are a number of different ways that guest instances can obtain an IPv6 address, and this attribute exposes these choices to users of the Networking API.

Router advertisements¶

The ipv6_ra_mode attribute is used to control router advertisements for a subnet.

The IPv6 Protocol uses Internet Control Message Protocol packets (ICMPv6) as a way to distribute information about networking. ICMPv6 packets with the type flag set to 134 are called “Router Advertisement” packets, which contain information about the router and the route that can be used by guest instances to send network traffic.

The ipv6_ra_mode is used to specify if the Networking service should generate Router Advertisement packets for a subnet.

ipv6_ra_mode and ipv6_address_mode combinations¶

Dataplane¶

Both the Linux bridge and the Open vSwitch dataplane modules support forwarding IPv6 packets amongst the guests and router ports. Similar to IPv4, there is no special configuration or setup required to enable the dataplane to properly forward packets from the source to the destination using IPv6. Note that these dataplanes will forward Link-local Address (LLA) packets between hosts on the same network just fine without any participation or setup by OpenStack components after the ports are all connected and MAC addresses learned.

Addresses for subnets¶

There are three methods currently implemented for a subnet to get its cidr in OpenStack:

1. Direct assignment during subnet creation via command line or Horizon
2. Referencing a subnet pool during subnet creation
3. Using a Prefix Delegation (PD) client to request a prefix for a subnet from a PD server

In the future, additional techniques could be used to allocate subnets to projects, for example, use of an external IPAM module.

Address modes for ports¶

IPv6 supports three different addressing schemes for address configuration and for providing optional network information.

Stateless Address Auto Configuration (SLAAC)

Address configuration using Router Advertisement (RA).

DHCPv6-stateless

Address configuration using RA and optional information using DHCPv6.

DHCPv6-stateful

Address configuration and optional information using DHCPv6.

OpenStack can be setup such that OpenStack Networking directly provides RA, DHCP relay and DHCPv6 address and optional information for their networks or this can be delegated to external routers and services based on the drivers that are in use. There are two neutron subnet attributes - ipv6_ra_mode and ipv6_address_mode – that determine how IPv6 addressing and network information is provided to project instances:

• ipv6_ra_mode: Determines who sends RA.
• ipv6_address_mode: Determines how instances obtain IPv6 address, default gateway, or optional information.

For the above two attributes to be effective, enable_dhcp of the subnet object must be set to True.

Using SLAAC for addressing¶

When using SLAAC, the currently supported combinations for ipv6_ra_mode and ipv6_address_mode are as follows.

Setting ipv6_ra_mode to slaac will result in OpenStack Networking routers being configured to send RA packets, when they are created. This results in the following values set for the address configuration flags in the RA messages:

• Auto Configuration Flag = 1
• Managed Configuration Flag = 0
• Other Configuration Flag = 0

New or existing neutron networks that contain a SLAAC enabled IPv6 subnet will result in all neutron ports attached to the network receiving IPv6 addresses. This is because when RA broadcast messages are sent out on a neutron network, they are received by all IPv6 capable ports on the network, and each port will then configure an IPv6 address based on the information contained in the RA packet. In some cases, an IPv6 SLAAC address will be added to a port, in addition to other IPv4 and IPv6 addresses that the port already has been assigned.

DHCPv6¶

For DHCPv6, the currently supported combinations are as follows:

Setting DHCPv6-stateless for ipv6_ra_mode configures the neutron router with radvd agent to send RAs. The list below captures the values set for the address configuration flags in the RA packet in this scenario. Similarly, setting DHCPv6-stateless for ipv6_address_mode configures neutron DHCP implementation to provide the additional network information.

• Auto Configuration Flag = 1
• Managed Configuration Flag = 0
• Other Configuration Flag = 1

Setting DHCPv6-stateful for ipv6_ra_mode configures the neutron router with radvd agent to send RAs. The list below captures the values set for the address configuration flags in the RA packet in this scenario. Similarly, setting DHCPv6-stateful for ipv6_address_mode configures neutron DHCP implementation to provide addresses and additional network information through DHCPv6.

• Auto Configuration Flag = 0
• Managed Configuration Flag = 1
• Other Configuration Flag = 1

Neutron’s Distributed Router feature and IPv6¶

IPv6 does work when the Distributed Virtual Router functionality is enabled, but all ingress/egress traffic is via the centralized router (hence, not distributed). More work is required to fully enable this functionality.

VPNaaS¶

VPNaaS supports IPv6, but support in Kilo and prior releases will have some bugs that may limit how it can be used. More thorough and complete testing and bug fixing is being done as part of the Liberty release. IPv6-based VPN-as-a-Service is configured similar to the IPv4 configuration. Either or both the peer_address and the peer_cidr can specified as an IPv6 address. The choice of addressing modes and router modes described above should not impact support.

FWaaS¶

FWaaS allows creation of IPv6 based rules.

NAT & Floating IPs¶

At the current time OpenStack Networking does not provide any facility to support any flavor of NAT with IPv6. Unlike IPv4 there is no current embedded support for floating IPs with IPv6. It is assumed that the IPv6 addressing amongst the projects is using GUAs with no overlap across the projects.

Configuring interfaces of the guest¶

OpenStack currently doesn’t support the privacy extensions defined by RFC 4941. The interface identifier and DUID used must be directly derived from the MAC as described in RFC 2373. The compute hosts must not be setup to utilize the privacy extensions when generating their interface identifier.

There is no provisions for an IPv6-based metadata service similar to what is provided for IPv4. In the case of dual stacked guests though it is always possible to use the IPv4 metadata service instead.

Unlike IPv4 the MTU of a given network can be conveyed in the RA messages sent by the router as well as in the DHCP messages.

Configuring OpenStack Networking for prefix delegation¶

To enable prefix delegation, edit the /etc/neutron/neutron.conf file.

ipv6_pd_enabled = True

pd_dhcp_driver = <class path to driver>

This tells OpenStack Networking to use the prefix delegation mechanism for subnet allocation when the user does not provide a CIDR or subnet pool id when creating a subnet.

Requirements¶

To use this feature, you need a prefix delegation capable DHCPv6 server that is reachable from your OpenStack Networking node(s). This could be software running on the OpenStack Networking node(s) or elsewhere, or a physical router. For the purposes of this guide we are using the open-source DHCPv6 server, Dibbler. Dibbler is available in many Linux package managers, or from source at tomaszmrugalski/dibbler.

When using the reference implementation of the OpenStack Networking prefix delegation driver, Dibbler must also be installed on your OpenStack Networking node(s) to serve as a DHCPv6 client. Version 1.0.1 or higher is required.

This guide assumes that you are running a Dibbler server on the network node where the external network bridge exists. If you already have a prefix delegation capable DHCPv6 server in place, then you can skip the following section.

Configuring the Dibbler server¶

After installing Dibbler, edit the /etc/dibbler/server.conf file:

script "/var/lib/dibbler/pd-server.sh"

iface "br-ex" {
    pd-class {
        pd-pool 2001:db8:2222::/48
        pd-length 64
    }
}

The options used in the configuration file above are:

• script Points to a script to be run when a prefix is delegated or released. This is only needed if you want instances on your subnets to have external network access. More on this below.
• iface The name of the network interface on which to listen for prefix delegation messages.
• pd-pool The larger prefix from which you want your delegated prefixes to come. The example given is sufficient if you do not need external network access, otherwise a unique globally routable prefix is necessary.
• pd-length The length that delegated prefixes will be. This must be 64 to work with the current OpenStack Networking reference implementation.

To provide external network access to your instances, your Dibbler server also needs to create new routes for each delegated prefix. This is done using the script file named in the config file above. Edit the /var/lib/dibbler/pd-server.sh file:

if [ "$PREFIX1" != "" ]; then
    if [ "$1" == "add" ]; then
        sudo ip -6 route add ${PREFIX1}/64 via $REMOTE_ADDR dev $IFACE
    fi
    if [ "$1" == "delete" ]; then
        sudo ip -6 route del ${PREFIX1}/64 via $REMOTE_ADDR dev $IFACE
    fi
fi

The variables used in the script file above are:

• $PREFIX1 The prefix being added/deleted by the Dibbler server.
• $1 The operation being performed.
• $REMOTE_ADDR The IP address of the requesting Dibbler client.
• $IFACE The network interface upon which the request was received.

The above is all you need in this scenario, but more information on installing, configuring, and running Dibbler is available in the Dibbler user guide, at Dibbler – a portable DHCPv6.

To start your Dibbler server, run:

# dibbler-server run

Or to run in headless mode:

# dibbler-server start

When using DevStack, it is important to start your server after the stack.sh script has finished to ensure that the required network interfaces have been created.

User workflow¶

First, create a network and IPv6 subnet:

$ openstack network create ipv6-pd
+---------------------------+--------------------------------------+
| Field                     | Value                                |
+---------------------------+--------------------------------------+
| admin_state_up            | UP                                   |
| availability_zone_hints   |                                      |
| availability_zones        |                                      |
| created_at                | 2017-01-25T19:26:01Z                 |
| description               |                                      |
| headers                   |                                      |
| id                        | 4b782725-6abe-4a2d-b061-763def1bb029 |
| ipv4_address_scope        | None                                 |
| ipv6_address_scope        | None                                 |
| mtu                       | 1450                                 |
| name                      | ipv6-pd                              |
| port_security_enabled     | True                                 |
| project_id                | 61b7eba037fd41f29cfba757c010faff     |
| provider:network_type     | vxlan                                |
| provider:physical_network | None                                 |
| provider:segmentation_id  | 46                                   |
| revision_number           | 3                                    |
| router:external           | Internal                             |
| shared                    | False                                |
| status                    | ACTIVE                               |
| subnets                   |                                      |
| tags                      | []                                   |
| updated_at                | 2017-01-25T19:26:01Z                 |
+---------------------------+--------------------------------------+

$ openstack subnet create --ip-version 6 --ipv6-ra-mode slaac \
--ipv6-address-mode slaac --use-default-subnet-pool \
--network ipv6-pd ipv6-pd-1
+------------------------+--------------------------------------+
| Field                  | Value                                |
+------------------------+--------------------------------------+
| allocation_pools       | ::2-::ffff:ffff:ffff:ffff            |
| cidr                   | ::/64                                |
| created_at             | 2017-01-25T19:31:53Z                 |
| description            |                                      |
| dns_nameservers        |                                      |
| enable_dhcp            | True                                 |
| gateway_ip             | ::1                                  |
| headers                |                                      |
| host_routes            |                                      |
| id                     | 1319510d-c92c-4532-bf5d-8bcf3da761a1 |
| ip_version             | 6                                    |
| ipv6_address_mode      | slaac                                |
| ipv6_ra_mode           | slaac                                |
| name                   | ipv6-pd-1                            |
| network_id             | 4b782725-6abe-4a2d-b061-763def1bb029 |
| project_id             | 61b7eba037fd41f29cfba757c010faff     |
| revision_number        | 2                                    |
| service_types          |                                      |
| subnetpool_id          | prefix_delegation                    |
| tags                   | []                                   |
| updated_at             | 2017-01-25T19:31:53Z                 |
| use_default_subnetpool | True                                 |
+------------------------+--------------------------------------+

The subnet is initially created with a temporary CIDR before one can be assigned by prefix delegation. Any number of subnets with this temporary CIDR can exist without raising an overlap error. The subnetpool_id is automatically set to prefix_delegation.

To trigger the prefix delegation process, create a router interface between this subnet and a router with an active interface on the external network:

$ openstack router add subnet router1 ipv6-pd-1

The prefix delegation mechanism then sends a request via the external network to your prefix delegation server, which replies with the delegated prefix. The subnet is then updated with the new prefix, including issuing new IP addresses to all ports:

$ openstack subnet show ipv6-pd-1
+-------------------+--------------------------------------+
| Field             | Value                                |
+-------------------+--------------------------------------+
| allocation_pools  | 2001:db8:2222:6977::2-2001:db8:2222: |
|                   | 6977:ffff:ffff:ffff:ffff             |
| cidr              | 2001:db8:2222:6977::/64              |
| created_at        | 2017-01-25T19:31:53Z                 |
| description       |                                      |
| dns_nameservers   |                                      |
| enable_dhcp       | True                                 |
| gateway_ip        | 2001:db8:2222:6977::1                |
| host_routes       |                                      |
| id                | 1319510d-c92c-4532-bf5d-8bcf3da761a1 |
| ip_version        | 6                                    |
| ipv6_address_mode | slaac                                |
| ipv6_ra_mode      | slaac                                |
| name              | ipv6-pd-1                            |
| network_id        | 4b782725-6abe-4a2d-b061-763def1bb029 |
| project_id        | 61b7eba037fd41f29cfba757c010faff     |
| revision_number   | 4                                    |
| service_types     |                                      |
| subnetpool_id     | prefix_delegation                    |
| tags              | []                                   |
| updated_at        | 2017-01-25T19:35:26Z                 |
+-------------------+--------------------------------------+

If the prefix delegation server is configured to delegate globally routable prefixes and setup routes, then any instance with a port on this subnet should now have external network access.

Deleting the router interface causes the subnet to be reverted to the temporary CIDR, and all ports have their IPs updated. Prefix leases are released and renewed automatically as necessary.

References¶

The following link provides a great step by step tutorial on setting up IPv6 with OpenStack: Tenant IPV6 deployment in OpenStack Kilo release.

Extra configuration¶

pd_dhcp_driver = neutron_pd_agent

pd_interface = br-ex

neutron-pd-agent --config-file /etc/neutron/neutron.conf