Jump to: navigation, search

SR-IOV-Passthrough-For-Networking

Revision as of 15:11, 8 October 2014 by Baoli (talk | contribs) (Nova)

SR-IOV Networking in OpenStack Juno

OpenStack Juno adds inbox support to request VM access to virtual network via SR-IOV NIC. With the introduction of SR-IOV based NIC, the traditional virtual bridge is no longer required. Each SR-IOV port is associated with a virtual function (VF). SR-IOV ports may be provided by Hardware-based Virtual Eithernet Bridging (HW VEB); or they may be extended to an upstream physical switch (IEEE 802.1br). There are two ways that SR-IOV port may be connected:

  • directly connected to its VF
  • connected with a macvtap device that resides on the host, which is then connected to the corresponding VF.

Nova

Nova support for SR-IOV enables to schedule an instance with SR-IOV ports based on their network connectivity. The neutron ports' associated physical networks have to be used in making the scheduling decision. PCI Whitelist was enchanced to allow tags to be associated with PCI devices. PCI devices available for SR-IOV networking should be tagged with physical_network label.

For SR-IOV networking, a pre-defined tag "physical_network" is used to define the physical network that the devices are attached to. A whitelist entry is defined as:

   ["vendor_id": "<id>",] ["product_id": "<id>",]
   ["address": "[[[[<domain>]:]<bus>]:][<slot>][.[<function>]]" |
    "devname": "Ethernet Interface Name",]  
   "physical_network":"name string of the physical network"

<id> can be a '*' or a valid device/product id as displayed by the linux utility lspci. The address uses the same syntax as it's in lspci. The devname can be a valid PCI device name. The only device names that are supportedare those that are displayed by the linux utility ifconfig -a and correspond to either a PF or a VF on a vNIC.

If the device defined by the address or devname corresponds to a SR-IOV PF, all the VFs under the PF will match the entry.

Multiple whitelist entries per host are supported.

Neutron

Neutron support for SR-IOV requires ML2 Plugin with SR-IOV supporting Mechanism Driver. Currently there is ML2 Mechanism Driver for SR-IOV capable NIC based switching (HW VEB). There are Network Adapters from different vendors that may differ by supporting various functionality. If VF link state update is supported by vendor Network Adapter, the SR-IOV NIC L2 Agent should be deployed to leverage this functionality .

VM creation flow with SR-IOV vNIC

  • Create one or more neutron ports
  neutron port-create <net-id> --binding:vnic-type direct
  • Boot VM with one or more neutron ports
  nova boot --flavor m1.large --image <image>
         --nic port-id=<port1> --nic port-id=<port2>

Note that in the nova boot API, users can specify either a port-id or a net-id. If it's the latter case, it's assumed that the user is requesting a normal virtual port (which is not a SR-IOV port).

SR-IOV Configuration

Neutron Server

Using ML2 Neutron plugin modify /etc/neutron/plugins/ml2/ml2_conf.ini:

[ml2]
tenant_network_types = vlan
type_drivers = vlan
mechanism_drivers = openvswitch,sriovnicswitch
[ml2_type_vlan]
network_vlan_ranges = default:2:100

Make sure /etc/neutron/plugins/ml2/ml2_conf_sriov.ini has the following section:

[ml2_sriov]
agent_required = True

Neutron server should be run with the two configuration files /etc/neutron/plugins/ml2/ml2_conf.in and /etc/neutron/plugins/ml2/ml2_conf_sriov.ini

neutron-server --config-file /etc/neutron/neutron.conf --config-file /etc/neutron/plugins/ml2/ml2_conf.ini --config-file /etc/neutron/plugins/ml2/ml2_conf_sriov.ini

Compute

Nova

On each compute you have to associate the Virtual Functions available to each Physical Network. You do it by configuring pci_passthrough_whitelist in /etc/nova/noca.conf

For example:

pci_passthrough_whitelist = {"address":"*:0a:00.*","physical_network":"physnet1"}

This will associate any VF with address that includes ':0a:00.' in it's address to the Physical network physnet1

After configuring the white list you have to restart nova-compute service.

Neutron

If the hardware supports and and you want to enable changing the port admin_state you have to run the Neutron SR-IOV agent.

Note:If you configured agent_required=True on the Neutron server you must run the Agent on each compute node.

In /etc/neutron/plugins/ml2/ml2_conf.ini make sure you have the following:

[securitygroup]
firewall_driver = neutron.agent.firewall.NoopFirewallDriver

Modify /etc/neutron/plugins/ml2/ml2_conf_sriov.ini as follows

[sriov_nic]
physical_device_mappings = physnet1:eth1
exclude_devices =

physnet1 is the physical network eth1 if the Physical Function (PF) exclude_devices is empty so All the VFs associated with eth1 are allowed to be configured by the agent.

After modifying the configuration file start the Neutron SR-IOV agent:

neutron-sriov-nic-agent --config-file /etc/neutron/neutron.conf --config-file /etc/neutron/plugins/ml2/ml2_conf.ini --config-file /etc/neutron/plugins/ml2/ml2_conf_sriov.ini

Exclude VFs

If you want to exclude some of the VFs so the agent won't configure them you need to list them in the sriov_nic section:

Example: exclude_devices = eth1:0000:07:00.2; 0000:07:00.3, eth2:0000:05:00.1; 0000:05:00.2

References

Openstack ML2 SR-IOV driver support