Ansible NFV playbooks

Ansible NFV repository holds various playbooks for installation, configuration, tuning, testing and day to day tasks related to NFV and Openstack.

TReX Execution

• “pip install netaddr”

• Initialize ansible inventory.

  For hybrid environment: Run below command from Hypervisor host with "-e host=" set to Hypervisor host IP.

  ansible-playbook initialize.yaml -e host=HypervisorIP -e user=root -e ssh_key=~/.ssh/id_rsa -e setup_type=virt

• To start execution of trex benchmark, run:

  ansible-playbook -i inventory trex-run.yaml --extra @user-specific/trex.yaml

  user-specific/trex.yaml defines the minimal set of parameters related to external, management and provider network.

  • Only 3 mandatory variables need to be set:
    • extern to specify the external network configured in Openstack,
    • physical_network_pf for specifying the SRIOV physical interface for TRex.
    • Interface for testpmd ( physical_network_dpdk or physical_network_vf).

• To cleanup trex resources, run: trex_cleanup.sh

  Tempest Execution

• “pip install netaddr”

• Initialize ansible inventory.

  For hybrid environment: Run below command from Hypervisor host with "-e host=" set to Hypervisor host IP.

  ansible-playbook initialize.yaml -e host=HypervisorIP -e user=root -e ssh_key=~/.ssh/id_rsa -e setup_type=virt

• To start execution of tempest tests, run:

  ansible-playbook -i inventory tempest-run.yaml --extra @user-specific/tempest.yaml

  user-specific/trex.yaml defines the minimal set of parameters related to external, management and provider network.

  • Only 2 mandatory variables need to be set:
    • extern to specify the external network configured in Openstack,
    • Interfaces to use with tempest ( physical_network_dpdk or physical_network_vf or physical_network_pf).

• To cleanup tempest resources,run: tempest_cleanup.sh

Examples of Physical to Virtual back to Physical(PVP) senarios

Senarios 1 - Device Under Test (DUT) backed by DPDK

|  +----------------+  |
|  |  Testpmd VM    |  |
|  +-+------------+-+  |
|    |            |    |
| +--+------------+--+ |
| |    OVS DPDK      | |
| +--+------------+--+ |
|    |            |    |
|    | Compute 1  |    |
|  +-+-+        +-+-+  |
|  |   |        |   |  |
+--+-+-+--------+-+-+--+
     |            |
+----+------------+----+
|      TOR Switch      |
|                      |
+----+------------+----+
     |            |
+----+------------+----+
| || P ||      || P || |
| || F ||      || F || |
| +--+--+      +--+--+ |
|    | Compute 2  |    |
| +--+------------+--+ |
| |                  | |
| |       TReX VM    | |
| +------------------+ |
+----------------------+

"OSP Network Info:",
"  Bridge Mappings:      datacentre:br-ex,data1:br-link0,data2:br-link1",
"  Network VLANRanges:   datacentre:1205:1210,data1:201:220,data2:201:220,sriov1:201:220,sriov2:201:220",
"  Flat Networks:        datacentre,data1,data2",
"  Physical Dev Mapping: sriov1:em1,sriov2:em2",

For the above network config, user-specific/trex.yml would be set as follows:

extern: 'datacentre' #[1]
mgmt: 'data1' #[2]
physical_network_dpdk: ['data1','data2'] #[3]
physical_network_pf: ['sriov1', 'sriov2'] #[4]

[1] The external floating network is mapped to "datacenter", so "extern" is set to datacentre.
[2] Since the DPDK Compute1 node does not have DPDK-VXLAN network, one of provider network "data1" is used as management network for the Testpmd VM, hence "mgmt" is set to data1. On compute2 a VXLAN network would be created for use as management.
[3] Provider network based on "data1" and "data2", within the VLANRanges, will be created on Compute1.
[4] Provider network based on "sriov1" and "sriov2" will be created on Compute2.

Senarios 2 - Device Under Test (DUT) backed by SRIOV-VF

|  +----------------+  |
|  |   Testpmd VM   |  |
|  +-+------------+-+  |
|    |            |    |
|    | Compute 1  |    |
| +--+--+      +--+--+ |
| ||V  ||      || V || |
| ||F  ||      || F || |
+----+------------+----+
     |            |
+----+------------+----+
|         TOR Switch   |
|                      |
+----+------------+----+
     |            |
+----+------------+----+
| || P ||      || P || |
| || F ||      || F || |
| +--+--+      +--+--+ |
|    | Compute 2  |    |
| +--+------------+--+ |
| |                  | |
| |    TReX  VM      | |
| +------------------+ |
+----------------------+

Documentation

For the repository documentation, refer to the docs directory, which provide explanation regarding the playbooks.
In order to start working with TripleO environment, refer to the following doc - docs/tripleo/post_install/tripleo_inventory.md.
The playbook will generate an inventory file for the tripleo environment.

The minimum required version of Ansible for the playbooks >=2.7.5,<2.8.0
Note - In order to work properly with the selinux, make sure the libselinux-python package is installed on your host.

Playbooks

• TripleO
  • Tester
    • Tempest
    • Testpmd deploy and MoonGen run
    • Testpmd deploy and MoonGen run on secondary NUMA node
  • Tuning
    • CPU pinning and Huge pages
    • Tuned
  • Post install
    • Openstack tasks
    • Overcloud repo install
    • TripleO Inventory
  • Configuration
    • Config NFV Lab switches
    • Interact with Dell EMC iDRAC:
      • Power Actions
      • Query information
      • Set device boot order
• Moongen
  • Moongen install
  • Moongen run

How to contribue

Patches should be submitted using git review to the GerritHub.

Note

Please, modify the playbooks variables to meet your needs before running the playbooks.

The variables could be applied to the playbook run, by saving them into a separate yml file and include the file during the playbook execution.
Note the '@' sign, which is used to apply the variables located within the provided file.

ansible-playbook playbooks/path/to/the/playbook.yml -e @/path/to/the/variable/file.yml

For any question, refer to the NFV Cloud QE team - nfv-cloud-qe-team@redhat.com