[DOCS] Refactor of nova role docs

Code Issues Proposed changes

Update the role documentation to match the newly proposed format. In
the new format, each role project should maintain the configuration
variations for the role project itself.
Change-Id: I165891189ffa67680980a70178fd55951cad293b

This commit is contained in:

Drew Thorstensen

2016年08月10日 16:54:31 -05:00

parent 039458e0e2

commit 7777d81f1a

5 changed files with 137 additions and 27 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

36
doc/source/app-powervm.rst Normal file

View File

@@ -0,0 +1,36 @@

`Home <index.html>`_ OpenStack-Ansible Neutron

=====================================

Scenario - Using PowerVM Nova plugin

=====================================

Prerequisites

~~~~~~~~~~~~~

In order to use the PowerVM OpenStack drivers with OpenStack-Ansible (OSA), the

following pre-requisites must be fulfilled:

- At least one of the repo-build servers must be ppc64le. Can mix and match

repo-build servers between x86 and ppc64le.

- The compute nodes should be pre-configured for PowerVM with the NovaLink_

feature.

- The NovaLink Management VM needs at least one direct attach I/O card.

OpenStack Ansible is currently able to deploy the PowerVM drivers when

paired with the Linux Bridge or Open vSwitch agent. The traditional PowerVM

Shared Ethernet Adapter networking agent is not yet supported.

- The network topology on the NovaLink must match a supported OpenStack

Ansible network configuration.

.. _NovaLink: http://www.ibm.com/support/knowledgecenter/POWER8/p8eig/p8eig_kickoff.htm?cp=POWER8

PowerVM configuration

~~~~~~~~~~~~~~~~~~~~~

The Compute driver for OpenStack-Ansible should automatically detect that it

is of type PowerVM. If the user has specified a specific compute type, that

is applicable to the whole cloud. It is advised that the deployer allows

OSA to detect the appropriate compute node type.

39
doc/source/index.rst

View File

@@ -1 +1,38 @@

.. include:: ../../README.rst

===============================

Nova role for OpenStack-Ansible

===============================

.. toctree::

:maxdepth: 2

overview.rst

app-powervm.rst

:tags: openstack, nova, cloud, ansible

:category: \*nix

This role will install the following Upstart services:

* nova-server

* nova-compute

Default Variables

^^^^^^^^^^^^^^^^^

.. literalinclude:: ../../defaults/main.yml

:language: yaml

:start-after: under the License.

Example playbook

~~~~~~~~~~~~~~~~

.. literalinclude:: ../../examples/playbook.yml

:language: yaml

Tags

~~~~

This role supports two tags: ``nova-install`` and ``nova-config``

The ``nova-install`` tag can be used to install and upgrade.

The ``nova-config`` tag can be used to manage configuration.

40
doc/source/overview.rst Normal file

View File

@@ -0,0 +1,40 @@

CPU Platform Compatibility

~~~~~~~~~~~~~~~~~~~~~~~~~~

This role supports multiple CPU architecture types. At least one repo_build

node must exist for each CPU type that is in use in the deployment.

Currently supported CPU architectures:

- x86_64 / amd64

- ppc64le

Compute Driver Compatibility

~~~~~~~~~~~~~~~~~~~~~~~~~~~~

This role supports multiple nova compute driver types. The following

compute drivers are supported:

- libvirt (default)

- ironic

- lxd (via nova-lxd)

- powervm (via nova-powervm)

The driver type will automatically be detected by the OpenStack Ansible Nova

role for the following compute driver types:

- libvirt

- ironic

- powervm

Any mix and match of compute node types can be used for those platforms.

If using the lxd driver, the compute type must be specified using the

``nova_virt_type`` variable in the ``/etc/openstack_deploy/user_variables.yml``

file.

.. code-block:: shell-session

nova_virt_type: lxd

It should be noted that if the ``nova_virt_type`` variable is set, then all

nodes in the deployment will be set to that hypervisor type. It is recommended

to allow the automatic hypervisor detection.