openstack
/
tripleo-quickstart
Code Issues Proposed changes
tripleo-quickstart/roles/tripleo-inventory
History
Attila Darazs 7d32b55df7 Use cachable facts to allow multiple runs 
* add "cachable: true" to every set_fact invocation to include them in
  the cached facts; most of these are not necessary as they are not used
  outside the roles, but it would be difficult to guess which ones are
  really necessary, so using the shotgun approach to avoid days or weeks
  of debugging
* disable the expiration of facts so that we can rely on them being
  there even if days pass between the reruns
* move the fact cache into the virtual environment directory (this is
  usually the local_working_dir if quickstart.sh is used to set things
  up); there can be some issues when no clean virtenv is created for new
  deployments and resulting in using stale facts

Change-Id: I987d51cc252b7ebc3972d60497234fa676864882
 		2017-12-05 15:22:11 +01:00
..
defaults Add support for openstack inventory provider 2017-08-16 10:04:24 +02:00
meta Add blank newline at the end of file 2017-03-27 10:40:32 +02:00
tasks Use cachable facts to allow multiple runs 2017-12-05 15:22:11 +01:00
templates Don't hardcode .quickstart since this is configurable 2017-11-27 15:09:54 -05:00
tests Use dynamic include for conditional/tagged tasks 2017-05-12 17:02:34 +02:00
README.md Update documentation around tripleo-inventory 2017-10-18 13:53:17 -06:00

README.md

Configure a static inventory and SSH config to access nodes

The tripleo-inventory role generates a static inventory and SSH config, based on a deployment mode and a type of undercloud.

Openstack inventory

When cluster nodes are pre-provisioned and hosted on an OpenStack host cloud, use the inventory: openstack mode. Fetched data from the dynamic shade-inventory will be filtered by the clusterid, if defined, and by node types, like overcloud/undercloud/bastion. Then the filtered data is stored into the static inventory and SSH config is generated to access overcloud nodes via a bastion, which is the undercloud node as well.

The following variables should be customized to match the host OpenStack cloud configuration:

  • clusterid: -- an optional Nova servers' metadata parameter to identify your environment VMs.
  • openstack_private_network_name:<'private'> -- defines a private network name used as an admin/control network. Ansible and SSH users will use that network when connecting to the inventory nodes via the undercloud/bastion.
  • os_username: -- credentials to connect the OpenStack cloud hosting your pre-provisioned Nova servers.
  • os_password: -- credentials to connect the host OpenStack cloud.
  • os_project_name: -- The project name for use with the OpenStack Cloud (Identity API v3).
  • os_tenant_name: -- The tenant name for use with the OpenStack Cloud (Identity API v2).
  • os_auth_url: -- The authentication url for the OpenStack cloud.
  • cloud_name: -- The name of the cloud for the os-cloud-config configuration file.
  • os_identity_api_version: -- Identity API version to use when accessing the OpenStack cloud.
  • os_region_name: -- Region name to be used when accessing the OpenStack cloud.

TODO(bogdando) document remaining modes 'all', 'undercloud'.