Git repository structure¶

You can have the following configuration priorities in the Git repository:

• Cluster - configuration is applied to all nodes. All cluster level configuration files must be placed in the root directory in the Git repository.
• Role - configuration is applied to all nodes with the selected role. Parameters with the Role level override parameters with the Cluster level.
• Node - configuration is applied to the selected node ID. Parameters with the Node level override the parameters with Global and Role levels.

The following text is an example of a Git repository that you can create:

.
|-- cluster.yaml
    |`-- glance-api.conf
|-- nodes
    |`-- node_1.domain.local.yaml
|-- roles
    |`-- compute.yaml
    |`-- controller.yaml
    |`-- primaray-controller.yaml

The example above illustrates what types of files can be stored in the repository. However, you can modify this structure as needed.

The configuration files you create must use Puppet’s resource types to describe the required configuration. For more information about the parameters and attributes that you can specify in the configuration files, see Puppet Resource Type Reference.

The following table describes the Git repository structure.

configuration:
    nova_config:
        'DEFAULT/nova_test':
            value: cluster_param
        'DEFAULT/another_param':
            value: another_param_value

configuration:
    nova_config:
        'DEFAULT/debug':
            value: True
        'DEFAULT/nova_cluster_override':
            value: node_param

    package:
        'mc':
            ensure: absent
    exec:
        'some_test':
            command: '/bin/touch /tmp/test'
            unless: '/bin/test -f /tmp/test'
            path: '/bin:/sbin'

configuration:
    nova_config:
        'DEFAULT/nova_test':
            value: controller_param