Define a new role

Define a new roleΒΆ

You must define a new role in node_roles.yaml.

The automatically generated plugin contains one example role in the file, but you can delete the contents of the file and start with a new role:

fuel-plugin-example_role:
  name: Example plugin role
  description: Just a test role
  weight: 1000

The only required field besides the ID here is name, but this is always a good idea to provide some description of the role’s purpose. The parameter weight defines relative placement of the role in the list of roles in the Fuel web UI. Standard roles have weights from 10 to 100, with the weight 1000 your role will appear at the end of the list. You can look up the exact values as well as the definitions for the standard roles at openstack.yaml on GitHub .

Like plugin tasks and unlike plugin settings, role definitions are affected by the fuel --sync` command and can be easily updated.

To verify that your new role is defined and can be assigned to new nodes use the button Add nodes in the Nodes UI tab: url http://<fuel ip>:8000/#cluster/<cluster id>/nodes/add.

Note

Keep in mind that a plugin should be enabled for its roles to appear in the Fuel web UI.

You must define a new group for the tasks that belong to the new role in deployment_tasks.yaml:

- id: fuel-plugin-example_role
  type: group
  role: [fuel-plugin-example_role]
  tasks: [globals]
  parameters:
   strategy:
    type: parallel

The ID of the group must be the same as the ID of the new role. You define the list of the core tasks, which run on the new role, in the field tasks. The parameter strategy is required and it defines how the deployment should proceed if there are multiple nodes to be deployed with the role.

The possible values for strategy are:

  • one-by-one – all the nodes with the role will be deployed
    sequentially, one after another
  • parallel – nodes will be deployed in parallel, at the same time

Note

Fuel 8.0 Maintenance Update 1 runs hooks on the target nodes with a new role even if you defined no tasks for the role. One of such hooks creates the file /etc/hosts on targets and it will fail with the error Could not find data item node_name_prefix_for_ messaging in any Hiera data file and no default supplied at /etc/puppet/modules/osnailyfacter/modular/hosts/hosts.pp:6 if the task globals was not run before it. You should always add globals or the one that depends on it in the list of the tasks for a new role; the task global, in turn, depends on tasks hiera, setup_repostories, and fuel_pkgs, which provide some basic configuration of the node: for example, the hiera task configures hiera, so all of the settings defined via the Fuel web UI are available through hiera.

The next step after determining core tasks that should be run on your role, will be to define plugin tasks and the process is the same as the one that described in Actions for existing roles. The only difference is that you will use the ID of the new role in the task definition:

- id: fuel-plugin-example-iotop
  type: puppet
  role: [fuel-plugin-example_role]
  requires: [post_deployment_start]
  required_for: [post_deployment_end]
  parameters:
   puppet_manifest: puppet/manifests/install_iotop.pp
   puppet_modules: puppet/modules:/etc/puppet/modules

The difference between the deployment of the nodes with the new role and the pre-existing ones is that only the core tasks that you explicitly defined in the tasks field of the role’s group definition will run on the node with the new role, while there is a predefined list of core tasks to run on the nodes with pre-existing roles.

Tip

For debugging purposes, deploy the environment consisting only of the node with the role defined by a plugin. By default, you cannot deploy an environment without a controller from the Fuel web UI, so here is an example showing how to do this with the help of Fuel CLI:

# fuel env create --name test-plugin-role-1 --rel 2
Environment 'test-plugin-role-1' with id=3 was created!
# id=`fuel env | grep test-plugin-role-1 | tr -d  ' ' \
  |cut -d\| -f1`
# fuel --env $id settings download
# awk -F: 'BEGIN {OFS = FS} $1 ~ "fuel-plugin-example" \
  { f=1 } f && $1 ~ "enabled" { f=0; $2=" true"} \
  { print } ' settings_${id}.yaml > tmp && mv \
  tmp settings_${id}.yaml
# fuel --env $id settings upload
# fuel --env $id node set --node 10 --role
  fuel-plugin-example_role
Nodes [10] with roles ['fuel-plugin-example_role'] were added
to environment 3
# fuel --env $id deploy-changes

This example does the following:

  • Creates a new environment.
  • Downloads the settings.
  • Changes the plugin to be enabled in the settings by changing the plugin enabled parameter from false to true through AWK.
  • Uploads the settings back to the Nailgun database.
  • Adds a new node with the role fuel-plugin-example_role to the environment.
  • Deploys this new environment consisting only of one node with the new role.

You can create a new environment, enable the plugin in it, and add a single node with the new role from the Fuel web UI, using only the last command to deploy the environment then.

Creative Commons Attribution 3.0 License

Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.

Contents