Project Tripleo-Heatable

This project aims to provide an interface that will consume heat templates and produce Ansible playbooks.

Interface

CLI interface which uses the heat template arguments to read templates and parse them. The heatable command is the user interface for interacting with the parser.

Installation

python -m venv heatable
heatable/bin/pip install tripleo-heatable

Once installed, the heatable command will be available within the virtual environment.

Usage

The heatable command will parse heat templates and generate a set of Ansible playbooks. These playbooks will map all items discovered in the heat templates to appropriate modules within the Ansible paradigm. In the event that an interaction does not have a valid mapping, an error will be raised.

heatable --stack overcloud \
         --templates $THT/ \
         --environment-file $THT/environments/disable-telemetry.yaml \
         --environment-file $THT/environments/enable-swap.yaml \
         --environment-file $THT/environments/ssl/inject-trust-anchor.yaml \
         --environment-file /home/cloud-user/local_images.yaml \
         --environment-file /tmp/templates/environments/docker.yaml \
         --environment-file /home/cloud-user/parameters.yaml \
         -role $THT/roles_data.yaml \
         --disable-validations \
         --validation-errors-nonfatal
         --output-directory ~/generated-playbooks

This invocation will generate a set of Ansible playbooks based on the provided environment and heat template options. The output playbooks will be monolythic in nature but will remain separated based on the original template context. To use these playbooks, the entry point is the site.yml however, all generated playbook should be executable on their own.

Generated Playbooks

The provided output directory will contain the set of ansible playbooks required to perform a given heat interaction. The generated playbooks are a direct translation of the heat templates which means the heatable command is not a new deployment method and will not create inventory. The output structure will be something similar to this.

ls -1 ~/generated-playbooks
site.yml
host-prep.yml
OS--TripleO--Services--Aide.yml
OS--TripleO--Services--Pacemaker.yml
OS--TripleO--Services--Timesync.yml
...

The site.yaml file will contain the order in which all playbooks should be executed for a successful deployment.

All instances of ::SoftwareConfig and ::SoftwareDeployment will be converted into stand-alone shell scripts which are then executed by Ansible on a defined "server" using the script module. This will ensure expected software deployments continue to function as expected.