The convergence spec is really big to be seen and understood in entirety
without going through multiple iterations. It is probably a good idea to
break the spec into multiple implementable specs and move a chunk of
material from main convergence spec to individual implementable specs
like engine or observer or even newer specs.

One of the steps in the direction of convergence is to enable Heat
engine to persist dependency task graph and version the resources. The
main convergence spec talks about it. This spec elaborates it and
discusses what needs to be done and why back-up stack needs to be

Convergence: https://review.openstack.org/#/c/95907/7

Persisting dependency graph and resource version:

Please review it and let us know your thoughts.

- Anant

OpenStack-dev mailing list

Reply via email to