-
Story
-
Resolution: Done
-
Undefined
-
None
-
None
The default deployment model of tripleo collapses the super-conductor into the cell1 conductor. We need to make sure that we can adopt such deployment properly, while keeping the nova-operator's default deployment mode which is using a super-conductor (in cell0) and a separate cell1 conductor.
The expected outcomes:
* Alongside https://github.com/openstack-k8s-operators/data-plane-adoption/blob/main/docs/openstack/mariadb_copy.md, create a nova_adoption.md file
that describes how to adopt the 3+ nova databases ensuring the api and cell0 db are using the shared openstack db instance and cell1+ are using seperate db instnaces.
* Part of the procedure is stopping the contol plan services for the service being adopted: list how to stop the nova services on the 3 contolers
* Explain how to enable the adopted service in openshift, following the placement example https://github.com/openstack-k8s-operators/data-plane-adoption/blob/main/docs/openstack/placement_adoption.md#procedure---placement-adoption
* modify https://github.com/openstack-k8s-operators/data-plane-adoption/tree/main/tests/roles to actully test the procedure
- is depended on by
-
OSPRH-1491 As a cloud operator, I want to adopt my existing RHOSP 17.1-based compute nodes without disrupting the existing workloads, so that they continue to work with RHOSP 18.0.
-
- Closed
-
- links to