Goal

Migrate single-stack and dual-stack IPv6 jobs to the Vexxhost shared cloud. This may require reworking our jobs depending on the networking configuration available.

Why is this important?

This allows us to remove the users of the hwoffload self-managed cloud, allowing us to free up this hardware and transfer it to the managed cloud.

Scenarios

• As a CI user, I would like to benefit from the scalability and stability that the managed private cloud provides.

Acceptance Criteria

• All jobs currently targeting the hwoffload cloud should be migrated to the vexxhost shared cloud.

Dependencies (internal and external)

None.

Previous Work (Optional):

I (rh-ee-stephenfin) took a look at this previously. My notes:

• These jobs appear to be the only users of the hwoffload cloud right now (based on running ag -l --ignore openshift-priv hwoffload ci-operator/config/ in the release repo)
• The hwoffload cloud is currently configured with dev-install using the configuration file found here. It is worth diff this file with the configuration for nfv and mecha-central as this will identify specific configuration that is needed from the underlying cloud. We will almost certainly need to talk to vexxhost to establish the viability of this exercise once we know what we actually need.
• The openshift-e2e-openstack-dualstack workload if used for all of these jobs. This is effectively identical to the openshift-e2e-openstack-ipi workflow except for some additions. openshift-e2e-openstack-ipi workflow, openshift-e2e-openstack-dualstack workflow.

Open questions::

None.