To ensure that OCL doesn't adversely affect other OpenShift components, we should write a periodic job that performs the following steps:

• Bring up an OCP cluster.
• Opt both the control plane and worker pools into OCL and wait for that process to complete.
• Run the openshift-e2e and openshift-e2e-upgrade tests against the cluster.
• Collect any useful information for debugging jobs and clean up after itself.

In the past, I adapted one of the MCO's devex helpers to serve in this role and a jobs were created that did basically this. See: https://github.com/openshift/release/blob/dfcbe41f07483038c067045a8d97f1f26f67f7ab/ci-operator/config/openshift/machine-config-operator/openshift-machine-config-operator-main.yaml#L478-L525 for reference.

However, it has come to light that there are reusable Prow CI steps that can perform this same task, specifically:

• enable-ocl
• Example test job

For now, we should disable tests which target MachineConfigNode within this test job.

Done When:

• The aforementioned jobs in the MCO's config are modified to use the enable-ocl step to opt into OCL.
• Those jobs are modified to be periodic jobs executed at some predetermined interval.
• A mechanism to alert us to failures of those jobs is surfaced. Care should be taken to avoid alarm fatigue.