When we create a MOSC to enable OCL in a pool, and then we delete the MOSC resource to revert it, then the MOSB and CMs are garbage collected but we need to wait a long and random time until the nodes are updated with the new config.
    

$ oc get clusterversion
NAME      VERSION                                                AVAILABLE   PROGRESSING   SINCE   STATUS
version   4.18.0-0.test-2024-10-15-080246-ci-ln-0gsqflb-latest   True        False         8h      Cluster version is 4.18.0-0.test-2024-10-15-080246-ci-ln-0gsqflb-latest

    

Always
    

    1. Create a MOSC to enable OCL in the worker pool
    2. Wait until the new OCL image is applied to all worker nodes
    3. Remove the MOSC resource created in step 1

    

MOSB and CMs are cleaned, but the nodes are not updated. After a random amount of time  the nodes are updated. (Somewhere around 10-20 minutes)
    

There should be no long pause between the deletion of the MOSC resource and the beginning of the nodes update process.
    

As a workaround, if we add any label to the worker pool to force a sync operation the worker nodes start updating immediately.