If a ROSA HCP customer uses the default worker security group that the CPO creates for some other purpose (i.e. creates their own VPC Endpoint or EC2 instance using this security group) and then starts an uninstallation - the uninstallation will hang indefinitely because the CPO is unable to delete the security group.

https://github.com/openshift/hypershift/blob/9e6255e5e44c8464da0850f8c19dc085bdbaf8cb/control-plane-operator/controllers/hostedcontrolplane/hostedcontrolplane_controller.go#L317-L331    

4.14.8    

100%    

    1. Create a ROSA HCP cluster
    2. Attach the default worker security group to some other object unrelated to the cluster, like an EC2 instance or VPC Endpoint
    3. Uninstall the ROSA HCP cluster

The uninstall hangs without much feedback to the customer    

Either that the uninstall gives up and moves on eventually, or that clear feedback is provided to the customer, so that they know that the uninstall is held up because of an inability to delete a specific security group id. If this feedback mechanism is already in place, but not wired through to OCM, this may not be an OCPBUGS and could just be an OCM bug instead!