When GKM is undeployed/uninstalled while workload pods still have extracted kernel cache volume mounted, need to determine how GKM should behave.
- There is no way to keep the Operator/Agent from being removed, so how will the extracted cache get cleaned up?
- The GKMCache/ClusterGKMCache and GKMCacheNode/ClusterGKMCacheNode already have Finalizers attached, so the CRDs won't be deleted when GKM is undeployed/uninstalled. How will the FInalizers be removed once the workload pods are stopped?
- If GKM is reinstalled, need to reconcile the usage and cache database files that are now stale.