RBD pool always deleted even when use_existing_volumes is true #171
Description
When running a librbd fio test with "use_existing_volumes = true", the RBD volumes still get deleted because the librbd client endpoint calls create_rbd() in CephClientEndpoint which recreates the pool and volumes. Maybe we need an option to "use existing endpoints" to skip this process if the pool and volumes already exist?