infinite reconciling loop between kubefed namespace controller and ks namespace controller

Created by: swiftslee

Describe the Bug Ks ns controller will add ownerReference field to a ns if the ns has label such as kubesphere.io/workspace: xccc. However, if the ns is created by kubefed(e.g. federatednamespace), the kubefed ns controller will remove the ownerReference field. In this situation, it can cause competition between the two controllers. Finally, we can not create federatednamespace anymore.

Versions Used KubeSphere: latest

How To Reproduce Steps to reproduce the behavior:

1. create a multi-cluster-project that contains host cluster.
2. create another multi-cluster-project contains host and member cluster.
3. can not create ns in the member cluster.

Expected behavior The ns should be created in the member cluster.