Failed to delete Internal LB resources when k8s service was deleted on GKE

[yandongxiao]

Describe the bug

This only happens on internal LB (not on external LB). When we uninstall SR cluster, the LB related GCP resource is not cleaned up, making the next round of SR helm install failing. It complains the LB IP I pre-alloc and put in helm yaml file - still IN USE.

I had to delete the related resources to enable SR cluster reinstall. Including:

1. the fwd rules. gcloud compute forwarding-rules delete xxx
2. the backend-services. get from gcloud compute backend-services list.

[yandongxiao]

StarRocks Operator now does not handle GCP related resources. It just delete the resource it created.
StarRocks Operator does not do anything adaptation on public provider.