Refactor: CTS: Split crm_standby test out of tools.

This one is a little harder to verify against existing regression test
output because it starts with a blank CIB and only creates a single node
to manipulate. However, the expected output has not changed.