Steps to reproduce:
1. Create cluster of 2 managers
2. Remove the passive manager
3. Check if you still connected to the active manager
The cli will disconnect from the manager, but you can workaround it by reconnecting to the manager (cfy profiles use)
Removing managers from cluster will now correctly remove the nodes from the profile
Also, updating the profile (like `cfy cluster update-profile`) was also added to every `cfy cluster nodes list` call, so the profile should be now always up to date in realistic cases