replication manager startup/shutdown

I’m experimenting with the percona replication manager using the documentation here: nts/blob/master/doc/PRM-setup-guide.rst

What seems to be missing is a description of how to shutdown/startup the cluster as a whole though. It seems that if you don’t shutdown all the slave and later start up the master first this can create problems with the cluster.
Is there a way to make this more robust?

thanks for suggestions, I’ll add a start/stop section. So for a clean stop, you should stop the slaves first and then the master. Upon restart, restart the previous master first. You can also do “crm resource demote ms_MySQL” before stopping and then, the order you stop or start the nodes no longer matter.