From the Riak CS docs, I understand that only one Stanchion server should be used for a cluster. In my cluster I'm spinning up identical machines that all have the same setup, so that theoretically any machine could takeover if needed.
Ideally I would like to have Stanchion also running on all servers and if the main Stanchion server fails, I want another server to take over that role. Can this work or do I need to isolate the Stanchion server on to a separate machine (with a failover machine)? Or can I configure a list of IPs for the Stanchion server, so that if the main one becomes unavailable automatically the next one is tried?