http://bugzilla.suse.com/show_bug.cgi?id=1135973 http://bugzilla.suse.com/show_bug.cgi?id=1135973#c1 --- Comment #1 from Thorsten Kukuk <kukuk@suse.com> --- I discussed already with rbrown that we need a better description, as I was afraid that people will get this mixed up with velum. The "Admin Node" role is the node running "KubicD". Today we can say, that the "master node" and "admin node" are identical. But if you think about multi-master configurations (currently experimental in kubeadm), then an admin node is always a master node, but a master node does not need to be necessarily an admin node. So I don't like "Master Node", because then users would install every master node with this system role, and we would have e.g. three salt master and three kubicd, but none of them works together. What we need to make clear to the user: - We need exactly one admin node (which will also be a master node) - All other machines have to be installed with the second system role (currently worker node), and with kubicctl you will define which ones will become additional kubernetes master and which ones worker. Similar to velum. I'm happy about every better suggestion. -- You are receiving this mail because: You are the assignee for the bug.