Suppose that I have a cluster with an odd-number of nodes - say 3 nodes.
In the event of a network partition, split brain is avoided by moving (or
keeping) leadership to the partition that contains two nodes.
In the event of routine maintenance, where one of the members needs to be
brought down in order to be replaced or upgraded, I now have a 2-node system
which presumably is at risk of split brain in the event of a network partition
happening during the maintenance window.
What is the best way of reducing the risk of split brain during the maintenance
window when the cluster is not at full strength ?