Tweak the unsafe failover warning

The unsafe failover warning should be updated. In particular, we should mention:

• that this happens only when a majority of nodes in the cluster can't be reached
• the requirement that all nodes that will survive the unsafe failover must be reachable
• that all unsafely failed over nodes will be immediately removed from the cluster and should be brought down and re-initialized
• we should link to the docs

My suggestion is as follows:

Failover cannot be performed safely as a majority of servers in the cluster can't be reached. If you wish you may unsafely failover the unreachable nodes. If so, all nodes that survive the unsafe failover must be reachable. Additionally, unsafely failed over nodes will be immediately removed from the cluster and administrators should ensure they are brought down and reinitialized. Do you wish to perform an unsafe failover?

The doc page we should link to is this:
https://docs.couchbase.com/server/7.0/learn/clusters-and-availability/hard-failover.html#performing-an-unsafe-failover