In the wild, there will be many scenarios for which consensus, mining diversity and permissions will leave the chain in a precarious position.
for example, a node with admin going down, or 2 nodes from a partnership, take themselves offline unexpectedly.
In my latest scenario, a node sitting at AWS was unrecoverable due to a service not running to connect to the key. The hdd was detached, and put into another instance, where the IP address changed, but I brought it up anyhow. For some reason, mining stopped. We then found an address that belonged to none of the nodes, being given mining permission. It could be an accident on my part, as in copying the wrong address or something, but I have no understanding of any action that would have caused this.
In this scenario, an address that has mining permission needs to be kicked via whatever means, consensus, rollback or something, but the manual process of doing so is complicated and not foolproof.
I would run commands in the same order, sometimes bantx would work, sometimes not, and extreme difficulty to coordinate this with multiple people with different nodes. For me it is hard, just coordinating 3 in my realm, bit what in a multi-party scenario with dozens or hundreds of nodes?
If possible, running a command like, revoke mining on an address, run on all admins, could be messaged to all nodes and then perhaps the problematic address could be skipped automatically for n rounds, which would then confirm.