Use cases are really needed for novices and for those who do not practice these kind of solutions in production. That's why you'll see schema with explanations of what could happen in production. Now, let's discuss outages and how to resolve them.
In a replication mode, you can encounter errors if you do not respect what has been established. A common mistake is for the privileges to be configured on a slave node and not the master.
Here is a scenario with a database as an example:
Master: This is the server used as read/write
Slave1: This is the server used as read only
A junior database administrator connects to the slave server as someone asked for a new database and creates a database crap. The database is present on the slave node, but not the master. Two days later, another database administrator connects on the master node because he needs to perform an urgent insertion in that database. As he doesn't see it, he creates it now and performs the...