Algorithm of MySQL Master-Slave Replication

Question

I've configured a master-slave structure for my database by following this tutorial. It works well and all is fine. I mean, any new change that happens in the master, will be applied to the slave in a few secs.

Just I have 5 questions about how it works.

1. In the master server MySQL config file, I replaced the bind-address value from 127.0.0.1 to the server IP itself. Why? The localhost and the server's IP aren't the same? Don't get me wrong, I'm saying the IP of the server itself, not the IP of the slave server.

2. I want to know, either does the master send new logs to the slave (using a cron job or something) in a specific interval or the slave is keep calling the master (using a cron job or something) to see if there is any new change in the binlog file to apply on the replicated database?

3. What's the mysql-relay-bin file? I saw it on the slave server. I know there is a mysql-bin file on the master that contains the database-write-changes and the slave server uses it to update itself. But, what's the usage of mysql-relay-bin?

4. There are lots of mysql-bin files created also in the slave server. Why? I think just those files should be in the master server. Can I delete them all from the slave server? Also, the last file is mysql-bin.000043 on the slave and the last file is mysql-bin.000004 on the master server. So, why 43 binlog files on the slave and just 4 binlog files on the master?

5. Is there any approach to delete binlog files from the master server that have been applied to the slave server? Since they have been successfully replicated/copied to the slave server and generally we don't need them anymore.

Answer 1

1. Bind address -- (I don't have the answer.) @ErgestBasha suggests https://superuser.com/questions/897699/what-is-the-difference-between-127-0-0-1-and-my-assigned-ipv4-address
2. The Replica "pulls" the data from the "Primary". Well, actually, the Primary simultaneously sends stuff to both its binlog and to each Replica.
3. The Replica writes the incoming data to its "relay log", from which it reads stuff to apply to its copy of the databases. "Cron" is not invloved.
4. The Replica may have its own "binlog" -- this is in case it is a step toward further Replicas downstream. They are not identical files between any two servers in the replication topology. They may have the same data, but probably spread among the mysql-bin.* files differently.
5. Use am expire_log_... setting to get the binlogs purged automatically, rather than building up forever. Set this separately on each server.