Re: update slave after failover

[Rick James]

SLAVE START;

Replication is designed to recover from network failures. Even the SLAVE START should not be necessary, but perhaps it gave up after some number of retries.

SHOW SLAVE STATUS \G
will show more info; let's see that.

Replication works like this:
1. The Master executes a write statement.
2. The Master writes that statement to its binlog (a rotating set of files).
3. The statement is also sent to the Slave(s).
4. The Slave (assuming the network is up, and the Slave is up, etc) write the statement to its "relay log" (a different set of files).
5. The Slave reads from the relay log and executes the statements.

If the network goes down, #3 is delayed, not lost. Statements remain in the binlog(s) until they are purged, which is either manual, or configured via a setting. I like to keep at least 7 days' worth.