Replication needs to be rest when When source/master and replica/slave have got out of sync. Execute steps 1 and 2 on the source/master server; 3, 4 and 5 on the replica/slave. Note that since 8.0, the preferred reference word for a "master" server is "source", and the preferred word for a "slave" server is "replica".
1 Reset source/master: On the source/master server, in the mysql client program, execute ... RESET MASTER; FLUSH TABLES WITH READ LOCK; SHOW MASTER STATUS\GNote the File , Position values, save them somewhere.
2 Dump databases: If the databases are not too large, the traditional mysqldump utility should suffice. If all master databases are being replicated, from the master terminal prompt execute ... mysqldump -uUSR -pPWD --add-drop-database -K -E -R --all-databases >BACKUPPATH... or if only some databases are being replicated then ... mysqldump -uUSR -pPWD --add-drop-database -K -E -R --databases DBLIST >BACKUPPATH... in each case substituting your values for USR, PWD, DBLIST, BACKUPPATH .
For larger databases, consider using one of the tools described here. All these dump tools are more than an order of magnitude faster. After source/master data has been dumped and saved, on source/master in the mysql client program or in mysql shell, execute ... UNLOCK TABLES;3 Stop replica/slave, import the dump file: On the replica/slave ... STOP SLAVE;Import the synchronising dump file on the replica/slave, if mysqldump or mysqlpump was used then in the mysql client program with the source command, otherwise with the appropriate recovery method for the dump tool you used.
4 Reset the replica/slave: In the replica's client program execute ... RESET SLAVE; CHANGE MASTER TO MASTER_LOG_FILE='FILE', MASTER_LOG_POS=POSITION;... substituting for FILE, POSITION the values from step 1.
5 Restart the replica/slave: START SLAVE; SHOW SLAVE STATUS\G Interruptions for maintenanceIf source and replicas are in sync but replication needs to be suspended, e.g., for maintenance, consider this less time-consuming sequence ...1 Take all servers offline: On the source, execute ... FLUSH TABLES WITH READ LOCK; 2 Stop replication on replicas: Ensure that STOP SLAVE;3 Reconfigure and restart the source: Restart source with skip-networking=ON and session sql_log_bin=OFF, then do the maintenance on source and replicas.
4 Restart source: Restart source with code>skip-networking=OFF and session sql_log_bin=ON. 5 Restart replicas: On each replica execute ... START SLAVE; SHOW SLAVE STATUS; Last updated 16 Jun 2022 |
![]() |