Resyncing Broken MySQL Replication
- Prepare the slave
- Prepare the master, copy databases and restart master
- Restart Replication Slave
Prepare the slave
SSH into the slave server.
Log into MySQL on the slave as root
mysql>STOP SLAVE;
mysql>RESET SLAVE;
mysql> exit;
Now exit and shutdown slave
$ mysqladmin shutdown -u root -p
Delete the databases on slave (DANGER … this deletes ALL databases on slave)
$ cd /usr/local/mysql/
$ sudo rm -r data
Prepare the master, copy databases and restart master
SSH into the master as root
log into mysql as root
Check users and kick them all off if more than yourself is connected
mysql> show processlist;
Reset the Master
mysql> RESET MASTER;
exit and shutdown immediately
$ mysqladmin shutdown -p
Now copy all the data from this master to the slave (may take a while)
# cd /usr/local/mysql
# scp -r data root@slavehost:/usr/local/mysql
Note: If you have a really huge amount of data and you have the drive space, it may be faster to make a local copy of the data directory on the master which can be then copied to the slave after you have restarted the master.
When copying is done, restart the master
$ sudo echo
$ sudo mysqld_safe &
Log into master and make sure it is logging.
mysql> show master status\G
*************************** 1. row ***************************
File: binary-log.002
Position: 280Binlog_do_db:
Binlog_ignore_db:
1 row in set (0.00 sec)
Make sure Position is increasing if insert and update activity is taking place on master.
Now you can breathe for a while!
Restart Replication Slave
Log into slave server and fix privileges on the data folder
$ cd /usr/local/mysql
$ sudo chown -R mysql. data
Restart server
$ sudo echo
$ sudo mysqld_safe &
Log into mysql as root
mysql> SHOW SLAVE STATUS\G;
Verify that it is replicating and you are done.
Great References:
MySQL: How do you set up master-slave replication in MySQL?
Resyncing Broken MySQL Replication
How to make MySQL replication reliable
