假设三节点MGR某个节点异常，需要重新把这个节点加入到MGR集群中，具体操作过程如下：
贡献者端执行（192.168.1.11）
DROP USER ‘donor_clone_user‘@‘192.168.1.12‘;
CREATE USER ‘donor_clone_user‘@‘192.168.1.12‘ IDENTIFIED BY ‘password‘;
GRANT BACKUP_ADMIN on *.* to ‘donor_clone_user‘@‘192.168.1.12‘;
INSTALL PLUGIN clone SONAME ‘mysql_clone.so‘;
接收者端执行（192.168.1.12）
DROP USER ‘recipient_clone_user‘@‘192.168.1.12‘;
CREATE USER ‘recipient_clone_user‘@‘192.168.1.12‘ IDENTIFIED BY ‘password‘;
GRANT CLONE_ADMIN on *.* to ‘recipient_clone_user‘@‘192.168.1.12‘;
INSTALL PLUGIN clone SONAME ‘mysql_clone.so‘;
SET GLOBAL clone_valid_donor_list = ‘192.168.1.11:3311‘;
exit
mysql -urecipient_clone_user -ppassword -h 192.168.1.12 -P3311
CLONE INSTANCE FROM ‘donor_clone_user‘@‘192.168.1.11‘:3311 IDENTIFIED BY ‘password‘;
exit
mysql -uroot -p
start group_replication;
select * from performance_schema.replication_group_members;