1、需要提升为主的从库，停止io线程等待slave数据全部更新完毕
stop slave IO_THREAD
#show processlist的输出，直到看到状态是Slave has read all relay log; waiting for the slave I/O thread to update it，表示更新都执行完毕
2、需要提升为主的从库：
stop slave 停止slave
reset master
grant replication slave on *.* to 'repl'@'192.168.0.100' identified by '123456';
reset slave all
3、其他从库
stop slave
change master #到新主库
start slave
删除新主服务器上的master.info和relay-log.info文件