标签：load day 恢复 inno 数据库 nlp dev targe star
一、背景
工作中有需求数据库需要迁移，但是不能停服，不能锁库锁表影响业务的正常运行，所以使用XtraBackup
二、环境：
操作系统：CentOS Linux release 7.4.1708 (Core)
mysql：5.6.35
IP：192.168.0.131
xtrabackup版本：percona-xtrabackup-24-2.4.2-1.el7.x86_64.rpm
xtrabackup下载地址：https://www.percona.com/downloads/XtraBackup/LATEST/
三、安装xtrabackup
[root@host1 ~]# yum -y install perl perl-devel libaio libaio-devel perl-Time-HiRes perl-DBD-MySQL libev numactl [root@host1 ~]# rpm -Uvh percona-xtrabackup-24-2.4.2-1.el7.x86_64.rpm
四、备份和恢复
1、使用环境：本人测试使用的是为环境新搭建的mysql_multi多实例方式，具体安装步骤见：https://www.cnblogs.com/01-single/p/9051412.html
my.cnf配置：
[root@host1 ~]# cat /usr/local/mysql/my.cnf [client] [mysqld] [mysqld_multi] user=mysql password=mysql mysqld = /usr/local/mysql/bin/mysqld_safe mysqladmin = /usr/local/mysql/bin/mysqladmin [mysqld3307] socket = /tmp/mysql3307.sock port = 3307 user=mysql pid-file = /alidata1/mysql_multi/mysql3307/mysqld.pid datadir = /alidata1/mysql_multi/mysql3307/mydata log-bin= /alidata1/mysql_multi/mysql3307/log/binlog server-id = 3307001 innodb_buffer_pool_size = 256M log_error = /alidata1/mysql_multi/mysql3307/log/log-err expire_logs_days = 5 bind-address = 192.168.0.131 skip-name-resolve [mysqld3308] socket = /tmp/mysql3308.sock port = 3308 user=mysql pid-file = /alidata1/mysql_multi/mysql3308/mysqld.pid datadir = /alidata1/mysql_multi/mysql3308/mydata log-bin= /alidata1/mysql_multi/mysql3308/log/binlog server-id = 3308001 innodb_buffer_pool_size = 256M log_error = /alidata1/mysql_multi/mysql3308/log/log-err expire_logs_days = 5 bind-address = 192.168.0.131 skip-name-resolve
一定注意需要做主从的server-id不能是一样的，否则会报错
通常一般都直接使用innobackupex，因为它能同时备份InnoDB和MyISAM引擎的表
重点关注Slave_IO_Running和Slave_SQL_Runningd的状态是否为YES
2、备份：
[root@host1 ~]# innobackupex --socket=/tmp/mysql3307.sock --user=root --password=123456 --defaults-file=/usr/local/mysql/my.cnf /mysqlbackup [root@host1 ~]# innobackupex --defaults-file=/usr/local/mysql/my.cnf --socket=/tmp/mysql3307.sock --user=root --password=123456 --apply-log /mysqlbackup/2018-06-21_10-35-09/ #保持事务一致性
如果使用另外一台服务器做主从，需要传输备份的数据：
[root@host2 ~]# scp -r 192.168.0.131:/mysqlbackup/2018-06-21_10-35-09 /tmp/backup
3、恢复：
[root@host1 mysql3308]# pwd /alidata1/mysql_multi/mysql3308 [root@host1 mysql3308]# mv mydata mydatabak #备份原有的数据库 [root@host1 mysql3308]# mkdir mydata #新建数据库目录
#恢复数据库：
[root@host1 mysql3308]# innobackupex --defaults-file=/usr/local/mysql/my.cnf --datadir=/alidata1/mysql_multi/mysql3308/mydata/ --socket=/tmp/mysql3308.sock --user=root --password=123456 --copy-back /mysqlbackup/2018-06-21_10-35-09/ [root@host1 mysql3308]# chown -R mysql:mysql mydata #还原权限 [root@host1 mysql3308]# cd /usr/local/mysql/bin/
#重新启动3308数据库 [root@host1 bin]# ./mysqld_multi --defaults-file=../my.cnf stop 3308 --user=root --password=123456 [root@host1 bin]# ./mysqld_multi --defaults-file=/usr/local/mysql/my.cnf start 3308 [root@host1 bin]# netstat -nlpt | grep mysql tcp 0 0 192.168.0.131:3307 0.0.0.0:* LISTEN 35205/mysqld tcp 0 0 192.168.0.131:3308 0.0.0.0:* LISTEN 37161/mysqld
五、开启主从同步：
一、 #主：192.168.0.131 3307 mysql> GRANT REPLICATION SLAVE ON *.* TO ‘slaveuser‘@‘%‘ IDENTIFIED BY ‘slave123‘; mysql> show master status \G #从：192.168.0.131 3308 [root@host1 ~]# cat /mysqlbackup/2018-06-21_10-35-09/xtrabackup_binlog_info binlog.000001 32399093 mysql> CHANGE MASTER TO MASTER_HOST=‘192.168.0.131‘, MASTER_USER=‘slaveuser‘, MASTER_PASSWORD=‘slave123‘, MASTER_PORT=3307, MASTER_LOG_FILE=‘binlog.000001‘, MASTER_LOG_POS=32399093; mysql> start slave; mysql> show slave status\G *************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.131 Master_User: slaveuser Master_Port: 3307 ………… ………… Slave_IO_Running: Yes Slave_SQL_Running: Yes ………… ………… 1 row in set (0.00 sec) 二、 #主：192.168.0.131 3308 mysql> GRANT REPLICATION SLAVE ON *.* TO ‘slaveuser‘@‘%‘ IDENTIFIED BY ‘slave123‘; mysql> show master status \G *************************** 1. row *************************** File: binlog.000003 Position: 592 Binlog_Do_DB: Binlog_Ignore_DB: Executed_Gtid_Set: 1 row in set (0.00 sec) #从：192.168.0.131 3307 mysql> CHANGE MASTER TO MASTER_HOST=‘192.168.0.131‘, MASTER_USER=‘slaveuser‘, MASTER_PASSWORD=‘slave123‘, MASTER_PORT=3308, MASTER_LOG_FILE=‘binlog.000003‘, MASTER_LOG_POS=592; mysql> start slave; mysql> show slave status\G *************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: 192.168.0.131 Master_User: slaveuser Master_Port: 3308 ………… ………… Slave_IO_Running: Yes Slave_SQL_Running: Yes ………… ………… 1 row in set (0.00 sec)
至此互为主从已配置好，两个库均可执行读写操作，且互相同步
如果是两个不同IP的服务器做主从，只需改下IP就行，操作方式一样
标签：load day 恢复 inno 数据库 nlp dev targe star
原文地址：https://www.cnblogs.com/01-single/p/9210347.html
