xtrabackup 在线主从搭建

因为意外导致某个MySQL的从服务器宕机，且不可修复，因为是业务数据库，不能停机和锁表进行从库的搭建，所以考虑了使用xtrabackup 进行在线主从搭建。

一、数据库环境

二、准备安装包

percona-xtrabackup-2.1.9-744-Linux-x86_64.tar.gz

yum install perl-DBI

yum install perl-DBD -MySQL

yum install perl-Time-HiRes

yum install perl-lO-Socket-SSL

三、安装

1.1、安装 xtrabackup 包

[[email protected]] #tar -xvf percona-xtrabackup-2.1.9-744-Linux-x86_64.tar.gz

[[email protected] soft]# cd percona-xtrabackup-2.1.9-Linux-x86_64/

[[email protected] percona-xtrabackup-2.1.9-Linux-x86_64]# ls

bin  share

[[email protected] percona-xtrabackup-2.1.9-Linux-x86_64]# cd bin/

[[email protected] bin]# ls

innobackupex  innobackupex-1.5.1  xbcrypt  xbstream  xtrabackup  xtrabackup_55  xtrabackup_56

[[email protected] bin]# pwd

/data/soft/percona-xtrabackup-2.1.9-Linux-x86_64/bin

1.2、将innobackupex,xtrabackup等文件拷贝到*/mysql/bin 的目录下

[[email protected] bin]# cp * /data/soft/mysql/bin/

[[email protected] bin]# ln -s /data/soft/mysql/bin/* /usr/bin/

主从环境都要搭建到此位置。

2、数据库 创建复制用户和备份用户

2.1 在主库上创建备份用户

mysql> CREATE USER ‘bkpuser‘@‘localhost‘ IDENTIFIED BY ‘bkppass‘;

mysql> GRANT RELOAD, LOCK TABLES, REPLICATION CLIENT,PROCESS,SUPER ON *.* TO ‘bkpuser‘@‘localhost‘;

mysql>  FLUSH PRIVILEGES;

2.2在主库上创建复制用户

mysql>grant replication slave,reload,super on  *.* to [email protected] identified by ‘xxxxx‘;

mysql>flush privileges;

3.1 开始全量备份

需要修改下配置文件，但是不需要重启；加上以下俩个参数，否则在备份时候会报错。主从都要添加两行参数。

[[email protected] bin]# cat /etc/my.cnf

innodb_log_file_size = 5M

datadir=/data/mysql5

3.2 开始备份

[[email protected] soft]# export BKP_PASS="bkppass"

[[email protected] soft]# innobackupex --defaults-file=/etc/my.cnf --host=‘localhost‘ --port=3306 --user=bkpuser --password=${BKP_PASS}   /data/soft/（备份位置）

备份成功显示如下字样：

171024 09:07:45  innobackupex: Connection to database server closed

171024 09:07:45  innobackupex: completed OK!

3.3  把备份传到从库的服务器上

[[email protected] soft]# scp -r   2017-10-26_21-49-35   192.168.0.13:/data/soft

4、从库操作

4.1 全量恢复

[[email protected] soft]#  innobackupex --use-memory=2G --apply-log  /data/soft/2017-10-26_21-49-35/

成功显示如下字样

171024 17:11:04  innobackupex: completed OK!

4.2  确认从库的是关闭的（停掉从库），datadir 目录是空的,开始恢复数据

[[email protected] bin]# cat /etc/my.cnf

server-id = 2       #不要和主库的重复

[[email protected] mysql5]#  /etc/init.d/mysqld stop

[[email protected] mysql5]# cd /data/mysql5/

[[email protected] mysql5]# rm -rf *

[[email protected] mysql5]#  innobackupex  --defaults-file=/etc/my.cnf --copy-back    /data/soft/2017-10-26_21-49-35/

71024 17:13:09  innobackupex: completed OK!

4.3  修改数据的属组 属主

[[email protected] mysql5]# chown mysql:mysql  /data/mysql5 -R

4.4  启动数据库

[[email protected] mysql5]# /etc/init.d/mysqld start

Starting MySQL..                                           [确定]

[[email protected] mysql5]# mysql -uroot -pxxxxxx

5.1 、修改复制起始位置

[email protected] mysql5]# cd   /data/soft/2017-10-26_21-49-35/

[email protected] mysql5]#   cat xtrabackup_binlog_info

mysql-bin.000002    532     （记住文件偏移量和位置）

mysql> change master to master_host=‘192.168.0.123‘,master_user=‘rep‘,master_password=‘xxxxxx‘,master_log_file=‘mysql-bin.000002‘,master_log_pos=532;

Query OK, 0 rows affected, 2 warnings (0.23 sec)

5.2  查看从库同步状态

mysql>  show slave status\G

*************************** 1. row ***************************

Slave_IO_State:

Master_Host: 192.168.20.3

Master_User: rep

Master_Port: 3306

Connect_Retry: 60

Master_Log_File: mysql-bin.000005

Read_Master_Log_Pos: 1390

Relay_Log_File: localhost-relay-bin.000001

Relay_Log_Pos: 4

Relay_Master_Log_File: mysql-bin.000005

Slave_IO_Running: No

Slave_SQL_Running: No

注意：要在主库赋予新的从库复制权限

在主库上执行：

mysql> grant all privileges on *.* to [email protected]‘192.168.0.13‘ identified by ‘xxxxxx‘;

在从库上开启复制进程：

mysql> start slave;

Query OK, 0 rows affected (0.00 sec)

mysql> show slave status\G

*************************** 1. row ***************************

Slave_IO_State: Waiting for master to send event

Master_Host: 192.168.20.3

Master_User: rep

Master_Port: 3306

Connect_Retry: 60

Master_Log_File: mysql-bin.000005

Read_Master_Log_Pos: 1390

Relay_Log_File: localhost-relay-bin.000002

Relay_Log_Pos: 267

Relay_Master_Log_File: mysql-bin.000005

Slave_IO_Running: Yes

Slave_SQL_Running: Yes

1 row in set (0.00 sec)

6、施工后校验

1: 主库新建测试表rep_test，并在从库查看是否创建成功

【MASTER】

#mysql -u root -p****** -h127.0.0.1

mysql>use laclocal;

mysql>create table rep_test (id int,name varchar(20));

【SLAVE】

#mysql -u root -p******

mysql>use laclocal;

mysql>show create table rep_test;

2: 主库对rep_test测试表进行增删改查测试，并在从库查看该表是否进行了相应的改动

3: 主库对rep_test进行删除表操作，并在从库查看是否删除成功

【MASTER】

#mysql -u root -p******

mysql>use laclocal;

mysql>drop table rep_test;

【SLAVE】

#mysql -u root -p******

mysql>use laclocal;

mysql>show columns from rep_test;

7、基于  xtrabackup 进行在线主从搭建施工完毕！

原文地址：https://www.cnblogs.com/lixinliang/p/8672462.html