时间:2021-03-28 09:17:08 | 栏目:Mysql | 点击:次
本实验中分别针对空库、脱机、联机三种方式,配置一主两从的mysql标准异步复制。只做整服务器级别的复制,不考虑对个别库表或使用过滤复制的情况。
实验环境
[root@slave2 ~]# cat /etc/hosts 192.168.2.138 master 192.168.2.192 slave1 192.168.2.130 slave2 mysql> select version(); +-----------+ | version() | +-----------+ | 8.0.16 | +-----------+ 1 row in set (0.00 sec)
一、空库
1.查看主库二进制信息
mysql> show master status; +------------------+----------+--------------+------------------+-------------------+ | File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set | +------------------+----------+--------------+------------------+-------------------+ | mysql-bin.000004 | 155 | | | | +------------------+----------+--------------+------------------+-------------------+ 1 row in set (0.00 sec)
2.在主库上建立复制用户
mysql> create user 'repl'@'%' identified with mysql_native_password by 'wwwwww'; Query OK, 0 rows affected (0.03 sec) mysql> grant replication client,replication slave on *.* to 'repl'@'%'; Query OK, 0 rows affected (0.04 sec) mysql> flush privileges; Query OK, 0 rows affected (0.00 sec)
3.在从库创建主库信息
mysql> stop slave; mysql> change master to master_host='192.168.2.138', master_port=3306, master_user='repl', master_password='wwwwww', master_log_file='mysql-bin.000004', master_log_pos=155; Query OK, 0 rows affected, 2 warnings (0.00 sec) mysql> start slave; Query OK, 0 rows affected (0.00 sec) mysql> show slave status\G
刚才我们并没有在从库上建立repl用户,但由于create user语句是在起始位置点后执行的,因此可以正常复制到从库,查询mysql.user表即可确认。
sql> select * from mysql.user where user='repl'\G
二、脱机
如果数据库已经存在应用数据,但允许一个可接受的脱机时间窗口做复制,这种场景下常用的做法是先直接将主库的数据目录整体拷贝到从库,再启动复制。具体步骤如下。
1.在master节点创建测试库和测试表
CREATE DATABASE test; Query OK, 1 row affected (0.04 sec) mysql> USE test; Database changed mysql> CREATE TABLE t(id int(10)); Query OK, 0 rows affected (0.09 sec) mysql> INSERT INTO t VALUES (111); Query OK, 1 row affected (0.05 sec) mysql> INSERT INTO t VALUES (222); Query OK, 1 row affected (0.00 sec) mysql> INSERT INTO t VALUES (333); Query OK, 1 row affected (0.00 sec)
2.在主库创建复制用户
mysql> create user 'repl'@'%' identified with mysql_native_password by 'wwwwww'; Query OK, 0 rows affected (0.03 sec) mysql> grant replication client,replication slave on *.* to 'repl'@'%'; Query OK, 0 rows affected (0.04 sec) mysql> flush privileges; Query OK, 0 rows affected (0.00 sec)
3.停止复制的所有实例,在master、slave1、slave2分别执行
[root@master ~]# ln -s /usr/local/mysql/bin/mysqladmin /usr/bin/mysqladmin [root@master ~]# mysqladmin -hlocalhost -uroot -pwwwwww shutdown
4.复制数据至slave1、slave2
[root@master data]# cd /data [root@master data]# scp -r mysql/ slave1:/data/ [root@master data]# scp -r mysql/ slave2:/data/
5.在slave1、slave2从库执行命令,删除auto.cnf文件
[root@slave1 mysql]# cd /data/mysql [root@slave1 mysql]# rm -rf auto.cnf [root@slave2 mysql]# cd /data/mysql [root@slave2 mysql]# rm -rf auto.cnf
6.重启实例,在三个节点都需要执行
[root@master data]# service mysqld start Starting MySQL.. SUCCESS!
7.在主库查看二进制日志
mysql> show master status; +------------------+----------+--------------+------------------+-------------------+ | File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set | +------------------+----------+--------------+------------------+-------------------+ | mysql-bin.000005 | 155 | | | | +------------------+----------+--------------+------------------+-------------------+ 1 row in set (0.00 sec)
8.在slave1、slave2从库执行命令
mysql> stop slave; Query OK, 0 rows affected (0.01 sec) mysql> change master to master_host='192.168.2.138', master_port=3306, master_user='repl', master_password='wwwwww', master_log_file='mysql-bin.000005', master_log_pos=155; Query OK, 0 rows affected, 2 warnings (0.02 sec) mysql> start slave; Query OK, 0 rows affected (0.03 sec) mysql> show slave status\G
9.在slave1、slave2从库执行命令查看库和表是否同步过来
mysql> use test; Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Database changed mysql> show tables; +----------------+ | Tables_in_test | +----------------+ | t | +----------------+ 1 row in set (0.00 sec) mysql> select * from t; +------+ | id | +------+ | 111 | | 222 | | 333 | +------+ 3 rows in set (0.00 sec)
三、mysqldump联机
脱机建立复制的需求太过理想化,大多数情况下,复制是被要求在不影响线上业务的情况下,联机创建的,而且还要求对线上库的影响越小越好。例如,复制过程化中对主库加锁会影响对主库的访问,因此通常是不被允许的。这种场景下有两种备选的复制方案:使用mysqldump程序或使用如XtraBackup的第三方工具。这两种方案有各自的适用场合。使用mysqldump联机建立复制的过程如下。
1.在主库创建测试的数据库和表
mysql> CREATE DATABASE test; Query OK, 1 row affected (0.04 sec) mysql> use test; Database changed mysql> CREATE TABLE t(id int(10)); Query OK, 0 rows affected (0.10 sec) mysql> INSERT INTO t VALUES(111); Query OK, 1 row affected (0.09 sec) mysql> INSERT INTO t VALUES(222); Query OK, 1 row affected (0.00 sec) mysql> INSERT INTO t VALUES(333); Query OK, 1 row affected (0.05 sec) mysql> INSERT INTO t VALUES(444); Query OK, 1 row affected (0.00 sec)
2.在主库创建复制用户
mysql> create user 'repl'@'%' identified with mysql_native_password by 'wwwwww'; Query OK, 0 rows affected (0.01 sec) mysql> grant replication client,replication slave on *.* to 'repl'@'%'; Query OK, 0 rows affected (0.01 sec) mysql> flush privileges; Query OK, 0 rows affected (0.01 sec)
3.在slave1、slave2从库创建主库信息
mysql> change master to master_host='192.168.2.138', master_port=3306, master_user='repl', master_password='wwwwww'; Query OK, 0 rows affected, 2 warnings (0.04 sec)
4.在slave1、slave2从库使用mysqldump命令复制数据
[root@slave2 ~]# mysqldump --single-transaction --all-databases --master-data=1 --host=192.168.2.138 --user=root --password=wwwwww --apply-slave-statements | mysql -uroot -pwwwwww -hlocalhost mysql: [Warning] Using a password on the command line interface can be insecure. mysqldump: [Warning] Using a password on the command line interface can be insecure.
参数说明
?Csingle-transaction参数可以对Innodb表执行非锁定导出。此选项将事务隔离模式设置为REPEATABLE READ,并在转储数据之前向服务器发送START TRANSACTION SQL语句。它仅适用于Innodb等事务表,因为它会在发出START TRANSACTION时转储数据库的一致状态,而不会阻塞任何应用程序。因此这里假定:1. 所有的应用数据表都使用Innodb引擎。2. 所有系统表数据在备份过程中不会发生变化。
?Cmaster-data参数会导致转储输出包含类似 CHANGE MASTER TO MASTER_LOG_FILE=‘binlog.000004', MASTER_LOG_POS=1480; 的SQL语句,该语句指示主库的二进制日志坐标(文件名和位置)。如果选项值为2,则CHANGE MASTER TO语句将写为SQL注释,因此仅提供信息,不会执行。如果参数值为1,则该语句不会写为注释,并在重新加载转储文件时执行。如果未指定选项值,则默认值为1。
?Capply-slave-statements参数会在CHANGE MASTER TO语句之前添加STOP SLAVE语句,并在输出结尾处添加START SLAVE语句,用来自动开启复制。
通过管道操作符,导出导入一步进行,不需要中间落盘生成文件。
5.在从库确认复制状态
mysql> show slave status\G
6.在从库查看库和表是否复制成功
use test; Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Database changed mysql> select * from t; +------+ | id | +------+ | 111 | | 222 | | 333 | | 444 | | 555 | +------+ 5 rows in set (0.00 sec)
mysqldump方式的优点是可以进行部分复制,如在配置文件中定义replicate-do-table=db1.*,则用这种方法可以只复制db1库而忽略其它复制事件。缺点是由于mysqldump会生成主库转储数据的SQL语句,实际是一种逻辑备份方式所以速度较慢,不适用于大库。
四、XtraBackup联机复制
联机建立复制的另一种可选方案是使用XtraBackup。XtraBackup是Percona公司的开源项目,用以实现类似Innodb官方的热备份工具InnoDB Hot Backup的功能,它支持在线热备份,备份时不影响数据读写。到目前为止,最新的版本为Percona XtraBackup 8.0.6,可以从https://www.percona.com/downloads/下载安装包。XtraBackup有很多功能和优点,例如支持全备、增量备份、部分备份;支持压缩备份;备份不影响数据读写、事务等,但是也有缺陷不足:例如不支持脱机备份、不支持直接备份到磁带设备、不支持Cloud Back,MyISAM的备份也会阻塞。不过这些小瑕疵不影响XtraBackup成为一款流行的MySQL备份工具。另外,注意XtraBackup只支持Linux平台,不支持Windows平台。下面演示用XtraBackup联机搭建主从复制的过程,主库已经建立了用于执行复制的用户repl。
在主库创建复制用户
mysql> create user 'repl'@'%' identified with mysql_native_password by 'wwwwww'; Query OK, 0 rows affected (0.01 sec) mysql> grant replication client,replication slave on *.* to 'repl'@'%'; Query OK, 0 rows affected (0.01 sec) mysql> flush privileges; Query OK, 0 rows affected (0.01 sec)
1.在主库和从库都安装XtraBackupv
[root@master ~]# yum -y install libev [root@master home]# yum localinstall percona-xtrabackup-80-8.0.6-1.el7.x86_64.rpm -y
2.配置主库到从库的SSH免密码连接
[root@master home]# ssh-keygen Generating public/private rsa key pair. Enter file in which to save the key (/root/.ssh/id_rsa): Enter passphrase (empty for no passphrase): Enter same passphrase again: Your identification has been saved in /root/.ssh/id_rsa. Your public key has been saved in /root/.ssh/id_rsa.pub. The key fingerprint is: SHA256:GBLbrw17UVck8RyCa/fbYyLkSNZIRc5p+jPQmpkD+bI root@master The key's randomart image is: +---[RSA 2048]----+ | . .o+o+ | | + +..* . | | o o o*. o | | . +.o*.. | | ooS+oo . | | =o=Bo . | | o.=B++ o | | .o..oo..o.| | E . o .| +----[SHA256]-----+ [root@master home]# ssh-copy-id 192.168.2.138 [root@master home]# ssh-copy-id 192.168.2.192 [root@master home]# ssh-copy-id 192.168.2.130
3.停止从库,并删除从库里面的数据
[root@slave1 home]# service mysql stop [root@slave2 home]# service mysql stop [root@slave1 home]# rm -rf /data/mysql/* [root@slave2 home]# rm -rf /data/mysql/*
4.备份数据并传输
[root@master tmp]# xtrabackup -uroot -pwwwwww --socket=/data/mysql/mysql.sock --no-lock --backup --compress --stream=xbstream --parallel=4 --target-dir=./ | ssh root@192.168.2.192 "xbstream -x -C /data/mysql/ --decompress"
执行过程中报错,
190606 01:21:47 >> log scanned up to (19597291)
190606 01:21:47 Selecting LSN and binary log position from p_s.log_status
Error: failed to fetch query result SELECT server_uuid, local, replication, storage_engines FROM performance_schema.log_status: Access denied; you need (at least one of) the BACKUP_ADMIN privilege(s) for this operation
mysql> grant BACKUP_ADMIN on *.* to 'root'@'%'; Query OK, 0 rows affected (0.01 sec)
行如下命令,删除192.168.2.192:/data/mysql/*的内容,再次执行命令,发现已经正确了。成功执行如下所示:
这条命令连接主库,进行并行压缩流式备份,同时将备份通过管道操作符传输到从库,并直接解压缩到从库的数据目录。所有操作一条命令完成,不需要中间落盘生成文件。
5.在从库恢复备份
[root@slave1 /]# xtrabackup --prepare --target-dir=/data/mysql [root@slave2 /]# xtrabackup --prepare --target-dir=/data/mysql
6.在从库查看二进制bin-log日志
[root@slave1 mysql]# cat xtrabackup_binlog_info mysql-bin.000008 155 [root@slave2 mysql]# cat xtrabackup_binlog_info mysql-bin.000009 155
7.启动从库
[root@slave1 data]# service mysqld start Starting MySQL... SUCCESS! [root@slave2 data]# service mysqld start Starting MySQL... SUCCESS!
8.创建主库信息,其中的master_log_file和master_log_pos值来自第6步
mysql> change master to master_host='192.168.2.138', master_port=3306, master_user='repl', master_password='wwwwww', master_log_file='mysql-bin.000008', master_log_pos=155; Query OK, 0 rows affected, 2 warnings (0.04 sec) mysql> start slave; Query OK, 0 rows affected (0.00 sec) mysql> show slave status\G
9.在从库测试数据
mysql> use test; Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Database changed mysql> select * from t; +------+ | id | +------+ | 111 | | 222 | | 333 | | 444 | | 555 | +------+ 5 rows in set (0.00 sec)
XtraBackup是物理复制,性能比mysqldump高的多,而且对主库的影响极小,非常适用于从头联机创建高负载、大数据量、全实例从库的场景。