MySQL主从复制结构是基于mysql bin-log日志基础上从库通过打开IO进程收到主库的bin-log日志增量信息,并保存到本地relay log,而后再通过打开MYSQL进程从relay log上获取的增量信息并翻译成SQL语句后写到从数据库。
主从复制结构实际上可以实现两个功能
1.从库充当主库的数据库备份实例
2.读写分离主库负责正常读写数据从库只负责读数据
实际生产环境因为很多应用实际读数据库的次数远大于写数据库的次数所以在项目开发初期编写程序时做一个判断对所有读的操作全部推到从库若从库无法获取数据则再向主库获取数据从而在一定意义上实现读写分离缓解主库的IO压力。
所以生产环境下推荐使用这种架构。
系统环境centos6.3 x64
数据库 mysql-5.6.10
mysql master:192.168.100.90
mysql slave:192.168.100.91
一.部署环境
1.关闭iptables和SELINUX
# service iptables stop
# setenforce 0
# vi /etc/sysconfig/selinux
---------------
SELINUX=disabled
---------------
2.安装配置mysql传送门http://showerlee.blog.51cto.com/2047005/1174141
二.主mysql配置:(mysql master)
修改mysql配置文件
# vi /etc/my.cnf
添加
-----------------
# Replication Master Server
# bin日志路径
log-bin = /usr/local/mysql/log/bin.log
# 服务器ID号
server-id = 1
# 忽略mysql系统库复制
binlog-ignore-db=mysql
binlog-ignore-db=information_schema
------------
重启服务
# service mysqld restart
登录mysql后台
# mysql -u root -p123456
查看此刻登录账号
> select user();
在master为slave添加同步帐号
> grant replication slave on *.* to 'slave'@'192.168.100.91' identified by '123456';
查看创建的用户
> select user.host from mysql.user;
查看权限
> show grants for 'slave'@'192.168.100.91';
mysql锁表只读(其他账户登录mysql后无法进行写表操作防止备份数据库后主mysql表更新导致和从数据库内容不一致)
> flush tables with read lock;
查看锁表倒计时时间
> show variables like '%timeout%';
------------------------
....
wait_timeout | 28800
------------------------
将master的数据库表全部备份导出并传送到slave服务器上。
# /usr/local/mysql/bin/mysqldump -u root -p123456 --opt --flush-logs --all-database > /root/allbak.sql
# cd ~
# scp allbak.sql root@192.168.100.91:/root
查看mysql偏移量(数据库如果有写操作偏移值会递增)
# mysql -u root -p123456 -e "show master status"
----------------------
+------------+----------+--------------+------------------+-------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------+----------+--------------+------------------+-------------------+
| bin.000009 | 120 | | | |
+------------+----------+--------------+------------------+-------------------+
----------------------
保证FILE列和Position列与从库配置一致
三.从mysql配置mysql slave
修改mysql配置文件
# vi /etc/my.cnf
添加
--------------------
# Replication Slave Server
# bin日志路径无需开bin-log日志
#log-bin = /usr/local/mysql/log/bin.log
server-id=2
# 只读
read-only
# 忽略mysql系统库复制
binlog-ignore-db=mysql
binlog-ignore-db=information_schema
---------------------
重启服务
# service mysqld restart
恢复server的数据库到slave
#/usr/local/mysql/bin/mysql -u root -p123456 < /root/allbak.sql
配置连接同步到server端
# mysql -u root -p123456;
> stop slave;
> reset slave;
> change master to master_host='192.168.100.90',master_user='slave',master_password='123456',master_log_file="bin.000009",master_log_pos= 120 ;
> start slave;
注master_log_file表示从主数据库哪个bin-log文件开始同步
master_log_pos表示从该bin-log文件哪条记录点开始同步
需与主库偏移值保持同步
回到主mysql数据库解锁mysql master
# mysql -u root -p123456
> unlock tables;
最后登陆从mysql后台查看主从连接状态
# mysql -u root -p123456 -e "show slave status\G;"
找到这五行如下则主从配置成功
------------------------
Slave_IO_State: Waiting for master to send event
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
Read_Master_Log_Pos: 120
Relay_Master_Log_File: bin.000009
-------------------------
四.测试主从是否同步
(server)
# mysql -u root -p123456 -e "create database test02;"
# mysql -u root -p123456 -e "show databases like 'test02';"
------------
+-------------------+
| Database (test02) |
+-------------------+
| test02 |
+-------------------+
------------
(cilent)
# mysql -u root -p123456 -e "show databases like 'test02';"
-------------
+-------------------+
| Database (test02) |
+-------------------+
| test02 |
+-------------------+
-------------
测试成功。。
当server端的数据库数据发生变化时client端会同步更新从而实现主库备份和读写分离作用。
这里注意几点
1.mysql从库中需在my.cnf配置文件中加入 read-only参数保证从库只读。
# echo "read-only" >> /etc/my.cnf
2.忽略从库mysql与information_schema系统表同步
# echo "binlog-ignore-db=mysql" >> /etc/my.cnf
# echo "binlog-ignore-db=information_schema" >> /etc/my.cnf
生产环境主库用户的授权,授权增删改查权限。
> GRANT SELECT,INSERT,UPDATE,DELETE ON *.* to 'user'@'%' identified by '123456';
生产环境从库的授权,仅授权查权限。
> GRANT SELECT ON *.* to 'user'@'%' identified by '123456';
3.从库默认不开启bin-log日志功能除非做下级从库级联同步才需开启从库的bin-log日志。
4.主库由于硬件故障如何将从库提升为主库(一主多从)
(mysql slave)
(1) 确保从机没有再同步的SQL语句即出现Has read all relay log再关闭从库IO_Threat进程
# mysql -uroot -p123456
> stop slave IO_THREAD
(2) 关闭从库slave服务然后将其提升为主库
> stop slave
> reset master
(3) 更换从库IP为故障主库IP(配置方法略)
(4) 删除新的主库master.info和relay-log.info,防止下次重启还会按照从库启动
# cd /usr/local/mysql/log
# rm -rf master.info relay-log.info
(5) 重新配置从库连接主库的账号同步信息以及在下级从库重新设置偏移量保持与新的主库一致即可。
最后待主库硬件恢复将其再设置为从库并更换为上述从库IP地址,完成主从切换。
五.配置脚本
1.如果想实现无人值守备份主数据库可添加如下脚本并在凌晨执行定时任务
# vi /etc/rc.d/mysql_bak.sh
---------------------
#!/bin/sh
MYSQL_USER=root
MYSQL_PW="123456"
LOG_PATH=/usr/local/mysql/log
DATA_PATH=/usr/local/mysql/data
LOG_FILE=${LOG_PATH}/mysqllog_`date +%F`.log
DATA_FILE=${DATA_PATH}/mysql_backup_`date +%F`.sql.gz
BIN_PATH=/usr/local/mysql/bin
MYSQL_CMD="$BIN_PATH/mysql -u$MYSQL_USER -p$MYSQL_PW"
MYSQL_DUMP="$BIN_PATH/mysqldump -u$MYSQL_USER -p$MYSQL_PW --opt --flush-logs --all-database"
$MYSQL_CMD -e "flush tables with read lock;"
echo "-------show master status result-------" >> $LOG_FILE
$MYSQL_CMD -e "show master status" >> $LOG_FILE
${MYSQL_DUMP}|gzip > $DATA_FILE
$MYSQL_CMD -e "unlock tables"
mail -s "mysql slave log" 1234567@qq.com < $LOG_FILE
---------------------
凌晨3:30执行备份数据库操作
# crontab -e
---------------------
30 3 * * * /bin/sh /etc/rc.d/mysql_bak.sh > /dev/mull 2>&1
---------------------
# service crond restart
2.如果想实现将主库备份的数据分发到从库恢复从库数据库以及开启从库功能可添加如下脚本
注这里建议主从先做好秘钥认证
详见传送门http://showerlee.blog.51cto.com/2047005/1217651
# vi /etc/rc.d/mysql_bak1.sh
------------------
#!/bin/sh
MYSQL_USER=root
MYSQL_PW="123456"
MYSQL_SLAVE_IP="192.168.100.91"
SSH_PATH="/usr/bin/ssh"
SSH_CMD="${SSH_PATH} ${MYSQL_SLAVE_IP}"
GZIP_CMD="/bin/gzip"
LOG_PATH=/usr/local/mysql/log
DATA_PATH=/usr/local/mysql/data
LOG_FILE=${LOG_PATH}/mysqllog_`date +%F`.log
DATA_FILE=${DATA_PATH}/mysql_backup_`date +%F`.sql.gz
BIN_PATH=/usr/local/mysql/bin
MYSQL_CMD="$BIN_PATH/mysql -u$MYSQL_USER -p$MYSQL_PW"
MYSQL_DUMP="$BIN_PATH/mysqldump -u$MYSQL_USER -p$MYSQL_PW --opt --flush-logs --all-database"
$MYSQL_CMD -e "flush tables with read lock;"
echo "-------show master status result-------" >> $LOG_FILE
$MYSQL_CMD -e "show master status" >> $LOG_FILE
${MYSQL_DUMP}|gzip > $DATA_FILE
#config slave
cd ${DATA_PATH}
scp "mysql_backup_`date +%F`.sql.gz" $MYSQL_SLAVE_IP:/tmp/
${SSH_CMD} "${GZIP_CMD} -d /tmp/mysql_backup_`date +%F`.sql.gz"
${SSH_CMD} "${MYSQL_CMD} < /tmp/mysql_backup_`date +%F`.sql"
${SSH_CMD} "cat |$MYSQL_CMD" << EOF
stop slave;
change master to
master_host='192.168.100.90',
master_user='slave',
master_password='123456',
master_log_file="bin.000009",
master_log_pos= 120;
start slave;
EOF
$SSH_CMD $MYSQL_CMD -e "show slave status\G;"|egrep "IO_Running|SQL_Running" >> $LOG_FILE
$MYSQL_CMD -e "unlock tables"
mail -s "mysql slave log" 1234567@qq.com < $LOG_FILE
$SSH_CMD mail -s "mysql slave log" 1234567@qq.com < $LOG_FILE
------------------
mysql主从my.cnf参数配置:
--------------------------------
[mysqld]
port= 3306
datadir=/usr/local/mysql/data
socket=/var/lib/mysql/mysql.sock
user=mysql
symbolic-links=0
max_connections = 16384
skip-external-locking
skip-name-resolve
key_buffer_size = 256M
query_cache_limit = 1M
query_cache_size = 64M
max_allowed_packet = 16M
#table_cache = 8
thread_concurrency = 8
sort_buffer_size = 8M
read_buffer_size = 4M
read_rnd_buffer_size = 4M
net_buffer_length = 1M
thread_stack = 1M
log-error=/usr/local/mysql/log/error.log
log=/usr/local/mysql/log/mysql.log
long_query_time=2
log-slow-queries= /usr/local/mysql/log/slowquery.log
# 服务器ID号(主为1从为2)
server-id = 1
# 忽略mysql系统库复制
binlog-ignore-db=mysql
binlog-ignore-db=information_schema
log-bin = /usr/local/mysql/log/bin.log
# 删除10天之前的二进制日志
expire_logs_days = 10
[mysqldump]
quick
max_allowed_packet = 16M
[mysql]
no-auto-rehash
#safe-updates
[isamchk]
key_buffer = 8M
sort_buffer_size = 8M
[myisamchk]
key_buffer = 8M
sort_buffer_size = 8M
[mysqlhotcopy]
interactive-timeout
--------------------------------
-------大功告成---------
