使用percona-xtrabackup工具对mysql数据库的备份方案

使用percona-xtrabackup工具对mysql数据库的备份方案

需要备份mysql的主机 172.16.155.23
存放备份mysql的主机 172.16.155.22

目的:将155.23主机上mysql数据库每天全量备份一次,并且同步到远程主机中

1.安装备份工具

yum localinstall -y percona-xtrabackup-2.2.9-5067.el6.x86_64.rpm

2.编写备份数据库脚本:

vim /data/scripts/backdb.sh

#!/bin/bash

back_time=`date +%Y%m%d%H%M%S`
backpath="/app/mysqlbak"
cd $backpath && /bin/mkdir $back_time

echo "start back ${back_time}" >> /app/mysqlbak/backdb.log
# completely backup
/usr/bin/innobackupex --defaults-file="/etc/my.cnf" --user=root --password=pass --socket=/tmp/mysql.sock $backpath/$back_time

/usr/bin/scp -P 58422 -l 200000 -r $backpath/$back_time 172.16.155.22:/app/mysqlbak/

# delete 15 days ago backup
/bin/find /app/mysqlbak -type d -mtime +15 -exec rm -rf {} \;

echo "end back ${back_time}" >> /app/mysqlbak/backdb.log

chmod +x /data/scripts/backdb.sh

3.配置155.23免密码访问155.22
[root@sdtw04 20171127]# ssh-keygen -t rsa -P ''
如果ssh端口非默认的22端口,使用如下命令:
ssh-copy-id -i /root/.ssh/id_rsa.pub "-p 58422 [email protected]"

计划任务,每天晚上1点1分备份一次:
1 1 * * * /bin/bash /data/scripts/bakdb.sh >/dev/null 2>&1

4.目标备份服务器配置,删除7天以前的备份:
mkdir -p /app/mysqlbak/

2 2 * * * /bin/find /app/mysqlbak/ -mtime +7 -exec rm -rf {} \;


数据库的临时方案:
1.配置两台服务器都启动mysql服务,其中一台是目前正在运行的服务器,另外一台作为备用
2.每天备份一次数据,然后传送到远程服务器,删除10天以前的备份
3.如果正在使用的mysql宕机则恢复最新数据到备用mysql,然后恢复最新数据到备用mysql
记得修改解析/etc/hosts文件
172.16.155.23 db.mysql.com

恢复的具体操作:

恢复准备工作:
目标服务器安装和原服务器一样的Mysql版本(见最下面的附录)

# 1.完整备份现有mysql的数据库
# completely backup
/usr/bin/innobackupex --defaults-file="/etc/my.cnf" --user=root --password=pass --socket=/tmp/mysql.sock /app/mysqlbak/back_time

# 2.传送备份数据到目标机器
scp -P 58422 -r 2017-11-27_16-52-50 172.16.155.22:/app/mysqlbak/

# 3.目标机器执行:

# 因此,我们现在就是要通过回滚未提交的事务及同步已经提交的事务至数据文件也使得数据文件处于一致性状态
innobackupex --user=root --password=pass --defaults-file=/tmp/my.cnf --apply-log /app/mysqlbak/2017-11-27_16-52-50

# 4.关掉服务,迁移已有的数据目录
service mysqld stop
# cd /app/data
[root@sdtw03 data]# mv mydata mydata_old
[root@sdtw03 data]# mkdir mydata
[root@sdtw03 data]# chown -R mysql.mysql mydata
# 5.执行innobackupex恢复命令
innobackupex --defaults-file=/etc/my.cnf --user=root --password=pass --copy-back /app/mysqlbak/2017-11-27_16-52-50

# 6.修改权限,启动服务
chown -R mysql.mysql /app/data/mydata
service mysqld start


安装数据库mysql5.6二进制版本
安装包mysql-5.6.36-linux-glibc2.5-x86_64.tar.gz

groupadd -r mysql
useradd -g mysql -r -s /sbin/nologin mysql
tar -zxvf mysql-5.6.36-linux-glibc2.5-x86_64.tar.gz -C /usr/local

cd /usr/local
ln -sv mysql-5.6.36-linux-glibc2.5-x86_64 mysql

cd mysql
chown -R root.mysql ./*

mkdir /app/data/mydata -p
chown -R mysql.mysql /app/data/mydata
mkdir /app/data/binlogs
chown -R mysql.mysql /app/data/binlogs

cd /usr/local/mysql
scripts/mysql_install_db --datadir=/app/data/mydata --user=mysql

将服务脚本拷贝到启动项
cp support-files/mysql.server /etc/rc.d/init.d/mysqld
chkconfig --add mysqld
chkconfig --list mysqld

\cp my.cnf /etc/my.cnf

vim /etc/my.cnf

[client]
port = 3306
socket = /tmp/mysql.sock
default-character-set = utf8

[mysqld]
port = 3306
innodb_file_per_table = 1
init-connect = 'SET NAMES utf8mb4'
character-set-server = utf8mb4
default_storage_engine = InnoDB
skip-name-resolve
skip-external-locking

max_connections = 2000  # 最大连接数

datadir = /app/data/mydata
log-bin=/app/data/binlogs/master-bin
binlog_format=row

socket=/tmp/mysql.sock

interactive_timeout = 28800
wait_timeout = 28800

sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES

[mysqldump]
quick
max_allowed_packet = 16M

[myisamchk]
key_buffer_size = 8M
sort_buffer_size = 8M
read_buffer = 4M
write_buffer = 4M

# 启动服务
service mysqld start

删除匿名用户,配置root密码
mysql> delete from user where host='::1';
mysql> delete from user where host='sdtw03';
mysql> delete from user where host='localhost' and user='';

mysql> update user set password=PASSWORD('pass') where user='root';
mysql> flush privileges;

添加yun具有所有权限
mysql>grant all privileges on *.* to [email protected] identified by 'pass';
mysql>flush privileges;

 

修改主机名后,mysql无法正常重启,报错如下:

[root@sdtw04 ~]# /etc/init.d/mysqld restart
ERROR! MySQL server PID file could not be found!
Starting MySQL.........

是因为mysql的错误日志绑定在了原来的机器名上,需要强制关闭mysql
kill -9 pid
然后再次启动即可

你可能感兴趣的:(使用percona-xtrabackup工具对mysql数据库的备份方案)