搭建keepalived+mysql主从复制高可用
准备工作:
- 完成keepalived的安装
- 完成docker的安装
- docker镜像里面自行安装
iproute2
,vim
,iputils-ping
(可选)等工具,便于测试
apt-get install iproute2 apt-get install vim apt-get install iputils-ping
主数据库master
1. 使用docker安装mysql
mkdir -p ~/compose/mysql-master cd ~/compose/mysql-master cat docker-compose.yml version: '2' services: mysql-master: image: mysql:5 restart: always container_name: mysql-master ports: - 3306:3306 volumes: - ./conf.d:/etc/mysql/conf.d - /data/docker/mysql-master/data:/var/lib/mysql environment: - MYSQL_ROOT_PASSWORD=123456 networks: default: external: name: service
假如没有把3306端口映射到宿主机,在宿主机上可通过docker-ip:3306来访问。
2. mysql的配置
cat conf.d/lowercase.cnf [mysqld] lower_case_table_names = 1 default-time-zone = '+08:00' character-set-server = utf8 event_scheduler = on log-bin = mysql-bin server-id = 1
参数说明:
lower_case_table_names
设置不区分大小写default-time-zone
设置时区为东八区character-set-server
修改字符集为utf8log-bin
开启二进制日志server-id
设置server-id
master开启二进制日志后默认记录所有库所有表的操作,可以通过配置来指定只记录指定的数据库甚至指定的表的操作,具体在mysql配置文件的[mysqld]可添加修改如下选项:
# 不同步哪些数据库
binlog-ignore-db = mysql binlog-ignore-db = test binlog-ignore-db = information_schema
# 只同步哪些数据库,除此之外,其他不同步
binlog-do-db = mydatabase
3. 启动服务
docker-compose pull && docker-compose up -d
4. 在宿主机连接mysql
说明: 由于我是使用虚拟机安装的字符版Ubuntu系统,所以使用MyCli作为mysql命令行工具来连接mysql。
# 查看master主机的ip地址为192.168.11.188,使用MyCli连接mysql mycli -h 192.168.11.188 -u root -p 123456 # 给root用户分配远程访问权限: grant all on *.* to root@'%' identified by "123456"; flush privileges; # 查看master状态 mysql root@192.168.11.188:(none)> SHOW MASTER STATUS; +------------------+----------+--------------+------------------+-------------------+ | File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set | +------------------+----------+--------------+------------------+-------------------+ | mysql-bin.000003 | 586 | | | | +------------------+----------+--------------+------------------+-------------------+ 1 row in set Time: 0.004s
需要记录主数据库的二进制文件名(mysql-bin.000003)和位置586。
从数据库slave
1. docker-compose.yml
只需要从主数据库的配置名称由mysql-master
改为mysql-slave
即可。
2. mysql配置
cat conf.d/lowercase.cnf [mysqld] lower_case_table_names = 1 default-time-zone = '+08:00' character-set-server = utf8 event_scheduler = on server-id= 2
需添加server-id
并且与主数据库中不一致
3. 启动服务
docker-compose pull && docker-compose up -d
4. 在宿主机连接mysql
# 查看slave主机的ip地址为192.168.11.186,使用MyCli连接mysql mycli -h 192.168.11.186 -u root -p 123456 # 给root用户分配远程访问权限,略 # 执行同步SQL语句,参照MASTER配置: CHANGE MASTER TO MASTER_HOST='192.168.11.188',MASTER_USER='root',MASTER_PASSWORD='123456',MASTER_LOG_FILE='mysql-bin.000003',MASTER_LOG_POS=586; # 启动slave同步进程 start slave; # 查看slave状态: show slave status\G;
其中下面两项为YES则表示成功:
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
测试主从复制
在MASTER中新建数据库和表,发现数据在SLAVE中已经实时同步过来
Keepalived监控mysql服务
1. master主机上的配置
cat /etc/keepalived/keepalived.conf vrrp_script chk_mysql_port { #检测mysql服务是否在运行。有很多方式,比如进程,用脚本检测等等 script "/opt/chk_mysql.sh" #这里通过脚本监测 interval 2 #脚本执行间隔,每2s检测一次 weight -5 #脚本结果导致的优先级变更,检测失败(脚本返回非0)则优先级 -5 fall 2 #检测连续2次失败才算确定是真失败。会用weight减少优先级(1-255之间) rise 1 #检测1次成功就算成功。但不修改优先级 } vrrp_instance VI_1 { state MASTER interface ens33 #指定虚拟ip的网卡接口,不一定是eth0根据ifconfig确定 virtual_router_id 51 #路由器标识,MASTER和BACKUP必须是一致的 priority 100 #定义优先级,数字越大,优先级越高,在同一个vrrp_instance下,MASTER的优先级必须大于BACKUP的优先级。这样MASTER故障恢复后,就可以将VIP资源再次抢回来 advert_int 1 authentication { auth_type PASS auth_pass 123456 } virtual_ipaddress { 192.168.11.25 } track_script { chk_mysql_port } }
需要配置的地方有:script
,state
,interface
,virtual_router_id
,priority
,virtual_ipaddress
等
2. slave主机上的配置
cat /etc/keepalived/keepalived.conf vrrp_script chk_mysql_port { script "/opt/chk_mysql.sh" interval 2 weight -5 fall 2 rise 1 } vrrp_instance VI_1 { state BACKUP interface ens33 virtual_router_id 51 priority 99 advert_int 1 authentication { auth_type PASS auth_pass 123456 } virtual_ipaddress { 192.168.11.25 } track_script { chk_mysql_port } }
只需要设置state
为BACKUP
, priority
比MASTER低即可。
3. 监测监本的配置
cat /opt/chk_mysql.sh #!/bin/bash counter=$(netstat -na|grep "LISTEN"|grep "3306"|wc -l) if [ "${counter}" -eq 0 ] then /etc/init.d/keepalived stop else echo "running..." >> /opt/keepalived-running-info.log sleep 5000 fi
Keepalived监测Mysql测试
先要保证两台服务器的mysql服务正常启动哦~
1. 启动Keepalived
# 在master和slave上执行 sudo /etc/init.d/keepalived start # 查看脚本是否正常执行 tail -f /opt/keepalived-running-info.log # 查看master的ip,发现虚拟ip绑定成功 ip addr 2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:64:35:17 brd ff:ff:ff:ff:ff:ff inet 192.168.11.188/24 brd 192.168.11.255 scope global ens33 valid_lft forever preferred_lft forever inet 192.168.11.25/32 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::20c:29ff:fe64:3517/64 scope link valid_lft forever preferred_lft forever
2. 高可用测试
在任意一台主机执行以下命令测试:
mycli -h 192.168.11.25 -u root -p 123456 #ok mycli -h 192.168.11.188 -u root -p 123456 #ok mycli -h 192.168.11.186 -u root -p 123456 #ok
再次查看master主机的ip,发现虚拟ip不见了:
ip addr 2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:64:35:17 brd ff:ff:ff:ff:ff:ff inet 192.168.11.188/24 brd 192.168.11.255 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::20c:29ff:fe64:3517/64 scope link valid_lft forever preferred_lft forever
此时查看slave主机的ip定, 发现ip漂移情况,虚拟ip自动绑定到到了slave主机上:
ip addr 2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:f0:00:ad brd ff:ff:ff:ff:ff:ff inet 192.168.11.186/24 brd 192.168.11.255 scope global ens33 valid_lft forever preferred_lft forever inet 192.168.11.25/32 scope global ens33 valid_lft forever preferred_lft forever inet6 fe80::20c:29ff:fef0:ad/64 scope link valid_lft forever preferred_lft forever
此时查看slave主机情况,Slave_IO_Running
变成了Connecting
:
mycli -h 192.168.11.186 -u root -p 123456 show slave status\G;
Slave_IO_Running | Connecting
Slave_SQL_Running | Yes
继续测试
现在把master重新启动
mycli -h 192.168.11.188 -u root -p 123456 #ok show master status; +------------------+----------+--------------+-------------------------------+-------------------+ | File | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set | +------------------+----------+--------------+-------------------------------+-------------------+ | mysql-bin.000009 | 154 | | mysql,test,information_schema | | +------------------+----------+--------------+-------------------------------+-------------------+ 1 row in set Time: 0.004s
会发现信息发生了改变,再次查看slave的状态恢复正常:
Slave_IO_Running | Yes
Slave_SQL_Running | Yes
继续查看master主机发现未绑定vip,vip依然存在于slave所属主机上面。
那么现在把slave停掉试试看:
docker stop mysql-slave mycli -h 192.168.11.186 -u root -p 123456 #error mycli -h 192.168.11.25 -u root -p 123456 # error mycli -h 192.168.11.188 -u root -p 123456 # ok
出现只有master主机的mysql服务能访问的情况,是因为上面测试把两个mysql服务停止,脚本监测不到3306端口执行了/etc/init.d/keepalived stop
,所以需要重新启动keepalived :
# master主机 sudo /etc/init.d/keepalived start # slave主机 docker start mysql-slave sudo /etc/init.d/keepalived start
这个时候发现vip又重新绑定到master上面,OK,一切正常。
总结:
写的可能有些啰嗦,但是每一步的操作和测试又是必要的,只为记录自己的一次学习心得。