centos7下mysql双主+keepalived

一、keepalived简介

keepalived是vrrp协议的实现，原生设计目的是为了高可用ipvs服务，keepalived能够配置文件中的定义生成ipvs规则，并能够对各RS的健康状态进行检测；通过共用的虚拟IP地址对外提供服务；每个热备组内同一时刻只有一台主服务器提供服务，其他服务器处于冗余状态，若当前在线的服务器宕机，其虚拟IP地址将会被其他服务器接替（优先级决定接替顺序），实现高可用为后端主机提供服务。

   二、keepalived组件

Keepalived组件介绍

core:keepalived核心组件，主进程的启动和维护，全局配置等。

vrrp stack:keepalived是基于vrrp协议实现高可用vps服务，vrrp则为相关子进程为其提供服务

check:检测keepalived的健康状态相关进程

system call:系统调用

watch dog:监控check和vrrp进程的看管者，check负责检测器子进程的健康状态，当其检测到master上的服务不可用时则通告vrrp将其转移至backup服务器上。

三 环境准备

操作系统：centos7.1.1511（core）

数据库： mysql5.7.21 社区版

master1 ：10.0.0.11　　      安装mysql 和keeplived

master2 ：10.0.0.12             安装mysql 和keeplived

VIP：10.0.0.20

要实现互为主从，就必须 mster1-->master2设置主从同步 同时 master2--->master1 也设置主从同步

四、Mysql主主同步环境部署

---------------master1服务器操作记录---------------
在my.cnf文件的[mysqld]配置区域添加下面内容：
[root@master1 ~]# vim /usr/local/mysql/my.cnf
server-id =
log-bin = mysql-bin
sync_binlog =
binlog_checksum = none
binlog_format = mixed
auto-increment-increment =
auto-increment-offset =
slave-skip-errors = all      

[root@master1 ~]# /etc/init.d/mysql restart
Shutting down MySQL. SUCCESS!
Starting MySQL.. SUCCESS!

创建一个复制用户

出了小问题，由于之前root用户的密码设置过于简单在创建复制用户时报如下错误

mysql> grant replication slave,replication client on *.* to repl@'10.0.0.%' identified by '1qaz@WSX';
ERROR 1820 (HY000): You must reset your password using ALTER USER statement before executing this statement.

按照提示将密码设置的复杂一点 在授权创建就没有问题了

mysql> alter user 'root'@'localhost' identified by '1qaz@WSX';
Query OK,  rows affected (0.00 sec)

mysql> flush privileges;
Query OK,  rows affected (0.00 sec)

mysql> grant replication slave,replication client on *.* to repl@'10.0.0.%' identified by '1qaz@WSX';
Query OK,  rows affected,  warning (0.00 sec)

mysql> flush privileges;
Query OK,  rows affected (0.00 sec)

锁表，待同步配置完成在解锁

mysql> flush tables with read lock;
Query OK,  rows affected (0.00 sec)

查看当前的binlog以及数据所在位置

mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin. |       |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
 row in set (0.00 sec)

---------------master2服务器操作记录---------------
在my.cnf文件的[mysqld]配置区域添加下面内容：
[root@master2 ~]# vim /usr/local/mysql/my.cnf
server-id =
log-bin = mysql-bin
sync_binlog =
binlog_checksum = none
binlog_format = mixed
auto-increment-increment =
auto-increment-offset =
slave-skip-errors = all

[root@master2 ~]# /etc/init.d/mysql restart
Shutting down MySQL.. SUCCESS!
Starting MySQL.. SUCCESS!

mysql> grant replication slave,replication client on *.* to repl@'10.0.0.%' identified by '1qaz@WSX';
Query OK,  rows affected,  warning (0.01 sec)

mysql> flush privileges;
Query OK,  rows affected (0.00 sec)

mysql> flush tables with read lock;
  Query OK, 0 rows affected (0.00 sec)

查看 master情况

mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin. |       |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
 row in set (0.00 sec)

分别开启同步对方

---------------master1服务器做同步操作---------------
mysql> unlock tables;     //先解锁，将对方数据同步到自己的数据库中
mysql> slave stop；
mysql> change  master to master_host='10.0.0.12',master_user='repl',master_password='1qaz@WSX',master_log_file='mysql-bin.000001',master_log_pos=;
Query OK,  rows affected,  warnings (0.01 sec)

mysql> start slave;
Query OK,  rows affected (0.01 sec)

 

查看两个线程状态是否为YES 

mysql> show slave status \G;

Slave_IO_Running: Yes
Slave_SQL_Running: Yes

-------------master2服务器做同步操作---------------
mysql> unlock tables;     //先解锁，将对方数据同步到自己的数据库中
mysql> slave stop；
mysql> change  master to master_host='10.0.0.11',master_user='repl',master_password='1qaz@WSX',master_log_file='mysql-bin.000006',master_log_pos=;
Query OK,  rows affected,  warnings (0.06 sec)

mysql> start slave;
Query OK,  rows affected (0.01 sec)

mysql> show slave status \G;

Master_Log_File: mysql-bin.000006
Read_Master_Log_Pos: 996
Relay_Log_File: master2-relay-bin.000002
Relay_Log_Pos: 312
Relay_Master_Log_File: mysql-bin.000006
Slave_IO_Running: Yes
Slave_SQL_Running: Yes

以上表明双方已经实现了mysql主主同步。

当运行一段时间后，要是发现同步有问题，比如只能单向同步，双向同步失效。可以重新执行下上面的change master同步操作，只不过这样同步后，只能同步在此之后的更新数据。下面开始进行数据验证：

-----------------主主同步效果验证---------------------
）在master1数据库上写入新数据
mysql> unlock tables;
Query OK,  rows affected (0.00 sec)

mysql> create database huanqiu;
Query OK,  row affected (0.01 sec)

mysql> use huanqiu;
Database changed

mysql> create table if not exists haha ( id int(10) PRIMARY KEY AUTO_INCREMENT, name varchar(50) NOT NULL);
Query OK, 0 rows affected (0.04 sec)

mysql> insert into haha values(2,'guojing');
Query OK, 1 row affected (0.00 sec)

mysql> insert into haha values(1,"huangrong");
Query OK, 1 row affected (0.00 sec)

mysql> select * from haha;
+----+-----------+
| id | name |
+----+-----------+
| 1 | huangrong |
| 2 | guojing |
+----+-----------+
2 rows in set (0.00 sec)

然后在master2数据库上查看，发现数据已经同步过来了!

mysql> select * from huanqiu.haha;
+----+-----------+
| id | name |
+----+-----------+
| 1 | huangrong |
| 2 | guojing |
+----+-----------+
2 rows in set (0.00 sec)

）在master2数据库上写入新数据
mysql> create database hehe;
Query OK,  row affected (0.00 sec)

mysql> insert into huanqiu.haha values(3,"haha"),(4,"haha");
Query OK, 2 rows affected (0.00 sec)
Records: 2 Duplicates: 0 Warnings: 0

然后在master1数据库上查看，发现数据也已经同步过来了!
mysql> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| hehe               |
| huanqiu            |
| mysql              |
| performance_schema |
| test               |
+--------------------+
 rows in set (0.00 sec)

mysql> select * from huanqiu.haha;
+----+-----------+
| id | name |
+----+-----------+
| 1 | huangrong |
| 2 | guojing |
| 3 | haha |
| 4 | haha |
+----+-----------+
4 rows in set (0.00 sec)

至此，Mysql主主同步环境已经实现。

五 配置mysql+keepalived 高可用环境

）安装keepalived并将其配置成系统服务。master1和master2两台机器上同样进行如下操作：
[root@master1 ~]# yum install -y openssl-devel
[root@master1 ~]# cd /usr/local/src/
[root@master1 src]# wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz
[root@master1 src]# tar -zvxf keepalived-1.3..tar.gz
[root@master1 src]# cd keepalived-1.3.
[root@master1 keepalived-1.3.]# ./configure --prefix=/usr/local/keepalived
[root@master1 keepalived-1.3.]# make && make install

[root@master1 keepalived-1.3.]# cp /usr/local/src/keepalived-1.3./keepalived/etc/init.d/keepalived /etc/rc.d/init.d/
[root@master1 keepalived-1.3.]# cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
[root@master1 keepalived-1.3.]# mkdir /etc/keepalived/
[root@master1 keepalived-1.3.]# cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
[root@master1 keepalived-1.3.]# cp /usr/local/keepalived/sbin/keepalived /usr/sbin/
[root@master1 keepalived-1.3.]# echo "/etc/init.d/keepalived start" >> /etc/rc.local

2）master1机器上的keepalived.conf配置。（下面配置中没有使用lvs的负载均衡功能，所以不需要配置虚拟服务器virtual server）

[root@master1 ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf.bak
[root@master1 ~]# vim /etc/keepalived/keepalived.conf       ＃清空默认内容，直接采用下面配置：
! Configuration File for keepalived

global_defs {
notification_email {
ops@wangshibo.cn
tech@wangshibo.cn
}

notification_email_from ops@wangshibo.cn
smtp_server 127.0.0.1
smtp_connect_timeout
router_id MASTER-HA
}

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                    #脚本执行间隔，每2s检测一次
    weight -                    #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 -
    fall                     #检测连续2次失败才算确定是真失败。会用weight减少优先级（-255之间）
    rise                     #检测1次成功就算成功。但不修改优先级
}

vrrp_instance VI_1 {
    state MASTER
    interface eth0      #指定虚拟ip的网卡接口
    mcast_src_ip 10.0.0.11
    virtual_router_id     #路由器标识，MASTER和BACKUP必须是一致的
    priority             #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
    authentication {
        auth_type PASS
        auth_pass
    }
    virtual_ipaddress {
        10.0.0.20
    }

track_script {
   chk_mysql_port
}
}

编写切换脚本。KeepAlived做心跳检测，如果Master的MySQL服务挂了(3306端口挂了),那么它就会选择自杀。Slave的KeepAlived通过心跳检测发现这个情况，就会将VIP的请求接管

[root@master1 ~]# vim /opt/chk_mysql.sh
#!/bin/bash
counter=$(netstat -na|grep "LISTEN"|grep ""|wc -l)
if [ "${counter}" -eq  ]; then
    /etc/init.d/keepalived stop
fi

[root@master1 ~]# chmod  /opt/chk_mysql.sh

启动keepalived服务
[root@master1 ~]# /etc/init.d/keepalived start
正在启动 keepalived：                                      [确定]

4）master2机器上的keepalived配置。master2机器上的keepalived.conf文件只修改priority为90、nopreempt不设置、real_server设置本地IP。

[root@master2 ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf.bak
[root@master2 ~]# >/etc/keepalived/keepalived.conf
[root@master2 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
notification_email {
ops@qq.com
tech@qq.com
}

notification_email_from ops@wangshibo.cn
smtp_server 127.0.0.1
smtp_connect_timeout
router_id MASTER-HA
}

vrrp_script chk_mysql_port {
    script "/opt/chk_mysql.sh"
    interval
    weight -
    fall
    rise
}

vrrp_instance VI_1 {
    state BACKUP
    interface eth0
    mcast_src_ip 10.0.0.12
    virtual_router_id
    priority
    advert_int
    authentication {
        auth_type PASS
        auth_pass
    }
    virtual_ipaddress {
        10.0.0.20
    }

track_script {
   chk_mysql_port
}
}

[root@master2 ~]# cat /opt/chk_mysql.sh
#!/bin/bash
counter=$(netstat -na|grep "LISTEN"|grep ""|wc -l)
if [ "${counter}" -eq  ]; then
    /etc/init.d/keepalived stop
fi

[root@master2 ~]# chmod  /opt/chk_mysql.sh

[root@master2 ~]# /etc/init.d/keepalived start
正在启动 keepalived：                                      [确定]

我这里启动时出现了问题，分析日志

tail -f /var/log/message
Mar  :: master1 systemd: Configuration file /usr/lib/systemd/system/ebtables.service is marked executable. Please remove executable permission bits. Proceeding anyway.

查看keepalived.service

# vi /lib/systemd/system/keepalived.service  

[Unit]
Description=LVS and VRRP High Availability Monitor
After=syslog.target network-online.target

[Service]
Type=forking
# PIDFile=/usr/local/keepalived/var/run/keepalived.pid

# 上面这个注释掉 改成下面 因为这个默认路径不存在所以就无法写入进程ID文件

PIDFile=/var/run/keepalived.pid 
KillMode=process
EnvironmentFile=-/usr/local/keepalived/etc/sysconfig/keepalived
ExecStart=/usr/local/keepalived/sbin/keepalived $KEEPALIVED_OPTIONS
ExecReload=/bin/kill -HUP $MAINPID

[Install]
WantedBy=multi-user.target

5）master1和master2两台服务器都要授权允许root用户远程登录，用于在客户端登陆测试！

mysql> grant all on *.* to root@'10.0.0.%' identified by "1qaz@WSX";
Query OK,  rows affected (0.00 sec)

mysql> flush privileges;
Query OK,  rows affected (0.01 sec)

   

6）在master1和master2两台机器上设置iptables防火墙规则，如下：

[root@master1 ~]# cat /etc/sysconfig/iptables
........
-A INPUT -s 10.0.0.0/ -d 224.0.0.18 -j ACCEPT       #允许组播地址通信
-A INPUT -s 10.0.0.0/ -p vrrp -j ACCEPT             #允许VRRP（虚拟路由器冗余协）通信
-A INPUT -m state --state NEW -m tcp -p tcp --dport  -j ACCEPT    #开放mysql的3306端口

[root@master1 ~]# /etc/init.d/iptables restart

六 Mysql＋keepalived故障转移的高可用测试

1）通过Mysql客户端通过VIP连接，看是否连接成功。

比如，在远程一台测试机上连接，通过vip地址可以正常连接（下面的连接权限要是在服务端提前授权的）

[root@master1 ~]# mysql -uroot -p1qaz@WSX -h10.0.0.
mysql: [Warning] Using a password on the command line interface can be insecure.
Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is
Server version: 5.7.-log MySQL Community Server (GPL)

Copyright (c) , , Oracle and/or its affiliates. All rights reserved.

Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

mysql> select * from huanqiu.haha;
+----+-----------+
| id | name |
+----+-----------+
| 1 | huangrong |
| 2 | guojing |
| 3 | haha |
| 4 | haha |
+----+-----------+
4 rows in set (0.02 sec)

2）默认情况下，vip是在master1上的。使用"ip addr"命令查看vip切换情况 

[root@master1 ~]# ip addr |grep 10.0
: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP qlen
    inet 10.0.0.11/ brd 10.255.255.255 scope global eth0
    inet 10.0.0.20/ scope global eth0

[root@master2 ~]# ip addr |grep 10.0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
inet 10.0.0.12/8 brd 10.255.255.255 scope global eth0

停止master1机器上的mysql服务，根据配置中的脚本，mysql服务停了，keepalived也会停，从而vip资源将会切换到master2机器上。（mysql服务没有起来的时候，keepalived服务也无法顺利启动！）

[root@master1 ~]# systemctl stop mysqld

[root@master1 ~]# ps -ef|grep mysql
root             : pts/    :: grep --color=auto mysql
[root@master1 ~]# ps -ef|grep keepalived
root             : pts/    :: grep --color=auto keepalived

[root@master1 ~]# ip addr |grep 10.0
: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP qlen
    inet 10.0.0.11/ brd 10.255.255.255 scope global eth0

查看master2主机

[root@master2 ~]# ip addr |grep 10.0
: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP qlen
    inet 10.0.0.12/ brd 10.255.255.255 scope global eth0
    inet 10.0.0.20/ scope global eth0

3）再次启动master1的mysql和keepalived服务。（注意：如果restart重启mysql，那么还要启动下keepalived，因为mysql重启，根据脚本会造成keepalived关闭）

注意：一定要先启动mysql服务，然后再启动keepalived服务。如果先启动keepalived服务，按照上面的配置，mysql没有起来，就会自动关闭keepalived。

[root@master1 ~]# systemctl start mysqld
[root@master1 ~]# /etc/init.d/keepalived start
Starting keepalived (via systemctl):                       [  OK  ]

[root@master1 ~]# ip addr |grep 10.0
: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP qlen
    inet 10.0.0.11/ brd 10.255.255.255 scope global eth0
    inet 10.0.0.20/ scope global eth0

[root@master2 ~]# ip addr |grep 10.0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
inet 10.0.0.12/8 brd 10.255.255.255 scope global eth0

此时虚拟ip又回到了 master1 主机上了

以上在vip资源切换过程中，对于客户端连接mysql（使用vip连接）来说几乎是没有任何影响的。

---------------------------------温馨提示（Keepalived的抢占和非抢占模式）---------------------------------------

keepalive是基于vrrp协议在linux主机上以守护进程方式，根据配置文件实现健康检查。

VRRP是一种选择协议，它可以把一个虚拟路由器的责任动态分配到局域网上的VRRP路由器中的一台。

控制虚拟路由器IP地址的VRRP路由器称为主路由器，它负责转发数据包到这些虚拟IP地址。

一旦主路由器不可用，这种选择过程就提供了动态的故障转移机制，这就允许虚拟路由器的IP地址可以作为终端主机的默认第一跳路由器。

keepalive通过组播，单播等方式（自定义），实现keepalive主备推选。工作模式分为抢占和非抢占（通过参数nopreempt来控制）。

1）抢占模式：

主服务正常工作时，虚拟IP会在主上，备不提供服务，当主服务优先级低于备的时候，备会自动抢占虚拟IP，这时，主不提供服务，备提供服务。

也就是说，工作在抢占模式下，不分主备，只管优先级。

如上配置，不管keepalived.conf里的state配置成master还是backup，只看谁的priority优先级高（一般而言，state为MASTER的优先级要高于BACKUP）。

priority优先级高的那一个在故障恢复后，会自动将VIP资源再次抢占回来！！

2）非抢占模式：

这种方式通过参数nopreempt（一般设置在advert_int的那一行下面）来控制。不管priority优先级，只要MASTER机器发生故障，VIP资源就会被切换到BACKUP上。

并且当MASTER机器恢复后，也不会去将VIP资源抢占回来，直至BACKUP机器发生故障时，才能自动切换回来。

千万注意：

nopreempt这个参数只能用于state为backup的情况，所以在配置的时候要把master和backup的state都设置成backup，这样才会实现keepalived的非抢占模式！

也就是说：

a）当state状态一个为master，一个为backup的时候，加不加nopreempt这个参数都是一样的效果。即都是根据priority优先级来决定谁抢占vip资源的，是抢占模式！

b）当state状态都设置成backup，如果不配置nopreempt参数，那么也是看priority优先级决定谁抢占vip资源，即也是抢占模式。

c）当state状态都设置成backup，如果配置nopreempt参数，那么就不会去考虑priority优先级了，是非抢占模式！即只有vip当前所在机器发生故障，另一台机器才能接管vip。即使优先级高的那一台机器恢复  后也不会主动抢回vip，只能等到对方发生故障，才会将vip切回来。

---------------------------------mysql状态检测脚本优化---------------------------------

上面的mysql监测脚本有点过于简单且粗暴，即脚本一旦监测到Master的mysql服务关闭，就立刻把keepalived服务关闭，从而实现vip转移！

 

下面对该脚本进行优化，优化后，当监测到Master的mysql服务关闭后，就会将vip切换到Backup上（但此时Master的keepalived服务不会被暴力kill）

当Master的mysql服务恢复后，就会再次将VIP资源切回来！

[root@master ~]# cat /opt/chk_mysql.sh
#!/bin/bash
MYSQL=/usr/bin/mysql
MYSQL_HOST=localhost
MYSQL_USER=root
MYSQL_PASSWORD=1qaz@WSX
CHECK_TIME=

#mysql  is working MYSQL_OK is  , mysql down MYSQL_OK is 

MYSQL_OK=

function check_mysql_helth (){
    $MYSQL -h $MYSQL_HOST -u $MYSQL_USER -p${MYSQL_PASSWORD} -e "show status;" >/dev/null >&
    if [ $? =  ] ;then
    MYSQL_OK=
    else
    MYSQL_OK=
    fi
    return $MYSQL_OK
}
while [ $CHECK_TIME -ne  ]
do
    let "CHECK_TIME -= 1"
    check_mysql_helth
if [ $MYSQL_OK =  ] ; then
    CHECK_TIME=
    exit
fi
if [ $MYSQL_OK -eq  ] &&  [ $CHECK_TIME -eq  ]
then
    pkill keepalived
    exit
fi
sleep
done