Linux下keepalived+mysql实现高可用

首先在搭建前分析，想用keepalived实现mysql高可用，那么先的对keepalived+lvs跟mysql主主复制有所了解。因为mysql+keepalived架构是在keepalived+lvs跟mysql主主复制的基础上实现的lvs+keepalived跟mysql主主复制，在前面的博文中已经介绍过了这里不啰嗦了。

lvs+keepalived链接：http://duyunlong.blog.51cto.com/1054716/1118447,

mysql主主复制链接：http://duyunlong.blog.51cto.com/1054716/1306841

我们的目标是，两台mysql服务器如果其中有一台mysql服务器挂掉后，另外一台能立马接替其进行工作。因此我们就必须保证两台mysql数据库的数据完全一样，而且当挂掉的那一台重新启动的话，不再会被客户端继被访问，而是会充当备机跟现在工作的mysql进行数据同步，一直到提供服务的那台挂掉后再接替其工作。如此周而复始的实现了mysql的高可用。在通常情况下实现这种模式的keepalived无疑是最好的选择。因为有虚拟IP的原因，如果有一台mysql挂掉了，keepalived会从服务器群中剔除，而客户端访问会被切换到另外一台接替其工作的机器上。

在搭建的时候一些注意事项，mysql主主复制跟lvs+keepalived搭建的注意事情前面博文已经提高这里就不再啰嗦。在搭建完毕后，因为arp的原因不能够实现自由切换。这就是要注意的重点。

vim /etc/sysctl.conf
#.............前面部分省去
#######以下是文件末尾添加的部分######
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.lo.arp_announce = 2
##这4行的主要作用是本地arp不作回应##

添加完后保存退出，并执行：sysctl -p使其生效

[root@mysql-ha1 ~]# sysctl -p
net.ipv4.ip_forward = 0
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.default.accept_source_route = 0
kernel.sysrq = 0
kernel.core_uses_pid = 1
kernel.msgmnb = 65536
kernel.msgmax = 65536
kernel.shmmax = 68719476736
kernel.shmall = 4294967296
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
net.ipv4.conf.eth0.send_redirects = 0
net.ipv4.conf.eth0.send_redirects = 0
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.lo.arp_announce = 2

下面具体操作，首先有两台mysql机器已经搭建好了mysql的主主复制

vip:192.168.5.55

mysql-ha1:192.168.5.234

mysql-ha2:192.168.5.155

在两台mysql机器上都安装ipvsadm,keepalived，首先在mysql-ha1(192.168.5.234)安装

[root@mysql-ha1 src]# wget http://www.keepalived.org/software/keepalived-1.2.2.tar.gz
[root@mysql-ha1 src]# wget http://www.linuxvirtualserver.org/software/kernel-2.6/ipvsadm-1.26.tar.gz
[root@mysql-ha1 src]# yum -y install kernel-devel make gcc openssl-devel libnl* popt*
[root@mysql-ha1 src]# ln -s /usr/src/kernels/2.6.32-220.13.1.el6.x86_64/ /usr/src/linux
[root@mysql-ha1 src]# tar -zxvf ipvsadm-1.26.tar.gz
[root@mysql-ha1 src]# cd ipvsadm-1.26/
[root@mysql-ha1 ipvsadm-1.26]# make
[root@mysql-ha1 ipvsadm-1.26]# make install
[root@mysql-ha1 ipvsadm-1.26]# cd ..
[root@mysql-ha1 src]# tar -zxvf keepalived-1.2.2.tar.gz
[root@mysql-ha1 src]# cd keepalived-1.2.2/
[root@mysql-ha1 keepalived-1.2.2]# ./configure --with-kernel-dir=/usr/src/kernels/2.6.32-358.2.1.el6.x86_64/
[root@mysql-ha1 keepalived-1.2.2]# make && make install
[root@mysql-ha1 keepalived-1.2.2]# cp /usr/local/etc/rc.d/init.d/keepalived /etc/rc.d/init.d/
[root@mysql-ha1 keepalived-1.2.2]# cp /usr/local/etc/sysconfig/keepalived /etc/sysconfig/
[root@mysql-ha1 keepalived-1.2.2]# mkdir /etc/keepalived
[root@mysql-ha1 keepalived-1.2.2]# cp /usr/local/etc/keepalived/keepalived.conf /etc/keepalived/
[root@mysql-ha1 keepalived-1.2.2]# cp /usr/local/sbin/keepalived /usr/sbin/

在mysql-ha2（192.168.5.155）也同样进行安装

然后我们编辑keepalived配置文件，下面是mysql-ha1（192.168.5.234）上的配置文件内容

[root@mysql-ha1 ~]# >/etc/keepalived/keepalived.conf
[root@mysql-ha1 ~]# vim /etc/keepalived/keepalived.conf
#! Configuration File for keepalived
global_defs {
router_id mysql-ha1 #修改为自己的主机名
             }
##################第一部分###################
vrrp_instance VI_1 {
     state BACKUP #都修改成BACKUP
     interface eth0
     virtual_router_id 60 #默认51 主从都修改为60
     priority 100 #在mysql-ha2上LVS上修改成80
     advert_int 1
     nopreempt #不抢占资源，意思就是它活了之后也不会再把主抢回来
     authentication {
     auth_type PASS
     auth_pass 1111
     }
virtual_ipaddress {
     192.168.5.55
     }
}
##################第二部分###################
virtual_server 192.168.5.55 3306 {
     delay_loop 6
     lb_algo wrr
     lb_kind DR
     nat_mask 255.255.255.0
     persistence_timeout 50
     protocol TCP
 real_server 192.168.5.234 3306 {
     weight 1
     notify_down /usr/local/mysql/bin/mysql.sh
     TCP_CHECK {
         connect_timeout 10
         nb_get_retry 3
         connect_port 3306
         }
     }
}

这里需要注意的是，notify_down /usr/local/mysql/bin/mysql.sh这个选项，这个是是在keepalived检测不到mysql的时候要执行的脚本，从上面的配置文件来看real服务器只有本机。那么，keeaplived如果启动，客户端也只是访问本机的mysql。nopreempt这个选下也得注意，这个是不抢占资源在优先级高的机器上配置就可以。

看下这个脚本的内容：

[root@mysql-ha1 ~]# vim /usr/local/mysql/bin/mysql.sh
#!/bin/bash
pkill keepalived
[root@mysql-ha1 ~]# chmod +x /usr/local/mysql/bin/mysql.sh

脚本内容就一条命令：pkill keepalived，主要作用是如果本机的mysql挂掉了，那么同时会杀死本机的keepalived，这样另外一台就会接替他工作，虚拟IP也会被另一台接管，如果不杀死keepalived虚拟IP不会被另一台接管，mysql访问也就不会切换过去。

测试下如果，mysql服务器挂掉，脚本是否可以杀死keepalived

[root@mysql-ha1 ~]# ps aux | grep keepalived
root      2518  0.0  0.0  41796   936 ?        Ss   19:35   0:00 keepalived -D
root      2519  0.0  0.1  43900  2152 ?        S    19:35   0:00 keepalived -D
root      2520  0.0  0.0  43900  1572 ?        S    19:35   0:00 keepalived -D
root      2610  0.0  0.0 103240   860 pts/0    S+   23:07   0:00 grep keepalived
[root@mysql-ha1 ~]# netstat -ntlp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
tcp        0      0 0.0.0.0:3306                0.0.0.0:*                   LISTEN      2237/mysqld
tcp        0      0 0.0.0.0:22                  0.0.0.0:*                   LISTEN      1041/sshd
tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      1416/master
tcp        0      0 0.0.0.0:5672                0.0.0.0:*                   LISTEN      1429/qpidd
tcp        0      0 :::22                       :::*                        LISTEN      1041/sshd
tcp        0      0 ::1:25                      :::*                        LISTEN      1416/master
[root@mysql-ha1 ~]# service mysqld stop
Shutting down MySQL... SUCCESS!
[root@mysql-ha1 ~]# ps aux | grep keepalived
root      2640  0.0  0.0 103240   856 pts/0    S+   23:07   0:00 grep keepalived
[root@mysql-ha1 ~]# netstat -ntlp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
tcp        0      0 0.0.0.0:22                  0.0.0.0:*                   LISTEN      1041/sshd
tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      1416/master
tcp        0      0 0.0.0.0:5672                0.0.0.0:*                   LISTEN      1429/qpidd
tcp        0      0 :::22                       :::*                        LISTEN      1041/sshd
tcp        0      0 ::1:25                      :::*                        LISTEN      1416/master
[root@mysql-ha1 ~]#

可以看到当mysqld服务器挂掉后，脚本生效同时杀死本机keepalived使得mysql故障转移。这时另一台keepalived会接替它工作。

同样看下mysql-ha2(192.168.5.155)上keepalived配置文件

[root@mysql-ha2 ~]# vim /etc/keepalived/keepalived.conf
#! Configuration File for keepalived
global_defs {
router_id mysql-ha1 #修改为自己的主机名
             }
##################第一部分###################
vrrp_instance VI_1 {
     state BACKUP #都修改成BACKUP
     interface eth1
     virtual_router_id 60 #默认51 主从都修改为60
     priority 80 #在mysql-ha1上LVS上修改成100
     advert_int 1
     authentication {
     auth_type PASS
     auth_pass 1111
     }
virtual_ipaddress {
     192.168.5.55
     }
}
##################第二部分###################
virtual_server 192.168.5.55 3306 {
     delay_loop 6
     lb_algo wrr
     lb_kind DR
     nat_mask 255.255.255.0
     persistence_timeout 50
     protocol TCP
 real_server 192.168.5.155 3306 {
     weight 1
     notify_down /usr/local/mysql/bin/mysql.sh
     TCP_CHECK {
         connect_timeout 10
         nb_get_retry 3
         connect_port 3306
         }
     }
}

配置完毕后可以查看下keealived有没有检测到本机的mysql。（启动的时候先启动mysql在启动keepalived,要不mysql没起来启动keepalived会被脚本杀死。）

先看下mysql-ha1(192.168.5.234)

[root@mysql-ha1 ~]# service mysqld start
Starting MySQL.. SUCCESS!
[root@mysql-ha1 ~]# /etc/init.d/keepalived start
正在启动 keepalived：
[root@mysql-ha1 ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.5.55:mysql wrr persistent 50
  -> 192.168.5.234:mysql          Local   1      0          0

然后再看下mysql-ha1(192.168.5.155)

[root@mysql-ha2 ~]# service mysqld start
Starting MySQL.. SUCCESS!
[root@mysql-ha2 ~]# /etc/init.d/keepalived start
正在启动 keepalived：                                      [确定]
[root@mysql-ha2 ~]# ipvsadm -Ln      
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.5.55:3306 wrr persistent 50
  -> 192.168.5.155:3306           Local   1      0          0

测试只开启mysql-ha1(192.168.5.234)mysql服务跟keepalived用客户端链接虚拟ip

[root@mysql-ha1 ~]# ps aux | grep keepalived
root      2977  0.0  0.0  41796   936 ?        Ss   23:14   0:00 keepalived -D
root      2978  0.0  0.1  43900  2156 ?        S    23:14   0:00 keepalived -D
root      2979  0.0  0.0  43900  1576 ?        S    23:14   0:00 keepalived -D
root      2982  0.0  0.0 103240   860 pts/0    S+   23:16   0:00 grep keepalived
[root@mysql-ha1 ~]# netstat -ntlp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
tcp        0      0 0.0.0.0:3306                0.0.0.0:*                   LISTEN      2917/mysqld
tcp        0      0 0.0.0.0:22                  0.0.0.0:*                   LISTEN      1041/sshd
tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      1416/master
tcp        0      0 0.0.0.0:5672                0.0.0.0:*                   LISTEN      1429/qpidd
tcp        0      0 :::22                       :::*                        LISTEN      1041/sshd
tcp        0      0 ::1:25                      :::*                        LISTEN      1416/master

mysql-ha2(192.168.5.155)

[root@mysql-ha2 ~]# ps aux | grep keepalived
root     10576  0.0  0.0   5980   756 pts/4    S+   15:24   0:00 grep keepalived
[root@mysql-ha2 ~]# netstat -ntlp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
tcp        0      0 0.0.0.0:22                  0.0.0.0:*                   LISTEN      4776/sshd
tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      1101/master
tcp        0      0 0.0.0.0:5672                0.0.0.0:*                   LISTEN      1114/qpidd
tcp        0      0 :::22                       :::*                        LISTEN      4776/sshd
tcp        0      0 ::1:25                      :::*                        LISTEN      1101/master

然后用客户端：192.168.5.10连接

C:\Users\Administrator>mysql -uduyunlong -p123456 -h192.168.5.55
Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 33
Server version: 5.5.22-log Source distribution
Type 'help;' or '\h' for help. Type '\c' to clear the buffer.
mysql> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
| test               |
+--------------------+
4 rows in set (0.02 sec)
mysql>

然后把mysql-ha1（192.168.5.234）mysql服务关掉，开启mysql-ha2(192.168.5.155)上的mysql服务跟keepalived。然后不退出mysql，继续查看数据库：

mysql-ha1(192.168.5.234)

[root@mysql-ha1 ~]# service mysqld stop
Shutting down MySQL... SUCCESS!

mysql-ha2(192.168.5.155)

[root@mysql-ha2 ~]# service mysqld start
Starting MySQL.. SUCCESS!
[root@mysql-ha2 ~]# /etc/init.d/keepalived start
正在启动 keepalived：                                      [确定]
[root@mysql-ha2 ~]#

客户端：192.168.5.10

mysql> show databases;
ERROR 2006 (HY000): MySQL server has gone away
No connection. Trying to reconnect...
Connection id:    4
Current database: *** NONE ***
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
| test               |
+--------------------+
4 rows in set (0.01 sec)
mysql>

切换很快，大概在2到3秒之间！

转载于:https://blog.51cto.com/duyunlong/1310405