用keepalived配置高可用集群

• 准备两台机器130和132，130作为master，132作为backup
• 两台机器都执行yum install -y keepalived
• 两台机器都安装nginx，其中130上已经编译安装过nginx，132上需要yum安装nginx: yum install -y nginx
• 设定vip为100
• 编辑130上keepalived配置文件，内容获取地址
• 130编辑监控脚本，内容获取地址
• 给脚本755权限
• systemctl start keepalived 130启动服务
• 132上编辑配置文件，内容获取地址
• 132上编辑监控脚本，内容获取地址
• 给脚本755权限
• 132上也启动服务 systemctl start keepalived

keepalived配置高可用集群

1. 首先准备两台机器，都安装keepalived

• keepalived，实际是包含一个服务的，也就是说这个服务用来实现高可用

A机器，安装keepalived
[root@hanfeng ~]# yum install -y keepalivedB机器，安装keepalived
[root@hf-01 ~]# yum install -y keepalived

1. 这里使用 nginx ，把它作为一个高可用的对象——>使用nginx作为演示对象的原因，因为nginx在工作中，在生产环境中，很多企业把nginx做一个负载均衡器

• 假设nginx一旦挂掉，那么后端所有的web，即使说是正常的，那也无法访问到

1. 若是A、B机器没有装nginx服务，可以直接 yum安装

• 若是lnmp安装过nginx，则不需要安装了（源码包安装的nginx）
• 直接yum安装的nginx，两者很容易区分（PS：有时直接yum安装不了，需要安装yum扩展源——>yum install -y epel-release）
• 源码包安装nginx
• 源码安装nginx经常出现的错误

A机器源码安装nginx
（PS：有时初始化的时候，会看到无法初始化，是因为缺少一些包yum install  -y gcc）B机器yum安装nginx
[root@hf-01 ~]# yum install -y epel-release
[root@hf-01 ~]# yum install -y nginx
[root@hf-01 ~]# systemctl start nginx
[root@hf-01 ~]# ps aux |grep nginx
root      2825  0.0  0.2 123300  2100 ?        Ss   11:40   0:00 nginx: master process /usr/sbin/nginx
nginx     2826  0.0  0.3 123764  3120 ?        S    11:40   0:00 nginx: worker process
root      2828  0.0  0.0 112656   992 pts/0    R+   11:40   0:00 grep --color=auto nginx
[root@hf-01 ~]#

1. 更改keepalived配置文件了，内容地址

• 默认的配置文件路径在/etc/keepalived/keepalived.conf
• 清空文件的快捷键方法 > !$

A机器更改配置文件
[root@hanfeng ~]# ls /etc/keepalived/keepalived.conf
/etc/keepalived/keepalived.conf
[root@hanfeng ~]# > !$        //直接清空文件内容了
> /etc/keepalived/keepalived.conf
[root@hanfeng ~]# cat /etc/keepalived/keepalived.conf
[root@hanfeng ~]# vim /etc/keepalived/keepalived.conf   //去文件地址去下载内容
将拷贝的内容复制进去
只需要改网卡名字和飘逸IP为192.168.133.100
####################### #  全局配置 #######################
global_defs {                            //global_defs 全局配置标识notification_email {               //notification_email用于设置报警邮件地址
han1118feng@163.com           //可以设置多个，每行一个}notification_email_from root@aminglinux.com    //设置邮件发送地址  smtp_server 127.0.0.1                //设置邮件的smtp server地址smtp_connect_timeout 30            //设置连接smtp sever超时时间router_id LVS_DEVEL
}###################### #  VRRP配置 ######################
vrrp_script chk_nginx {               script "/usr/local/sbin/check_ng.sh"   //检查服务是否正常，通过写脚本实现，脚本检查服务健康状态interval 3                      //检查的时间间断是3秒
}
vrrp_instance VI_1 {                        //VRRP配置标识 VI_1是实例名称 state MASTER                           //定义master相关interface eno16777736                        //通过vrrp协议去通信、去发广播。配置时，需注意自己的网卡名称virtual_router_id 51                   //定义路由器ID ，配置的时候和从机器一致priority 100                              //权重，主角色和从角色的权重是不同的advert_int 1                            //设定MASTER与BACKUP主机质检同步检查的时间间隔，单位为秒authentication {                       //认证相关信息auth_type PASS                   //这里认证的类型是PASSauth_pass aminglinux>com   //密码的形式是一个字符串}virtual_ipaddress {                  //设置虚拟IP地址 (VIP),又叫做漂移IP地址192.168.74.100   //更改为192.168.74.100}track_script {               //加载脚本 chk_nginx            }
}保存退出

• virtual_ipaddress：简称VIP，这个vip，两台机器，一个主，一个从，正常的情况是主在服务，主宕掉了，从起来了，从启动服务，从启动nginx以后，，启动以后，访问那个IP呢？把域名解析到那个IP上呢？假如解析到主上，主宕掉了，所以这个，需要定义一个公有IP（主上用的IP，从上也用的IP）；这个IP是随时可以夏掉，去配置的

1. 定义监控脚本，脚本内容获取地址

• 脚本路径在keepalived配置文件中有定义，路径为/usr/local/sbin/check_ng.sh

A机器定义监控脚本
[root@hanfeng ~]# vim /usr/local/sbin/check_ng.sh#!/bin/bash
#时间变量，用于记录日志
d=`date --date today +%Y%m%d_%H:%M:%S`
#计算nginx进程数量
n=`ps -C nginx --no-heading|wc -l`
#如果进程为0，则启动nginx，并且再次检测nginx进程数量，
#如果还为0，说明nginx无法启动，此时需要关闭keepalived
if [ $n -eq "0" ]; then/etc/init.d/nginx startn2=`ps -C nginx --no-heading|wc -l`if [ $n2 -eq "0"  ]; thenecho "$d nginx down,keepalived will stop" >> /var/log/check_ng.logsystemctl stop keepalivedfi
fi保存退出

• “脑裂”，在高可用（HA）系统中，当联系2个节点的“心跳线”断开时，本来为一整体、动作协调的HA系统，就分裂成为2个独立的个体。由于相互失去了联系，都以为是对方出了故障。两个节点上的HA软件像“裂脑人”一样，争抢“共享资源”、争起“应用服务”，就会发生严重后果——或者共享资源被瓜分、2边“服务”都起不来了；或者2边“服务”都起来了，但同时读写“共享存储”，导致数据损坏
• 如何判断脑裂？
  • 分别在两台机查看当前服务器是否拥有虚拟IP，如果两台服务器都拥有，则说明发生了脑裂，证明目前双机通信出现问题，产生此问题的原有在于 两台服务器都探测不到组内其他服务器的状态（心跳请求无法正常响应），私自判定另一台服务器挂起，则抢占虚拟IP，脑裂的出现是不被允许的，解决此问题的方法为检查防火墙设置（关闭防火墙）或者使用串口通信。

1. 脚本创建完之后，还需要改变脚本的权限（不更改权限，就无法自动加载脚本，那就无法启动keepalived服务）

[root@hanfeng ~]# chmod 755 /usr/local/sbin/check_ng.sh
[root@hanfeng ~]#

1. 启动keepalived服务，并查看是否启动成功（PS：启动不成功，有可能是防火墙未关闭或者规则限制导致的）

• systemctl stop firewalld 关闭firewalld
• iptables -nvL
• setenforce 0 临时关闭selinux
• getenforce命令查看是否为Permissive
• 这时再来启动keepalived，就会看到keepalived进程服务了

[root@hanfeng ~]# systemctl start keepalived
[root@hanfeng ~]# ps aux |grep keepalived
root      2970  0.0  0.1 121324  1404 ?        Ss   07:14   0:00 /usr/sbin/keepalived -D
root      2971  0.0  0.2 123396  2356 ?        S    07:14   0:00 /usr/sbin/keepalived -D
root      2972  0.0  0.2 123396  2384 ?        S    07:14   0:00 /usr/sbin/keepalived -D
root      2974  0.0  0.0 112672   988 pts/1    R+   07:14   0:00 grep --color=auto keepalived
[root@hanfeng ~]#

1. 查看nginx服务进程

[root@hanfeng ~]# ps aux |grep nginx
root      3004  0.0  0.2 123372  2108 ?        Ss   07:18   0:00 nginx: master process /usr/sbin/nginx
nginx     3005  0.0  0.3 123836  3148 ?        S    07:18   0:00 nginx: worker process
root      3007  0.0  0.0 112672   984 pts/1    R+   07:19   0:00 grep --color=auto nginx
[root@hanfeng ~]#

1. 这时停止nginx服务

• /etc/init.d/nginx stop

[root@hanfeng ~]# /etc/init.d/nginx stop
Stopping nginx (via systemctl):                            [  确定  ]
[root@hanfeng ~]#

1. 再来查看nginx服务进程，会看到自动加载了

[root@hanfeng ~]# ps aux |grep nginx
root      6238  0.0  0.0  20996   628 ?        Ss   08:07   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody    6242  0.0  0.3  23440  3212 ?        S    08:07   0:00 nginx: worker process
nobody    6243  0.0  0.3  23440  3212 ?        S    08:07   0:00 nginx: worker process
root      6263  0.0  0.0 112676   980 pts/0    R+   08:07   0:00 grep --color=auto nginx
[root@hanfeng ~]# 

1. keepalived日志文件路径

• /var/log/messages

1. 查看ip地址，使用 ip add 命令，而不能使用ifconfig命令，因为 ifconfig命令 是无法查看到vip192.168.133.100这个IP的

[root@hanfeng ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:c7:05:28 brd ff:ff:ff:ff:ff:ffinet 192.168.74.130/24 brd 192.168.74.255 scope global dynamic eno16777736valid_lft 1158sec preferred_lft 1158secinet 192.168.133.100/32 scope global eno16777736valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fec7:528/64 scope link valid_lft forever preferred_lft forever
[root@hanfeng ~]#

1. 检查A、B 机器下防火墙和selinux是否关闭了，若没有关闭，可能会导致实验失败

• systemctl stop firewalld 关闭firewalld
• iptables -nvL
• setenforce 0 临时关闭selinux
• getenforce命令查看是否为Permissive

以上就是主机器A的配置

backup 机器配置

1. 在B机器yum安装nginx和keepalived

[root@hf-01 ~]# yum install -y epel-release
[root@hf-01 ~]# yum install -y nginx

1. 关闭B机器的防火墙和selinux

• iptables -F 清空规则
• setenforce 0 临时关闭selinux

1. 自定义B机器keepalived配置文件，内容获取地址，更改虚拟IP和主一样的

首先清空B机器keepalived里面自带的配置文件
[root@hf-01 ~]# > /etc/keepalived/keepalived.conf
[root@hf-01 ~]# cat !$
cat /etc/keepalived/keepalived.conf
[root@hf-01 ~]# 然后复制配置文件并粘贴进去，更改虚拟IP和主一样的
[root@hf-01 ~]# vim /etc/keepalived/keepalived.confglobal_defs {notification_email {aming@aminglinux.com}notification_email_from root@aminglinux.comsmtp_server 127.0.0.1smtp_connect_timeout 30router_id LVS_DEVEL
}
vrrp_script chk_nginx {script "/usr/local/sbin/check_ng.sh"interval 3
}
vrrp_instance VI_1 {state BACKUP        //这里 和master不一样的名字interface eno16777736        //网卡和当前机器一致，否则无法启动keepalived服务virtual_router_id 51        //和主机器 保持一致priority 90            //权重，要比主机器小的数值advert_int 1authentication {auth_type PASSauth_pass aminglinux>com}virtual_ipaddress {192.168.74.100        //这里更改为192.168.74.100}track_script {chk_nginx}
}保存退出

1. 定义监控脚本，路径再keepalived里面已定义过，脚本内容地址

• 这个脚本和主上的脚本有一点区别，启动nginx的命令不同，因为一个是yum安装，一个是源码包安装

[root@hf-01 ~]# vim /usr/local/sbin/check_ng.sh#时间变量，用于记录日志
d=`date --date today +%Y%m%d_%H:%M:%S`
#计算nginx进程数量
n=`ps -C nginx --no-heading|wc -l`
#如果进程为0，则启动nginx，并且再次检测nginx进程数量，
#如果还为0，说明nginx无法启动，此时需要关闭keepalived
if [ $n -eq "0" ]; thensystemctl start nginxn2=`ps -C nginx --no-heading|wc -l`if [ $n2 -eq "0"  ]; thenecho "$d nginx down,keepalived will stop" >> /var/log/check_ng.log                systemctl stop keepalivedfi
fi
保存退出

1. 改动脚本的权限，设置为755权限

[root@hf-01 ~]# chmod 755 /usr/local/sbin/check_ng.sh
[root@hf-01 ~]# 

1. B机器启动keepalived服务

• systemctl start keepalived

[root@hf-01 ~]# systemctl start keepalived
[root@hf-01 ~]# ps aux |grep keep
root      2814  0.0  0.1 121324  1396 ?        Ss   07:10   0:00 /usr/sbin/keepalived -D
root      2815  0.0  0.2 121324  2740 ?        S    07:10   0:00 /usr/sbin/keepalived -D
root      2816  0.0  0.2 121324  2324 ?        S    07:10   0:00 /usr/sbin/keepalived -D
root      2827  0.0  0.0 112672   980 pts/0    R+   07:10   0:00 grep --color=auto keep
[root@hf-01 ~]#

如何区分主和从上的nginx？

• A机器，是源码包安装的nginx（PS：这是lnmp配置好的环境虚拟主机内容）

[root@hanfeng ~]# cat /usr/local/nginx/conf/vhost/aaa.com.conf
server
{listen 80 default_server; server_name aaa.com;index index.html index.htm index.php;root /data/wwwroot/default;location ~ \.php${include fastcgi_params;fastcgi_pass unix:/tmp/champ.sock;#fastcgi_pass 127.0.0.1:9000;fastcgi_index index.php;fastcgi_param SCRIPT_FILENAME /data/wwwroot/default$fastcgi_script_name;}}[root@hanfeng ~]# cat /data/wwwroot/default/index.html
This is the default sete.
[root@cham002 ~]# vim /data/wwwroot/default/index.html
master This is the default sete.
[root@cham002 ~]#

• B机器是yum安装的nginx
  • 默认的索引页在 /usr/share/nginx/html/index.html

[root@hf-02~]# cat /usr/share/nginx/html/index.html
[root@hf-02 ~]# vim /usr/share/nginx/html/index.html
backup backup.

![输入图片说明](https://static.oschina.net/uploads/img/201801/28233716_ovDw.png "backup机器访问IP")

• 访问192.168.74.100这个VIP会看到和主机器（即A机器相同的内容），说明现在访问到的是机器master，VIP在master上

问题-B机器无法调用nginx服务？

• B机器关闭nginx服务，keepalived无法拉动nginx服务起来
• 解决方法：
  • 再次设置755权限，就可以拉动nginx服务了

测试高可用

1. 模拟线上生产环境，主机器宕机环境，最简单直接的方法，就是直接关闭keepalived服务
2. 关闭master机器（即A机器）上的keepalived服务关闭

[root@hanfeng ~]# systemctl stop keepalived
[root@hanfeng ~]#

1. 查看A机器上的VIP被已经释放掉了

[root@hanfeng ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:c7:05:28 brd ff:ff:ff:ff:ff:ffinet 192.168.74.130/24 brd 192.168.74.255 scope global dynamic eno16777736valid_lft 1219sec preferred_lft 1219secinet6 fe80::20c:29ff:fec7:528/64 scope link valid_lft forever preferred_lft forever
[root@hanfeng ~]#

1. 查看backup机器（即B机器）在监听VIP

[root@hf-01 ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:ff:fe:93 brd ff:ff:ff:ff:ff:ffinet 192.168.74.129/24 brd 192.168.74.255 scope global eno16777736valid_lft forever preferred_lft foreverinet 192.168.74.100/32 scope global eno16777736valid_lft forever preferred_lft foreverinet 192.168.74.150/24 brd 192.168.74.255 scope global secondary eno16777736:0valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:feff:fe93/64 scope link valid_lft forever preferred_lft forever
[root@hf-01 ~]#

1. 查看B机器日志

[root@hf-01 ~]# tail /var/log/messages
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng avahi-daemon[568]: Registering new address record for 192.168.74.100 on eno16777736.IPv4.
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
[root@hf-01 ~]#

1. 浏览器访问vip，会看到已经变成backup机器上的了

   

2. 在master机器（即A机器）启动keepalived服务，会看到vip这个IP立刻回来了

[root@hanfeng ~]# systemctl start keepalived
[root@hanfeng ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00inet 127.0.0.1/8 scope host lovalid_lft forever preferred_lft foreverinet6 ::1/128 scope host valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:c7:05:28 brd ff:ff:ff:ff:ff:ffinet 192.168.74.130/24 brd 192.168.74.255 scope global dynamic eno16777736valid_lft 1577sec preferred_lft 1577secinet 192.168.74.100/32 scope global eno16777736valid_lft forever preferred_lft foreverinet6 fe80::20c:29ff:fec7:528/64 scope link valid_lft forever preferred_lft forever
[root@hanfeng ~]#

1. 查看B机器日志变化

[root@hf-01 ~]# tail /var/log/messages
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:39 hanfeng avahi-daemon[568]: Registering new address record for 192.168.74.100 on eno16777736.IPv4.
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
[root@hf-01 ~]# tail /var/log/messages
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:20:44 hanfeng Keepalived_vrrp[7886]: Sending gratuitous ARP on eno16777736 for 192.168.74.100
Jan 29 08:30:01 hanfeng systemd: Started Session 49 of user root.
Jan 29 08:30:01 hanfeng systemd: Starting Session 49 of user root.
Jan 29 08:30:51 hanfeng Keepalived_vrrp[7886]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 90
Jan 29 08:30:51 hanfeng Keepalived_vrrp[7886]: VRRP_Instance(VI_1) Entering BACKUP STATE
Jan 29 08:30:51 hanfeng Keepalived_vrrp[7886]: VRRP_Instance(VI_1) removing protocol VIPs.
Jan 29 08:30:51 hanfeng avahi-daemon[568]: Withdrawing address record for 192.168.74.100 on eno16777736.
[root@hf-01 ~]#

##总结

• 在生产环境中，可能会用到2-3台backup角色, vim /etc/keepalived/keepalived.conf 这里面的权重调成不通级别，权重越高优先级越高！除了nginx服务的话，还可以做MySQL的高可用集群服务。（做mysql的高可用，一定要保证两边的数据一致）