第十八章 Linux集群

18.1 集群介绍

Linux集群根据功能划分为两大类：高可用和负载均衡。

高可用集群

高可用集群通常为2台服务器，一台工作，另一台作为冗余，当提供服务的机器宕机时，冗余将接替继续提供服务。
实现高可用的开源软件有：heartbeat、Keepalived，其中heartbeat已无人维护，所以今后使用Keepalived来搭建高可用集群。

负载均衡集群

负载均衡集群，需要有1台服务器作为分发器，它负责把用户的请求分发给后端的服务器处理，在这个集群里，除了分发器外就是给用户提供服务的服务器了，这些服务器数量至少为2台，所以负载均衡集群至少要3台服务器。
实现负载均衡的开源软件有：LVS、Keepalived、haproxy、nginx；商业软件有：F5、Netscaler。

18.2 Keepalived介绍

Keepalived通过VRRP（virtual router redundancy protocol）虚拟路由冗余协议来实现高可用。
在这个协议里会将多台功能相同的路由器组成一个小组，这个小组会有1个master角色和N（N≥1）个backup角色。
Keepalived有三个模块：core、check、vrrp。其中core模块为Keepalived的核心，负责主进程的启动、维护以及全局配置文件的加载和解析；check模块负责健康检查，vrrp负责实现vrrp协议。

18.3-18.5 用Keepalived配置高可用集群（上）

操作环境：

两台虚拟机：

master：192.168.230.135；
backup：192.168.230.145；

准备工作

master：

[root@cham002 ~]# yum install -y keepalived[root@cham002 ~]# /usr/local/nginx/sbin/nginx -v
nginx version: nginx/1.12.1

backup：

[root@test ~]# yum install -y keepalived[root@test ~]# yum install -y nginx[root@test ~]# nginx -v
nginx version: nginx/1.12.2

开始配置

配置master机器：

在此使用自定义的Keepalived配置文件脚本(先清空或删除自带配置文件脚本)。

[root@cham002 ~]# vim /etc/keepalived/keepalived.conf
global_defs {notification_email {aming@aminglinux.com                          }                                               #定义故障提醒邮件接收地址（本实验用不到）notification_email_from root@aminglinux.comsmtp_server 127.0.0.1smtp_connect_timeout 30                         router_id LVS_DEVEL                             #定义故障提醒邮件发送地址（本实验用不到）
}
vrrp_script chk_nginx {                            #定义监控脚本script "/usr/local/sbin/check_ng.sh"           #指定服务健康检测（check）脚本interval 3                                     #检测时间间隔
}
vrrp_instance VI_1 {                               state MASTER                                   #定义该服务的角色 interface ens33                                #定义在本机中监听VIP的网卡virtual_router_id 51                           #虚拟路由id（同组中的服务器保持该id一致）priority 100                                   #指定本机权重（决定优先级）advert_int 1                                   authentication {                               #定义认证相关信息  auth_type PASS                             #认证类型为密码形式auth_pass 123456                           #定义认证密码}virtual_ipaddress {                            #定义VIP192.168.230.100}track_script {                                 #加载监控服务（脚本）chk_nginx                                  #注意此处服务名称要与上面监控脚本名称一致}
}

创建监控脚本：

[root@cham002 ~]# 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 keepalived    #在此停止Keepalived服务是为了避免发生脑裂fi
fi更改脚本权限：
[root@cham002 ~]# chmod 755 !$
chmod 755 /usr/local/sbin/check_ng.sh
启动Keepalived服务：
[root@cham002 ~]# systemctl start keepalived
检测服务状态：
[root@cham002 ~]# ps aux |grep keep
root      4915  0.0  0.1 120720  1400 ?        Ss   18:51   0:00 /usr/sbin/keepalived -D
root      4916  0.0  0.3 127460  3276 ?        S    18:51   0:00 /usr/sbin/keepalived -D
root      4917  0.0  0.2 131656  3024 ?        S    18:51   0:00 /usr/sbin/keepalived -D
root      4966  0.0  0.0 112680   976 pts/1    S+   18:52   0:00 grep --color=auto keep
[root@cham002 ~]# ps aux |grep nginx
root      2302  0.0  0.2  46840  2956 ?        Ss   01:08   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody    4763  0.0  0.4  48764  4164 ?        S    18:25   0:00 nginx: worker process
nobody    4764  0.0  0.4  48764  4164 ?        S    18:25   0:00 nginx: worker process
root      4986  0.0  0.0 112684   976 pts/1    S+   18:52   0:00 grep --color=auto nginx
[root@cham002 ~]# date
2018年 01月 24日 星期三 18:54:09 CST
注：由以上检测状态可知关闭nginx服务之后Keepalived会再次自动启动nginx服务[root@cham002 ~]# /etc/init.d/nginx stop
Stopping nginx (via systemctl):                            [  确定  ]
[root@cham002 ~]# !ps
ps aux |grep nginx
root      2302  0.0  0.2  46840  2956 ?        Ss   01:08   0:00 nginx: master process /usr/local/nginx/sbin/nginx -c /usr/local/nginx/conf/nginx.conf
nobody    4763  0.0  0.4  48764  4164 ?        S    18:25   0:00 nginx: worker process
nobody    4764  0.0  0.4  48764  4164 ?        S    18:25   0:00 nginx: worker process
root      5175  0.0  0.0 112684   976 pts/1    S+   18:53   0:00 grep --color=auto nginx
注：由以上检测状态可知关闭nginx服务之后Keepalived会再次自动启动nginx服务

说明： “脑裂”，即当master宕机后仍然未释放VIP，同时backup接替master提供服务要使用同一VIP，因而导致两台机器争占同一VIP导致服务紊乱，所以当master宕机后需要关闭其Keepalived服务来避免脑裂现象发生。Keepalived服务的日志位置：/var/log/messages。

查看网卡信息：

[root@cham002 ~]# ip add
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1link/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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:e3 brd ff:ff:ff:ff:ff:ffinet 192.168.230.135/24 brd 192.168.230.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.188.100/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.230.150/24 brd 192.168.230.255 scope global secondary ens33:0valid_lft forever preferred_lft foreverinet6 fe80::6f15:52d3:ebeb:e193/64 scope link valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:ed brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global ens37valid_lft forever preferred_lft foreverinet6 fe80::1801:cbbb:ebcc:89a3/64 scope link valid_lft forever preferred_lft forever

说明： 在此不能使用ifconfig查看，ens33中增加了一个IP：192.168.188.100/32。

检查防火墙

[root@cham002 ~]# iptables -nvL
Chain INPUT (policy ACCEPT 0 packets, 0 bytes)pkts bytes target     prot opt in     out     source               destination         Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)pkts bytes target     prot opt in     out     source               destination         Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes)pkts bytes target     prot opt in     out     source               destination
[root@cham002 ~]# getenforce
Disabled

配置backup机器：

检查防火墙

[root@test ~]# systemctl stop firewalld
[root@test ~]# iptables -nvL
Chain INPUT (policy ACCEPT 92 packets, 6136 bytes)pkts bytes target     prot opt in     out     source               destination         Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)pkts bytes target     prot opt in     out     source               destination         Chain OUTPUT (policy ACCEPT 50 packets, 5200 bytes)pkts bytes target     prot opt in     out     source               destination
[root@test ~]# setenforce 0
[root@test ~]# getenforce
Permissive

配置Keepalived：
同样使用自定义脚本。

[root@test ~]# vim /etc/keepalived/keepalived.conf
global_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 {       #只有该部分与master有区别state BACKUP           #角色interface ens33        virtual_router_id 51   priority 90            #权重advert_int 1authentication {auth_type PASSauth_pass 123456}virtual_ipaddress {192.168.230.100}   track_script {chk_nginx}
}   

说明： 如果是多台服务器，权重决定了master宕机后，backup接管master角色的顺序。

创建监控脚本：

[root@test ~]# 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.logsystemctl stop keepalivedfi
fi
~
[root@test ~]# chmod 755 !$
chmod 755 /usr/local/sbin/check_ng.sh

说明： 该检测脚本中与master检测脚本的区别在于：systemctl start nginx，即nginx的启动方式，因为在backup服务器中是使用yum安装的nginx，默认添加到系统服务中。

[root@test ~]# systemctl start keepalived
[root@test ~]# ps aux |grep keep
root     10695  0.0  0.1 120704  1396 ?        Ss   20:19   0:00 /usr/sbin/keepalived -D
root     10696  0.0  0.3 127444  3260 ?        S    20:19   0:00 /usr/sbin/keepalived -D
root     10697  0.0  0.2 131640  2840 ?        S    20:19   0:00 /usr/sbin/keepalived -D
root     10728  0.0  0.0 112664   968 pts/1    R+   20:19   0:00 grep --color=auto keep

测试

宕机前

master：

[root@cham002 ~]# 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@cham002 ~]# 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 ~]#

backup：

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

使用VIP访问：

此时默认访问到master服务器。

master宕机后：

[root@cham002 ~]# systemctl stop keepalived
[root@cham002 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1link/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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:e3 brd ff:ff:ff:ff:ff:ffinet 192.168.230.135/24 brd 192.168.230.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.230.150/24 brd 192.168.230.255 scope global secondary ens33:0valid_lft forever preferred_lft foreverinet6 fe80::6f15:52d3:ebeb:e193/64 scope link valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:ed brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global ens37valid_lft forever preferred_lft foreverinet6 fe80::1801:cbbb:ebcc:89a3/64 scope link valid_lft forever preferred_lft forever
[root@cham002 ~]#

backup:

[root@test ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1link/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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:50:f9:44 brd ff:ff:ff:ff:ff:ffinet 192.168.230.145/24 brd 192.168.230.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.230.100/32 scope global ens33valid_lft forever preferred_lft foreverinet6 fe80::9b07:b28d:f5e9:d107/64 scope link valid_lft forever preferred_lft forever

[root@test ~]# tail /var/log/messages
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100

注： 恢复master上的Keepalived服务后,会由master服务器为该集群提供服务。如下：

[root@cham002 ~]# systemctl start keepalived
[root@cham002 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1link/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: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:e3 brd ff:ff:ff:ff:ff:ffinet 192.168.230.135/24 brd 192.168.230.255 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.230.100/32 scope global ens33valid_lft forever preferred_lft foreverinet 192.168.230.150/24 brd 192.168.230.255 scope global secondary ens33:0valid_lft forever preferred_lft foreverinet6 fe80::6f15:52d3:ebeb:e193/64 scope link valid_lft forever preferred_lft forever
3: ens37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000link/ether 00:0c:29:b6:9f:ed brd ff:ff:ff:ff:ff:ffinet 192.168.100.1/24 brd 192.168.100.255 scope global ens37valid_lft forever preferred_lft foreverinet6 fe80::1801:cbbb:ebcc:89a3/64 scope link

在看看日志变化

[root@test ~]# tail /var/log/messages
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
[root@test ~]# tail /var/log/messages
Jan 28 20:44:17 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:44:22 test Keepalived_vrrp[12877]: Sending gratuitous ARP on ens33 for 192.168.230.100
Jan 28 20:47:29 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 90
Jan 28 20:47:29 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) Entering BACKUP STATE
Jan 28 20:47:29 test Keepalived_vrrp[12877]: VRRP_Instance(VI_1) removing protocol VIPs.

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。