一 环境介绍

1.操作系统
CentOS Linux release 7.2.1511 (Core)

2.服务
keepalived+nginx双主高可用负载均衡集群及LAMP应用
keepalived-1.2.13-7.el7.x86_64
nginx-1.10.2-1.el7.x86_64
httpd-2.4.6-45.el7.centos.x86_64

二 原理及拓扑图

1.vrrp协议
在现实的网络环境中，两台需要通信的主机大多数情况下并没有直接的物理连接。对于这样的情况，它们之间路由怎样选择？主机如何选定到达目的主机的下一跳路由，这个问题通常的解决方法有二种：
 在主机上使用动态路由协议(RIP、OSPF等) 
 在主机上配置静态路由 
很明显，在主机上配置动态路由是非常不切实际的，因为管理、维护成本以及是否支持等诸多问题。配置静态路由就变得十分流行，但路由器(或者说默认网关default gateway)却经常成为单点故障。VRRP的目的就是为了解决静态路由单点故障问题，VRRP通过一竞选(election)协议来动态的将路由任务交给LAN中虚拟路由器中的某台VRRP路由器。

2.nginx反代
nginx是以反向代理的方式进行负载均衡的。反向代理（Reverse Proxy）方式是指以代理服务器来接受Internet上的连接请求，然后将请求转发给内部网络上的服务器，并将从服务器上得到的结果返回给Internet上请求连接的客户端，此时代理服务器对外就表现为一个服务器。(为了理解反向代理，这里插播一条什么是正向代理：正向代理指的是，一个位于客户端和原始服务器之间的服务器，为了从原始服务器取得内容，客户端向代理发送一个请求并指定目标(原始服务器)，然后代理向原始服务器转交请求并将获得的内容返回给客户端。)
3.拓扑图

三 配置

1.后端RS配置

1
2
[root@inode4 ~]# yum install httpd -y
[root@inode5 ~]# yum install httpd -y

2.Nginx反代配置
MASTER:

1
2
3
4
5
6
7
8
9
10
upstream websrvs {
 server 172.18.67.11:80;
 server 172.18.67.12:80;
 server 127.0.0.1:80 backup;
}
server {
    listen       80 ;
    location / {
    proxy_pass http://websrvs;
    }

BACKUP:

1
2
3
4
5
6
7
8
9
10
upstream websrvs {
 server 172.18.67.11:80;
 server 172.18.67.12:80;
 server 127.0.0.1:80 backup;
}
server {
    listen       80 ;
    location / {
    proxy_pass http://websrvs;
    }

3.keepalived高可用配置
MASTER:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
! Configuration File for keepalived
global_defs {
    notification_email {
 root@localhost
    }
    notification_email_from keepalived@localhost
    smtp_server 127.0.0.1
    smtp_connect_timeout 30
    router_id node1
    vrrp_mcast_group4 224.0.67.67
}
vrrp_script chk_down {
    script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0"
    interval 1
    weight -5
}
vrrp_script chk_nginx {
    script "killall -0 nginx && exit 0 || exit 1"
    interval 1
    weight -5
    fall 2
    rise 1
}
vrrp_instance myr {
    state MASTER
    interface eno16777736
    virtual_router_id 167
    priority 100
    advert_int 1
    authentication {
 auth_type PASS
 auth_pass 571f97b2
    }
    virtual_ipaddress {
 172.18.67.33/16 dev eno16777736
    }
    track_script {
 chk_down
 chk_nginx
    }
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

BACKUP:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
! Configuration File for keepalived
global_defs {
    notification_email {
 root@localhost
    }
    notification_email_from keepalived@localhost
    smtp_server 127.0.0.1
    smtp_connect_timeout 30
    router_id node1
    vrrp_mcast_group4 224.0.67.67
}
vrrp_script chk_down {
    script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0"
    interval 1
    weight -5
}
vrrp_script chk_nginx {
    script "killall -0 nginx && exit 0 || exit 1"
    interval 1
    weight -5
    fall 2
    rise 1
}
vrrp_instance myr {
    state BACKUP
    interface eno16777736
    virtual_router_id 167
    priority 95
    advert_int 1
    authentication {
 auth_type PASS
 auth_pass 571f97b2
    }
    virtual_ipaddress {
 172.18.67.33/16 dev eno16777736
    }
    track_script {
 chk_down
 chk_nginx
    }
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

4.通知脚本示例

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
[root@inode2 nginx]# vim notify.sh
#!/bin/bash
#
contact='root@localhost'
notify() {
 mailsubject="$(hostname) to be $1, vip floating"
 mailbody="$(date +'%F %T'): vrrp transition, $(hostname) changed to be $1"
 echo "$mailbody" | mail -s "$mailsubject" $contact
}
case $1 in
master)
 notify master
 ;;
backup)
 notify backup
 ;;
fault)
 notify fault
 ;;
*)
 echo "Usage: $(basename $0) {master|backup|fault}"
 exit 1
 ;;
esac

节点二同样配置

四 启动服务并测试

1.启动后端web服务器

1
2
[root@inode4 ~]# systemctl start httpd
[root@inode5 ~]# systemctl start httpd

为了测试显示效果明显一点，自定义一个访问页面

1
2
[root@inode4 ~]# echo "RS1:172.18.67.11" > /var/www/html/index.html
[root@inode5 ~]# echo "RS2:172.18.67.12" > /var/www/html/index.html

2.测试
MASTER:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
[root@inode2 ~]# systemctl start  keepalived
[root@inode2 ~]# systemctl status -l  keepalived
● keepalived.service - LVS and VRRP High Availability Monitor
   Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled)
   Active: active (running) since Mon 2017-05-15 15:45:20 CST; 3s ago
  Process: 20971 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS)
 Main PID: 20972 (keepalived)
   CGroup: /system.slice/keepalived.service
           ├─20972 /usr/sbin/keepalived -D
           ├─20973 /usr/sbin/keepalived -D
           └─20974 /usr/sbin/keepalived -D
May 15 15:45:20 inode2 Keepalived_healthcheckers[20973]: Opening file '/etc/keepalived/keepalived.conf'.
May 15 15:45:20 inode2 Keepalived_healthcheckers[20973]: Configuration is using : 7521 Bytes
May 15 15:45:20 inode2 Keepalived_healthcheckers[20973]: Using LinkWatch kernel netlink reflector...
May 15 15:45:20 inode2 Keepalived_vrrp[20974]: VRRP_Script(chk_nginx) succeeded
May 15 15:45:21 inode2 Keepalived_vrrp[20974]: VRRP_Instance(myr) Transition to MASTER STATE
May 15 15:45:22 inode2 Keepalived_vrrp[20974]: VRRP_Instance(myr) Entering MASTER STATE
May 15 15:45:22 inode2 Keepalived_vrrp[20974]: VRRP_Instance(myr) setting protocol VIPs.
May 15 15:45:22 inode2 Keepalived_vrrp[20974]: VRRP_Instance(myr) Sending gratuitous ARPs on eno16777736 for 172.18.67.33
May 15 15:45:22 inode2 Keepalived_vrrp[20974]: Opening script file /etc/keepalived/notify.sh
May 15 15:45:22 inode2 Keepalived_healthcheckers[20973]: Netlink reflector reports IP 172.18.67.33 added
[root@inode2 ~]# ip a l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::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 1000
    link/ether 00:0c:29:8b:08:6f brd ff:ff:ff:ff:ff:ff
    inet 172.18.67.13/16 brd 172.18.255.255 scope global eno16777736
       valid_lft forever preferred_lft forever
    inet 172.18.67.33/16 scope global secondary eno16777736
       valid_lft forever preferred_lft forever
    inet6 fe80::20c:29ff:fe8b:86f/64 scope link tentative dadfailed
       valid_lft forever preferred_lft forever

主节点启动，Entering MASTER STATE，此时我们在客户端进行测试访问

1
2
3
4
5
[root@inode1 ~]# for i in {1..4};do curl http://172.18.67.33;done
RS1:172.18.67.11
RS2:172.18.67.12
RS1:172.18.67.11
RS2:172.18.67.12

访问正常，接下来我们启动备用节点的服务器

BACKUP:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
[root@inode3 keepalived]# systemctl start keepalived
[root@inode3 keepalived]# systemctl status -l keepalived
● keepalived.service - LVS and VRRP High Availability Monitor
   Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled)
   Active: active (running) since Mon 2017-05-15 15:46:51 CST; 3s ago
  Process: 24329 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS)
 Main PID: 24330 (keepalived)
   CGroup: /system.slice/keepalived.service
           ├─24330 /usr/sbin/keepalived -D
           ├─24331 /usr/sbin/keepalived -D
           └─24332 /usr/sbin/keepalived -D
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Registering Kernel netlink command channel
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Registering gratuitous ARP shared channel
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Opening file '/etc/keepalived/keepalived.conf'.
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Configuration is using : 66427 Bytes
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Using LinkWatch kernel netlink reflector...
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) Entering BACKUP STATE
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: Opening script file /etc/keepalived/notify.sh
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP_Script(chk_down) succeeded
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP_Script(chk_nginx) succeeded
[root@inode3 keepalived]# ip a l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::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 1000
    link/ether 00:0c:29:78:24:c3 brd ff:ff:ff:ff:ff:ff
    inet 172.18.67.14/16 brd 172.18.255.255 scope global eno16777736
       valid_lft forever preferred_lft forever
    inet6 fe80::20c:29ff:fe78:24c3/64 scope link tentative dadfailed
       valid_lft forever preferred_lft forever

此时，我们可以看到备用节点服务器启动后进入了BACKUP状态，Entering BACKUP STATE。接下来我们测试主节点宕机的情形下，我们的服务是否还可用

1
[root@inode2 ~]# systemctl stop keepalived

主节点宕机后我们查看备用节点的状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
[root@inode3 keepalived]# systemctl status -l keepalived
● keepalived.service - LVS and VRRP High Availability Monitor
   Loaded: loaded (/usr/lib/systemd/system/keepalived.service; disabled; vendor preset: disabled)
   Active: active (running) since Mon 2017-05-15 15:46:51 CST; 2min 19s ago
  Process: 24329 ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS (code=exited, status=0/SUCCESS)
 Main PID: 24330 (keepalived)
   CGroup: /system.slice/keepalived.service
           ├─24330 /usr/sbin/keepalived -D
           ├─24331 /usr/sbin/keepalived -D
           └─24332 /usr/sbin/keepalived -D
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP_Script(chk_down) succeeded
May 15 15:46:51 inode3 Keepalived_vrrp[24332]: VRRP_Script(chk_nginx) succeeded
May 15 15:48:35 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) Transition to MASTER STATE
May 15 15:48:36 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) Entering MASTER STATE
May 15 15:48:36 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) setting protocol VIPs.
May 15 15:48:36 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) Sending gratuitous ARPs on eno16777736 for 172.18.67.33
May 15 15:48:36 inode3 Keepalived_vrrp[24332]: Opening script file /etc/keepalived/notify.sh
May 15 15:48:36 inode3 Keepalived_healthcheckers[24331]: Netlink reflector reports IP 172.18.67.33 added
May 15 15:48:41 inode3 Keepalived_vrrp[24332]: VRRP_Instance(myr) Sending gratuitous ARPs on eno16777736 for 172.18.67.33
[root@inode3 keepalived]# ip a l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::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 1000
    link/ether 00:0c:29:78:24:c3 brd ff:ff:ff:ff:ff:ff
    inet 172.18.67.14/16 brd 172.18.255.255 scope global eno16777736
       valid_lft forever preferred_lft forever
    inet 172.18.67.33/16 scope global secondary eno16777736
       valid_lft forever preferred_lft forever
    inet6 fe80::20c:29ff:fe78:24c3/64 scope link tentative dadfailed
       valid_lft forever preferred_lft forever

我们发现备用节点由备用状态进入了主状态，并且IP地址也成功绑定至备用节点下。再次进行测试访问

1
2
3
4
5
[root@inode1 ~]# for i in {1..4};do curl http://172.18.67.33;done
RS1:172.18.67.11
RS2:172.18.67.12
RS1:172.18.67.11
RS2:172.18.67.12

测试一台web服务器宕机

1
2
3
4
5
6
[root@inode4 ~]# systemctl stop httpd
[root@inode1 ~]# for i in {1..4};do curl http://172.18.67.33;done
RS2:172.18.67.12
RS2:172.18.67.12
RS2:172.18.67.12
RS2:172.18.67.12

在实际生产环境中后端两台web服务器的内容应该一样的，在这里我们可认为客户端已成功访问到服务器，因此我们可认为这样的架构体现了高可用负载均衡。

Keepalived+Nginx实现高可用负载均衡集群的更多相关文章

  1. LVS+Keepalived+Nginx+Tomcat高可用负载均衡集群配置（DR模式，一个VIP，多个端口）

    一．概述 LVS作用:实现负载均衡 Keepalived作用:监控集群系统中各个服务节点的状态,HA cluster. 配置LVS有两种方式: 1. 通过ipvsadm命令行方式配置 2. 通过Red ...

  2. LVS+Keepalived搭建MyCAT高可用负载均衡集群

    LVS+Keepalived 介绍 LVS LVS是Linux Virtual Server的简写,意即Linux虚拟服务器,是一个虚拟的服务器集群系统.本项目在1998年5月由章文嵩博士成立,是中国 ...

  3. Haproxy+Keepalived搭建Weblogic高可用负载均衡集群

    配置环境说明: KVM虚拟机配置 用途 数量 IP地址 机器名 虚拟IP地址 硬件 内存3G  系统盘20G cpu 4核 Haproxy keepalived 2台 192.168.1.10 192 ...

  4. Keepalived+LVS(dr)高可用负载均衡集群的实现

    一 环境介绍 1.操作系统CentOS Linux release 7.2.1511 (Core) 2.服务keepalived+lvs双主高可用负载均衡集群及LAMP应用keepalived-1.2 ...

  5. docker下用keepalived+Haproxy实现高可用负载均衡集群

    启动keepalived后宿主机无法ping通用keepalived,报错: [root@localhost ~]# ping 172.18.0.15 PING () bytes of data. F ...

  6. Linux 笔记 - 第十八章 Linux 集群之（三）Keepalived+LVS 高可用负载均衡集群

    一.前言 前两节分别介绍了 Linux 的高可用集群和负载均衡集群,也可以将这两者相结合,即 Keepalived+LVS 组成的高可用负载均衡集群,Keepalived 加入到 LVS 中的原因有以 ...

  7. 测试LVS+Keepalived高可用负载均衡集群

    测试LVS+Keepalived高可用负载均衡集群 1. 启动LVS高可用集群服务 此时查看Keepalived服务的系统日志信息如下: [root@localhost ~]# tail -f /va ...

  8. Dubbo入门到精通学习笔记（二十）：MyCat在MySQL主从复制的基础上实现读写分离、MyCat 集群部署(HAProxy + MyCat)、MyCat 高可用负载均衡集群Keepalived

    文章目录 MyCat在MySQL主从复制的基础上实现读写分离 一.环境 二.依赖课程 三.MyCat 介绍 ( MyCat 官网:http://mycat.org.cn/ ) 四.MyCat 的安装 ...

  9. LVS+Keepalived 实现高可用负载均衡集群

    LVS+Keepalived  实现高可用负载均衡集群     随着网站业务量的增长,网站的服务器压力越来越大?需要负载均衡方案!商业的硬件如 F5 ,Array又太贵,你们又是创业型互联公司如何有效 ...

随机推荐

  1. ARC068E - Snuke Line

    原题链接 题意简述 给出个区间和.求对于任意,有多少个区间包含的倍数. 题解 考虑怎样的区间不包含的倍数. 对于的倍数和,满足的区间不包含任何的倍数. 于是转化为二维数点问题,可以用可持久化线段树解决 ...

  2. typeahead + JDK 8 并行流 + redis 高速即时查询.

    感谢JDK8,让我们JAVA 程序员暂时不用担心失业. 有些情况,需要根据用户输入值,即时查询数据库,MYSQL显然不再适合这种业务. mongoDB看似最适合,但是为了这么一个破功能,也不值得特意去 ...

  3. nginx的环境配置的问题

    在安装好nginx之后,运行nginx,报错: nginx dyld: Library not loaded: /usr/local/lib/libpcre.1.dylib Referenced fr ...

  4. 关联规则—频繁项集Apriori算法

    频繁模式和对应的关联或相关规则在一定程度上刻画了属性条件与类标号之间的有趣联系,因此将关联规则挖掘用于分类也会产生比较好的效果.关联规则就是在给定训练项集上频繁出现的项集与项集之间的一种紧密的联系.其 ...

  5. TableLayoutPanel居中和单元格内元素居中

    在后台程序新建一个TableLayoutPanel 添加到form中,默认显示在左上角,想了很多让它居中的办法,在网上找了不少 最好的是: winform要设置控件的位置有3种: 1.控件的ancho ...

  6. Enable multi-tenancy on ironic

    Multi-tenancy 是openstack ironic从Ocata版本开始支持的新特性,通过network-generic-switch插件控制交换机,Ironic可以实现在不同租户间机网络隔 ...

  7. java读取文件乱码

    List<String> lines=new ArrayList<String>(); BufferedReader br = new BufferedReader(new F ...

  8. sqoop:mysql和Hbase/Hive/Hdfs之间相互导入数据

    1.安装sqoop 请参考http://www.cnblogs.com/Richardzhu/p/3322635.html 增加了SQOOP_HOME相关环境变量:source ~/.bashrc  ...

  9. HTML入门标签汇总

    HTML入门标签汇总 1.<div></div>用于定义文档的区块,用来划分出独立不同的部分. 2.<h1></h1>数字1-6定义从大到小的标题. 3 ...

  10. 用dd实现linux硬盘备份

    一个去年的老本,500G硬盘,空间各种捉急,准备迁移到公司的台式机上,却发现Linux上迁移环境没有Windows下那么方便,只能复制整块硬盘了. 从公司拿了一块1T的硬盘,插入移动硬盘盒(淘宝上搞的 ...