一環境介紹

1.作業系統

CentOS Linux release 7.2.1511 (Core)

2.服務

keepalived+nginx雙主高可用負載均衡叢集及LAMP應用

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

二原理及拓撲圖

1.vrrp協定

在現實的網絡環境中，兩台需要通信的主機大多數情況下并沒有直接的實體連接配接。對于這樣的情況，它們之間路由怎樣選擇？主機如何標明到達目的主機的下一跳路由，這個問題通常的解決方法有二種：

在主機上使用動态路由協定(RIP、OSPF等)

在主機上配置靜态路由

很明顯，在主機上配置動态路由是非常不切實際的，因為管理、維護成本以及是否支援等諸多問題。配置靜态路由就變得十分流行，但路由器(或者說預設網關default gateway)卻經常成為單點故障。VRRP的目的就是為了解決靜态路由單點故障問題，VRRP通過一競選(election)協定來動态的将路由任務交給LAN中虛拟路由器中的某台VRRP路由器。

2.nginx反代

nginx是以反向代理的方式進行負載均衡的。反向代理（Reverse Proxy）方式是指以代理伺服器來接受Internet上的連接配接請求，然後将請求轉發給内部網絡上的伺服器，并将從伺服器上得到的結果傳回給Internet上請求連接配接的用戶端，此時代理伺服器對外就表現為一個伺服器。(為了了解反向代理，這裡插播一條什麼是正向代理：正向代理指的是，一個位于用戶端和原始伺服器之間的伺服器，為了從原始伺服器取得内容，用戶端向代理發送一個請求并指定目标(原始伺服器)，然後代理向原始伺服器轉交請求并将獲得的内容傳回給用戶端。)

3.拓撲圖

三配置

1.後端RS配置

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

2.Nginx反代配置

MASTER:

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:

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

! 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"
}

! 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.通知腳本示例

[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

節點二同樣配置

四啟動服務并測試

1.啟動後端web伺服器

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

為了測試顯示效果明顯一點，自定義一個通路頁面

[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.測試

[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

主節點啟動，Entering MASTER STATE，此時我們在用戶端進行測試通路

[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

通路正常，接下來我們啟動備用節點的伺服器

[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

此時，我們可以看到備用節點伺服器啟動後進入了BACKUP狀态，Entering BACKUP STATE。接下來我們測試主節點當機的情形下，我們的服務是否還可用

[root@inode2 ~]# systemctl stop keepalived

主節點當機後我們檢視備用節點的狀态

[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

我們發現備用節點由備用狀态進入了主狀态，并且IP位址也成功綁定至備用節點下。再次進行測試通路

[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

測試一台web伺服器當機

[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

在實際生産環境中後端兩台web伺服器的内容應該一樣的，在這裡我們可認為用戶端已成功通路到伺服器，是以我們可認為這樣的架構展現了高可用負載均衡。

Keepalived+Nginx實作高可用負載均衡叢集一環境介紹二原理及拓撲圖三配置四啟動服務并測試

一環境介紹

二原理及拓撲圖

三配置

四啟動服務并測試

繼續閱讀

2.6核心netfilter包截獲

歡迎使用CSDN-markdown編輯器Nginx配置檔案nginx.conf中文詳解（總結）

雲伺服器部署項目：vue-cli 部署服務配置

win7 localhost 配置 nginx ssl自簽名證書步驟報錯相關連結

（Nginx）03_Nginx原理與優化一、Nginx原理二、master-workers機制三、面試題：

（Nginx）02_Nginx配置執行個體一、Nginx的應用：反向代理實作一：二、反向代理執行個體二：三、Nginx的配置執行個體三–負載均衡：四、Nginx的配置執行個體四–動靜分離：Nginx在實際開發中的應用：

nginx配置示例

niginx的location的location詳細規則和優先級關系以及測試案例2.1比對文法2.2優先級判斷2.3練習3.1處理邏輯如何執行3.2功能測試3.3流程解析

Puppet自動化Nginx+Mongrel負載均衡配置

Ubuntu修改nginx将多個域名分别配置到伺服器不同端口

修改or隐藏nginx版本号

Nginx安全改動：隐藏Nginx版本号背景解決serve_tokens指令參考

版本号隐藏

Nginx服務優化（版本隐藏、版本号僞裝、逾時管理、程序管理、日志分割）Nginx服務優化未完待續

nginx location中斜線的位置的重要性

Keepalived+Nginx實作高可用負載均衡叢集一 環境介紹 二 原理及拓撲圖 三 配置 四 啟動服務并測試

一 環境介紹

二 原理及拓撲圖

三 配置

四 啟動服務并測試

繼續閱讀

Keepalived+Nginx實作高可用負載均衡叢集一環境介紹二原理及拓撲圖三配置四啟動服務并測試

一環境介紹

二原理及拓撲圖

三配置

四啟動服務并測試