一 環境介紹
1.操作系統
CentOS Linux release 7.2.1511 (Core)
2.服務
keepalived+nginx雙主高可用負載均衡集群及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協議
在現實的網絡環境中，兩台需要通信的主機大多數情況下並沒有直接的物理連接。對於這樣的情況，它們之間路由怎樣選擇？主機如何選定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：
在主機上使用動態路由協議(RIP、OSPF等)
在主機上配置靜態路由
很明顯，在主機上配置動態路由是非常不切實際的，因為管理、維護成本以及是否支持等諸多問題。配置靜態路由就變得十分流行，但路由器(或者說默認網關default gateway)卻經常成為單點故障。VRRP的目的就是為了解決靜態路由單點故障問題，VRRP通過一競選(election)協議來動態的將路由任務交給LAN中虛擬路由器中的某台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）方式是指以代理服務器來接受Internet上的連接請求，然后將請求轉發給內部網絡上的服務器，並將從服務器上得到的結果返回給Internet上請求連接的客戶端，此時代理服務器對外就表現為一個服務器。(為了理解反向代理，這里插播一條什么是正向代理：正向代理指的是，一個位於客戶端和原始服務器之間的服務器，為了從原始服務器取得內容，客戶端向代理發送一個請求並指定目標(原始服務器)，然后代理向原始服務器轉交請求並將獲得的內容返回給客戶端。)
3.拓撲圖
三 配置
1.后端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
MASTER:
! 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:
! 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 啟動服務並測試
1.啟動后端web服務器
[root@inode4 ~]# systemctl start httpd [root@inode5 ~]# systemctl start httpd
為了測試顯示效果明顯一點，自定義一個訪問頁面
[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
MASTER:
[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，此時我們在客戶端進行測試訪問
[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
訪問正常，接下來我們啟動備用節點的服務器
BACKUP:
[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服務器啟動后進入了BACKUP狀態，Entering BACKUP STATE。接下來我們測試主節點宕機的情形下，我們的服務是否還可用
[root@inode2 ~]# systemctl stop keepalived
主節點宕機后我們查看備用節點的狀態
[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態進入了主狀態，並且IP地址也成功綁定至備用節點下。再次進行測試訪問
[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服務器宕機
[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
在實際生產環境中后端兩台web服務器的內容應該一樣的，在這里我們可認為客戶端已成功訪問到服務器，因此我們可認為這樣的架構體現了高可用負載均衡。