HA Cluster的目的，為了防止重要的服務(wù)器在提供服務(wù)時，出現(xiàn)不可抗力的因素，例如硬件故障

創(chuàng)新互聯(lián)是一家以網(wǎng)絡(luò)技術(shù)公司，為中小企業(yè)提供網(wǎng)站維護、成都網(wǎng)站設(shè)計、做網(wǎng)站、網(wǎng)站備案、服務(wù)器租用、域名注冊、軟件開發(fā)、小程序開發(fā)等企業(yè)互聯(lián)網(wǎng)相關(guān)業(yè)務(wù)，是一家有著豐富的互聯(lián)網(wǎng)運營推廣經(jīng)驗的科技公司，有著多年的網(wǎng)站建站經(jīng)驗，致力于幫助中小企業(yè)在互聯(lián)網(wǎng)讓打出自已的品牌和口碑，讓企業(yè)在互聯(lián)網(wǎng)上打開一個面向全國乃至全球的業(yè)務(wù)窗口：建站歡迎聯(lián)系：18980820575

自然災(zāi)害，斷電，軟件bug，操作系統(tǒng)bug等，導(dǎo)致提供服務(wù)的主機出現(xiàn)宕機，死機，從而影響正

常業(yè)務(wù)，因此采用高可用的方案，實現(xiàn)持續(xù)性穩(wěn)定的提供主機服務(wù)的方案稱為HA

A=可用

MTBF=平均無故障時間

MTTR=平均修復(fù)時長

換算公式

A=MTBF/(MTBF+MTTR)

注意提升A值的方法：

1.分子足夠大，但是不劃算，成本太高

2.降低分母，即降低平均修復(fù)時長，成本較低效果出色

如何降低平均修復(fù)時長？

建立備用服務(wù)器，實現(xiàn)Failover功能。

原理：在主服務(wù)器發(fā)生故障時，快速轉(zhuǎn)移IP地址（floating ip），以及快速轉(zhuǎn)移應(yīng)用程序至備用服務(wù)器

需要相關(guān)軟件轉(zhuǎn)移IP（轉(zhuǎn)移IP即實現(xiàn)ip地址從新配置），轉(zhuǎn)移服務(wù)（即在備用主機上啟動相關(guān)服務(wù)應(yīng)用程序）

總體來說HA 即為在主服務(wù)器宕機時，實現(xiàn)備用服務(wù)器的快速切換

關(guān)鍵點：IP地址轉(zhuǎn)移，數(shù)據(jù)共享

HA Cluster （ip，nginx）ip與nginx進程通常稱為HA資源

備用服務(wù)器使用“心跳”檢測，向主服務(wù)器發(fā)起udp報文（UDP報文不需要三次握手等),

根據(jù)主服務(wù)器的響應(yīng)來判斷主服務(wù)器是否正常工作

關(guān)鍵點：響應(yīng)時間，資源爭用共享存儲

假如鏈接主從服務(wù)器的網(wǎng)線連接中斷，則會導(dǎo)致，ip爭奪，而且最主要的是存儲的混亂（同一文件

一邊在增加，另一邊再刪除）

會導(dǎo)致源數(shù)據(jù)的損壞，損害很大，解決方法，使用爆頭設(shè)備（STONITH），在補刀，將未徹底斷電的還在運行設(shè)備斷電

當(dāng)處理完主服務(wù)器的故障后，還要將服務(wù)器上線，即為Failback

Failover<------->Failback

HA Cluster實現(xiàn)方案

1、vrrp協(xié)議的實現(xiàn)

keepalived

2、ais（available Interface standard）：可用接口標(biāo)準(zhǔn)，完備HA集群

RHCS（cmam）

heartbeat

corosync

Keepalived：

vrrp協(xié)議：Virtual Redundant Protocol

術(shù)語:

虛擬路由：virtual router

虛擬路由器標(biāo)識：VRID（0-255）

物理路由：

master：主設(shè)備

back：備用設(shè)備

priority：優(yōu)先級

VIP：virtual Ip

VMAC:Virtual MAC (00-00-5e-00-01-VRID)

GraciousARP(免費arp)

通告：心跳，優(yōu)先級等；周期性；

搶占式，非搶占式；

安全工作：

  認證：

無認證

簡單字符認證

MD5

工作模式

主/備：單虛擬路由器；

主/主：主/備（虛擬路由器1），備/主（虛擬路徑器2）

特點：

vrrp協(xié)議的軟件實現(xiàn)，原生設(shè)計的目的為了高可用ipvs服務(wù)：

vrrp協(xié)議完成地址流動；

為vip地址所在的節(jié)點生成ipvs規(guī)則（在配置文件中預(yù)先定義）；

為ipvs集群的各RS做健康狀態(tài)檢測；

基于腳本調(diào)用接口通過執(zhí)行腳本完成腳本中定義的功能，進而影響集群事務(wù)；

組件：

核心組件：

vrrp stack

ipvs wrapper

checkers

控制組件：配置文件分析器

IO復(fù)用器

內(nèi)存管理組件

HA Cluster的配置前提：

（1）各節(jié)點時間必須同步

（2）確保iptables及selinux不會成為阻礙；

（3）各節(jié)點之間可通過主機名互相通信(對KA并非必須)；

建議使用/etc/hosts文件實現(xiàn)；

（4）各節(jié)點之間的root用戶可以基于密鑰認證的ssh服務(wù)完成互相通信（并非必須）

keepalived安裝配置：

CentOS 6.4+隨base倉庫提供；

1、同步時間

配置chronyd服務(wù)器172.18。200.100

yum安裝chrony，并啟動服務(wù)

[root@localhost ~]# service chronyd start

Starting chronyd:                      [  OK  ]

使用ntpdate命令，同步172.18.10.10以及172.18.10.11的時間

[root@localhost ~]# ntpdate 172.18.200.100

2、清空iptables和selinux

iptables -F

setenforce 0

3、配置hosts文件（非必須）

4、安裝keepalived

[root@localhost ~]# yum install keepalived

[root@localhost ~]# cd /etc/keepalived/

[root@localhost keepalived]# ls

keepalived.conf

[root@localhost keepalived]# cp keepalived.conf keepalived.conf.bak

[root@localhost keepalived]# ls

keepalived.conf  keepalived.conf.bak

[root@localhost keepalived]# vim keepalived.conf

主配置文件：/etc/keepalived/keepalived.conf

配置文件組成部分及相關(guān)選項解釋

TOP HIERACHY

GLOBAL CONFIGURATION

Global definitions

Static routes/addresses

VRRPD CONFIGURATION

VRRP synchronization group(s)：vrrp同步組；

VRRP instance(s)：每個vrrp instance即一個vrrp路由器；

LVS CONFIGURATION

Virtual server group(s)

Virtual server(s)：ipvs集群的vs和rs；

global_defs {   ###全局定義

  notification_email {

   acassen@firewall.loc

   failover@firewall.loc  ####定義出現(xiàn)問題后發(fā)送郵箱的地址

   sysadmin@firewall.loc

  }

  notification_email_from Alexandre.Cassen@firewall.loc  ##從哪里發(fā)過來

  smtp_server 192.168.200.1  ###郵件服務(wù)器地址

  smtp_connect_timeout 30#####超時時間

  router_id LVS_DEVEL###路由器IP

  vrrp_mcast_group4  224.0.100.5###ipv4多播地址

}

vrrp_instance VI_1 {  ##vrrp配置段

  state MASTER###表示是主還是從這里顯示主，另一個則為從

  interface eth0###表明工作從哪個網(wǎng)卡發(fā)出 “多波心跳信息”

  virtual_router_id 51###虛擬路由ID

  priority 100###主的優(yōu)先級

  advert_int 1    ##通告時間間隔

  authentication###認證

    auth_type PASS####認證類型：簡單密鑰認證

    auth_pass 1111#####認證密碼：最多不能超過8位

  }

  virtual_ipaddress {##虛擬IP地址配在哪個網(wǎng)卡上

    192.168.200.16/24 dev eth0  ##定義配置在哪個網(wǎng)卡的別名上

    192.168.200.17

    192.168.200.18

  }

}

track_interface {  ##配置要監(jiān)控的網(wǎng)絡(luò)接口，一旦接口出現(xiàn)故障，則轉(zhuǎn)為FAULT狀態(tài)；即接口跟蹤

eth0

eth2

...

}

nopreempt：定義工作模式為非搶占模式；

preempt_delay 300：搶占式模式下，節(jié)點上線后觸發(fā)新選舉操作的延遲時長；

5、修改配置文件

[root@localhost keepalived]# vim keepalived.conf

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  224.0.100.50

}

vrrp_instance myroute {

  state MASTER

  interface eth2

  virtual_router_id 50

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

    172.18.50.50/16 dev eth2

  }

}

6、將配置文件發(fā)送到另一臺機器10上

[root@localhost keepalived]# scp keepalived.conf 172.18.10.10:/etc/keepalived/

修改配置文件

[root@localhost keepalived]# vim keepalived.conf

global_defs {

  notification_email {

    root@localhost

  }

  notification_email_from keepalived@localhost

  smtp_server 127.0.0.1

  smtp_connect_timeout 30

  router_id node2

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute {

  state BACKUP

  interface eth2

  virtual_router_id 50

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

    172.18.50.50/16 dev eth2

  }

}

7、啟動服務(wù)

啟動備用服務(wù)器11

[root@localhost ~]# service keepalived start

查看地址

[root@localhost ~]# ip a

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

發(fā)現(xiàn)地址已經(jīng)添加，這是若開啟主服務(wù)器，由于沒有設(shè)置搶斷延遲，則會立刻搶斷

8、啟動主服務(wù)器

[root@localhost keepalived]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost keepalived]# ip a

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:99:76:84 brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.11/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe99:7684/64 scope link

    valid_lft forever preferred_lft forever

發(fā)現(xiàn)地址已經(jīng)添加

而從服務(wù)器11上

[root@localhost ~]# ip a

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

IP地址已經(jīng)刪除

9、使用tcpdump抓包工具查看主從服務(wù)器的相應(yīng)心跳測試

[root@localhost keepalived]# tcpdump -i eth2 host 224.0.100.50  ###在主服務(wù)器端抓包

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

16:39:33.357307 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:34.358905 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:35.360605 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:36.362301 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:37.363904 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:38.365658 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:39.367266 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:40.368921 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:41.370599 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

[root@localhost ~]#  tcpdump -i  eth2  -nn host 224.0.100.50  ###在從服務(wù)器端抓包

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

16:39:40.367044 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:41.368741 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:42.370289 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:43.371983 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:44.373750 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:45.375413 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:46.377092 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

16:39:47.378760 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析說明。實現(xiàn)簡單的vrrp

即從服務(wù)器每隔一秒向主服務(wù)器發(fā)送1個通報報文。探測主服務(wù)器是否存活，實現(xiàn)具體實施軟件keepalived

###############################################################################################################################

雙主模型

1、我們在172.18.10.11上配置了主服務(wù)器配置，雙主服務(wù)可在配置文件下面繼續(xù)添如下內(nèi)容，配置如下

[root@localhost keepalived]# vim keepalived.conf

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

    172.18.51.51/16 dev eth2

  }

}

2、將內(nèi)容服務(wù)至粘貼至172.18.10.10服務(wù)器的keepalived.conf配置文件中,然后需要在state和priority上進行相應(yīng)修改

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

    172.18.51.51/16 dev eth2

  }

}

保存并退出，實現(xiàn)雙主模型的設(shè)置

3、從啟服務(wù)并測試

service keepalived restart

Stopping keepalived:                    [  OK  ]

Starting keepalived:                    [  OK  ]

使用tcpdump抓包，結(jié)果如下

172.18.10.11端

[root@localhost keepalived]#  tcpdump -i  eth2  -nn host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:50:20.150330 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:20.521639 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:21.151175 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:21.522539 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:22.152517 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:22.523232 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:23.154334 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:23.524046 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

172.18.10.10端

[root@localhost keepalived]# tcpdump -i eth2 host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:54:01.436075 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:01.437266 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:54:02.437295 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:02.438831 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:54:03.438695 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:54:03.439205 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析每次都會收到兩次信息，一次發(fā)送，一次接收

使用iptable設(shè)置規(guī)則，拒絕172.18.10.11向224.0.100.50發(fā)送通知報文

[root@localhost keepalived]# iptables -A OUTPUT -s 172.18.10.11 -d 224.0.100.50 -j REJECT

在172.18.10.10端使用tcpdump抓包

[root@localhost keepalived]#  tcpdump -i  eth2  -nn host 224.0.100.50

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode

listening on eth2, link-type EN10MB (Ethernet), capture size 65535 bytes

00:50:20.150330 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:20.521639 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:21.151175 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:21.522539 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

00:50:22.152517 IP 172.18.10.10 > 224.0.100.50: VRRPv2, Advertisement, vrid 51, prio 100, authtype simple, intvl 1s, length 20

00:50:22.523232 IP 172.18.10.11 > 224.0.100.50: VRRPv2, Advertisement, vrid 50, prio 100, authtype simple, intvl 1s, length 20

分析發(fā)現(xiàn)發(fā)送兩條通告，因為172.18.10.11不通告，便認為172.18.10.11掛掉了，因此搶斷，讓自己變?yōu)橹鳈C。即別人不通告則認為對方掛掉了

可以使用ip a l 查看相應(yīng)的ip地址獲?。?/p>

[root@localhost keepalived]# ip a l

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.51.51/16 scope global secondary eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

再次在172.18.10.11服務(wù)器上，清空iptables規(guī)則

[root@localhost keepalived]# iptables -F

再回到172.18.10.10服務(wù)器上使用ip a l 查詢

[root@localhost keepalived]# ip a l

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:07:27:ff brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.10/16 brd 172.18.255.255 scope global eth2

  inet 172.18.51.51/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe07:27ff/64 scope link

    valid_lft forever preferred_lft forever

發(fā)現(xiàn)地址已經(jīng)立馬被奪回，是因為工作在搶占模式下。沒有設(shè)置preempt_delay 300搶占延遲時間，

結(jié)論：實現(xiàn)雙主模型實驗

##################################################################################################################

如何實現(xiàn)自定義通知腳本

一、在172.18.10.11服務(wù)器上添加腳本，實現(xiàn)自動發(fā)郵件

1.編寫郵件腳本

vim notify.sh

#!/bin/bash

#

contact='root@localhost'

notify() {

    mailsubject="vrrp： $(hostname) to be $1"

    mailbody="$(hostname) to be $1,vrrp transition, $(date)."

    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

2、測試腳本

語法檢測

[root@localhost keepalived]# bash -n notify.sh

運行腳本測試

[root@localhost keepalived]# bash -x notify.sh master

+ contact=root@localhost

+ case $1 in

+ notify master

++ hostname

+ mailsubject='localhost.localdomain to be master'

++ hostname

++ date

+ mailbody='localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.'

+ echo 'localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.'

+ mail -s mailsubject root@localhost

[root@localhost keepalived]# vim notify.sh

You have mail in /var/spool/mail/root

3、查看收到的郵件

[root@localhost keepalived]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 1 message 1 new

>N  1 root          Mon May 15 01:36  18/696  "mailsubject"

&

Message  1:

From root@localhost.localdomain  Mon May 15 01:36:34 2017

Return-Path:

X-Original-To: root@localhost

Delivered-To: root@localhost.localdomain

Date: Mon, 15 May 2017 01:36:33 +0800

To: root@localhost.localdomain

Subject: mailsubject

User-Agent: Heirloom mailx 12.4 7/29/08

Content-Type: text/plain; charset=us-ascii

From: root@localhost.localdomain (root)

Status: R

localhost.localdomain to be master,vrrp transition, Mon May 15 01:36:33 CST 2017.

&

4、將腳本發(fā)送至172.18.10.10端

[root@localhost keepalived]# scp notify.sh 172.18.10.10:/etc/keepalived/

root@172.18.10.10's password:

notify.sh                                    100%  367   0.4KB/s  00:00

5、調(diào)用腳本

[root@localhost keepalived]# vim keepalived.conf

在172.18.10.11上的vrrp_instance myrouter1下面添加如下內(nèi)容,注意是放在vrrp_instance myrouter1上下文中調(diào)用

    notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

在172.18.10.10上的vrrp_instance myrouter2下面添加如下內(nèi)容

    notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

6，為了實現(xiàn)測試效果，將之前定義的雙主模型刪除，并停止服務(wù)（在10.10和10.11上做同樣的操作）

：.,$d  表示從當(dāng)前行都最后一行全部刪除

[root@localhost keepalived]# service keepalived stop

Stopping keepalived:                    [  OK  ]

7、給之前編寫的腳本加上執(zhí)行權(quán)限

[root@localhost keepalived]# chmod +x  notify.sh

[root@localhost keepalived]# ll

total 8

-rw-r--r-- 1 root root 658 May 15 02:01 keepalived.conf

-rwxr-xr-x 1 root root 367 May 15 01:41 notify.sh

8、啟動服務(wù)

在172.18.10.11端

[root@localhost keepalived]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost keepalived]# ip a l

1: lo: 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

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth2: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:99:76:84 brd ff:ff:ff:ff:ff:ff

  inet 172.18.10.11/16 brd 172.18.255.255 scope global eth2

  inet 172.18.50.50/16 scope global secondary eth2

  inet6 fe80::20c:29ff:fe99:7684/64 scope link

    valid_lft forever preferred_lft forever

[root@localhost keepalived]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 3 messages 2 unread

  1 root          Mon May 15 01:36  19/707  "mailsubject"

>U  2 root          Mon May 15 11:03  19/735  "vrrp: localhost.localdomain to be master"

&

9、啟動172.18.10.10端的keepalived，并且再次到172.18.10.11端查看郵件

[root@localhost ~]# mail

Heirloom Mail version 12.4 7/29/08.  Type ? for help.

"/var/spool/mail/root": 7 messages 5 new 7 unread

 U  1 root          Mon May 15 11:09  19/735  "vrrp: localhost.localdomain to be backup"

 U  2 root          Mon May 15 11:11  19/735  "vrrp: localhost.localdomain to be backup"

>N  3 root          Mon May 15 11:11  18/725  "vrrp: localhost.localdomain to be master"

 N  4 root          Mon May 15 11:11  18/725  "vrrp: localhost.localdomain to be backup"

 N  5 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be backup"

 N  6 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be master"

 N  7 root          Mon May 15 11:26  18/725  "vrrp: localhost.localdomain to be backup"

&

結(jié)論：通知腳本功能實現(xiàn)

######################################################################################################

如何實現(xiàn) keepalived 高可用LVS （重點）

實驗準(zhǔn)備：4臺虛擬主機

其中172.18.10.10和172.18.10.11做為VS端分別為VS2和VS1

172.18.200.100和172.18.249.57做為RS分別為RS1和RS2

首先分別再RS1和RS2端安裝httpd

1、進行如下配置

[root@localhost ~]# cat /var/www/html/index.html

RS1:172.18.200.100

[root@localhost ~]# cat /var/www/html/index.html

RS2:172.18.249.57

2、編寫VIP配置腳本

vim setparam.sh

#!/bin/bash

#

vip='172.18.50.50'

netmask='255.255.255.255'

iface='lo:0'

case $1 in

start)

    echo 1 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 1 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    echo 2 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 2 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    ifconfig $iface $vip netmask $netmask broadcast $vip up

    route add -host $vip dev $iface

    ;;

stop)

    ifconfig $iface down

    echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

    echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

    ;;

esac

3、測試腳本

[root@localhost ~]# bash -n setparam.sh

[root@localhost ~]# bash -x setparam.sh start

+ vip=172.18.50.50

+ netmask=255.255.255.255

+ iface=lo:0

+ case $1 in

+ echo 1

setparam.sh: line 9: /pro/sys/net/ipv4/conf/all/arp_ignore: No such file or directory

+ echo 1

setparam.sh: line 10: /pro/sys/net/ipv4/conf/lo/arp_ignore: No such file or directory

+ echo 2

setparam.sh: line 11: /pro/sys/net/ipv4/conf/all/arp_announce: No such file or directory

+ echo 2

setparam.sh: line 12: /pro/sys/net/ipv4/conf/lo/arp_announce: No such file or directory

+ ifconfig lo:0 172.18.50.50 netmask 255.255.255.255 broadcast 172.18.50.50 up

+ route add -host 172.18.50.50 dev lo:0

4、使用scp將腳本分發(fā)至RS2

[root@localhost ~]# scp setparam.sh 172.18.249.57:/root

root@172.18.249.57's password:

setparam.sh                                          100%  610   0.6KB/s  00:00

5、在RS2端執(zhí)行腳本，并查看是否生成VIP

[root@localhost ~]# bash -x setparam.sh start

+ vip=172.18.50.50

+ netmask=255.255.255.255

+ iface=lo:0

+ case $1 in

+ echo 1

setparam.sh: line 9: /pro/sys/net/ipv4/conf/all/arp_ignore: No such file or directory

+ echo 1

setparam.sh: line 10: /pro/sys/net/ipv4/conf/lo/arp_ignore: No such file or directory

+ echo 2

setparam.sh: line 11: /pro/sys/net/ipv4/conf/all/arp_announce: No such file or directory

+ echo 2

setparam.sh: line 12: /pro/sys/net/ipv4/conf/lo/arp_announce: No such file or directory

+ ifconfig lo:0 172.18.50.50 netmask 255.255.255.255 broadcast 172.18.50.50 up

+ route add -host 172.18.50.50 dev lo:0

[root@localhost ~]# ip a

1: lo: 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

  inet 172.18.50.50/32 brd 172.18.50.50 scope global lo:0

  inet6 ::1/128 scope host

    valid_lft forever preferred_lft forever

2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000

  link/ether 00:0c:29:b2:ca:ea brd ff:ff:ff:ff:ff:ff

  inet 172.18.249.57/16 brd 172.18.255.255 scope global eth0

  inet6 fe80::20c:29ff:feb2:caea/64 scope link

    valid_lft forever preferred_lft forever

6、啟動RS1和RS2的httpd服務(wù)，并查看端口，兩端都要查看，這里只演示一端的

[root@localhost ~]# service httpd start

[root@localhost ~]# ss -tnl

State    Recv-Q Send-Q                Local Address:Port                 Peer Address:Port

LISTEN    0    128                       :::80                       :::*

LISTEN    0    128                       :::22                       :::*

LISTEN    0    128                        *:22                        *:*

LISTEN    0    100                       ::1:25                       :::*

LISTEN    0    100                    127.0.0.1:25

7、在兩個前段節(jié)點生成ipvs規(guī)則

在VS2端

停止keepalived服務(wù)

配置sorry server頁面

vim /var/www/html/index.html

Director2 sorry server2

啟動httpd服務(wù)

[root@localhost ~]# service httpd start

在VS1端

首先停止keepalived服務(wù)

[root@localhost ~]#  service keepalived stop

vim /var/www/html/index.html

Director1

啟動httpd服務(wù)

[root@localhost ~]# service httpd start

在VS1端編輯keepalived配置文件，添加如下內(nèi)容：

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.10.10 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

在VS2端，同樣編輯keepalived.conf文件，添加如下內(nèi)容

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

  persistence_timeout 0

  protocol TCP

  real_server 172.18.10.11 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.10.10 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

在VS2上啟動keepalived服務(wù)

[root@localhost ~]# service keepalived start

Starting keepalived:                    [  OK  ]

[root@localhost ~]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.200.100:80       Route  1    0      0

 -> 172.18.249.57:80       Route  1    0      0

在客戶端使用curl進行訪問測試（配置完有一定延遲，稍等片刻在訪問）

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost ~]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost ~]# curl http://172.18.50.50

RS1:172.18.200.100

在172.18.200.100端停止httpd服務(wù)

[root@localhost ~]# service httpd stop

Stopping httpd:                       [  OK  ]

在VS2端使用ipvsadm觀察

[root@localhost keepalived]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.249.57:80       Route  1    0      2

在172.18.200.100端停止httpd服務(wù)

[root@localhost ~]# service httpd start

在VS2端使用ipvsadm觀察

[root@localhost keepalived]# ipvsadm -ln

IP Virtual Server version 1.2.1 (size=4096)

Prot LocalAddress:Port Scheduler Flags

 -> RemoteAddress:Port      Forward Weight ActiveConn InActConn

TCP  172.18.50.50:80 wrr

 -> 172.18.200.100:80       Route  1    0      0

 -> 172.18.249.57:80       Route  1    0      0

啟動VS1上的keepalived服務(wù)，并且關(guān)閉VS2，客戶端使用curl測試發(fā)現(xiàn)，仍然能夠訪問

[root@localhost keepalived]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost keepalived]# curl http://172.18.50.50

RS1:172.18.200.100

[root@localhost keepalived]# curl http://172.18.50.50

RS2:172.18.249.57

[root@localhost keepalived]# curl http://172.18.50.50

RS1:172.18.200.100

更改配置文件，將之前刪除的雙主內(nèi)容添加進去

VS1端

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

VS2端

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

重啟keepalived服務(wù),相當(dāng)重要，，，，，不重啟不會有效果，這就是個坑

總結(jié)VS端

VS2端keepalived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 node2

  vrrp_mcast_group4  224.0.100.50

}

vrrp_instance myroute1 {

  state BACKUP

  interface eth2

  virtual_router_id 50

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

172.18.50.50/16 dev eth2

  }

notify_master "/etc/keepalived/notify.sh master"

notify_backup "/etc/keepalived/notify.sh backup"

notify_fault "/etc/keepalived/notify.sh fault"

}

vrrp_instance myroute2 {

  state MASTER

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

virtual_server 172.18.51.51 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

  sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

        status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

VS1端keepalived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  224.0.100.50

}

vrrp_instance myroute1 {

  state MASTER

  interface eth2

  virtual_router_id 50

  priority 100

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123456

  }

  virtual_ipaddress {

172.18.50.50/16 dev eth2

  }

notify_master "/etc/keepalived/notify.sh master"

    notify_backup "/etc/keepalived/notify.sh backup"

    notify_fault "/etc/keepalived/notify.sh fault"

}

vrrp_instance myroute2 {

  state BACKUP

  interface eth2

  virtual_router_id 51

  priority 98

  advert_int 1

  authentication {

    auth_type PASS

    auth_pass 123457

  }

  virtual_ipaddress {

172.18.51.51/16 dev eth2

  }

}

virtual_server 172.18.50.50 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

virtual_server 172.18.51.51 80 {

  delay_loop 6

  lb_algo wrr

  lb_kind DR

  persistence_timeout 0

  protocol TCP

sorry_server 127.0.0.1 80

  real_server 172.18.200.100 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

  real_server 172.18.249.57 80 {

    weight 1

    HTTP_GET {

      url {

       path /

status_code 200

      }

      connect_timeout 3

      nb_get_retry 3

      delay_before_retry 3

    }

  }

}

VIP配置腳本 （由用戶是雙主模型因此VIP有兩個）

#!/bin/bash

#

vip='172.18.50.50'

vip2='172.18.51.51'

netmask='255.255.255.255'

iface='lo:0'

iface2='lo:1'

case $1 in

start)

echo 1 > /pro/sys/net/ipv4/conf/all/arp_ignore

echo 1 > /pro/sys/net/ipv4/conf/lo/arp_ignore

echo 2 > /pro/sys/net/ipv4/conf/all/arp_announce

echo 2 > /pro/sys/net/ipv4/conf/lo/arp_announce

ifconfig $iface $vip netmask $netmask broadcast $vip up

ifconfig $iface2 $vip2 netmask $netmask broadcast $vip2 up

route add -host $vip dev $iface

;;

stop)

ifconfig $iface down

ifconfig $iface2 down

echo 0 > /pro/sys/net/ipv4/conf/all/arp_ignore

echo 0 > /pro/sys/net/ipv4/conf/lo/arp_ignore

echo 0 > /pro/sys/net/ipv4/conf/all/arp_announce

echo 0 > /pro/sys/net/ipv4/conf/lo/arp_announce

;;

esac

實驗結(jié)論：實現(xiàn)keepalived 高可用lvs負載均衡

另外有需要云服務(wù)器可以了解下創(chuàng)新互聯(lián)cdcxhl.cn，海內(nèi)外云服務(wù)器15元起步，三天無理由+7*72小時售后在線，公司持有idc許可證，提供“云服務(wù)器、裸金屬服務(wù)器、高防服務(wù)器、香港服務(wù)器、美國服務(wù)器、虛擬主機、免備案服務(wù)器”等云主機租用服務(wù)以及企業(yè)上云的綜合解決方案，具有“安全穩(wěn)定、簡單易用、服務(wù)可用性高、性價比高”等特點與優(yōu)勢，專為企業(yè)上云打造定制，能夠滿足用戶豐富、多元化的應(yīng)用場景需求。

標(biāo)題名稱：HA（高可用）Cluster實現(xiàn)-創(chuàng)新互聯(lián)
網(wǎng)站網(wǎng)址：http://weahome.cn/article/ehjso.html