Author: Jin
Date: 20130613
Title: Nginx + Keeplived 双主测试

前言:一年多前做过一次测试，时间久了忘记了，现在又重新做一次

一、环境
1、基本信息和规划
pxe-svr.skylog.cn CentOS release 6.2 (Final)
eth0 192.168.78.250
vip 192.168.78.215

vrrp_instance NGINX1

client0601.skylog.cn CentOS release 6.2 (Final)
eth0 192.168.78.110
vip 192.168.78.115

vrrp_instance NGINX2

2、web
[root@pxe-svr ~]# nginx -v
nginx version: nginx/1.0.15
http://192.168.78.250/index.html
[root@client0601 ~]# nginx -v
nginx version: nginx/1.0.15
http://192.168.78.110/index.html

二、 keeplived安装配置
1、安装
[root@client0601 ~]# yum -y install keepalived
[root@pxe-svr ~]# yum -y install keepalived
2、设置脚本
检测本机http状态，如果状态有问题，则stop keepalived
192.168.78.250上
[root@pxe-svr ~]# cat /root/bin/check_http.sh
#!/bin/bash
url="http://192.168.78.250/index.html"
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/etc/init.d/keepalived stop
fi

192.168.78.110上
[root@client0601 ~]# cat /root/bin/check_http.sh
#!/bin/bash
url="http://192.168.78.110/index.html"
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/etc/init.d/keepalived stop
fi

3、配置keepalived
pxe-svr.skylog.cn
[root@pxe-svr ~]# vim /etc/keepalived/keepalived.conf
#vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
router_id NGINX_ID_1 #定义本机的ID
}

vrrp_script chk_nginx { #检测脚本
script "/root/bin/check_http.sh"
interval 2 ##检测间隔
weight 2
}

vrrp_instance NGINX1 { #定义一个实例
state MASTER #定义为master
interface eth0
virtual_router_id 51 #0-255在同一个instance中一致，在整个vrrp唯一
priority 150 #优先级大的为master
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script { #检查脚本
chk_nginx
}

virtual_ipaddress { #此实例的浮动IP
192.168.78.215
}
}

vrrp_instance NGINX2 {
state BACKUP
interface eth0
virtual_router_id 52
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
track_script {
chk_nginx
}
virtual_ipaddress {
192.168.78.115
}
}

client0601.skylog.cn
将pxe-svr.skylog.cn 设置为backup

[root@client0601 ~]#
[root@client0601 ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived

global_defs {
router_id NGINX_ID_2
}

vrrp_script chk_nginx {
script "/root/bin/check_http.sh"
interval 2
weight 2
}
vrrp_instance NGINX1 {
state BACKUP #把另外一台的设置为BACKUP
interface eth0
virtual_router_id 51 #0-255在同一个instance中一致，在整个vrrp唯一，无变化
priority 110 #优先级大的为mastera,这里要降低，有变化
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script { #检查脚本
chk_nginx
}

virtual_ipaddress { #此实例的浮动IP
192.168.78.215
}
}

vrrp_instance NGINX2 { #本机
state MASTER #修改为MASTER
interface eth0
virtual_router_id 52 #和另外一台配置一
priority 150 #提升为150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script {
chk_nginx
}

virtual_ipaddress {
192.168.78.115
}
}

配置补充
man keepalived.conf
根据范例
# interface for inside_network, bound by vrrp
interface eth0

virtual_ipaddress {
<IPADDR>/<MASK> brd <IPADDR> dev <STRING> scope <SCOPE> label <LABEL>
192.168.200.17/24 dev eth1
192.168.200.18/24 dev eth2 label eth2:1
}
这两个设备不一样

4、启动服务

[root@pxe-svr ~]# /etc/init.d/keepalived start
Starting keepalived: [ OK ]
[root@pxe-svr ~]# /etc/init.d/keepalived status
keepalived (pid 2896) is running...

[root@client0601 ~]# /etc/init.d/keepalived restart
Stopping keepalived: [ OK ]
Starting keepalived: [ OK ]
[root@client0601 ~]# /etc/init.d/keepalived status
keepalived (pid 3024) is running...

[root@client0601 ~]# ps xua|grep keepalived
root 3024 0.0 0.0 16456 1020 ? Ss 12:20 0:00 /usr/sbin/keepalived -D
root 3026 0.0 0.2 16512 2432 ? S 12:20 0:00 /usr/sbin/keepalived -D
root 3027 0.1 0.1 16512 1916 ? S 12:20 0:12 /usr/sbin/keepalived -D
PID
111 Keepalived <-- Parent process monitoring children
112 \_ Keepalived <-- VRRP child
113 \_ Keepalived <-- Healthchecking child

三、测试
测试前的状态
[root@pxe-svr ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:81:78:c1 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.250/24 brd 192.168.78.255 scope global eth0
inet 192.168.1.251/24 brd 192.168.1.255 scope global eth0:1
inet 192.168.78.215/32 scope global eth0
inet6 fe80::20c:29ff:fe81:78c1/64 scope link
valid_lft forever preferred_lft forever
[root@client0601 ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.110/24 brd 192.168.78.255 scope global eth0
inet 192.168.78.115/32 scope global eth0
inet6 fe80::20c:29ff:fe97:c0f8/64 scope link
valid_lft forever preferred_lft forever

C:\Users\diege>ping 192.168.78.215

正在 Ping 192.168.78.215 具有 32 字节的数据:
来自 192.168.78.215 的回复: 字节=32 时间<1ms TTL=64

192.168.78.215 的 Ping 统计信息:
数据包: 已发送 = 1，已接收 = 1，丢失 = 0 (0% 丢失)，
往返行程的估计时间(以毫秒为单位):
最短 = 0ms，最长 = 0ms，平均 = 0ms
Control-C
^C
C:\Users\diege>ping 192.168.78.115

正在 Ping 192.168.78.115 具有 32 字节的数据:
来自 192.168.78.115 的回复: 字节=32 时间<1ms TTL=64
来自 192.168.78.115 的回复: 字节=32 时间<1ms TTL=64

192.168.78.115 的 Ping 统计信息:
数据包: 已发送 = 2，已接收 = 2，丢失 = 0 (0% 丢失)，
往返行程的估计时间(以毫秒为单位):
最短 = 0ms，最长 = 0ms，平均 = 0ms

http访问
http://192.168.78.115/
This is client0601
http://192.168.78.215/
This is pxe-svr.skylog.cn

1.将pxe-svr.skylog.cn上的nginx杀掉后观察
[root@pxe-svr ~]# pkill nginx
先访问
http://192.168.78.115/
This is client0601
没变化
http://192.168.78.215/
切换到
This is client0601

都能访问 ping肯定也会同的
观察ip

[root@pxe-svr ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:81:78:c1 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.250/24 brd 192.168.78.255 scope global eth0
inet 192.168.1.251/24 brd 192.168.1.255 scope global eth0:1
inet6 fe80::20c:29ff:fe81:78c1/64 scope link
valid_lft forever preferred_lft forever
192.168.78.215 这个IP从 pxe-svr.skylog.cn 拿掉了
[root@pxe-svr ~]# /etc/init.d/keepalived status
keepalived is stopped
服务也停掉了
[root@client0601 ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.110/24 brd 192.168.78.255 scope global eth0
inet 192.168.78.115/32 scope global eth0
inet 192.168.78.215/32 scope global eth0
inet6 fe80::20c:29ff:fe97:c0f8/64 scope link
valid_lft forever preferred_lft forever
192.168.78.215 绑到了client0601.skylog.cn的eth0了

注意 inet 192.168.78.215/32 scope global eth0

开启nginx看看
[root@pxe-svr ~]# /etc/init.d/nginx start

开启nginx并不能恢复

2.开启nginx后再开启keepalived
[root@pxe-svr ~]# /etc/init.d/keepalived start
Starting keepalived: [ OK ]
访问http://192.168.78.215/
This is pxe-svr.skylog.cn

[root@pxe-svr ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:81:78:c1 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.250/24 brd 192.168.78.255 scope global eth0
inet 192.168.1.251/24 brd 192.168.1.255 scope global eth0:1
inet 192.168.78.215/32 scope global eth0
inet6 fe80::20c:29ff:fe81:78c1/64 scope link
valid_lft forever preferred_lft forever
[root@client0601 ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.110/24 brd 192.168.78.255 scope global eth0
inet 192.168.78.115/32 scope global eth0
inet6 fe80::20c:29ff:fe97:c0f8/64 scope link
valid_lft forever preferred_lft forever

vip从client0601.skylog.cn切换到pxe-svr.skylog.cn

3、直接关掉其中一台服务器
[root@pxe-svr ~]# shutdown -h now
访问
http://192.168.78.215/
This is client0601
这个没有靠脚本检查关闭keepalived因为直接关机就关闭了服务器，

[root@client0601 ~]# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.110/24 brd 192.168.78.255 scope global eth0
inet 192.168.78.115/32 scope global eth0
inet 192.168.78.215/32 scope global eth0
inet6 fe80::20c:29ff:fe97:c0f8/64 scope link
valid_lft forever preferred_lft forever
[root@client0601 ~]#
也正常了

LOG
初次启动keepalived的时候
[root@pxe-svr ~]#
Jun 13 12:19:13 pxe-svr Keepalived_healthcheckers[2897]: Opening file '/etc/keepalived/keepalived.conf'.
Jun 13 12:19:13 pxe-svr Keepalived_vrrp[2899]: Configuration is using : 42289 Bytes
Jun 13 12:19:13 pxe-svr Keepalived_healthcheckers[2897]: Configuration is using : 5609 Bytes
Jun 13 12:19:13 pxe-svr Keepalived_vrrp[2899]: Using LinkWatch kernel netlink reflector...
Jun 13 12:19:14 pxe-svr Keepalived_healthcheckers[2897]: Using LinkWatch kernel netlink reflector... #healthcheckers 健康检查
Jun 13 12:19:14 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Entering BACKUP STATE
Jun 13 12:19:14 pxe-svr Keepalived_vrrp[2899]: VRRP sockpool: [ifindex(2), proto(112), fd(11,12)]
Jun 13 12:19:14 pxe-svr Keepalived_vrrp[2899]: VRRP_Script(chk_nginx) succeeded #检查脚本成功
Jun 13 12:19:15 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) Transition to MASTER STATE
Jun 13 12:19:16 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) Entering MASTER STATE
Jun 13 12:19:16 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) setting protocol VIPs.
Jun 13 12:19:16 pxe-svr Keepalived_healthcheckers[2897]: Netlink reflector reports IP 192.168.78.215 added
Jun 13 12:19:16 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) Sending gratuitous ARPs on eth0 for 192.168.78.215
Jun 13 12:19:17 pxe-svr ntpd[1534]: Listening on interface #9 eth0, 192.168.78.215#123 Enabled
Jun 13 12:19:17 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Transition to MASTER STATE
Jun 13 12:19:18 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Entering MASTER STATE
Jun 13 12:19:18 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) setting protocol VIPs.
Jun 13 12:19:18 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Sending gratuitous ARPs on eth0 for 192.168.78.115
Jun 13 12:19:18 pxe-svr Keepalived_healthcheckers[2897]: Netlink reflector reports IP 192.168.78.115 added
Jun 13 12:19:20 pxe-svr ntpd[1534]: Listening on interface #10 eth0, 192.168.78.115#123 Enabled
Jun 13 12:19:21 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) Sending gratuitous ARPs on eth0 for 192.168.78.215
Jun 13 12:19:23 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Sending gratuitous ARPs on eth0 for 192.168.78.115
Jun 13 12:19:33 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Received higher prio advert
Jun 13 12:19:33 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) Entering BACKUP STATE
Jun 13 12:19:33 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX2) removing protocol VIPs.
Jun 13 12:19:33 pxe-svr Keepalived_healthcheckers[2897]: Netlink reflector reports IP 192.168.78.115 removed

VIP被删除
Jun 13 12:24:52 pxe-svr Keepalived[2896]: Stopping Keepalived v1.2.7 (02/21,2013)
Jun 13 12:24:52 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) sending 0 priority
Jun 13 12:24:52 pxe-svr Keepalived_vrrp[2899]: VRRP_Instance(NGINX1) removing protocol VIPs.
Jun 13 12:24:54 pxe-svr ntpd[1534]: Deleting interface #9 eth0, 192.168.78.215#123, interface stats: received=0, sent=0, dropped=0, active_time=337 secs
Jun 13 13:53:10 pxe-svr Keepalived[4402]: Starting Keepalived v1.2.7 (02/21,2013)

另外一台VIP被添加
Jun 13 12:20:21 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX2) Sending gratuitous ARPs on eth0 for 192.168.78.115
Jun 13 12:25:27 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Transition to MASTER STATE
Jun 13 12:25:28 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Entering MASTER STATE
Jun 13 12:25:28 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) setting protocol VIPs.
Jun 13 12:25:28 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Sending gratuitous ARPs on eth0 for 192.168.78.215
Jun 13 12:25:28 client0601 Keepalived_healthcheckers[3026]: Netlink reflector reports IP 192.168.78.215 added
Jun 13 12:25:29 client0601 ntpd[1299]: Listening on interface #10 eth0, 192.168.78.215#123 Enabled
Jun 13 12:25:33 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Sending gratuitous ARPs on eth0 for 192.168.78.215
Jun 13 13:18:10 client0601 kernel: hrtimer: interrupt took 2965114 ns
恢复正常时被删除
Jun 13 13:53:45 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Received higher prio advert
Jun 13 13:53:45 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) Entering BACKUP STATE
Jun 13 13:53:45 client0601 Keepalived_vrrp[3027]: VRRP_Instance(NGINX1) removing protocol VIPs.
Jun 13 13:53:45 client0601 Keepalived_healthcheckers[3026]: Netlink reflector reports IP 192.168.78.215 removed
Jun 13 13:53:46 client0601 ntpd[1299]: Deleting interface #10 eth0, 192.168.78.215#123, interface stats: received=0, sent=0, dropped=0, active_time=5297 secs

前端做dns轮训到，这个vip即可

四、将VIP绑定到另外一个块网卡试试看
配置补充
man keepalived.conf
根据范例
# interface for inside_network, bound by vrrp
interface eth0

virtual_ipaddress {
<IPADDR>/<MASK> brd <IPADDR> dev <STRING> scope <SCOPE> label <LABEL>
192.168.200.17/24 dev eth1
192.168.200.18/24 dev eth2 label eth2:1
}
这两个设置不一样

[root@pxe-svr ~]#
vrrp_instance NGINX1 { #定义一个实例,本机
state MASTER #定义为master
interface eth0
virtual_router_id 51 #0-255在同一个instance中一致，在整个vrrp唯一
priority 150 #优先级大的为master
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script { #检查脚本
chk_nginx
}

virtual_ipaddress { #此实例的浮动IP
#192.168.78.215
172.16.9.70/24 dev eth1
}
}

vrrp_instance NGINX2 { #另外一台
state BACKUP
interface eth0
virtual_router_id 52
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
track_script {
chk_nginx
}
virtual_ipaddress {
#192.168.78.115
172.16.9.69/24 dev eth1
}
}
}
[root@client0601 ~]#

vrrp_instance NGINX1 {
state BACKUP #把另外一台的设置为BACKUP
interface eth0
virtual_router_id 51 #0-255在同一个instance中一致，在整个vrrp唯一，无变化
priority 110 #优先级大的为master,这里要降低，有变化
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script { #检查脚本
chk_nginx
}

virtual_ipaddress { #此实例的浮动IP
#192.168.78.215
172.16.9.70/24 dev eth1
}
}

vrrp_instance NGINX2 { #本机
state MASTER #修改为MASTER
interface eth0
virtual_router_id 52 #和另外一台配置一
priority 150 #提升为150
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}

track_script {
chk_nginx
}

virtual_ipaddress {
#192.168.78.115
172.16.9.69/24 dev eth1
}
}
启动keepalived
[root@pxe-svr ~]# /etc/init.d/keepalived start
Starting keepalived: [ OK ]

[root@pxe-svr ~]# ip addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:81:78:c1 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.250/24 brd 192.168.78.255 scope global eth0
inet 192.168.1.251/24 brd 192.168.1.255 scope global eth0:1
inet6 fe80::20c:29ff:fe81:78c1/64 scope link
valid_lft forever preferred_lft forever
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:81:78:cb brd ff:ff:ff:ff:ff:ff
inet 172.16.9.204/24 brd 172.16.9.255 scope global eth1
inet 172.16.9.70/24 scope global secondary eth1
inet 172.16.9.69/24 scope global secondary eth1
inet6 fe80::20c:29ff:fe81:78cb/64 scope link
valid_lft forever preferred_lft forever
启动后访问
http://172.16.9.70/
http://172.16.9.69/
都是
This is pxe-svr.skylog.cn
[root@client0601 ~]# /etc/init.d/keepalived start
Starting keepalived: [ OK ]

再启动
[root@client0601 ~]# ip addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 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: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:f8 brd ff:ff:ff:ff:ff:ff
inet 192.168.78.110/24 brd 192.168.78.255 scope global eth0
inet6 fe80::20c:29ff:fe97:c0f8/64 scope link
valid_lft forever preferred_lft forever
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:97:c0:02 brd ff:ff:ff:ff:ff:ff
inet 172.16.9.251/16 brd 172.16.255.255 scope global eth1
inet 172.16.9.69/24 scope global eth1
inet6 fe80::20c:29ff:fe97:c002/64 scope link
valid_lft forever preferred_lft forever
http://172.16.9.70/
This is pxe-svr.skylog.cn

http://172.16.9.69/
This is client0601

总结
# interface for inside_network, bound by vrrp
#内部网络接口,受vrrp
interface eth0
如果不在virtual_ipaddress{}中指定dev 则默认为前面定义的interface
interface

五、改进
1、增加停止keepalived 时的log
这样知道什么时候nginx不正常 keepalived被停止的
#!/bin/bash
url="http://192.168.78.250/index.html"
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/etc/init.d/keepalived stop
echo "$(date) keepalived has been stopped" >> $(pwd -P)/$0.log
fi

[root@pxe-svr ~]# pkill nginx
[root@pxe-svr ~]# cat /root/bin/check_http.sh.log
Thu Jun 13 14:44:04 CST 2013 keepalived has been stopped

2、增加mail通知，或者调用http接口通知

3、启动keeplived就启动nginx
恢复时，需要先启动nginx,再启动keepalived
单独启动keepalived 无法启动nginx，keepalived也自动不起来
如果需要启动keepalived的时候，nginx也起来，则需要改进脚本，检测进程，如果没有启动则启动
http://deidara.blog.51cto.com/400447/302402/
#!/bin/bash
url="http://192.168.78.250/index.html"
if [ $(ps -C nginx --no-header |wc -l) -eq 0 ];then
/etc/init.d/nginx start
sleep 3
status=$(/usr/bin/curl -s --head "$url" | awk '/HTTP/ {print $2}')
if [ "$status" != "200" ]; then
/etc/init.d/keepalived stop
echo "$(date) keepalived has been stopped" >> $(pwd -P)/$0.log
fi
fi
原理，启动keepalived会执行检测脚本。执行检测脚本的时候就启动nginx,nginx启动异常的话keeplived无法启动

这个脚本，如果把nginx关掉后，会自动再起来，keepalive检查到不存在就执行脚本启动了

是否用脚本把keepalive和nginx 关联起来 看具体的情况

六、总结
其实这个是测试keepalived，web服务器是apache或者nginx都可以
双主为两台都使用，相互为备份

七、延伸
nginx做反向代理
代理到后面的多台服务器
nginx反向代理配置一样，代理多台后端服务器。
这个方式需要手动修改配置文件，手动重载，可能需要需要puppet去管理

八、配置文件内容

master

#vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {

    notification_email {

    wangkaijin@pxe-svr.skylog.cn

    }

    notification_email_from root@pxe-svr.skylog.cn

    smtp_server 127.0.0.1

    stmp_connect_timeout 

    router_id NGINX_ID_1

}

vrrp_script chk_nginx {    #简单脚本

    script "/root/bin/check_http.sh"

    interval         ##检测间隔

    weight

}

vrrp_instance NGINX1 {  #定义一个实例,本机

    state MASTER    #定义为master

    interface eth0

    virtual_router_id  #-255在同一个instance中一致，在整个vrrp唯一

    priority     #优先级大的为master

    advert_int

    authentication {

        auth_type PASS

        auth_pass

    }

    track_script {    #检查脚本

        chk_nginx

    }

    virtual_ipaddress {  #此实例的浮动IP

        #192.168.78.215

        172.16.9.70/ dev eth1

    }

    smtp_alert

}

vrrp_instance NGINX2 {  #另外一台

    state BACKUP

    interface eth0

    virtual_router_id

    priority

    advert_int

    authentication {

        auth_type PASS

        auth_pass

    }

    track_script {

        chk_nginx

    }

    virtual_ipaddress {

        #192.168.78.115

        172.16.9.69/ dev eth1

    }

    smtp_alert

}

slave

! Configuration File for keepalived

global_defs {

    notification_email {

        wangkaijin@pxe-svr.skylog.cn

    }

    notification_email_from root@pxe-svr.skylog.cn

    smtp_server 127.0.0.1

    stmp_connect_timeout 

    router_id NGINX_ID_2

}

vrrp_script chk_nginx {

    script "/root/bin/check_http.sh"

    interval

    weight

}

vrrp_instance NGINX1 {

    state BACKUP       #把另外一台的设置为BACKUP

    interface eth0

    virtual_router_id  #-255在同一个instance中一致，在整个vrrp唯一，无变化

    priority         #优先级大的为master,这里要降低，有变化

    advert_int

    authentication {

        auth_type PASS

        auth_pass

    }

    track_script {      #检查脚本

        chk_nginx

    }

    virtual_ipaddress {  #此实例的浮动IP

        #192.168.78.215

    172.16.9.70/ dev eth1

    }

}

vrrp_instance NGINX2 { #本机

    state MASTER    #修改为MASTER

    interface eth0

    virtual_router_id   #和另外一台配置一

    priority       #提升为150

    advert_int

    authentication {

        auth_type PASS

        auth_pass

    }

    track_script {

        chk_nginx

    }

    virtual_ipaddress {

        #192.168.78.115

    172.16.9.69/ dev eth1

    }

}

http://www.cnblogs.com/dkblog/archive/2011/07/06/2098949.html
网上有看到haproxy+keepalived 方式更好
http://andrewyu.blog.51cto.com/1604432/697466

Nginx + Keeplived双主测试的更多相关文章

  1. 企业Nginx+Keepalived双主架构案例实战

    通过上一次课程的学习,我们知道Nginx+keepalived主从配置,始终有一台服务器处于空余状态,那如何更好的利用起来呢,我们需要借助Nginx+keepalived双主架构来实现,如下图通过改装 ...

  2. nginx+mysql双主搭建

    说明:mysql双主架构经过测试在生产环境中使用没有问题,但是还是建议使用读写分离, Mysql主主同步环境部署: centos 7.4 三台云主机:mysql1 :10.1.1.142 mysql2 ...

  3. nginx+keepalived双主高可用负载均衡

    实验环境及软件版本:CentOS版本: 6.6(2.6.32.-504.el6.x86_64)nginx版本: nginx-1.6.3keepalived版本:keepalived-1.2.7 主LB ...

  4. Nginx+Keepalived双主架构实现

    Keepalived+Nginx实现高可用Web负载均衡 Master 192.168.0.69 nginx.keepalived Centos7.4backup 192.168.0.70 nginx ...

  5. 配置nginx + keepalived 双主模式（双机互为主备）

  6. nginx+keepalived实现nginx双主高可用的负载均衡

    http://kling.blog.51cto.com/3320545/1253474 一.前言: 在互联网上面,网站为用户提供原始的内容访问,同时为用户提供交互操作.提供稳定可靠的服务,可以给用户带 ...

  7. Centos7+nginx+keepalived集群及双主架构案例

    目录简介 一.简介 二.部署nginx+keepalived 集群 三.部署nginx+keepalived双主架构 四.高可用之调用辅助脚本进行资源监控,并根据监控的结果状态实现动态调整 一.简介 ...

  8. nginx+keepalived高可用及双主模式

    高可用有2中方式. 1.Nginx+keepalived 主从配置 这种方案,使用一个vip地址,前端使用2台机器,一台做主,一台做备,但同时只有一台机器工作,另一台备份机器在主机器不出现故障的时候, ...

  9. nginx+keepalived高可用及双主模式【h】

    高可用有2中方式. 1.Nginx+keepalived 主从配置 这种方案,使用一个vip地址,前端使用2台机器,一台做主,一台做备,但同时只有一台机器工作,另一台备份机器在主机器不出现故障的时候, ...

随机推荐

  1. Professional Linux Kernel Architecture 笔记 —— 中断处理（Part 2）【转】

    转自:http://blog.163.com/vic_kk/blog/static/494705242010719483774/ Table of Contents 1 中断 1.1 中断的类型 1. ...

  2. 图论-强连通分量-Tarjan算法

    有关概念: 如果图中两个结点可以相互通达,则称两个结点强连通. 如果有向图G的每两个结点都强连通,称G是一个强连通图. 有向图的极大强连通子图(没有被其他强连通子图包含),称为强连通分量.(这个定义在 ...

  3. hadoop InputFormat 类别

    FileInputFormat是所有使用文件作为数据源的InputFormat的积累.它提供两个功能:一个是定义哪些文件包含在一个作业的输入中:一个为输入文件生成分片的实现.自动将作业分块 作业分块大 ...

  4. JS常用操作方法

    1.splice() 方法向/从数组中添加/删除项目,然后返回被删除的项目. 注释:该方法会改变原始数组. 1 <script type="text/javascript"& ...

  5. 使用 Visual Studio 部署 .NET Core 应用 ——.Net Core 部署到Ubuntu 16.04

    .Net Core 部署到Ubuntu 16.04 中的步骤 1.安装工具 1.apache 2..Net Core(dotnet-sdk-2.0) 3.Supervisor(进程管理工具,目的是服务 ...

  6. java的关键字final

    final可以修饰类,成员方法,成员变量. 1.final修饰的类不能被继承,所以没有子类 final class First{ int num; } class Second extends Fir ...

  7. 百度之星资格赛--IP聚合

    IP聚合 Accepts: 1901 Submissions: 4979 Time Limit: 2000/1000 MS (Java/Others) Memory Limit: 65536/6553 ...

  8. maven新建web项目提示The superclass "javax.servlet.http.HttpServlet" was not found on the Java Build Path

    maven新建web项目提示The superclass "javax.servlet.http.HttpServlet" was not found on the Java Bu ...

  9. Python 什么是ORM?

    关系映射 性能比源生sql效率略差一些 操作性更简单,快捷 Django的orm和sqlalchamy 区别 sqlalchamy没有django的功能全,不支持双下划线的连表跨表操作 sqlalch ...

  10. hdu 4632(区间dp)

    Palindrome subsequence Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 131072/65535 K (Java/ ...