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. centos 安装flash

    linux系统中安装flash插件 linux中安装flashplayer插件的简单方法: 1.下载其中最新版本的播放器,下载地址: http://get.adobe.com/cn/flashplay ...

  2. C基础 一个可以改变linux的函数getch

    引言  -  getch简述 引用老的TC版本getch说明. (文章介绍点有点窄,  应用点都是一些恐龙游戏时代的开发细节) #include <conio.h> /* * 立即从客户端 ...

  3. FineReport——JS二次开发（CSS改变控件样式）

    FR提供一些选择器,可以改变部分控件的样式,那么对于如何书写自己的css: 可以自己写css文件导入,也可以在页面加载结束事件中添加JS方法来改变css样式. 以文本框为例: 书写css文件,保存至% ...

  4. FineReport——JS二次开发（局部刷新）

    在FR中,可以通过在form表单设置多个报表模板,然后通过对某一模板刷新实现局部刷新的功能,在cpt模板中,由于只存在一个模板,所以无法实现局部刷新. 其实,最好的局部刷新办法是自定义一个页面,然后添 ...

  5. 使用 Visual Studio 部署 .NET Core 应用 ——.Net Core 部署到SUSE Linux Enterprise Server 12 SP2 64 位（GNOME 版本3.20.2）

    SUSE Linux安装 apache 命令:sudo zypper in apache 设置apache 服务可用 命令:sudo systemctl enable apache2.service启 ...

  6. Freemarker的页面和JS遍历后台传入的Map

    后端传到前端的Map Freemarker页面遍历Map: JS遍历Map:

  7. 玩转RaspberryPi

    step1:烧制树莓派内存卡 可以用[Linux系统烧制]http://www.williamsang.com/archives/1764.html 如果用windows烧制的话,就用Win32 Di ...

  8. Centos7Yum安装Mysql8

    1.去官网下载rpm文件,该文件专门用于yum安装方式: https://dev.mysql.com/downloads/repo/yum/ 然后拉到最下面,我下载的是第一个:Red Hat Ente ...

  9. AC日记——可怜的狗狗 洛谷 P1533

    可怜的狗狗 思路: 主席树第k大: 代码: #include <bits/stdc++.h> using namespace std; #define maxn 300001 #defin ...

  10. Python基础系列----语法、数据类型、变量、编码

    1.基本语法                                                                                        Python ...