设为首页 收藏本站

运维网

 找回密码
 立即注册
查看: 876|回复: 0

[经验分享] haproxy + keepalived “一键安装包” 配置与测试HOWTO

[复制链接]
搜鞥都哦

尚未签到
发表于 2018-12-29 11:11:58 | 显示全部楼层 |阅读模式
　　本文为 “haproxy + keepalived 一键安装包”  的 第四篇
　　目的:帮助快速设置haproxy 与 keepalived 配置文档
　　第一篇 haproxy 安装与配置
　　第二篇 keepalived 安装与配置
　　第三篇 haproxy + keepalived 一键安装包(keepalived 为haproxy 做 HA 高可用）
　　第四篇 haproxy + keepalived “一键安装包”  配置与测试HOWTO
　　一 haproxy.sh 与 keepalived.sh 服务脚本


  • [root@four cluster]# ls /usr/local/sbin/*.ah
  • haproxy.sh  keepalived.sh

　　二 设置自动启动 haproxy.sh 与 keepalived.sh


  • [root@four cluster]# cat /etc/rc.local  
  • #!/bin/sh
  • #
  • # This script will be executed *after* all the other init scripts.
  • # You can put your own initialization stuff in here if you don't
  • # want to do the full Sys V style init stuff.

  • touch /var/lock/subsys/local
  • ulimit -SHn 65535

  • /usr/local/sbin/haproxy.sh start
  • /usr/local/sbin/keepalived.sh start

　　三 haproxy 相关配置


  • 3.1 haroxy 程序目录
  • #ls /usr/local/haproxy
  • doc  haproxy.cfg  html  logs  sbin  share

  • 3.2 haproxy 配置文档
  • [root@four cluster]# file /usr/local/haproxy/haproxy.cfg  
  • /usr/local/haproxy/haproxy.cfg: UTF-8 Unicode text

  • 3.3 编辑haproxy 配置文档
  • [root@four cluster]# vim /usr/local/haproxy/haproxy.cfg  
  • 第37 行 定义健康检查html文档，所以你需要在你的web服务上建立check.html文件用于haproxy
  • 健康检查，也可以跟具需要自定义检查文档。
  • option  httpchk HEAD /check.html #用来做健康检查html文档
  • 第38 39 行修10.0.1.252/253 替换成你的web 服务器ip地址
  • server server1 10.0.1.252:80 cookie server1 check inter 2000 rise 3     fall 3 weight 3
  • server server2 10.0.1.253:80 cookie server2 check inter 2000 rise 3     fall 3 maxconn 120 weight 3

  • 3.4 启动haproxy 输入如下命令  
  • [root@four cluster]# haproxy.sh start

  • 3.5 检查是否有haproxy 进程
  • [root@four cluster]# ps aux | grep haproxy | grep -v 'grep'
  • nobody    4216  0.0  0.2   7436  1108 ?        Ss   20:35   0:00 /usr/local/haproxy/sbin/haproxy -f /usr/local/haproxy/haproxy.cfg

  • 3.6 检查haproxy 是否监听80 端口
  • [root@four cluster]# lsof -i :80
  • COMMAND  PID   USER   FD   TYPE DEVICE SIZE NODE NAME
  • haproxy 4216 nobody    4u  IPv4   8229       TCP *:http (LISTEN)

  • 3.7 简单测试 haproxy 看看否返回 ’200 OK‘
  • [root@four cluster]# curl -IL http://10.0.1.254 (换成你的ip )
  • HTTP/1.1 200 OK
  • Server: nginx/1.0.5
  • Date: Wed, 28 Dec 2011 12:42:30 GMT
  • Content-Type: text/html
  • Content-Length: 159
  • Last-Modified: Sat, 03 Dec 2011 11:36:10 GMT
  • Connection: close
  • Accept-Ranges: bytes

　　四 keepalived 相关配置


  • 4.1 keepalived 程序目录
  • [root@four cluster]# ls /usr/local/keepalived
  • bin  etc  haproxy_check.sh  keepalived.conf  sbin  share

  • 4.2 haproxy_check.sh shell脚本监控程序,用于检查 80端口，来判断haproxy 是否可用
  • [root@four cluster]# file /usr/local/keepalived/haproxy_check.sh  
  • /usr/local/keepalived/haproxy_check.sh: Bourne-Again shell script text executable

  • 4.3 keepalived 配置文档
  • [root@four cluster]# file /usr/local/keepalived/keepalived.conf  
  • /usr/local/keepalived/keepalived.conf: ASCII text

  • 4.4 编辑keepalived 配置文档
  • [root@four cluster]# vim /usr/local/keepalived/keepalived.conf  
  • 第16 行 interface 默认值 eth1 ,改成你的haproxy 外网卡(公网ip网卡)
  • interface eth0

  • 第 34 行 默认值 10.0.1.250/24 dev eth1 ，10.0.1.250 替换成你virtual ip 地址,dev eth1 改成你的haproxy 外网卡(与第16行一致)
  • 10.0.1.250/24 dev eth0 scope global

  • 4.5 启动keepalived 输入如下命令  
  • [root@four cluster]# keepalived.sh start
  • [root@four cluster]# 4242                #此处返回的是keepalived 进程id，敲下回车键即可

  • 4.6 检查是否有keepalived 进程
  • # ps aux | grep keepalived | grep -v 'grep'  
  • root      4238  0.0  0.0  63844   464 pts/0    S    20:52   0:00 /bin/bash /usr/local/sbin/keepalived.sh start
  • root      4240  0.0  0.1  35776   608 ?        Ss   20:52   0:00 /usr/local/keepalived/sbin/keepalived -f /usr/local/keepalived/keepalived.conf
  • root      4241  0.0  0.1  39936  1048 ?        S    20:52   0:00 /usr/local/keepalived/sbin/keepalived -f /usr/local/keepalived/keepalived.conf
  • root      4242  0.0  0.2  63848  1152 pts/0    S    20:52   0:00 /bin/bash /usr/local/keepalived/haproxy_check.sh start

  • 4.7 查看keepalived 输出日志
  • [root@four cluster]# tail /var/log/messages  
  • Dec 28 20:52:28 four Keepalived_vrrp: Registering Kernel netlink reflector
  • Dec 28 20:52:28 four Keepalived_vrrp: Registering Kernel netlink command channel
  • Dec 28 20:52:28 four Keepalived_vrrp: Registering gratutious ARP shared channel
  • Dec 28 20:52:28 four Keepalived: Starting VRRP child process, pid=4241
  • Dec 28 20:53:08 four Keepalived_vrrp: Opening file '/usr/local/keepalived/keepalived.conf'.  
  • Dec 28 20:53:08 four Keepalived_vrrp: Configuration is using : 63577 Bytes
  • Dec 28 20:53:08 four Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 20:53:08 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 20:53:12 four Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 20:53:13 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE              #当前keepalived 状态为master

　　五 haproxy 与 keepalived 整体测试


  • 5.1 查看 haproxy 与 keepalived 进程树
  • [root@four cluster]# pstree  
  • init─┬─crond
  •      ├─events/0
  •      ├─events/1
  •      ├─haproxy                                      #haproxy 进程
  •      ├─keepalived───keepalived                      #keepalived 进程(2个)
  •      ├─keepalived.sh───haproxy_check.s───sleep      #keepalived.sh 调用 haproxy_check.sh (监控脚本) 调用 sleep （用于控检查间隔时间,2s）


  • 5.2 模拟网络故障 haproxy 切换测试  
  • haproxy 1 日志
  • [root@one ~]# tail /var/log/messages -f
  • Dec 28 21:09:43 one kernel: NFSD: starting 90-second grace period
  • Dec 28 21:09:44 one Keepalived: Starting Keepalived v1.2.2 (12/25,2011)  
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering Kernel netlink reflector
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering Kernel netlink command channel
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering gratutious ARP shared channel
  • Dec 28 21:09:44 one Keepalived: Starting VRRP child process, pid=1993
  • Dec 28 21:10:24 one Keepalived_vrrp: Opening file '/usr/local/keepalived/keepalived.conf'.  
  • Dec 28 21:10:24 one Keepalived_vrrp: Configuration is using : 63597 Bytes
  • Dec 28 21:10:24 one Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 21:10:24 one Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE

  • haproxy 2 日志
  • [root@four cluster]# ifdown eth0
  • [root@four cluster]# tail /var/log/messages  
  • Dec 28 20:52:28 four Keepalived: Starting VRRP child process, pid=4241
  • Dec 28 20:53:08 four Keepalived_vrrp: Opening file '/usr/local/keepalived/keepalived.conf'.  
  • Dec 28 20:53:08 four Keepalived_vrrp: Configuration is using : 63577 Bytes
  • Dec 28 20:53:08 four Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 20:53:08 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 20:53:12 four Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 20:53:13 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: Kernel is reporting: interface eth0 DOWN
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering FAULT STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Now in FAULT state                 #keepalived fault 状态

  • haproxy 1 日志
  • [root@one ~]# tail /var/log/messages -f
  • Dec 28 21:09:43 one kernel: NFSD: starting 90-second grace period
  • Dec 28 21:09:44 one Keepalived: Starting Keepalived v1.2.2 (12/25,2011)  
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering Kernel netlink reflector
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering Kernel netlink command channel
  • Dec 28 21:09:44 one Keepalived_vrrp: Registering gratutious ARP shared channel
  • Dec 28 21:09:44 one Keepalived: Starting VRRP child process, pid=1993
  • Dec 28 21:10:24 one Keepalived_vrrp: Opening file '/usr/local/keepalived/keepalived.conf'.  
  • Dec 28 21:10:24 one Keepalived_vrrp: Configuration is using : 63597 Bytes
  • Dec 28 21:10:24 one Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 21:10:24 one Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 21:12:08 one Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 21:12:09 one Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE             #成功切换为 master  

  • haproxy 1 检查vip  
  • [root@one ~]# ip add
  • 1: lo:  mtu 16436 qdisc noqueue  
  •     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:  mtu 1500 qdisc pfifo_fast qlen 1000
  •     link/ether 08:00:27:74:3c:05 brd ff:ff:ff:ff:ff:ff
  •     inet 10.0.2.251/24 brd 10.0.2.255 scope global eth0
  •     inet6 fe80::a00:27ff:fe74:3c05/64 scope link  
  •        valid_lft forever preferred_lft forever
  • 3: eth1:  mtu 1500 qdisc pfifo_fast qlen 1000
  •     link/ether 08:00:27:3e:9a:2b brd ff:ff:ff:ff:ff:ff
  •     inet 10.0.1.251/24 brd 10.0.1.255 scope global eth1                                   

  •     inet 10.0.1.250/24 scope global secondary eth1           #virtual ip
  •     inet6 fe80::a00:27ff:fe3e:9a2b/64 scope link  
  •        valid_lft forever preferred_lft forever
  • 4: sit0:  mtu 1480 qdisc noop  
  •     link/sit 0.0.0.0 brd 0.0.0.0

  • 5.3 模拟haproxy 80端口故障 keepalived 切换测试  
  • haproxy 2 恢复
  • [root@four cluster]# ifup eth0
  • haproxy 2 日志
  • [root@four cluster]# tail /var/log/messages -f
  • Dec 28 20:53:08 four Keepalived_vrrp: Configuration is using : 63577 Bytes
  • Dec 28 20:53:08 four Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 20:53:08 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 20:53:12 four Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 20:53:13 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: Kernel is reporting: interface eth0 DOWN
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering FAULT STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Now in FAULT state
  • Dec 28 21:15:34 four kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
  • Dec 28 21:15:35 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE


  • haproxy 1 进程
  • [root@one ~]# pstree
  • init─┬─crond
  •      ├─dbus-daemon
  •      ├─events/0
  •      ├─haproxy
  •      ├─keepalived───keepalived
  •      ├─keepalived.sh───haproxy_check.s───sleep

  • 停止 haproxy 进程
  • [root@one ~]# killall haproxy                                        #停止haproxy 进程
  • [root@one ~]# tail /usr/local/keepalived/fault.log                   #fault log
  • 2011-12-28_21:17:50 http://10.0.1.251:80 1      fault
  • 2011-12-28_21:17:51 http://10.0.1.251:80 2      fault
  • 2011-12-28_21:17:52 http://10.0.1.251:80 3      fault
  • 2011-12-28_21:17:53 http://10.0.1.251:80 4      fault
  • 2011-12-28_21:17:54 http://10.0.1.251:80 5      fault
  • 2011-12-28_21:17:55 http://10.0.1.251:80 6      fault
  • 2011-12-28_21:17:56 http://10.0.1.251:80 7      fault
  • 2011-12-28_21:17:57 http://10.0.1.251:80 8      fault

  • haproxy 1 日志
  • [root@one ~]# tail /var/log/messages  
  • Dec 28 21:12:09 one Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE
  • Dec 28 21:17:58 one Keepalived: Terminating on signal                                      
  • Dec 28 21:17:58 one Keepalived: Stopping Keepalived v1.2.2 (12/25,2011)                   #keepalived 被haproxy_check.sh 监控脚本关闭
  • Dec 28 21:17:58 one Keepalived_vrrp: Terminating VRRP child process on signal
  • Dec 28 21:17:58 one kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready
  • Dec 28 21:17:58 one kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
  • Dec 28 21:17:58 one kernel: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
  • Dec 28 21:18:00 one kernel: ADDRCONF(NETDEV_UP): eth1: link is not ready
  • Dec 28 21:18:00 one kernel: e1000: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
  • Dec 28 21:18:00 one kernel: ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready


  • haproxy 2 日志
  • [root@four cluster]# tail /var/log/messages -f
  • Dec 28 20:53:08 four Keepalived_vrrp: Configuration is using : 63577 Bytes
  • Dec 28 20:53:08 four Keepalived_vrrp: Using LinkWatch kernel netlink reflector...
  • Dec 28 20:53:08 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 20:53:12 four Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 20:53:13 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: Kernel is reporting: interface eth0 DOWN
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering FAULT STATE
  • Dec 28 21:12:07 four Keepalived_vrrp: VRRP_Instance(VI_1) Now in FAULT state
  • Dec 28 21:15:34 four kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
  • Dec 28 21:15:35 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering BACKUP STATE
  • Dec 28 21:18:02 four Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE
  • Dec 28 21:18:03 four Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE          #keepalived 状态为 master            

  • [root@four cluster]# ip add                 
  • 1: lo:  mtu 16436 qdisc noqueue  
  •     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:  mtu 1500 qdisc pfifo_fast qlen 1000
  •     link/ether 08:00:27:27:ef:a6 brd ff:ff:ff:ff:ff:ff
  •     inet 10.0.2.254/24 brd 10.0.2.255 scope global eth0
  •     inet6 fe80::a00:27ff:fe27:efa6/64 scope link  
  •        valid_lft forever preferred_lft forever
  • 3: eth1:  mtu 1500 qdisc pfifo_fast qlen 1000
  •     link/ether 08:00:27:cd:f3:db brd ff:ff:ff:ff:ff:ff
  •     inet 10.0.1.254/24 brd 10.0.1.255 scope global eth1                                 

  •     inet 10.0.1.250/24 scope global secondary eth1            #virtual ip
  •     inet6 fe80::a00:27ff:fecd:f3db/64 scope link  
  •        valid_lft forever preferred_lft forever
  • 4: sit0:  mtu 1480 qdisc noop  
  •     link/sit 0.0.0.0 brd 0.0.0.0

　　六 haproxy_check.sh 监控脚本


  • 脚本默认监控 eth1 网卡，所以你需要根据自己的需要调整，如果你haproxy virtual ip 在eth0 网卡，那么你需要做如下更改:
  • 6.1 使用"#"号注释掉 第14行 第16行，取消掉 第15行 第17行
  • #eth1_ip=$(/sbin/ifconfig eth1  | sed -n '/inet addr:/ s/inet addr://pg' | awk -F" " '{print $1}')
  • eth0_ip=$(/sbin/ifconfig eth0  | sed -n '/inet addr:/ s/inet addr://pg' | awk -F" " '{print $1}')
  • #eth1_url="http://${eth1_ip}:80"
  • eth0_url="http://${eth0_ip}:80"

  • 6.2 将第35行 $eth1_url,替换成 $eth0_url
  • if curl -IL -A "haproxy_check" "$eth1_url" > /dev/null 2>&1 ;then
  • 6.3 将第48行 $eth1_url，替换成 $eth0_url
  • if ! curl -IL -A "haproxy_check" "$eth1_url" > /dev/null 2>&1 ;then


　　七 其它


  • 5.2 模拟网络故障 haproxy 切换测试 与 5.3 模拟haproxy 80端口故障 keepalived 切换测试  
  • 五 haproxy 与 keepalived 整体测试，反过来测试也是没有问题的，不过要保证 haproxy.sh keepalived.sh 这两个服务脚本已经启动了，可以使用pstree 查看
  • 如：[root@four cluster]# pstree
  • init─┬─crond
  •      ├─events/0
  •      ├─events/1
  •      ├─haproxy
  •      ├─keepalived───keepalived
  •      ├─keepalived.sh───haproxy_check.s───sleep





运维网声明 1、欢迎大家加入本站运维交流群:群②:261659950 群⑤:202807635 群⑦870801961 群⑧679858003
2、本站所有主题由该帖子作者发表，该帖子作者与运维网享有帖子相关版权
3、所有作品的著作权均归原作者享有，请您和我们一样尊重他人的著作权等合法权益。如果您对作品感到满意，请购买正版
4、禁止制作、复制、发布和传播具有反动、淫秽、色情、暴力、凶杀等内容的信息，一经发现立即删除。若您因此触犯法律，一切后果自负，我们对此不承担任何责任
5、所有资源均系网友上传或者通过网络收集，我们仅提供一个展示、介绍、观摩学习的平台，我们不对其内容的准确性、可靠性、正当性、安全性、合法性等负责，亦不承担任何法律责任
6、所有作品仅供您个人学习、研究或欣赏，不得用于商业或者其他用途，否则，一切后果均由您自己承担，我们对此不承担任何法律责任
7、如涉及侵犯版权等问题，请您及时通知我们，我们将立即采取措施予以解决
8、联系人Email：admin@iyunv.com 网址：www.yunweiku.com

所有资源均系网友上传或者通过网络收集,我们仅提供一个展示、介绍、观摩学习的平台,我们不对其承担任何法律责任,如涉及侵犯版权等问题,请您及时通知我们,我们将立即处理,联系人Email:kefu@iyunv.com,QQ:1061981298 本贴地址：https://www.yunweiku.com/thread-657187-1-1.html 上篇帖子： moosefs+keepalived双机高可用热备解决方案 下篇帖子： MySQL共享存储主备模式利用Keepalived实现双机高可用
回复

使用道具 举报

返回列表
高级模式
B Color Image Link Quote Code Smilies
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

扫码加入运维网微信交流群X

扫码加入运维网微信交流群

扫描二维码加入运维网微信交流群，最新一手资源尽在官方微信交流群！快快加入我们吧...

扫描微信二维码查看详情

客服 E-mail:iyunvcom@gmail.com

本站由青云提供云计算服务

运维网--中国最专业的运维工程师交流社区

豫ICP备20007574号-1 Copyright © 2012-2025

关于我们

发展历程

联系我们

关注我们

广告服务

广告案例

隐私条款

免责声明

用户登录

用户注册

版主守则

申请版主

手机版

金币获取

帮助中心

每日签到

Good Good

Study

Day Day

UP


客服E-mail:kefu@iyunv.com 客服QQ：1061981298


QQ群⑦：运维网交流群⑦ QQ群⑧：运维网交流群⑧ k8s群：运维网kubernetes交流群


提醒：禁止发布任何违反国家法律、法规的言论与图片等内容；本站内容均来自个人观点与网络等信息，非本站认同之观点.


本站大部分资源是网友从网上搜集分享而来，其版权均归原作者及其网站所有，我们尊重他人的合法权益，如有内容侵犯您的合法权益，请及时与我们联系进行核实删除！



合作伙伴: 青云cloud

快速回复 返回顶部 返回列表