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: <LOOPBACK,UP,LOWER_UP> 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: <BROADCAST,MULTICAST,UP,LOWER_UP> 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: <BROADCAST,MULTICAST,UP,LOWER_UP> 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: <NOARP> 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: <LOOPBACK,UP,LOWER_UP> 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: <BROADCAST,MULTICAST,UP,LOWER_UP> 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: <BROADCAST,MULTICAST,UP,LOWER_UP> 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: <NOARP> 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
本文转自 dongnan 51CTO博客,原文链接:
http://blog.51cto.com/dngood/753739