由于在生产环境使用了mysqlcluster,需要实现高可用负载均衡,这里提供了keepalived+haproxy来实现.

      keepalived主要功能是实现真实机器的故障隔离及负载均衡器间的失败切换.可在第3,4,5层交换.它通过VRRPv2(Virtual Router Redundancy Protocol) stack实现的.

      Layer3:Keepalived会定期向服务器群中的服务器.发送一个ICMP的数据包（既我们平时用的Ping程序）,如果发现某台服务的IP地址没有激活，Keepalived便报告这台服务器失效，并将它从服务器群中剔除，这种情况的典型例子是某台服务器被非法关机。Layer3的方式是以服务器的IP地址是否有效作为服务器工作正常与否的标准。

     Layer4:主要以TCP端口的状态来决定服务器工作正常与否。如web server的服务端口一般是80，如果Keepalived检测到80端口没有启动，则Keepalived将把这台服务器从服务器群中剔除。

     Layer5:在网络上占用的带宽也要大一些。Keepalived将根据用户的设定检查服务器程序的运行是否正常，如果与用户的设定不相符，则Keepalived将把服务器从服务器群中剔除。

Software Design

keepalived启动后会有单个进程

 8352 ?        Ss     0:00 /usr/sbin/keepalived   8353 ?        S      0:00  \_ /usr/sbin/keepalived   8356 ?        S      0:01  \_ /usr/sbin/keepalived

父进程:内存管理,子进程管理等等

子进程:VRRP子进程

子进程:Healthchecking 子进程

实例

2台mysqlcluster 10.1.6.203 master  10.1.6.205 backup

vip 10.1.6.173 

目的访问10.1.6.173 3366端口 分别轮询通过haproxy转发到10.1.6.203 3306 和10.1.6.205 3306

mysqlcluster搭建参照之前博客,这里在2台机上安装keepalived

root@10.1.6.203:~# apt-get install keepalived  root@10.1.6.203:~# cat /etc/keepalived/keepalived.conf   vrrp_script chk_haproxy {          script "killall -0 haproxy"   # verify the pid existance          interval 2                    # check every 2 seconds          weight -2                    # add 2 points of prio if OK  }    vrrp_instance VI_1 {          interface eth1                # interface to monitor          state MASTER                            virtual_router_id 51          # Assign one ID for this route          priority 101                  # 101 on master, 100 on backup          nopreempt          debug            virtual_ipaddress {                  10.1.6.173          }            track_script {                  chk_haproxy          }           notify_master /etc/keepalived/scripts/start_haproxy.sh  #表示当切换到master状态时,要执行的脚本         notify_fault  /etc/keepalived/scripts/stop_keepalived.sh #故障时执行的脚本         notify_stop   /etc/keepalived/scripts/stop_haproxy.sh # keepalived停止运行前运行notify_stop指定的脚本 }

VRRPD配置包括三个类:

VRRP同步组(synchroization group)

VRRP实例(VRRP Instance)

VRRP脚本

这里使用了 VRRP实例, VRRP脚本

注意配置选项: 

stat:指定instance(Initial)的初始状态，就是说在配置好后，这台服务器的初始状态就是这里指定的，但这里指定的不算，还是得要通过竞选通过优先级来确定，里如果这里设置为master，但如若他的优先级不及另外一台，那么这台在发送通告时，会发送自己的优先级，另外一台发现优先级不如自己的高，那么他会就回抢占为master

interface：实例绑定的网卡，因为在配置虚拟IP的时候必须是在已有的网卡上添加的

priority 101：设置本节点的优先级，优先级高的为master

debug：debug级别

nopreempt:设置为不抢占

vrrp_script chk_haproxy {
        script "killall -0 haproxy"   # verify the pid existance
        interval 2                    # check every 2 seconds 脚本执行间隔
        weight -2                    # add 2 points of prio if OK 脚本结果导致的优先级变更：2表示优先级+2；-2则表示优先级-2
}

然后在实例(vrrp_instance)里面引用，有点类似脚本里面的函数引用一样：先定义，后引用函数名
        track_script {
                chk_haproxy
        }
注意：VRRP脚本(vrrp_script)和VRRP实例(vrrp_instance)属于同一个级别

root@10.1.6.203:scripts# cat start_haproxy.sh   #!/bin/bash    sleep 5  get=`ip addr  |grep 10.1.6.173 |wc -l`  echo $get >> /etc/keepalived/scripts/start_ha.log    if [ $get -eq 1 ]  then          echo "`date +%c` success to get vip" >> /etc/keepalived/scripts/start_ha.log          /usr/local/sbin/haproxy -f /etc/haproxy/haproxy.cfg  else          echo "`date +%c` can not get vip" >> /etc/keepalived/scripts/start_ha.log  fi  root@10.1.6.203:scripts# cat stop_keepalived.sh   #!/bin/bash    pid=`pidof keepalived`  if [ $pid == "" ]  then   echo "`date +%c` no keepalived process id"  >> /etc/keepalived/scripts/stop_keep.log  else   echo "`date +%c` will stop keepalived "  >> /etc/keepalived/scripts/stop_keep.log   /etc/init.d/keepalived  stop  fi    /etc/init.d/keepalived  stop    root@10.1.6.203:scripts# cat stop_haproxy.sh   #!/bin/bash    pid=`pidof haproxy`  echo "`date +%c` stop haproxy" >> /etc/keepalived/scripts/stop_ha.log  kill -9 $pid

同理配置10.1.6.205

root@10.1.6.205:~# cat /etc/keepalived/keepalived.conf   vrrp_script chk_haproxy {     script "killall -0 haproxy"   # verify the pid existance     interval 2                    # check every 2 seconds     weight 2                      # add 2 points of prio if OK  }     vrrp_instance VI_1 {     interface eth1               # interface to monitor     state BACKUP     virtual_router_id 51          # Assign one ID for this route     priority 100                 # 101 on master, 100 on backup     virtual_ipaddress {         10.1.6.173     }       track_script {         chk_haproxy     }    notify_master /etc/keepalived/scripts/start_haproxy.sh  notify_fault  /etc/keepalived/scripts/stop_keepalived.sh  notify_stop /etc/keepalived/scripts/stop_haproxy.sh    }

下面再介绍下haproxy

       HAProxy是一款基于TCP(第四层)和HTTP(第七层)应用的代理软件,它也可作为负载均衡器.可以支持数以万计的并发连接.同时可以保护服务器不暴露到网络上,通过端口映射.它还自带监控服务器状态的页面.

      安装haproxy

wget -O/tmp/haproxy-1.4.22.tar.gz http://haproxy.1wt.eu/download/1.4/src/haproxy-1.4.22.tar.gz  tar xvfz /tmp/haproxy-1.4.22.tar.gz -C /tmp/  cd /tmp/haproxy-1.4.22  make TARGET=linux26  make install

      haproxy需要对每一个mysqlcluster服务器进行健康检查

1.在2台主机分别配置haproxy.cfg

root@10.1.6.203:scripts# cat /etc/haproxy/haproxy.cfg   global          maxconn 51200  #默认最大连接数           #uid 99          #gid 99          daemon        #以后台形式运行haproxy          #quiet          nbproc 1      #进程数量(可以设置多个进程提高性能)           pidfile /etc/haproxy/haproxy.pid  #haproxy的pid存放路径,启动进程的用户必须有权限访问此文件     defaults          mode tcp            #所处理的类别 (#7层 http;4层tcp  )           option redispatch   #serverId对应的服务器挂掉后,强制定向到其他健康的服务器           option abortonclose #当服务器负载很高的时候，自动结束掉当前队列处理比较久的连接           timeout connect 5000s   #连接超时          timeout client 50000s  #客户端超时          timeout server 50000s   #服务器超时          log 127.0.0.1 local0   #错误日志记录          balance roundrobin    #默认的负载均衡的方式,轮询方式     listen proxy          bind 10.1.6.173:3366   #监听端口           mode tcp               #http的7层模式          option  httpchk        #心跳检测的文件          server db1 10.1.6.203:3306  weight 1 check port 9222 inter 12000 rise 3 fall 3      #服务器定义,check inter 12000是检测心跳频率 rise 3是3次正确认为服务器可用， fall 3是3次失败认为服务器不可用，weight代表权重           server db2 10.1.6.205:3306  weight 1 check port 9222 inter 12000 rise 3 fall 3    listen  haproxy_stats          mode http          bind 10.1.6.173:8888          option httplog          stats refresh 5s             stats uri /status #网站健康检测URL，用来检测HAProxy管理的网站是否可以用，正常返回200，不正常返回503           stats realm Haproxy Manager          stats auth admin:p@a1SZs24 #账号密码

root@10.1.6.205:~$ cat /etc/haproxy/haproxy.cfg   global          maxconn 51200          #uid 99          #gid 99          daemon          #quiet          nbproc 1          pidfile /etc/haproxy/haproxy.pid    defaults          mode tcp          option redispatch             option abortonclose          timeout connect 5000s          timeout client 50000s          timeout server 50000s          log 127.0.0.1 local0          balance roundrobin     listen proxy          bind 10.1.6.173:3366          mode tcp          option  httpchk          server db1 10.1.6.203:3306  weight 1 check port 9222 inter 12000 rise 3 fall 3          server db2 10.1.6.205:3306  weight 1 check port 9222 inter 12000 rise 3 fall 3    listen  haproxy_stats          mode http          bind 10.1.6.173:8888          option httplog          stats refresh 5s             stats uri /status            stats realm Haproxy Manager          stats auth admin:p@a1SZs24

2.安装xinetd

root@10.1.6.203:~# apt-get install xinetd

3.在每个节点添加xinetd服务脚本和mysqlchk端口号

root@10.1.6.203:~# vim /etc/xinetd.d/mysqlchk   # default: on  # description: mysqlchk  service mysqlchk                  #需要在servive定义  {          flags           = REUSE          socket_type     = stream          port            = 9222          wait            = no          user            = nobody          server          = /opt/mysqlchk            log_on_failure  += USERID          disable         = no          per_source      = UNLIMITED          bind            = 10.1.6.173  }    root@10.1.6.203:~# vim /etc/services   mysqlchk        9222/tcp                        # mysqlchk

4.编写mysqlchk监控服务脚本

root@10.1.6.203:~# ls -l /opt/mysqlchk   -rwxr--r-- 1 nobody root 1994 2013-09-17 11:27 /opt/mysqlchk  root@10.1.6.203:~# cat /opt/mysqlchk   #!/bin/bash  #  # This script checks if a mysql server is healthy running on localhost. It will  # return:  # "HTTP/1.x 200 OK\r" (if mysql is running smoothly)  # - OR -  # "HTTP/1.x 500 Internal Server Error\r" (else)  #  # The purpose of this script is make haproxy capable of monitoring mysql properly  #    MYSQL_HOST="localhost"  MYSQL_SOCKET="/var/run/mysqld/mysqld.sock"   MYSQL_USERNAME="mysqlchkusr"  MYSQL_PASSWORD="secret"  MYSQL_OPTS="-N -q -A"  TMP_FILE="/dev/shm/mysqlchk.$$.out"  ERR_FILE="/dev/shm/mysqlchk.$$.err"  FORCE_FAIL="/dev/shm/proxyoff"  MYSQL_BIN="/opt/mysqlcluster/mysql-cluster-gpl-7.2.6-linux2.6-x86_64/bin/mysql"  CHECK_QUERY="select 1"    preflight_check()  {      for I in "$TMP_FILE" "$ERR_FILE"; do          if [ -f "$I" ]; then              if [ ! -w $I ]; then                  echo -e "HTTP/1.1 503 Service Unavailable\r\n"                  echo -e "Content-Type: Content-Type: text/plain\r\n"                  echo -e "\r\n"                  echo -e "Cannot write to $I\r\n"                  echo -e "\r\n"                  exit 1              fi          fi      done  }    return_ok()  {      echo -e "HTTP/1.1 200 OK\r\n"      echo -e "Content-Type: text/html\r\n"      echo -e "Content-Length: 43\r\n"      echo -e "\r\n"      echo -e "MySQL is running.\r\n"      echo -e "\r\n"      rm $ERR_FILE $TMP_FILE      exit 0  }  return_fail()  {      echo -e "HTTP/1.1 503 Service Unavailable\r\n"      echo -e "Content-Type: text/html\r\n"      echo -e "Content-Length: 42\r\n"      echo -e "\r\n"      echo -e "MySQL is *down*.\r\n"      sed -e 's/\n$/\r\n/' $ERR_FILE      echo -e "\r\n"      rm $ERR_FILE $TMP_FILE      exit 1  }  preflight_check  if [ -f "$FORCE_FAIL" ]; then          echo "$FORCE_FAIL found" > $ERR_FILE          return_fail;  fi  $MYSQL_BIN $MYSQL_OPTS --host=$MYSQL_HOST --socket=$MYSQL_SOCKET --user=$MYSQL_USERNAME --password=$MYSQL_PASSWORD -e "$CHECK_QUERY" > $TMP_FILE 2> $ERR_FILE  if [ $? -ne 0 ]; then          return_fail;  fi  return_ok;

测试

2个节点开启keepalived(主节点会获得vip,自动拉起haproxy),xinetd

root@10.1.6.203:~# ip add  1: lo:  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  2: eth0:    mtu 1500 qdisc pfifo_fast state DOWN qlen 1000      link/ether 00:26:b9:36:0f:81 brd ff:ff:ff:ff:ff:ff      inet 211.151.105.186/26 brd 211.151.105.191 scope global eth0  3: eth1:    mtu 1500 qdisc pfifo_fast state UP qlen 1000      link/ether 00:26:b9:36:0f:83 brd ff:ff:ff:ff:ff:ff      inet 10.1.6.203/24 brd 10.1.6.255 scope global eth1      inet 10.1.6.173/32 scope global eth1  4: eth2:    mtu 1500 qdisc noop state DOWN qlen 1000      link/ether 00:26:b9:36:0f:85 brd ff:ff:ff:ff:ff:ff  5: eth3:    mtu 1500 qdisc noop state DOWN qlen 1000      link/ether 00:26:b9:36:0f:87 brd ff:ff:ff:ff:ff:ff  root@10.1.6.203:~# netstat -tunlp | grep ha  tcp        0      0 10.1.6.173:3366     0.0.0.0:*               LISTEN      1042/haproxy      tcp        0      0 10.1.6.203:8888     0.0.0.0:*               LISTEN      1042/haproxy      udp        0      0 0.0.0.0:56562           0.0.0.0:*                           1042/haproxy      root@10.1.6.203:~# netstat  -tunlp | grep xine  tcp        0      0 10.1.6.203:9222     0.0.0.0:*               LISTEN      30897/xinetd      root@10.1.6.203:~# ps -ef | grep haproxy  root      1042     1  0 Sep17 ?        00:00:00 /usr/local/sbin/haproxy -f /etc/haproxy/haproxy.cfg

测试:

通过vip10.1.6.173 3366访问cluster数据库(注意账户dave权限需要加3个ip10.1.6.203,10.1.6.205,10.1.6.173)

root@10.1.6.203:mgm# mysql -udave -p -h 10.1.6.173 -P 3366  Enter password:   Welcome to the MySQL monitor.  Commands end with ; or \g.  Your MySQL connection id is 1344316  Server version: 5.5.22-ndb-7.2.6-gpl-log MySQL Cluster Community Server (GPL)    Type 'help;' or '\h' for help. Type '\c' to clear the buffer.    mysql> show databases;  +--------------------+  | Database           |  +--------------------+  | information_schema |   | dave              |   | test               |   +--------------------+  3 rows in set (0.01 sec)    mysql> 

手动分别使keepalive,haproxy,数据库挂掉.vip10.1.6.173会自动漂到10.1.6.205从上，并不影响vip的访问

通过vip,haproxy查看各节点状态

http://10.1.6.173:8888/status

参考:

http://www.keepalived.org/

http://haproxy.1wt.eu/