Keepalived
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 { #注意大括号空格
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:
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
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
}
HAProxy
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" #该账户密码需要在mysql里添加
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_/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 "<html><body>MySQL is running.</body></html>\r\n"
echo -e "\r\n"
rm $ERR_FILE $TMP_FILE
exit 0
}
return_fail()
{
echo -e "HTTP/1.1 503 Service Unavailable\r\n"
echo -e "Content-Type: text/html\r\n"
echo -e "Content-Length: 42\r\n"
echo -e "\r\n"
echo -e "<html><body>MySQL is *down*.</body></html>\r\n"
sed -e 's/\n$/\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: <LOOPBACK,UP,LOWER_UP> mtu 136 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: <BROADCAST,MULTICAST> 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: <BROADCAST,MULTICAST,UP,LOWER_UP> 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: <BROADCAST,MULTICAST> 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: <BROADCAST,MULTICAST> 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 307/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