结合keepalived实现redis群集高可用故障自动切换

转载时请务必以超链接形式标明文章  原始出处 、作者信息和本声明。否则将追究法律责任。 http://birdinroom.blog.51cto.com/7740375/1401663

   系统架构图:

我们所要实现的目的很简单,

104,107 为keepalive和redis主从架构，其余服务器比如105、106均为redis从库并且挂在vip 192.168.56.180下面。

主keepalive负责主要日常工作，从keepalive担任备机角色，一旦主keepalive挂掉，从keepalive服务器立即使从redis转变角色切换成master状态开始接管任务提供服务，实现业务的无缝切换,当挂掉的服务器修好上线后继续担任主的角色，从库会自动切换到slave状态并且不影响挂载在vip下的从redis的数据同步，一满足高并发架构的需求。

keepalive在ubuntu的安装很简单

1 2 3 4

apt-get  install  libssl-dev apt-get  install  openssl apt-get  install  libpopt-dev apt-get  install  keepalived

redis的安装也很简单:

网上很多方法，也可参考我的另一篇文章: redis的shell安装脚本,实现在linux下本机主从架构

网上有一篇郭冬的一篇文章给了我很大启发故拿来参考:通过Keepalived实现Redis Failover自动故障切换，

下面我们看192.168.56.104主keepalived的配置

/etc/keepalived/keepalived.conf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

global_defs {          notification_email {                  409011500@qq.com          }          notification_email_from  409011500@qq.com                  smtp_server 127.0.0.1  （如果本机配置的话）                  smtp_connect_timeout 30                  router_id redis-ha } vrrp_script chk_redis {                  script  "/home/lhb/sh/redis_check.sh"    ###监控脚本                  interval 2                                         ###监控时间 } vrrp_instance VI_1 {          state MASTER                             ###设置为MASTER          interface eth0                           ###监控网卡          virtual_router_id 52          priority 101                             ###权重值          authentication {                       auth_type PASS              ###加密                       auth_pass redis             ###密码          }          track_script {                  chk_redis                        ###执行上面定义的chk_redis          }          virtual_ipaddress {               192.168.56.180                         ###VIP          }          notify_master  /home/lhb/sh/redis_master .sh          notify_backup  /home/lhb/sh/redis_backup .sh }

/home/lhb/sh/redis_master.sh

1 2 3 4 5 6 7 8 9 10 11

#!/bin/bash REDISCLI= "/usr/local/redis/bin/redis-cli" LOGFILE= "/usr/local/redis/log/keepalived-redis-state.log" echo  "[master]"  >> $LOGFILE date  >> $LOGFILE echo  "Being master...."  >> $LOGFILE 2>&1 echo  "Run SLAVEOF cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF 192.168.56.107 6379 >> $LOGFILE  2>&1 sleep  10  #延迟10秒以后待数据同步完成后再取消同步状态 echo  "Run SLAVEOF NO ONE cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1

/home/lhb/sh/redis_backup.sh

1 2 3 4 5 6 7 8 9

#!/bin/bash REDISCLI= "/usr/local/redis/bin/redis-cli" LOGFILE= "/usr/local/redis/log/keepalived-redis-state.log" echo  "[backup]"  >> $LOGFILE date  >> $LOGFILE echo  "Being slave...."  >> $LOGFILE 2>&1 sleep  15  #延迟15秒待数据被对方同步完成之后再切换主从角色 echo  "Run SLAVEOF cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF 192.168.56.107 6379 >> $LOGFILE  2>&1

/usr/local/redis/etc/redis.conf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

daemonize  yes pidfile  /var/run/redis .pid port 6379 tcp-backlog 511 timeout 0 tcp-keepalive 0 loglevel notice logfile  "/usr/local/redis/log/redis.log" databases 16 save 900 1 save 300 10 save 60 10000 stop-writes-on-bgsave-error  yes rdbcompression  yes rdbchecksum  yes dbfilename dump.rdb dir  /usr/local/redis/data slave-serve-stale-data  yes slave- read -only no repl-disable-tcp-nodelay no slave-priority 100 appendonly  yes appendfilename  "appendonly.aof" appendfsync everysec no-appendfsync-on-rewrite no auto-aof-rewrite-percentage 100 auto-aof-rewrite-min-size 64mb lua- time -limit 5000 slowlog-log-slower-than 10000 slowlog-max-len 128 notify-keyspace-events  "" hash -max-ziplist-entries 512 hash -max-ziplist-value 64 list-max-ziplist-entries 512 list-max-ziplist-value 64 set -max-intset-entries 512 zset-max-ziplist-entries 128 zset-max-ziplist-value 64 activerehashing  yes client-output-buffer-limit normal 0 0 0 client-output-buffer-limit slave 256mb 64mb 60 client-output-buffer-limit pubsub 32mb 8mb 60 hz 10 aof-rewrite-incremental-fsync  yes

192.168.56.107从keepalived的配置

/etc/keepalived/keepalived.conf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

global_defs {          notification_email {                  409011500@qq.com          }          notification_email_from 409011500@qq.com                  smtp_server 127.0.0.1                  smtp_connect_timeout 30                  router_id redis-ha } vrrp_script chk_redis {                  script  "/home/lhb/sh/redis_check.sh"    ###监控脚本                  interval 2                                         ###监控时间 } vrrp_instance VI_1 {          state BACKUP                                 ###设置为BACKUP          interface eth0                               ###监控网卡          virtual_router_id 52          priority 100                                 ###比MASTRE权重值低          authentication {                       auth_type PASS                       auth_pass redis                 ###密码与MASTRE相同          }          track_script {                  chk_redis                        ###执行上面定义的chk_redis          }          virtual_ipaddress {               192.168.56.180                          ###VIP          }          notify_master  /home/lhb/sh/redis_master .sh          notify_backup  /home/lhb/sh/redis_backup .sh }

/home/lhb/sh/redis_master.sh

1 2 3 4 5 6 7 8 9 10 11

#!/bin/bash REDISCLI= "/usr/local/redis/bin/redis-cli" LOGFILE= "/usr/local/redis/log/keepalived-redis-state.log" echo  "[master]"  >> $LOGFILE date  >> $LOGFILE echo  "Being master...."  >> $LOGFILE 2>&1 echo  "Run SLAVEOF cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF 192.168.56.104 6379 >> $LOGFILE  2>&1 sleep  10  #延迟10秒以后待数据同步完成后再取消同步状态 echo  "Run SLAVEOF NO ONE cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1

/home/lhb/sh/redis_backup.sh

1 2 3 4 5 6 7 8 9

#!/bin/bash REDISCLI= "/usr/local/redis/bin/redis-cli" LOGFILE= "/usr/local/redis/log/keepalived-redis-state.log" echo  "[backup]"  >> $LOGFILE date  >> $LOGFILE echo  "Being slave...."  >> $LOGFILE 2>&1 sleep  15  #延迟15秒待数据被对方同步完成之后再切换主从角色 echo  "Run SLAVEOF cmd ..."  >> $LOGFILE $REDISCLI SLAVEOF 192.168.56.104 6379 >> $LOGFILE  2>&1

/home/lhb/sh/redis_check.sh

1 2 3 4 5 6 7 8 9

#!/bin/bash ALIVE=` /usr/local/redis/bin/redis-cli  PING` if  [  "$ALIVE"  ==  "PONG"  ];  then    echo  $ALIVE    exit  0 else    echo  $ALIVE    exit  1 fi

/usr/local/redis/etc/redis_slave.conf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

daemonize  yes pidfile  /var/run/redis_salve .pid port 6379 tcp-backlog 511 timeout 0 tcp-keepalive 0 loglevel notice logfile  "/usr/local/redis/log/redis_slave.log" databases 16 save 900 1 save 300 10 save 60 10000 stop-writes-on-bgsave-error  yes rdbcompression  yes rdbchecksum  yes dbfilename dump_salve.rdb dir  /usr/local/redis/data slave-serve-stale-data  yes slave- read -only no repl-disable-tcp-nodelay no slave-priority 100 appendonly  yes appendfilename  "appendonly.aof" appendfsync everysec no-appendfsync-on-rewrite no auto-aof-rewrite-percentage 100 auto-aof-rewrite-min-size 64mb lua- time -limit 5000 slowlog-log-slower-than 10000 slowlog-max-len 128 notify-keyspace-events  "" hash -max-ziplist-entries 512 hash -max-ziplist-value 64 list-max-ziplist-entries 512 list-max-ziplist-value 64 set -max-intset-entries 512 zset-max-ziplist-entries 128 zset-max-ziplist-value 64 activerehashing  yes client-output-buffer-limit normal 0 0 0 client-output-buffer-limit slave 256mb 64mb 60 client-output-buffer-limit pubsub 32mb 8mb 60 hz 10 aof-rewrite-incremental-fsync  yes SLAVEOF 192.168.56.104 6379

192.168.56.105、192.168.56.106 redis配置文件相同:

/usr/local/redis/etc/redis_salve.conf

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

daemonize  yes pidfile  /var/run/redis_salve .pid port 6379 tcp-backlog 511 timeout 0 tcp-keepalive 0 loglevel notice logfile  "/usr/local/redis/log/redis_slave.log" databases 16 save 900 1 save 300 10 save 60 10000 stop-writes-on-bgsave-error  yes rdbcompression  yes rdbchecksum  yes dbfilename dump_salve.rdb dir  /usr/local/redis/data slave-serve-stale-data  yes slave- read -only no repl-disable-tcp-nodelay no slave-priority 100 appendonly no appendfilename  "appendonly.aof" appendfsync everysec no-appendfsync-on-rewrite no auto-aof-rewrite-percentage 100 auto-aof-rewrite-min-size 64mb lua- time -limit 5000 slowlog-log-slower-than 10000 slowlog-max-len 128 notify-keyspace-events  "" hash -max-ziplist-entries 512 hash -max-ziplist-value 64 list-max-ziplist-entries 512 list-max-ziplist-value 64 set -max-intset-entries 512 zset-max-ziplist-entries 128 zset-max-ziplist-value 64 activerehashing  yes client-output-buffer-limit normal 0 0 0 client-output-buffer-limit slave 256mb 64mb 60 client-output-buffer-limit pubsub 32mb 8mb 60 hz 10 aof-rewrite-incremental-fsync  yes SLAVEOF 192.168.56.180 6379

好了，下面我们就来展示一下是否是我们期待的那样？

启动192.168.56.104，192.168.56.107上的redis,keepalvied，以及192.168.56.105,192.168.56.106上的redis

在192.168.56.104 我们看到以下结果，已经获取vip:192.168.56.180

在192.168.56.107 我们看到以下结果,并没有获得vip192.168.56.180

在192.168.56.105,192.168.56.106我们看到相同结果redis挂载在vip:192.168.56.180下

这时我们把192.168.56.104下的redis给关闭掉，看一下结果:vip已经释放

然后到192.168.56.107下看一下信息: 已经获取vip，并且redis已经从salve切换到master，并且从库105、106状态均为online

192.168.56.105、192.168.56.106 redis结果:一直挂载在vip下面并且link_status处于up状态

由此可见，当主redis挂掉后，备机上的redis立即切换为master,并且不影响业务正常运行。

我们然后把主上redis从新启动,看到下面结果: 主服务器获取vip

redis信息:我们可以看到192.168.56.107备机已经出现在master的slave列表中

然后切换到:192.168.56.107看一下信息: 已经释放vip，并且redis已经从master切换到slave状态，并且指向的master是192.168.56.104

在看192.168.56.105、192.168.56.106 redis信息,亦然指向192.168.56.180

看到这里，已经满足我们系统架构最初的所有设想功能。有感兴趣的同学可以一起交流。谢谢

本文出自 “宅鸟乐园” 博客，请务必保留此出处http://birdinroom.blog.51cto.com/7740375/1401663