Keepalive+redis主从复制实现redis高可用
一、背景描述
项目上使用的软件是使用的单机版本redis，客户质疑为什么现在还用单机呢？故障了怎么办，要求做高可用。
redis有三种方式
①、主从复制
优点：数据高可用。
缺点：主节点故障后无法自动转移到从节点。
②、哨兵模式
优点：主节点故障可以不影响业务使用，做到了高可用。
缺点：程序需要连接哨兵IP+端口，意味着现在的业务代码需要做一些修改（不知道准确不）。
③、集群
优点：三主三从可以扩展分片数。
缺点：①、是否程序需要连接所有集群的IP+端口？如果是那代码就要重新发版才能用。②、redis集群后是以分槽位的方式将数据分散在多个master节点。因为项目是微服务并不确定使用redis 0数据库后会不会有key 冲突的问题！再者有部分命令也不支持。
主要是哨兵模式、集群用的资源也较多。
所以这里考虑用两台机器做主从，通过Keepalive+redis主从复制实现redis高可用。
关于主从复制、哨兵、集群可以参考：/hanease/p/15916605.html
二、部署redis
注、Master、Slave安装部署相同。
①、资源规划
172.27.3.62 redis Master
172.27.3.63 redis Slave
172.27.3.66 keepalive VIP
②、安装redis
这里通过yum的方式部署
[root@zabbix-proxy opt]# yum -y install redis[root@zabbix-proxy opt]# redis-server --versionRedis server v=3.2.12 sha=00000000:0 malloc=jemalloc-3.6.0 bits=64 build=7897e7d0e13773f
③、修改配置文件
[root@zabbix-proxy opt]# cat /etc/redis.conf |grep -v '#'|grep -v '^$'bind 0.0.0.0protected-mode yesport 6379tcp-backlog 511timeout 0tcp-keepalive 300daemonize nosupervised nopidfile /var/run/redis_6379.pidloglevel noticelogfile /var/log/redis/redis.logdatabases 96save 900 1save 300 10save 60 10000stop-writes-on-bgsave-error yesrdbcompression yesrdbchecksum yesdbfilename dump.rdbdir /var/lib/redismasterauth lskjdlfkjaslfslave-serve-stale-data yesslave-read-only yesrepl-diskless-sync norepl-diskless-sync-delay 5repl-disable-tcp-nodelay noslave-priority 100requirepass lskjdlfkjaslfappendonly noappendfilename "appendonly.aof"appendfsync everysecno-appendfsync-on-rewrite noauto-aof-rewrite-percentage 100auto-aof-rewrite-min-size 64mbaof-load-truncated yeslua-time-limit 5000slowlog-log-slower-than 10000slowlog-max-len 128latency-monitor-threshold 0notify-keyspace-events ""hash-max-ziplist-entries 512hash-max-ziplist-value 64list-max-ziplist-size -2list-compress-depth 0set-max-intset-entries 512zset-max-ziplist-entries 128zset-max-ziplist-value 64hll-sparse-max-bytes 3000activerehashing yesclient-output-buffer-limit normal 0 0 0client-output-buffer-limit slave 256mb 64mb 60client-output-buffer-limit pubsub 32mb 8mb 60hz 10aof-rewrite-incremental-fsync yes
最主要的配置
bind 0.0.0.0requirepass lskjdlfkjaslf #密码验证masterauth lskjdlfkjaslf #master 、slave都需要配置主从验证密码，这个密码就是对方的requirepass
④、配置redis主从
配置redis主从有两种方式：
1、修改配置文件（永久生效）
在redis从修改配置文件添加以下配置即可。
slaveof 172.27.3.62 6379
2、命令行修改（重启后失效）。
命令行执行：
slaveof 172.27.3.62 6379
我们选择第二种方式。
以上redis安装部署完成后设置redis开机启动
systemctl enable redis && systemctl start redis
三、安装部署keepalived
注：这里keepalived部署为不抢占模式，记得关闭防火墙、selinux设置为disabled！
①、安装keepalived（主从）
yum -y install keepalivedmkdir /app/log/keepalived #创建keepalive脚本日志目录touch /app/log/keepalived/status #生产eepalive脚本日志文件
②、redis Master配置keepalived
[root@localhost opt]# cat /etc/keepalived/keepalived.conf! Configuration File for keepalivedglobal_defs {router_id redis01}vrrp_script chk_redis{script "/etc/keepalived/script/redis_check.sh" ##判断进程是否存在interval 2timeout 2fall 3}vrrp_instance redis {state BACKUPinterface ens192 ##需要修改为实际网卡名称nopreemptvirtual_router_id 60priority 100##权重，数字越大权重越大，主节点数值大于备节点advert_int 1authentication { #all node must sameauth_type PASSauth_pass gkogi38GIOWE8398jd}virtual_ipaddress {172.27.3.66##需要修改为实际分配的虚拟IP}track_script {chk_redis}notify_master "/etc/keepalived/script/redis_master.sh 127.0.0.1 172.27.3.63 6379 Mcloud"## 172.27.3.63 6379需要修改为实际的远端主机IP ## 修改auth为密码notify_backup "/etc/keepalived/script/redis_backup.sh 127.0.0.1 172.27.3.63 6379 Mcloud"## 172.27.3.63 6379需要修改为实际的远端主机IP ## 修改auth为密码}
③、redis Master检测脚本 redis_check.sh （主从相同）
keepalived会定时check redis进程是否存在，如果掉了会尝试拉服务起来，拉不起来就会关闭keepalived服务。
[root@zabbix-proxy script]# pwd/etc/keepalived/script[root@zabbix-proxy script]# cat redis_check.sh#!/bin/bash #ALIVE=`/usr/bin/redis-cli -h $1 -p $2 -a $3 PING` LOGFILE="/app/log/keepalived/status"echo "[CHECK]" >> $LOGFILEdate >> $LOGFILEd=`date --date today +%Y%m%d_%H:%M:%S`n=`netstat -lntp | grep redis | wc -l`if [ $n -eq 0 ]; thensystemctl start redisecho "Redis is Down,but I start it! Please check it" >> $LOGFILE 2>&1n2=`netstat -lntp | grep redis | wc -l`if [ $n2 -eq 0 ]; thenecho "$d redis down,keepalived will stop" >> $LOGFILE 2>&1systemctl stop keepalivedfielseecho "Success: redis-cli -h $1 -p $2 -a $3 PING $ALIVE" >> $LOGFILE 2>&1# exit 1fi
④、redis Master当keepalived状态为master是执行脚本redis_master.sh（主从相同）
本脚本不做过多的处理，当本机keepalived状态为master时清空redis从配置！
[root@zabbix-proxy script]# cat redis_master.sh#!/bin/bashREDISCLI="/usr/bin/redis-cli -h $1 -p $3 -a $4"LOGFILE="/app/log/keepalived/status"echo "[master]" >> $LOGFILEdate >> $LOGFILEecho "Being master...." >> $LOGFILEecho "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE${REDISCLI} SLAVEOF NO ONE >> $LOGFILE
⑤、redis Master当keepalived状态为backup是执行脚本redis_backup.sh（主从相同）
[root@zabbix-proxy script]# cat redis_backup.sh #!/bin/bashREDISCLI="/usr/bin/redis-cli -h $1 -p $3 -a $4"LOGFILE="/app/log/keepalived/status"echo "[backup]" >> $LOGFILEdate >> $LOGFILEecho "Run SLAVEOF cmd ..." >> $LOGFILE$REDISCLI SLAVEOF $2 $3 >> $LOGFILE 2>&1##判断偏移量，也可以使用下面注释掉的sleep进行同步数据的等待slave=$(/usr/bin/redis-cli -h $2 -a $4 info | grep slave0 | awk -F "=" '{print $5}' | awk -F "," '{print $1}') master=$(/usr/bin/redis-cli -h $2 -a $4 info | grep master_repl_offset | awk -F ":" '{print $2}' | awk -F "\r" '{print $1}' )while [ "$slave" != "$master" ];doslave=$(/usr/bin/redis-cli -h $2 -a $4 info | grep slave0 | awk -F "=" '{print $5}' | awk -F "," '{print $1}')master=$(/usr/bin/redis-cli -h $2 -a $4 info | grep master_repl_offset | awk -F ":" '{print $2}' | awk -F "\r" '{print $1}' )echo $slaveecho $masterdone #sleep 15#sleep 15 #delay 15 s wait data sync exchange role
⑥、redis Slave配置keepalived
[root@zabbix-proxy script]# cat /etc/keepalived/keepalived.conf! Configuration File for keepalivedglobal_defs {router_id redis02}vrrp_script chk_redis{script "/etc/keepalived/script/redis_check.sh" ##修改auth为密码interval 2timeout 2fall 3}vrrp_instance redis {state BACKUPinterface ens192 ##需要修改为实际网卡名称virtual_router_id 60nopreemptpriority 90advert_int 1authentication { #all node must sameauth_type PASSauth_pass gkogi38GIOWE8398jd}virtual_ipaddress {172.27.3.66##需要修改为实际分配的虚拟IP}track_script {chk_redis}notify_master "/etc/keepalived/script/redis_master.sh 127.0.0.1 172.27.3.62 6379 Mcloud"## 172.27.3.63 6379需要修改为实际的远端主机IP ## 修改auth为密码notify_backup "/etc/keepalived/script/redis_backup.sh 127.0.0.1 172.27.3.62 6379 Mcloud"## 172.27.3.63 6379需要修改为实际的远端主机IP ## 修改auth为密码}
注：两外三个脚本内容与master服务器相同。
四、测试redis故障切换
①、查看现在redis服务状态及角色。
1.1、3.62
[root@localhost opt]# /usr/bin/redis-cli -h 172.27.3.62 -a lskjdlfkjaslf172.27.3.62:6379> INFO Replication# Replicationrole:slavemaster_host:172.27.3.63master_port:6379master_link_status:upmaster_last_io_seconds_ago:7master_sync_in_progress:0slave_repl_offset:8695slave_priority:100slave_read_only:1connected_slaves:0master_repl_offset:0repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0172.27.3.62:6379>
1.2、3.63
[root@zabbix-proxy ~]# /usr/bin/redis-cli -h 172.27.3.63 -a lskjdlfkjaslf172.27.3.63:6379> info replication# Replicationrole:slavemaster_host:172.27.3.62master_port:6379master_link_status:upmaster_last_io_seconds_ago:10master_sync_in_progress:0slave_repl_offset:71slave_priority:100slave_read_only:1connected_slaves:0master_repl_offset:0repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0172.27.3.63:6379>
②、重启3.62（因为改服务器状态目前时Master）
查看3.63 redis状态
172.27.3.63:6379> info replication# Replicationrole:masterconnected_slaves:1slave0:ip=172.27.3.62,port=6379,state=online,offset=1,lag=0master_repl_offset:1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:2repl_backlog_histlen:0172.27.3.63:6379>
再查看3.62状态
172.27.3.62:6379> info replication# Replicationrole:slavemaster_host:172.27.3.63master_port:6379master_link_status:upmaster_last_io_seconds_ago:9master_sync_in_progress:0slave_repl_offset:43slave_priority:100slave_read_only:1connected_slaves:0master_repl_offset:0repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0
重启Master时Slave状态变为Master！
Master重启后keepalive状态为BACKUP！
数据验证自行验证1
