生产环境实践：redis高可用架设

Posted 2020-10-31

    经过测试，当主redis停掉的时候，可以5秒钟内切换到从redis，反之切换到主，也是一样的时间。

    在A服务器(192.168.16.29),B服务器(192.168.16.2)上均安装redis,keepalived。

    A作为默认的master,B作为slave(在redis的配置文件中加上 SLAVEOF 192.168.16.29 6379)即可。

    A,B上的Redis均开启本地化策略，appendonly yes。

Master 192.168.16.29

Slave 192.168.16.2

Vip 192.168.16.3

Master redis 端口：6379

Slave redis 端口：6379

安装程序包：

#yum -y install gcc gcc+ gcc-c++ openssl openssl-devel pcre pcre-devel

安装keepalived

Master 192.168.16.29部署

keepalived安装部署就不做详细说明了，在架构那块会做详细说明。

#wget http://www.keepalived.org/software/keepalived-1.2.23.tar.gz
#tar xf keepalived-1.2.23.tar.gz
#cd keepalived-1.2.23
#./configure --prefix=/usr/local/keepalived
#make && make install
#cp keepalived/etc/init.d/keepalived.sysconfig /etc/sysconfig/keepalived
#cp keepalived/etc/init.d/keepalived.rh.init /etc/init.d/keepalived
#chmod a+x /etc/init.d/keepalived
#ln -s /usr/local/keepalived/sbin/keepalived /sbin/keepalived
#chkconfig --add keepalived
#cp -a keepalived/etc/keepalived/ /etc/
#mkdir /usr/local/Script/ -p
#mkdir /usr/local/logs/ -p

keepalived配置

#vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
 
 
global_defs {
notification_email {
[email protected]
}
notification_email_from [email protected]
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id redis1.xxx.com
}
 
vrrp_script chk_redis {
script "/usr/local/Script/redis_check.sh"
interval 2
}
 
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 112
priority 101
advert_int 2
authentication {
auth_type PASS
auth_pass 1111
}
track_script {
chk_redis
}
track_interface {
        eth0
}
virtual_ipaddress {
192.168.16.3 dev eth0
}
notify_master /usr/local/Script/redis_master.sh
notify_backup /usr/local/Script/redis_backup.sh
notify_fault /usr/local/Script/redis_fault.sh
notify_stop /usr/local/Script/redis_stop.sh
}

编写以下负责运作的关键脚本：

notify_master /usr/local/Script/redis_master.sh
notify_backup /usr/local/Script/redis_backup.sh
notify_fault /usr/local/Script/redis_fault.sh
notify_stop /usr/local/Script/redis_stop.sh

因为keepalived在转换状态时会依照状态来呼叫：

当进入Master状态时会呼叫notify_master

当Backup状态时会呼叫notify_backup

当发现异常情况时进入Fault状态呼叫notify_fault

当keepalived程序终止时呼叫notify_stop。

Slave 192.168.16.2 keepalived部署

从节点安装方法一样

#wget http://www.keepalived.org/software/keepalived-1.2.23.tar.gz
#tar xf keepalived-1.2.23.tar.gz
#cd keepalived-1.2.23
#./configure --prefix=/usr/local/keepalived
#make && make install
#cp keepalived/etc/init.d/keepalived.sysconfig /etc/sysconfig/keepalived
#cp keepalived/etc/init.d/keepalived.rh.init /etc/init.d/keepalived
#chmod a+x /etc/init.d/keepalived
#ln -s /usr/local/keepalived/sbin/keepalived /sbin/keepalived
#chkconfig --add keepalived
#cp -a keepalived/etc/keepalived/ /etc/
#mkdir /usr/local/Script/ -p 
#mkdir /usr/local/logs/ -p
 
 
keepalived配置
#vim /etc/keepalived/keepalived.conf 
! Configuration File for keepalived
 
 
global_defs {
notification_email {
[email protected]
}
notification_email_from [email protected]
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id redis2.xxx.com
}
 
vrrp_script chk_redis {
script "/usr/local/Script/redis_check.sh"
interval 2
}
 
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 112
priority 99
authentication {
auth_type PASS
auth_pass 1111
}
track_script {
chk_redis       ##执行上面定义的chk_redis
}
track_interface {
        eth0
}
virtual_ipaddress {
192.168.16.3 dev eth0
}
notify_master /usr/local/Script/redis_master.sh
notify_backup /usr/local/Script/redis_backup.sh
notify_fault /usr/local/Script/redis_fault.sh
notify_stop /usr/local/Script/redis_stop.sh
}

Master 192.168.16.29 脚本编写

##成为master

#vim  /usr/local/Script/redis_master.sh
#!/bin/bash
REDISCLI="/usr/local/redis/src/redis-cli"
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[master]" >> $LOGFILE
date >> $LOGFILE
echo "Being master...." >> $LOGFILE 2>&1
echo "Run SLAVEOF cmd ..." >> $LOGFILE
$REDISCLI -a 123456 SLAVEOF 192.168.16.2 6379 >> $LOGFILE 2>&1
sleep 5    ##延迟5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 -p 6379 -a 123456 SLAVEOF NO ONE
$REDISCLI -p 6379 -a 123456 SLAVEOF NO ONE >> $LOGFILE 2>&1

#成为backup

#vim  /usr/local/Script/redis_backup.sh
#!/bin/bash
REDISCLI="/usr/local/redis/src/redis-cli"
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[backup]" >> $LOGFILE
date >> $LOGFILE
echo "Being slave...." >> $LOGFILE 2>&1
sleep 5   #延迟5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 -a 123456 SLAVEOF 192.168.16.2 6379 >> $LOGFILE 2>&1
 
#监控脚本 exit 1 时
 
#vim  /usr/local/Script/redis_fault.sh
#!/bin/bash
 
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[fault]" >> $LOGFILE
date >> $LOGFILE
sh /usr/local/Script/redis_backup.sh
 
#keepalived 服务停止时
 
#vim  /usr/local/Script/redis_stop.sh
#!/bin/bash
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[stop]" >> $LOGFILE
date >> $LOGFILE
 
#监控脚本的路径
 
#vim /usr/local/Script/redis_check.sh
#!/bin/bash
ALIVE=`/usr/local/redis/src/redis-cli -p 6379 -a 123456 PING`
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
if [ "$ALIVE" == "PONG" ]; then
echo $ALIVE
#echo "check master pong" >> $LOGFILE
exit 0
else
echo $ALIVE
exit 1
fi

 

给脚本授予权限

chmod a+x /usr/local/Script/*

Slave 192.168.16.2 脚本

#成为master
 
#vim /usr/local/Script/redis_master.sh
#!/bin/bash
REDISCLI="/usr/local/redis/src/redis-cli"
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[master]" >> $LOGFILE
date >> $LOGFILE
echo "Being master...." >> $LOGFILE 2>&1
echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE
$REDISCLI -p 6379 -a 123456  SLAVEOF NO ONE >> $LOGFILE 2>&1
 
#成为backup
 
#vim /usr/local/Script/redis_backup.sh
#!/bin/bash
REDISCLI="/usr/local/redis/src/redis-cli"
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[backup]" >> $LOGFILE
date >> $LOGFILE
echo "Being slave...." >> $LOGFILE 2>&1
#sleep 10
echo "backup Run SLAVEOF 192.168.16.29 cmd..." >> $LOGFILE
$REDISCLI -a 123456 SLAVEOF 192.168.16.29 6379 >> $LOGFILE 2>&1
 
#监控脚本 exit 1 时 
 
#vim /usr/local/Script/redis_fault.sh
#!/bin/bash
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[fault]" >> $LOGFILE
date >> $LOGFILE
sh /usr/local/Script/redis_slave.sh
 
#keepalived 服务停止时
 
#vim /usr/local/Script/redis_stop.sh
#!/bin/bash
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[stop]" >> $LOGFILE
date >> $LOGFILE

#监控脚本的路径

#vim /usr/local/Script/redis_check.sh
#!/bin/bash
ALIVE=`/usr/local/redis/src/redis-cli -a 123456 -h 192.168.16.29 -p 6379 PING`
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
if [ "$ALIVE" != "PONG" ]; then
echo $ALIVE
exit 0
else
echo $ALIVE
exit 1
fi

数据同步脚本

#vim /usr/local/Script/redis_slave.sh
#!/bin/bash
REDISCLI="/usr/local/redis/src/redis-cli"
LOGFILE="/usr/local/logs/keepalived-redis-state.log"
echo "[slave]" >> $LOGFILE
echo "Being slave...." >> $LOGFILE 2>&1
sleep 5
date >> $LOGFILE
echo "slave Run SLAVEOF 192.168.16.3 cmd..." >> $LOGFILE
$REDISCLI -a 123456 SLAVEOF 192.168.16.29 6379 >> $LOGFILE 2>&1  
 
给脚本授予权限
chmod a+x /usr/local/Script/*

192.168.16.29 redis部署

# wget http://download.redis.io/releases/redis-3.0.6.tar.gz
# tar xzf redis-3.0.6.tar.gz -C /usr/local/
# cd /usr/local
# mv redis-3.0.6 redis
# cd redis
# make
 
新建redis目录
#mkdir /usr/local/redis/logs
#mkdir /usr/local/redis/etc
#mkdir /usr/local/redis/data
#mkdir /var/run/redis/ -p
 
redis配置 
#vim /usr/local/redis/etc/redis.conf
daemonize yes
pidfile /var/run/redis/redis.pid
port 6379
tcp-backlog 511
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/usr/local/logs/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 yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
slave-priority 100
requirepass 123456
appendonly yes
appendfilename "appendonly.aof"
appendfsync everysec
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
aof-load-truncated yes
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
latency-monitor-threshold 0
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
hll-sparse-max-bytes 3000
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
 
redis启动脚本
#!/bin/bash  
#description:simple example service  
RETVAL=0
REDIS_COMMOND=/usr/local/redis/src/redis-server
REDIS_CONFIG_FILE=/usr/local/redis/etc/redis.conf
REDIS_PID=/var/run/redis/redis.pid
REDIS_XINXI=`ps -ef | grep redis | grep -v grep |wc -l`
start(){
        $REDIS_COMMOND $REDIS_CONFIG_FILE
        if [ $REDIS_XINXI -gt 1 ];then
                echo "redis Start-up success!"
        else
                echo "redis Start-up failure!"
        fi
}
 
stop(){
        /bin/kill -s QUIT `cat $REDIS_PID`
        if [ $? -eq 0 ];then
                echo "redis Stop success!"
        fi
}
 
reload(){
        /bin/kill -s HUP `cat $REDIS_PID`
        REDIS_XINXI_RELOAD=`ps -ef | grep redis| grep -v grep |wc -l`
        if [ $REDIS_XINXI_RELOAD -gt 1 ];then
                echo "redis reload success!"
        else
                exit 0
        fi
}
 
case $1 in
start)
        start
;;
stop)
        stop
;;
reload)
        reload
;;
*)
echo "error choice ! please input start|stop|reload";;
esac
exit $RETVA

授予执行权限

#chmod a+x /etc/init.d/redisd

启动redis 主服务器

#chmod a+x /etc/init.d/redisd

192.168.16.2 redis部署

# wget http://download.redis.io/releases/redis-3.0.6.tar.gz
# tar xzf redis-3.0.6.tar.gz -C /usr/local/
# cd /usr/local
# mv redis-3.0.6 redis
# cd redis
# make
 
新建redis目录
#mkdir /usr/local/redis/logs
#mkdir /usr/local/redis/etc
#mkdir /usr/local/redis/data
#mkdir /var/run/redis/ -p
复制代码
 
redis配置 
##vim /usr/local/redis/etc/redis.conf
daemonize yes
pidfile /var/run/redis/redis.pid
port 6379
tcp-backlog 511
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/usr/local/logs/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 yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
slave-priority 100
masterauth 123456
requirepass 123456
appendonly yes
appendfilename "appendonly.aof"
appendfsync everysec
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
aof-load-truncated yes
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
latency-monitor-threshold 0
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
hll-sparse-max-bytes 3000
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 192.168.16.29 6379
 
redis启动脚本
#!/bin/bash  
#description:simple example service  
RETVAL=0
REDIS_COMMOND=/usr/local/redis/src/redis-server
REDIS_CONFIG_FILE=/usr/local/redis/etc/redis.conf
REDIS_PID=/var/run/redis/redis.pid
REDIS_XINXI=`ps -ef | grep redis | grep -v grep |wc -l`
start(){
        $REDIS_COMMOND $REDIS_CONFIG_FILE
        if [ $REDIS_XINXI -gt 1 ];then
                echo "redis Start-up success!"
        else
                echo "redis Start-up failure!"
        fi
}
 
stop(){
        /bin/kill -s QUIT `cat $REDIS_PID`
        if [ $? -eq 0 ];then
                echo "redis Stop success!"
        fi
}
 
reload(){
        /bin/kill -s HUP `cat $REDIS_PID`
        REDIS_XINXI_RELOAD=`ps -ef | grep redis| grep -v grep |wc -l`
        if [ $REDIS_XINXI_RELOAD -gt 1 ];then
                echo "redis reload success!"
        else
                exit 0
        fi
}
 
case $1 in
start)
        start
;;
stop)
        stop
;;
reload)
        reload
;;
*)
echo "error choice ! please input start|stop|reload";;
esac
exit $RETVA
 
授予执行权限
#chmod a+x /etc/init.d/redisd
 
启动redis 主服务器
#chmod a+x /etc/init.d/redisd

做好以上事情，接下来我们可以启动keepalived了

在Master 192.168.16.29启动keepalived
#/etc/init.d/keepalived start
 
在Slave 192.168.16.2启动keepalived 
#/etc/init.d/keepalived start

故障转移测试

在Master 192.168.16.29输入ip a查看虚拟IP为192.168.2.3

[[email protected] ~]# ip a
1: lo: mtu 65536 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
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:84:40:c4 brd ff:ff:ff:ff:ff:ff
    inet 192.168.16.29/24 brd 192.168.16.255 scope global eth0
    inet 192.168.16.3/32 scope global eth0
    inet6 fe80::20c:29ff:fe84:40c4/64 scope link 
       valid_lft forever preferred_lft forever
 
 
关闭redis主服务器（192.168.16.29）
[[email protected] keepalived-1.2.23]# /etc/init.d/redisd stop
redis Stop success!
[[email protected] keepalived-1.2.23]# ip a
1: lo: mtu 65536 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
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:84:40:c4 brd ff:ff:ff:ff:ff:ff
    inet 192.168.16.29/24 brd 192.168.16.255 scope global eth0
    inet6 fe80::20c:29ff:fe84:40c4/64 scope link 
       valid_lft forever preferred_lft forever
[[email protected] keepalived-1.2.23]#
 
keepalived检测到redis异常，把主身份切换到从（192.168.16.2）
 
Redis Master 192.168.16.29日志
[[email protected] ~]# tail -f /usr/local/logs/keepalived-redis-state.log 
[fault]
2016年 10月 19日 星期三 14:40:39 CST
[backup]
2016年 10月 19日 星期三 14:40:39 CST
Being slave....
Run SLAVEOF cmd...
Could not connect to Redis at 127.0.0.1:6379: Connection refused
 
Redis Slave 192.168.16.2日志
[[email protected] Script]# tail -f /usr/local/logs/keepalived-redis-state.log
[backup]
2016年 10月 19日 星期三 14:43:24 CST
Being slave....
backup Run SLAVEOF 192.168.16.29 cmd...
OK Already connected to specified master
[master]
2016年 10月 19日 星期三 14:43:28 CST
Being master....
Run SLAVEOF NO ONE cmd ...
OK
 
可以看出Slave已经接管了Master角色
 
 
在Slave 192.168.16.2输入ip a查看虚拟IP
[[email protected] keepalived-1.2.23]# ip a
1: lo: mtu 65536 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
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth1: mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:aa:43:58 brd ff:ff:ff:ff:ff:ff
    inet 192.168.16.2/24 brd 192.168.16.255 scope global eth1
    inet 192.168.16.3/32 scope global eth1
    inet6 fe80::20c:29ff:feaa:4358/64 scope link 
       valid_lft forever preferred_lft forever
 
 
在Master 192.168.16.29启动redis，16.29已经恢复了Master的角色
 
[[email protected] ~]# /etc/init.d/redisd start
redis Start-up success!
[[email protected] ~]# ip a
1: lo: mtu 65536 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
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:84:40:c4 brd ff:ff:ff:ff:ff:ff
    inet 192.168.16.29/24 brd 192.168.16.255 scope global eth0
    inet 192.168.16.3/32 scope global eth0
    inet6 fe80::20c:29ff:fe84:40c4/64 scope link 
       valid_lft forever preferred_lft forever