Redis哨兵集群

作者: 平凡的运维之路 | 来源:发表于2018-10-08 14:20 被阅读29次

哨兵模式

原理

由一个或多个Sentinel实例组成的Sentinel系统可以监视任意多个主服务器，以及这些主服务器属下的所有从服务器，并在被监视的主服务器进入下线状态时，自动将下线主服务器属下的某个从服务器升级为新的主服务器。也就是应用程序连接redis哨兵的三个端口但最后返回redis主的ip和端口，进行读写redis数据。

部署

创建redis用户，创建如下目录,把下载redis复制为如下目录

[redis@redis  ~]$ ls
6379  6380  6381

6379 修改配置文件

redis.conf

daemonize yes
pidfile "redis6379.pid"
port 6379
tcp-backlog 511
bind 192.168.128.34
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/home/redis/redis-env/6379/6379.log"
databases 16
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename "dump.rdb"
dir "/home/redis/redis-env/6379"
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
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-tru

sentinel.conf

port 27379
bind 192.168.128.34
sentinel monitor server-1M 192.168.128.34 6379 2
sentinel failover-timeout server-1M 300000
sentinel config-epoch server-1M 11
sentinel leader-epoch server-1M 11
dir "/home/redis/redis-env/6379"
maxmemory 3gb
maxmemory-policy noeviction
sentinel known-slave server-1M 192.168.128.34 6380
sentinel known-slave server-1M 192.168.128.34 6381
sentinel known-sentinel server-1M 192.168.128.34 27381 6405d1ab6632fbc20a97f65430a66d211392bf12
sentinel known-sentinel server-1M 192.168.128.34 27380 3e6fda7297f80711df260d924c9bebbc004d999a
sentinel current-epoch 11

6380 修改配置文件

redis.conf

daemonize yes
pidfile "redis6380.pid"
port 6380
tcp-backlog 511
bind 192.168.128.34
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/home/redis/redis-env/6380/6380.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 "/home/redis/redis-env/6380"
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
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
maxclients 3000
maxmemory 3gb
maxmemory-policy noeviction
slaveof 192.168.128.34 6379

sentinel.conf

port 27380
bind 192.168.128.34
sentinel monitor server-1M 192.168.128.34 6379 2
sentinel failover-timeout server-1M 300000
sentinel config-epoch server-1M 11
sentinel leader-epoch server-1M 11
dir "/home/redis/redis-env/6380"
maxmemory 3gb
maxmemory-policy noeviction
sentinel known-slave server-1M 192.168.128.34 6381
sentinel known-slave server-1M 192.168.128.34 6380
sentinel known-sentinel server-1M 192.168.128.34 27379 49922ded8bc33f23e37fa36535cb99f4e95150c7
sentinel known-sentinel server-1M 192.168.128.34 27381 6405d1ab6632fbc20a97f65430a66d211392bf12
sentinel known-sentinel server-1M 192.168.128.34 27380 3e6fda7297f80711df260d924c9bebbc004d999a
sentinel current-epoch 11

6381 修改配置文件

redis.conf

daemonize yes
pidfile "redis6381.pid"
port 6381
tcp-backlog 511
bind 192.168.128.34
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/home/redis/redis-env/6381/6381.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 "/home/redis/redis-env/6381"
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
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
maxclients 3000
maxmemory 3gb
maxmemory-policy noeviction
slaveof 192.168.128.34 6379

sentinel.conf

port 27381
bind 192.168.128.34
sentinel monitor server-1M 192.168.128.34 6379 2
sentinel failover-timeout server-1M 300000
sentinel config-epoch server-1M 11
sentinel leader-epoch server-1M 11
dir "/home/redis/redis-env/6381"
maxmemory 3gb
maxmemory-policy noeviction
sentinel known-slave server-1M 192.168.128.34 6380
sentinel known-slave server-1M 192.168.128.34 6381
sentinel known-sentinel server-1M 192.168.128.34 27379 49922ded8bc33f23e37fa36535cb99f4e95150c7
sentinel known-sentinel server-1M 192.168.128.34 27380 3e6fda7297f80711df260d924c9bebbc004d999a
sentinel known-sentinel server-1M 192.168.128.34 27381 6405d1ab6632fbc20a97f65430a66d211392bf12
sentinel current-epoch 11

redis启动脚本

脚本,可直接复制使用

#!/bin/bash

#########################################################################################
################                 service info ###########################################
#########################################################################################
ALIAS[0]="redis";
ALIAS[1]="sentinel";

PROC[0]="redis-server";
PROC[1]="redis-sentinel"

DIR[0]="./";
DIR[1]="./";

ARGV[0]="redis.conf"
ARGV[1]="sentinel.conf"

PORT[0]=":"`grep port ${ARGV[0]} |grep -v '#' | awk '{print $2}'`
PORT[1]=":"`grep port ${ARGV[1]} |grep -v '#' | awk '{print $2}'`

PROC_COUNT=2
CHECK_TIME=10

CHECK_LOGFILE="check.log"
CHECK_LOGFILE="/dev/null"

#########################################################################################

################################################################################33

usage()
{
    echo  "Usage: start.sh [options] service1 [,service2..]";
    echo  "Start/stop services or show files version" 
    echo  "  -c    Get coredump, used with -k."
    echo  "  -d    Run a daemon script for service"
    echo  "  -f    Force kill service, used with -k" 
    echo  "  -h    Print help infomation"
    echo  "  -k    Kill service"
    echo  "  -o    Get owner authority [ -o user ]"                       
    echo  "  -r    Restart service "
    echo  "  -s    Show service status"
    echo  "  -v    Show file version"
    echo  "  list  Show all services status"
    echo  "  Note: If no option ,default to start a service. "
    echo  ""
    echo  "Valid shortcut for services:"
    echo  " redis" 
    echo  "  "
    echo  "Example: "
    echo  "  \"start.sh list\"    View status of all services."
    echo  "  \"start.sh redis\"     Start redis."
    echo  "  \"start.sh -d redis\"  Start redis with a daemon."
    echo  "  \"start.sh cut_table redis\" Start cut_table and redis."
    echo  "  \"start.sh -v redis\"  Show redis version."
    echo  "  \"start.sh -s redis\"  Show redis status."
    echo  "  \"start.sh -k redis\"  Stop redis."
    echo  "  \"start.sh -r redis\"  Restart redis."
    echo  "  \"start.sh -kd redis\" Stop redis and its daemon."
    echo  "  \"start.sh -kf redis\" Force,kill redis immediately."
    echo  "  \"start.sh all\"     Start all services at once."
    echo  "  \"start.sh -k all\"  Stop all services at once."
}

echo_w()
{
    width=`expr 60 - $1`
    f=`printf "%ds" $width`
    printf  "%$f\n"  "$2"
}
set_ulimit() 
{
    core_limit=`ulimit -c`
    [ $core_limit="0" ]
    if [ $? -eq "0" ]; then
        ulimit -c unlimited
        #echo "ulimit set"
    fi;
}

checkDaemonRunning()
{
#ret=`ps -u $WHO -o pid -o comm -o cmd|grep "$SCRIPT_NAME -a -d $1" | grep "$2" |grep -v grep|awk '{print $1}'`;
    ret=`ps -u $WHO elf|grep "$SCRIPT_NAME -a -d $1" | grep "$2" |grep -v grep|awk '{print $3}'`;
    echo $ret
    if [ "$ret" = "" ]; then
        return 0;
    else
        return 1;
    fi
}

checkRunning()
{
#ret=`ps -u $WHO -o pid -o comm|grep $1 | grep "$2" |awk '{print $1}'`;
    ret=`ps -u $WHO elf|grep $1 | grep "$2" |awk '{print $3}'`;
    echo $ret
    
    if [ -z "$ret" ]; then
        return 0;
    else
        return 1;
    fi

    #proc=`ps aux | grep $1 | grep -v grep`
    #echo $proc
    #if [ -z "$proc" ]; then
    #   return 0;
    #else
    #   return 1;   
    #fi
    
}


show_status()
{
    proc=${PROC[$1]}
    ret=`checkRunning $proc ${PORT[$1]}`;
    if [ $? -ne 0 ];then
        str="$proc (pid $ret ${PORT[$1]})";
        len=`expr length "$str"`;
        echo -n $str
        ret=`checkDaemonRunning ${ALIAS[$1]} ${PORT[$1]}`;
        if [ $? -ne 0 ];then
            echo_w `expr $len - 10` "[daemon][�[32mrunning�[0m]";
        else
            echo_w `expr $len - 10` "[�[32mrunning�[0m]";
        fi
    else
        echo -n $proc
        len=`expr length "$proc"`;
        echo_w `expr $len - 10` "[�[31mstopped�[0m]";
    fi
}

restart_proc()
{
    kill_proc $1;
#   [ $? -ne 0 ]&& return;      
    start_proc $1;
}

kill_daemon()
{
    len=`expr length "${PROC[$1]} daemon"`
    ret=`checkDaemonRunning ${ALIAS[$1]} ${PORT[$1]}`;  
    if [ $? -ne 0 ];then
        echo -n "stoping ${PROC[$1]} ${PORT[$1]} daemon ."; 
        for pid in "$ret"
        do
            kill -9 $pid;
        done
        echo_w $len "[�[31mstopped�[0m]";
    fi  
}
kill_proc()
{
    result=1;
    [ "$DAEMON" = "true" ] && kill_daemon $1;
    proc=${PROC[$1]}
    echo -n "stoping $proc ${PORT[$1]}.";   
    len=`expr length "$proc"`
    ret=`checkRunning $proc ${PORT[$1]}`;   
    if [ $? -ne 0 ];then
        for pid in "$ret"       
        do
            if [ "$FORCE" = "true" ];then
                kill -9 $pid;
            elif [ "$CORE" = "true" ];then
                kill -6 $pid;
            else
                kill -9 $pid;
            fi
            for((t=0; t<10; t++))
            do
                echo -n "."
                len=`expr $len + 1`
                ret=`checkRunning $proc ${PORT[$1]}`;   
                if [ $? -ne 0 ];then
                    sleep 1;
                else
                    echo_w $len "[�[31mstopped�[0m]";
                    result=0;
                    break;
                fi
                if [ $t -eq 8 ];then
                    echo_w $len "[�[32mrunning�[0m]";
                    KILLFAILED=true;
                fi
                    
            done
        done
    else
        echo -n ".."
        echo_w `expr $len + 2` "[�[31mstopped�[0m]";
    fi
    return $result;
}

start_daemon()
{
    $SCRIPT_NAME -a -d ${ALIAS[$1]} >/dev/null 2>&1 &
    echo  "starting daemon for ${PROC[$1]}... OK";              
#   ret=`checkDaemonRunning ${ALIAS[$1]}`;
#   if [ $? -ne 0 ];then
#       echo  "already a instance running ...";
#   else    
#       $SCRIPT_NAME -a -d ${ALIAS[$1]} >/dev/null 2>&1 &
#       echo  "starting daemon for ${PROC[$1]}... OK";              
#   fi
}

start_real_daemon()
{
    while true
    do
        start_proc $1
        if [ $? = 2 ];then
            echo "`date`: start daemon ${PROC[$1]} ok" >> daemon.log
        elif [ $? = 3 ];then
            echo "`date`: start daemon ${PROC[$1]} failed" >> daemon.log
        fi
        sleep $CHECK_TIME
    done        
}

start_proc()
{
    proc=${PROC[$1]}    
    echo -n "starting $proc ";
    len=`expr length "$proc"`
    ret=`checkRunning $proc ${PORT[$1]}`;
    cret=$?;

    if [ "$ISLIST" != "true" ]; then
        echo "`date` check $proc ${PORT[$1]} return=[$ret] [$cret]" >> $CHECK_LOGFILE
    fi

    if [ $cret -ne 0 ] || [ "$ret" != "" ]; then
        echo -n ".."
        echo_w `expr $len + 2` "[�[31mFAILED�[0m]"
        echo "Error:$proc ${PORT[$1]} already have a instance (pid $ret)";
        return 1
    else
        cd ${DIR[$1]}
        nohup ./$proc ${ARGV[$1]} >/dev/null 2>&1 &
        cd - >> /dev/null 2>&1
        for t in 1 2 3
        do 
            echo -n "." 
            len=`expr $len + 1`
            sleep 1
        done

        echo -n "."
        ret=`checkRunning $proc ${PORT[$1]}`;
        if [ $? -ne 0 ];then
            echo_w `expr $len + 1` "[�[32m  OK  �[0m]";
            return 2
        else
            echo_w `expr $len + 1` "[�[31mFAILED�[0m]";
            return 3
        fi
    fi
}

show_version()
{
    echo "====================== ${PROC[$1]} Version Info ======================";  
    cd bin
    ./${PROC[$1]} --version
    cd ..       
}



do_process()
{
    if [ "$KILL" = "true" ];then
        kill_proc $1;
        return;
    fi
    if [ "$RESTART" = "true" ];then
        restart_proc $1;    
        return;
    fi
        
    if [ "$VERSION" = "true" ];then
        show_version $1;
    fi
    
    if [ "$STATUS" = "true" ];then
        show_status $1;
    fi
    
    if [ "$START" = "true" ];then
        start_proc $1;
    fi

    if [ "$DAEMON" = "true" ] && [ "$EXPAND" != "true" ];then
        start_daemon $1
    fi

    if [ "$DAEMON" = "true" ] && [ "$EXPAND" == "true" ];then
        start_real_daemon $1;
    fi
}

##=========================================================================================================
##=========================================================================================================
##=
##=========================================================================================================
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:`pwd`/lib
KILL=false;
VERSION=false;
STATUS=false;
START=true;
FORCE=false;
CORE=false;
DAEMON=false;
RESTART=false;
EXPAND=false;
WHO=`whoami`;
OWN=`stat -c %U $0`
KILLFAILED=false;
SCRIPT_NAME=$0
ISLIST=false;


while getopts :krvsahfcdo: OPTION
do
    case $OPTION in
    a)
        EXPAND=true;
        START=false;;
    k)
        KILL=true;
        START=false;;
    v)
        START=false;
        VERSION=true;;
    s)
        START=false;
        STATUS=true;;
    f)
        FORCE=true;;
    r)
        START=false;
        RESTART=true;;
    c)
        START=false;
        CORE=true;;
    o)
        WHO=$OPTARG;;
    d)
        START=false;
        DAEMON=true;;
    h)
        usage;
        exit 0;;
    \?)
        echo "start.sh: invalid option"
        echo "Tyr \"start.sh -h\" for more infomation."
        exit;;
    esac
done
    

shift `expr $OPTIND - 1`;

if [ "$#" = "0" ];then
    echo "start.sh: missing operand." 
    echo "Try \"start.sh -h\" for more infomation."
    exit 1;
fi

if [ $OWN != $WHO ];then
    echo "start.sh:sorry [$WHO], the owner is [$OWN]."
    echo "Add option \"-o $OWN\" to ignore this."
    echo "Try \"start.sh -h\" for more infomation."
    exit 1;
fi

set_ulimit;

for proc in "$@"
do
    num=-1
    proc=`tr A-Z a-z <<< $proc`;
    if [ "$proc" = "list" ];then
        ISLIST=true;
        for((i=0; i<PROC_COUNT; i++))
        do
            show_status $i;
        done    
        exit 0;
        
    fi  
    if [ "$proc" = "all" ];then
        for((i=0; i<PROC_COUNT; i++))
        do
            do_process $i;  
        done    
        exit 0;
    fi
    for((i=0; i<PROC_COUNT; i++))
    do
        if [ "$proc" = "${ALIAS[$i]}" ]; then
            num=$i; 
            break;
        fi
    done
    if [ $num -ne -1 ];then 
        do_process $num;
    else
        echo "start.sh: wrong service name [$proc]. "
        echo "Try \"start.sh -h\" for more infomation."
        exit 1;
    fi
done

if [ "$KILLFAILED" = "true" ];then
    echo "  ----"
    echo "If they are still running , check it later use command \"start.sh list\"";
    echo "Also can use \"start.sh -kf SERVICES\" to kill them immediately"
fi

exit 0;

脚本说明

[redis@redis 6379]$ ./start.sh list
redis-server (pid 2620 :6379)                       [running]
redis-sentinel (pid 3164 :27379)                    [running]
[redis@redis 6379]$ ./start.sh redis
[redis@redis 6379]$ ./start.sh sentinel

Redis哨兵集群

哨兵模式

部署

6379 修改配置文件

6380 修改配置文件

6381 修改配置文件

redis启动脚本

相关文章

网友评论

延伸阅读

深度阅读

栏目导航

热点阅读