监控redis服务是否存活

#!/bin/bash
#监控redis是否存活的脚本
while [ 1 ]; 
do
a=`lsof -i:6379 |grep ‘redis-ser‘ |grep -v grep |grep -v agent|sort | tail -1 | cut -f 1 -d‘ ‘`
if [ "$a" != ‘redis-ser‘ ];then
{  /usr/local/redis/bin/redis-server /usr/local/redis/conf/redis.conf &
echo redis restart at `date` >> /data/redis/data/redisLive.log  
}
fi
sleep 5
done

###Redis监控和告警

###健康状态

redis_stat.sh

#!/bin/bash
host="127.0.0.1"
passwd1="123456"
passwd2="123456"
passwd3="123456"
if [ $# -lt 4 ];then
        echo "Usage: bash $0 -p [6379] -s section [-k] keys [-d] db [-o] opt"
        exit 1
else
        port=6379
        db=‘‘
        opt=‘‘
        while [ $# -gt 0 ]
        do
                if [ $1 == "-p" ];then
                        port=$2
                fi
                if [ $1 == "-s" ];then
                        section=$2
                fi
                if [ $1 == "-k" ];then
                        key=$2
                fi
                if [ $1 == "-d" ];then
                        db=$2
                fi
                if [ $1 == "-o" ];then
                        opt=$2
                fi
                shift 2
        done
fi
case $port in
        6379)
                passwd=$passwd1
                ;;
        6380)
                passwd=$passwd2
                ;;
        6381)
                passwd=$passwd3
                ;;
        *)
                passwd=""
                ;;
esac
if [[ $section == "Keyspace" && $key == "dbs" ]];then
        cmd="/usr/local/redis/bin/redis-cli -p $port -a $passwd info Keyspace |grep ‘^db[0-9]\{1,2\}:*‘|wc -l"
elif [[ ! -z $db && ! -z $opt ]];then
        cmd="/usr/local/redis/bin/redis-cli -p $port -a $passwd info Keyspace |grep $db |tr -s ‘,‘ ‘\n‘|grep $opt|cut -d ‘=‘ -f 2"
        #cmd="/usr/local/redis/bin/redis-cli -p $port -a $passwd info Keyspace |grep $db |tr -s ‘,‘ ‘\n‘ |awk -F [=] -v akey=$opt ‘/$akey/{print """$NF"""}‘"
else
        cmd="/usr/local/redis/bin/redis-cli -p $port -a $passwd info $section|grep ${key}: |cut -d ‘:‘ -f 2"
fi
#echo $cmd
eval $cmd

###
redis-cli -h 127.0.0.1 -p 6379 -a zvIZ31Yn3G info all | awk -F ‘:‘ 
redis-cli用法：

# redis-cli -h 127.0.0.1 -p 6379 -a passwoed     //连接指定的redis
# redis-cli info        //没有密码的情况下，打印本地redis-srv的状态信息
# redis-cli info Keyspace     //查看指定状态域的信息

###redis3.0集群监控脚本
http://blog.itpub.net/20625855/viewspace-1847453/


#!/bin/bash
######变量定义####################################
auth="ASF6tp73yF5VPGVhHJuF"               #认证密码
configdir=‘/data/redis_cluster‘          #配置目录
bindir=‘/usr/local/redis-3.0.3/bin/‘            #redis安装目录
client="$bindir/redis-cli "       #
max_memory=0                              #初始化最大内存
flag=0                                    #输入参数是否有误标识

#集群ip list
list=(
172.16.106.76:6379
172.16.106.78:6379
172.16.106.77:6379
)                                        


is_sendms=$2                              #是否发送短信报警:默认不发送；=1: 发送
###################################################




###报警函数定义####################################
sendms(){
if [ "$2" == "1" ]; then
#自己填写报警方式
echo "报警:$1"
fi
}


export -f sendms 
###################################################






###redis状态统计函数###############################
statistics_redis(){
local ip=`echo $1 | awk -F‘:‘ ‘{print $1}‘`
local port=`echo $1 | awk -F‘:‘ ‘{print $2}‘`
local is_slowlog=0


last_slowlog_time=`$client -h $ip -p $port slowlog get 1 |awk  ‘{if(NR==2) print $1}‘`
current_time=`date +%s`


if [ "$last_slowlog_time" != "" ]; then
  if [ $((last_slowlog_time+60*10)) -gt $current_time ]; then
   is_slowlog=1
  fi
fi




$client -h $ip -p $port info all | awk -F ‘:‘ -v max_memory=$max_memory -v addr=$1 -v is_sendms=$2 -v is_slowlog=$is_slowlog ‘{\
if($0~/uptime_in_seconds:/) uptime=$2;else if($0~/connected_clients:/) cnt_clients=$2;else if($0~/used_memory:/) used_memory=$2;else if($0~/used_memory_rss:/) used_memory_rss=$2;else if($0~/used_memory_peak:/) used_memory_peak=$2;else if($0~/mem_fragmentation_ratio:/) mem_ratio=$2;

else if($0~/rdb_last_bgsave_status:/) bgsave_status=$2;else if($0~/aof_last_write_status:/) aof_write_status=$2;else if($0~/aof_last_bgrewrite_status:/) aof_bgrewrite_status=$2;

else if($0~/instantaneous_ops_per_sec:/) ops_per_sec=$2;else if($0~/keyspace_hits:/) keyspace_hits=$2;else if($0~/keyspace_misses:/) keyspace_misses=$2;else if($0~/:keys=/) keys=keys"\n"$2;

}END{printf("\033[1;33;1m####概况:\033[0m\n");printf("    启动时间:%d\n",uptime);printf("    当前连接数:%d\n",cnt_clients);printf("    当前OPS:%d\n",ops_per_sec);printf("    当前key分布情况:%s\n",keys);

printf("\033[1;33;1m####命中情况：\033[0m\n");printf("    命中次数: %d\n",keyspace_hits);printf("    miss次数: %d\n",keyspace_misses);printf("    命中率：%d%\n",keyspace_hits/(keyspace_hits+keyspace_misses+0.1)*100);

printf("\033[1;33;1m####内存使用情况：\033[0m\n");printf("    1)分配总内存:%dMb\n",used_memory_rss/1024/1024);printf("    2)使用内存:%dMb\n",used_memory/1024/1024);printf("    3)峰值:%dMb\n",used_memory/1024/1024);printf("    4)最大内存:%dMb\n",max_memory/1024/1024);printf("    5)内存碎片率:%s\n",mem_ratio);

printf("\033[1;33;1m####持久化: \033[0m\n");printf("    上次bgsave状态:%s\n",bgsave_status);printf("    上次aof状态:%s\n",aof_write_status);printf("    上次rewrite状态:%s\n",aof_bgrewrite_status);

printf("\033[1;33;1m####报警信息: \033[0m\n");if(cnt_clients>=1000){printf("\033[1;31;1mwarning: 当前连接数:%d,超标\033[0m\n",cnt_clients);                      system("sendms "addr"_当前连接数:"cnt_clients" "is_sendms);                      };if(ops_per_sec>=50000){printf("\033[1;31;1mwarning: 当前OPS:%d,超标\033[0m\n",ops_per_sec);                       system("sendms "addr"_当前OPS:"ops_per_sec" "is_sendms);                      };

if(keyspace_hits/(keyspace_hits+keyspace_misses+0.1)*100<=50){                  printf("\033[1;31;1mwarning:当前命中率：%d%,过低\033[0m\n",keyspace_hits/(keyspace_hits+keyspace_misses+0.1)*100);                 };

if(used_memory/max_memory*100>=80){printf("\033[1;31;1mwarning:当前内存使用：%d%,过高\033[0m\n",used_memory/max_memory*100);                                   system("sendms "addr"_内存使用率:"used_memory/max_memory*100"% "is_sendms);                                  };

if(is_slowlog>0) {printf("\033[1;31;1mwarning：存在慢查询，请确认!\033[0m\n");                    system("sendms "addr"_存在慢查询 "is_sendms) };

#if(bgsave_status!~/ok/){printf("上次bgsave状态错误:%s\n",bgsave_status)};#if(aof_write_status!~/ok/){printf("上次aof状态错误:%s\n",aof_write_status)};#if (aof_bgrewrite_status!~/ok/){printf("上次rewrite状态错误:%s\n",aof_bgrewrite_status)};\
}‘


}
###############################################






####BEGIN:遍历list##################
for config in ${list[@]}
do
   ip=`echo $config| awk -F‘:‘ ‘{print $1}‘`
   port=`echo $config| awk -F‘:‘ ‘{print $2}‘`


####stop######################
   if [ "$1" == "stop" ]; then
      flag=1
      echo -n $config" "
      $client -h $ip -p $port shutdown5588 2>>/dev/null
      #if [ $? -eq 0 ]; then
         echo "shutdown success!"
      #else
      #   echo "shutdown meet error!"
      #fi
   fi


####start###################
   if [ "$1" == "start" ]; then
      flag=1
      echo -n  $config" "
      ssh $ip "${bindir}/redis-server  ${configdir}/${port}.cnf"
      if [ $? -eq 0 ]; then
         echo "started"
      else
        echo "starting meet error"
      fi
   fi
####status##################
   if [ "$1" == "status" ]; then
      flag=1
      ###找到一个存活的节点，根据它来查看集群状态
      isalive=`$client -h $ip -p $port  ping 2>>/dev/null`
      if [ "$isalive" == "PONG" ]; then
         echo -e  "\e[1;32;1m#----------------------------------#\e[0m"
         echo  -e "\e[1;32;1m#集群基本信息:                     #\e[0m"
         echo -e  "\e[1;32;1m#----------------------------------#\e[0m"
         cluster_is_ok=`$client  -h $ip -p $port  cluster info | grep cluster_state| cut -b 15-16`
         if [ "$cluster_is_ok" == "ok" ]; then
             echo -e "cluster_state:\e[1;32;1mok\e[0m"
         else
             echo -e "\033[1;31;1m$($client  -h $ip -p $port  cluster info | grep cluster_state)\e[0m"
             sendms "${config}_cluster_state:$cluster_is_ok" $is_sendms
         fi
         nodes_alive=`$client -h ${ip}  -p ${port}  cluster nodes | grep -vE ‘fail|disconnected‘ | wc -l`
        
         if [ ${#list[*]} -ne $nodes_alive ]; then
            echo -e "total nodes:${#list[*]}, \033[1;31;1malive nodes:${nodes_alive}!!\033[0m"
            echo -e "\033[1;31;1mWarning: some nodes have down!!\033[0m"
            sendms "${config}_cluster_state:some_nodes_is_down" $is_sendms
         else
            echo "total nodes:${#list[*]}, alive nodes:${nodes_alive}"
         fi
         
         max_memory=`$client -h $ip -p $port config5588 get maxmemory | awk ‘{if(NR>1)print $1}‘`
         ###使用循环匹配整理出目前cluster的M-s关系树
         echo -e  "\e[1;32;4m#####主从结构树:\e[0m"


         v_str=""
         cnt=1
         for master in `$client -h $ip -p $port cluster nodes|grep ‘master‘|grep -vE ‘fail|disconnected‘|awk ‘{print $1","$2}‘|sort -k 2,2 -t ‘,‘`
         do
           mid=`echo $master | awk -F‘,‘ ‘{print $1}‘`
           maddr=`echo $master | awk -F‘,‘ ‘{print $2}‘`
           mip=`echo $master | awk -F‘,|:‘ ‘{print $2}‘`
           echo -e  "\033[1;36;1mmaster${cnt}:"$maddr"\033[0m"
           $client -h $ip  -p $port cluster nodes | grep ‘slave‘|grep -vE ‘fail|disconnected‘ | grep $mid | awk ‘{print "                      |-->slave"NR":"$2}‘
           tmp=`$client -h $ip -p $port cluster nodes | grep  ‘slave‘|grep -vE ‘fail|disconnected‘  | grep $mid | grep $mip | wc -l`
           v_tmp=`$client -h $ip  -p $port cluster nodes | grep ‘slave‘|grep -vE ‘fail|disconnected‘ | grep $mid | awk ‘{printf $2" "}‘`


          if [ $tmp -ne 0 ]; then
             echo -e "\033[1;31;1mWarning: master‘s slave node is on the master‘s server!!\033[0m"
             sendms "${maddr}_cluster_state:M-S_is_on_same_server" $is_sendms 
          fi


           v_str=$v_str"\""$maddr" "$v_tmp"\" " 
           let cnt++
         done
         
         ###v_str变量记录("m1 s1" "m2 s2")类似的二维数组结构用于逐个分析每个存活状态的redis实例
         declare -a array="("$v_str")"


         
         ###使用双层嵌套循环遍历收集各个redis实例的状态
         n_array=${#array[*]}
         for((i=0; i<$n_array;i++))
         do
           inner_array=(${array[$i]})
           n_inner_array=${#inner_array[*]}
           echo ""
           echo -e  "\e[1;32;1m#----------------------------------#\e[0m"
           echo -e  "\e[1;32;1m#分片$((i+1)):                            #\e[0m"
           echo -e  "\e[1;32;1m#----------------------------------#\e[0m"
           for((j=0;j<$n_inner_array;j++))
           do
             echo -e "\e[1;35;1m+++++${inner_array[$j]}+++++\e[0m"
             statistics_redis ${inner_array[$j]} $is_sendms
           done
         done

         exit 0
      fi
   fi


done 


if [ "$flag" == "0" ]; then
  echo -e  "\e[1;31;1musage: sh cluster_control [start|stop|status] [1]\e[0m"
fi


if [ "$is_alive" != "PONG" -a "$1" == "status" ]; then
echo -e "\e[1;31;1mAll nodes is stopped.\e[0m"
fi