我的监控脚本，写的垃圾请大家不要笑我~~~

显示全部楼层 · 2021-12-13 21:56:40

虽然看不懂，不过贵在贴够长

千问 · 2021-12-13 21:56:40

看不懂的表示路过~~~~~~~~

千问 · 2021-12-13 21:56:40

最好不要用于监控有同名子进程的程序，侦测的值会比实际的值大~~~#!/bin/sh
PROCESS_NAME=$1
CPU_MAX_USERATE=$2
MEM_MAX=$3
#MEM_MAX_USERATE=$3
CYCLE=$4
MAX_EXCEED_SET_NUMBER=$5
MAX_RESTART_NUMBER=$6
CPU_MAX_USERATE_I=0
MEM_MAX_USERATE_I=0
MEM_MAX_I=0
MAX_RESTART_NUMBER_I=0
while true; do
MEM_USERATE=`ps -C $PROCESS_NAME -o %mem | awk 'BEGIN{a=0.0;}/[0-9]/{a+=$1;}END{print a;}'`
CPU_USERATE=`ps -C $PROCESS_NAME -o %cpu | awk 'BEGIN{a=0.0;}/[0-9]/{a+=$1;}END{print a;}'`
MEM=`ps -C $PROCESS_NAME -o rss | awk 'BEGIN{a=0.0;}/[0-9]/{a+=$1;}END{print a;}'`
((MEM=MEM/1024))
if (echo $MEM 0.1 $MAX_RESTART_NUMBER_I | awk '!(($1 >= $2) && ($3 >= 1)){exit 1}') then

MAX_RESTART_NUMBER_I=0

#echo $(date) $PROCESS_NAME' already strat' >> /log/$PROCESS_NAME.log
fi
if (echo $CPU_USERATE $MEM_USERATE $MEM | awk '!(($1 == 0) || ($2 == 0) || ($3 == 0)){exit 1}') then

# Program is not start or already shut down

# restart or start code

echo 'restart '$PROCESS_NAME

/sbin/service $PROCESS_NAME restart

#echo $(date) $PROCESS_NAME' restart' >> /log/$PROCESS_NAME.log

CPU_MAX_USERATE_I=0; MEM_MAX_USERATE_I=0; MEM_MAX_I=0;

((MAX_RESTART_NUMBER_I=MAX_RESTART_NUMBER_I+1))

sleep 10
fi
if (echo $CPU_USERATE $CPU_MAX_USERATE | awk '!($1 >= $2){exit 1}') then

((CPU_MAX_USERATE_I=CPU_MAX_USERATE_I+1))

echo 'CPU_MAX_USERATE_I='$CPU_MAX_USERATE_I 'CPU_USERATE='$CPU_USERATE
fi
# if (echo $MEM_USERATE $MEM_MAX_USERATE | awk '!($1 >= $2){exit 1}') then
#
((MEM_MAX_USERATE_I=MEM_MAX_USERATE_I+1))
#
echo 'MEM_MAX_USERATE_I='$MEM_MAX_USERATE_I 'MEM_USERATE='$MEM_USERATE
# fi
if (echo $MEM $MEM_MAX | awk '!($1 >= $2){exit 1}') then

((MEM_MAX_I=MEM_MAX_I+1))

echo 'MEM_MAX_I='$MEM_MAX_I 'MEM='$MEM'MB'
fi
if (echo $CPU_MAX_USERATE_I $MEM_MAX_USERATE_I $MEM_MAX_I $MAX_EXCEED_SET_NUMBER | awk '!(($1 >= $4) || ($2 >= $4) || ($3 >= $4)){exit 1}') then

echo 'restart '$PROCESS_NAME

/sbin/service $PROCESS_NAME restart

#echo $(date) $PROCESS_NAME' restart' >> /log/$PROCESS_NAME.log

CPU_MAX_USERATE_I=0; MEM_MAX_USERATE_I=0; MEM_MAX_I=0;

((MAX_RESTART_NUMBER_I=MAX_RESTART_NUMBER_I+1))

sleep 10
fi
if (echo $MAX_RESTART_NUMBER_I $MAX_RESTART_NUMBER | awk '!($1 >= $2){exit 1}') then

# Send mail code

#echo $(date) $PROCESS_NAME' gua le!!!' | /usr/sbin/sendmail -v your @xxx.com > /dev/null

#echo $(date) $PROCESS_NAME' gua le!!!' >> /log/$PROCESS_NAME.log

break 1
fi
sleep $CYCLE
done
exit
# ./test.sh lighttpd 30 900 5 5 5
# ./test.sh 进程名最大使用CPU百分比最大使用内存数(MB) 侦测周期(秒) 超过设置多少个周期后重起服务尝试重起服务失败多少次后发邮件复制代码监控评分参与人数 9[/td]威望 +149[/td][td]收起理由[/td]

dakai + 9很给力!

fhs + 20技术贴加分

fake + 10很给力!

一步一杀 + 20感谢分享！

lxfy + 20原创技术帖一律加分查看全部评分

千问 · 2021-12-13 21:56:40

支持

千问 · 2021-12-13 21:56:40

慢慢看

千问 · 2021-12-13 21:56:40

楼主不可妄自菲薄，只要是技术贴，大家都支持你

千问 · 2021-12-13 21:56:40

不懂顶字多

千问 · 2021-12-13 21:56:40

哪位老大测试下，看看好用不？？？

千问 · 2021-12-13 21:56:40

技术贴，一定要支持。

千问 · 2021-12-13 21:56:40

赞一个