1.同事跑了一个php的抓取, 这个会导致非常高的sleep值,大概到好几千的水平.进而mysql会down掉.

写了个crontab

#crontab -e

[php]
* */1 * * * /sh/detect_php.sh

* */1 * * * /sh/detect_php.sh

detech_php 的内容

[php]
#!/bin/bash

host_dir=`cd /sh`
proc_name="mysql"
bug_time=`date -R`
pid=0

proc_num()
{
num=`ps -ef | grep $proc_name | grep -v grep | wc -l`
return $num
}

proc_id()
{
pid=`ps -ef | grep $proc_name | grep -v grep | awk '{print $2}'`
}

proc_num
number=$?
if [ $number -eq 0 ]
then
./restart_php_mysql.sh
proc_id
echo " Kill the php and new mysql pid is : ${pid} : ${bug_time} " >> php_mysql.log 2>&1
else
proc_id
echo " The mysql is working , ${bug_time} " >> php_mysql.log 2>&1
echo "it's ok!"
fi

#!/bin/bash

host_dir=`cd /sh`
proc_name="mysql"
bug_time=`date -R`
pid=0

proc_num()
{
num=`ps -ef | grep $proc_name | grep -v grep | wc -l`
return $num
}

proc_id()
{
pid=`ps -ef | grep $proc_name | grep -v grep | awk '{print $2}'`
}

proc_num
number=$?
if [ $number -eq 0 ]
then
./restart_php_mysql.sh
proc_id
echo " Kill the php and new mysql pid is : ${pid} : ${bug_time} " >> php_mysql.log 2>&1
else
proc_id
echo " The mysql is working , ${bug_time} " >> php_mysql.log 2>&1
echo "it's ok!"
fi

脚本解释,这个脚本非常简单的. 就是判断mysql进程是否为0, 如果为0,那么就去杀掉php,然后重启mysql.并且写入一个日志.

restart_php_mysql.sh 的内容

[php]
#!/bin/sh
killall -9 php
service mysql restart

#!/bin/sh
killall -9 php
service mysql restart

pkill会导致脚本在杀完进程后直接terminated. 所以这里不使用pkill.