Linux下创建定时任务监控kafka进程，异常时自动启动

Posted 2023-03-02 夜灬狼丶

背景：在有一台服务器上的kafka进程会莫名其妙的消失，暂时没找到原因，放假前决定先临时写一个脚本进行监控，消失了就自动启动一下

目录

理论步骤

实现步骤

1、编写脚本并定义kafka的bin目录及配置文件以及定时任务周期

2、获取当前时间(打印监控日志用)以及根据kafkaBinPath等参数获取当前kafka进程id

3、判断kafkaPid参数是否为空，

4、添加系统定时任务

完整脚本

示例

---

理论步骤

     1、查询kafka进程id并赋值给变量kafkaPid

     2、判断kafkaPid是否为空

             为空则启动kafka

             不为空则不作任何操作

     3、对脚本添加定时任务

实现步骤

1、编写脚本并定义kafka的bin目录及配置文件以及定时任务周期

kafkaBinPath='/opt/app/kafka_2.11-2.1.0/bin'
kafkaConfPath='/opt/app/kafka_2.11-2.1.0/config/server.properties'
cronTab='* * * * *'

2、获取当前时间(打印监控日志用)以及根据kafkaBinPath等参数获取当前kafka进程id

curDate=`date +"%F %T"`
kafkaPid=`ps -ef|grep -v grep |grep "$kafkaBinPath"|grep "server.properties"|awk 'print $2'`

3、判断kafkaPid参数是否为空，

     为空则进入kafkaBinPath目录启动kafka并放入后台执行，等待3秒后查询新进程id并打印，

      不为空则直接打印当前进程id

if [ -z "$kafkaPid" ];then
  echo -e "kafka is not runing,\\nBegin start kafka......"
  cd $kafkaBinPath
  nohup ./kafka-server-start.sh -daemon $kafkaConfPath 1>/dev/null 2>&1 &
 sleep 3
  newKafkaPid=`ps -ef|grep -v grep |grep "$kafkaBinPath"|grep "server.properties"|awk 'print $2'` 
  echo "kafkaPid is:$newKafkaPid"
else
  echo "kafkaPid is:$kafkaPid"
fi

4、添加系统定时任务

      识别脚本目录及文件名，并根据cronTab参数自动添加系统定时任务

curPath=$(cd `dirname $0`; pwd)
fileName=`echo "$0"|awk -F/ 'print $NF'`
touch /var/spool/cron/root
sed -i "/$fileName/d" /var/spool/cron/root
echo "$cronTab $curPath/$fileName >> $curPath/$fileName".log >> /var/spool/cron/root
/sbin/service crond restart $1>/dev/null 2>&1

完整脚本

完整脚本功能实现：脚本通过定义的kafka的bin目录以及配置文件以及定时任务周期来监控kafka进程，如果进程存在则跳过，不存在则启动kafka并将监控日志写入到当前目录下【$fileName.log】日志文件中，最后将当前文件自动写入到系统任务表中并定时执行

#!/bin/bash
##############################################################
# File Name：monitr_kafka.sh
# Version :V1 . 
# Author:   
# Email:   
# Created Time ：2023-02-07 15:43:04
# Description：
#############################################################

#parameter:Please specify the home directory and configuration file of kafka 
kafkaBinPath='/opt/app/kafka_2.11-2.1.0/bin'
kafkaConfPath='/opt/app/kafka_2.11-2.1.0/config/server.properties'
cronTab='* * * * *'


curDate=`date +"%F %T"`
kafkaPid=`ps -ef|grep -v grep |grep "$kafkaBinPath"|grep "server.properties"|awk 'print $2'`


echo -e "curDate:$curDate==================================="
#checking kafka
if [ -z "$kafkaPid" ];then
  echo -e "kafka is not runing,\\nBegin start kafka......"
  cd $kafkaBinPath
  nohup ./kafka-server-start.sh -daemon $kafkaConfPath 1>/dev/null 2>&1 &
 sleep 3
  newKafkaPid=`ps -ef|grep -v grep |grep "$kafkaBinPath"|grep "server.properties"|awk 'print $2'` 
  echo "kafkaPid is:$newKafkaPid"
else
  echo "kafkaPid is:$kafkaPid"
fi
echo -e "\\n\\n"


#add crontab
curPath=$(cd `dirname $0`; pwd)
fileName=`echo "$0"|awk -F/ 'print $NF'`
touch /var/spool/cron/root
sed -i "/$fileName/d" /var/spool/cron/root
echo "$cronTab $curPath/$fileName >> $curPath/$fileName".log >> /var/spool/cron/root
/sbin/service crond restart $1>/dev/null 2>&1

示例

[root@localhost ~]# tail -40f /opt/shell/monit_kafka.sh.log 

curDate:2023-02-07 17:55:01===================================
zookeeperPid is:14516
kafkaPid is:15107

curDate:2023-02-07 18:00:01===================================
kafkaPid is:15107

curDate:2023-02-07 20:40:01===================================
kafka is not runing,
Begin start kafka......
kafkaPid is:2815

curDate:2023-02-07 20:45:01===================================...
kafkaPid is:2815