这一两天现场实施的同事,发现系统的tomcat会因为内存溢出的情况出现假死的情况。现场的同事一时查不出问题,最后一招了写个脚本监控一下。
1 使用环境
操作系统:CentOS 4.8
JDK版本:j2sdk1.4.2
Tomcat版本:tomcat-5.0.28
2 监控脚本
#!/bin/bash
#
# Keep watch at tomcat's status,
# automatically restart it if it dead or out of memory.
export PATH=$PATH:/sbin:/bin:/usr/sbin:/usr/bin;
tomcat_port=":8080 ";
tomcat_base_dir="/home/test/tomcat-5.0.28";
check_page_url=http://127.0.0.1:8080/testweb/test.jsp;
pid_filt_pattern="tomcat";
guarder_dir="/home/test";
log_file="/home/test/tomcat_run_log.log";
d=$(date +%F" "%T);
# init the log file
touch ${log_file}
do_restart_tomcat() {
# first, try to shutdown it anyway!
${tomcat_base_dir}/bin/shutdown.sh
# second, check if the tomcat pid still exist, if yes then kill it!
if ps -ef | grep ${pid_filt_pattern} | grep -v grep
then
kill -9 $(ps -ef | grep ${pid_filt_pattern} | grep -v grep | awk '{print $2}')
fi
# run start tomcat
${tomcat_base_dir}/bin/startup.sh
echo "$d success restart tomcat, the new pid is " >> ${log_file}
ps -ef | grep ${pid_filt_pattern} | grep -v grep | awk '{print $2}' >> ${log_file}
echo >> ${log_file}
}
# first, check if the tomcat si listen on the port
if netstat -ln | grep ${tomcat_port}
then
# init the check result file
if [ -e ${guarder_dir}/checkResult.tmp ]
then
cat /dev/null > ${guarder_dir}/checkResult.tmp
else
touch ${guarder_dir}/checkResult.tmp
fi
# try to get the check result
wget -b -o wget.log -O ${guarder_dir}/checkResult.tmp ${check_page_url}
# wait 5 second to let the get check result job done.
sleep 5
# check the result
workflag=$(cat ${guarder_dir}/checkResult.tmp |grep ServerStillWorking)
memoryflag=$(cat ${guarder_dir}/checkResult.tmp |grep LessOfMemory)
if [ "$workflag" == "" ]; then
echo "$d can not found [ServerStillWorking] in the check result, try to restart tomcat ......" >> ${log_file}
do_restart_tomcat
elif [ "$memoryflag" == "" ]; then
echo "$d can not found [LessOfMemory] in the check result, the tomcat server may out of memory, try to restart it ......" >> ${log_file}
do_restart_tomcat
fi
else
echo "$d found the tomcat not listen on ${tomcat_port}, try to restart it ......" >> ${log_file}
do_restart_tomcat
fi
3 加入定时任务crontab
3.1 编辑任务
[test@cent4 ~]$ crontab -e
进入编辑crontab的vi界面,按i键进入插入模式,将如下代码复制到输入界面
*/20 * * * * "/home/test/deamon2tomcat.sh" > /dev/null 2>&1
注:每20分钟检查一次(可修改),加“> /dev/null 2>&1 ”是为了不让它发邮件。
3.2 保存退出
按Esc键进入命令模式,输入wq,回车后则会保存退出。
3.3 查看任务
最后查看任务,确认是否编辑成功。
[test@cent4 ~]$ crontab -l
*/20 * * * * "/home/test/deamon2tomcat.sh" > /dev/null 2>&1
4 jsp的内容
<%@ page contentType="text/html;charset=GBK"%>
<%@ page import="java.util.*"%>
<%
out.println("<br>");
out.println("<center><h1>");
out.println("The Web Server is Running!<br><br>");
out.println("</h1></center>");
out.println("<br><br>");
out.println("ServerStillWorking");//标记字符!
long maxMemory = Runtime.getRuntime().maxMemory() / 1024 / 1024; //java虚拟机能取得的最大内存
long totalMemory = Runtime.getRuntime().totalMemory() / 1024 / 1024; //java虚拟机当前取得的内存大小
long freeMemory = Runtime.getRuntime().freeMemory() / 1024 / 1024; //java虚拟机所占用的内存中的空闲部分
long usedMemory = totalMemory - freeMemory; //java虚拟机当前实际使用的内存大小
out.println("<br><br>Max Momery is: " + maxMemory + "M");
out.println("<br>Total Memory is: " + totalMemory + "M");
out.println("<br>Used Memory is: " + usedMemory + "M");
out.println("<br>Free Memory is: " + freeMemory + "M");
out.println("<br><br>");
if (usedMemory < maxMemory) {
out.println("LessOfMemory"); //标记字符!
} else {
out.println("OutOfMemory"); //标记字符!
}
out.println("<br><br>");
out.println(new java.util.Date());
out.println("<br>");
out.println(TimeZone.getDefault().getDisplayName());
%>