监控oracle连接,巧用Nagios脚本监控Oracle连接

场景&＃xff1a;

程序修改完成后&＃xff0c;上传到应用服务器tomcat&＃xff0c;重启tomcat后&＃xff0c;发现启动异常&＃xff01;前端nginx负载均衡出现504错误&＃xff0c;网站无法打开&＃xff01;

排查&＃xff1a;

经过排查tomcat日志&＃xff0c;发现连接数据库异常&＃xff1b;登陆数据库后以"conn /as sysdba"方式登陆正常&＃xff0c;数据库轻载&＃xff0c;无压力&＃xff1b;于是检查数据库的监听器&＃xff0c;输入"lsntctl services"命令&＃xff0c;卡住&＃xff0c;半天不出结果…

故障处理&＃xff1a;

输入"lsntctl stop" 或者"lsntctl reload"命令&＃xff0c;卡住&＃xff0c;半天不出结果&＃xff0c;于是输入service network restart命令&＃xff0c;一样没有效果&＃xff01;于是出必杀技&＃xff0c;输入"netstat -ntpl|grep :1521"命令找出监听器的进程号&＃xff0c;kill -9结束掉&＃xff0c;然后重启监听器&＃xff0c;再启动tomcat正常&＃xff0c;网站访问畅通无阻&＃xff01;

原因分析&＃xff1a;

后期维护&＃xff1a;

为了尽量避免这种问题&＃xff0c;想到去写一个监控脚本配合nagios监控&＃xff0c;当数据库连接异常的时候&＃xff0c;可以第一时间收到报警短信&＃xff1b;Java程序连接数据库使用连接池&＃xff0c;所以不一定会及时暴露出问题&＃xff01;

写一个脚本放任务计划中运行&＃xff0c;定期去连接下数据库&＃xff0c;查询下系统时间和数据库的状态&＃xff0c;spool输出到临时文件上

[root&＃64;server199 ~]# crontab -l

*/5  *    *             *  *  /usr/local/nagios/libexec/connect_Oracle.sh

[root&＃64;server199 ~]# cat /usr/local/nagios/libexec/connect_oracle.sh

#!/bin/sh

#functions: connect oracle server test

#author:lw.yang

#modify_date: 2012-02-22

rm -rf /tmp/check_oracle.log

export ORACLE_HOME&＃61;/u01/app/oracle/product/10.2.0/db_1/

$ORACLE_HOME/bin/sqlplus username/password&＃64;ip:1521/services_name <

set echo off

set feedback off

spool /tmp/check_oracle.log

alter session set nls_date_format&＃61;&＃39;YYYY-MM-DD:HH24:MI:SS&＃39;;

select sysdate from dual;

select name,open_mode from v\$database;

spool off

set echo on

set feedback on

EOF