(转)CentOS搭建Nagios监控

A.Nagios服务端
1.安装软件包

1. yum install -y httpd

2.下载nagios

1. wget http://syslab.comsenz.com/downloads/linux/nagios-3.0.5.tar.gz
2. wget http://syslab.comsenz.com/downloads/linux/nagios-plugins-1.4.13.tar.gz
3. wget http://syslab.comsenz.com/downloads/linux/nrpe-2.12.tar.gz

3.添加nagios账号

1. useradd nagios

4.编译安装nagios

1. mkdir /opt/hadoop/
2. tar -xzvf nagios-3.0.5.tar.gz
3. cd nagios-3.0.5
4. ./configure --prefix=/opt/hadoop/nagios
5. make all
6. make fullinstall
7. mkdir /opt/hadoop/nagios/etc
8. mkdir /opt/hadoop/nagios/etc/objects
9. cp ./sample-config/cgi.cfg /opt/hadoop/nagios/etc/
10. cp ./sample-config/nagios.cfg /opt/hadoop/nagios/etc/
11. cp ./sample-config/resource.cfg /opt/hadoop/nagios/etc/
12. cp ./sample-config/template-object/commands.cfg /opt/hadoop/nagios/etc/objects/
13. cp ./sample-config/template-object/contacts.cfg /opt/hadoop/nagios/etc/objects/
14. cp ./sample-config/template-object/timeperiods.cfg /opt/hadoop/nagios/etc/objects/
15. cp ./sample-config/template-object/templates.cfg /opt/hadoop/nagios/etc/objects/
16. cp ./sample-config/template-object/localhost.cfg /opt/hadoop/nagios/etc/objects/
17. touch /opt/hadoop/nagios/var/nagios.log
18. chmod -R 755/opt/hadoop/nagios/etc/
19. chown -R nagios:nagios /opt/hadoop/nagios

5.编译安装nagios-plugins

1. tar zxvf nagios-plugins-1.4.13.tar.gz
2. cd nagios-plugins-1.4.13
3. ./configure --prefix=/opt/hadoop/nagios --with-nagios-user=nagios --with-nagios-group=nagios
4. make && make install

检查是否已经安装成功,看这个目录下是否有插件文件

1. ls /opt/hadoop/nagios/libexec/

6.安装nrpe

1. tar zxvf nrpe-2.12.tar.gz
2. cd nrpe-2.12
3. ./configure --prefix=/opt/hadoop/nagios --enable-ssl --enable-command-args
4. make all
5. make install-plugin
6. make install-daemon
7. make install-daemon-config

7.配置httpd
添加web账号

1. htpasswd -c /opt/hadoop/nagios/etc/htpasswd.users nagiosadmin

B.Nagios客户端
1.准备软件包

1. wget http://syslab.comsenz.com/downloads/linux/nagios-plugins-1.4.13.tar.gz
2. wget http://syslab.comsenz.com/downloads/linux/nrpe-2.12.tar.gz

2.添加nagios账号，准备安装目录

1. mkdir /opt/hadoop/nagios
2. useradd nagios

3.编译安装nrpe

1. tar -xzvf nrpe-2.12.tar.gz
2. cd nrpe-2.12
3. ./configure --prefix=/opt/hadoop/nagios --enable-ssl --enable-command-args
4. make all
5. make install-plugin
6. make install-daemon
7. make install-daemon-config

4.安装nagios-plugin

1. tar -xzvf nagios-plugins-1.4.13.tar.gz
2. cd nagios-plugins-1.4.13
3. ./configure --prefix=/opt/hadoop/nagios --with-nagios-user=nagios --with-nagios-group=nagios
4. make && make install

检查是否已经安装成功,看这个目录下是否有插件文件

1. ls /opt/hadoop/nagios/libexec/

5. 配置nrpe

1. vim /opt/hadoop/nagios/etc/nrpe.cfg
2. 找到”allowed_hosts=127.0.0.1”改成“allowed_hosts=127.0.0.1,10.130.2.72”,后边的IP是nagios服务端IP
3. 找到” dont_blame_nrpe=0”改成“dont_blame_nrpe=1”

6.一段nrpe启停脚本,放在/etc/init.d/nrpe里

1. #!/bin/bash
2. #
3. # chkconfig: 2345 55 25
4. # description: NRPE Daemon
5. #
6. # source function library
7. ./etc/rc.d/init.d/functions
8. RETVAL=0
9. prog='nrpe'
10. NRPE_CFG='/opt/hadoop/nagios/etc/nrpe.cfg'
11. NRPE_PRG='/opt/hadoop/nagios/bin/nrpe'
12. NRPE_OPT='-d'
13. PID_FILE='/var/run/nrpe.pid'
14. start()
15. {
16. echo -n $"Starting $prog: "
17. [-f $PID_FILE ]&& rm -f $PID_FILE
18. $NRPE_PRG -c $NRPE_CFG $NRPE_OPT
19. pid=`ps aux | grep -v grep | grep $NRPE_PRG | awk '{print $2}'`
20. echo $pid > $PID_FILE
21. if ps aux | grep -v grep | grep -q $NRPE_PRG ;then
22. RETVAL=0
23. success
24. else
25. RETVAL=1
26. failure
27. fi
28. echo
29. }
30. stop()
31. {
32. echo -n $"Stopping $prog: "
33. ps --pid=`cat $PID_FILE`&>/dev/null
34. if[ $?-eq 0];then
35. kill -9`cat $PID_FILE`
36. RETVAL=0
37. fi
38. success
39. echo
40. RETVAL=0
41. }
42. case"$1"in
43. start)
44. start
45. ;;
46. stop)
47. stop
48. ;;
49. restart)
50. stop
51. start
52. ;;
53. status)
54. status -p $PID_FILE $prog
55. RETVAL=$?
56. ;;
57. *)
58. echo $"Usage: $0 {start|stop|restart|status}"
59. RETVAL=1
60. esac
61. exit $RETVAL

6. 启动nrpe

1. /etc/init.d/nrpe start

C.Nagios服务端添加被监控机
1.配置监控机目录

1. mkdir /opt/hadoop/nagios/etc/servers
2. vim /opt/hadoop/nagios/etc/nagios.cfg 追加cfg_dir=/opt/hadoop/nagios/etc/servers

2.添加配置的机器

1. vim /opt/hadoop/nagios/etc/servers/10.130.2.22.cfg
2. define host{
3. use linux-server
4. host_name 10.130.2.22
5. alias 10.130.2.22
6. address 10.130.2.22
7. }
8. define service{
9. use generic-service
10. host_name 10.130.2.22
11. service_description check_ping
12. check_command check_ping!100.0,20%!200.0,50%
13. max_check_attempts 5
14. normal_check_interval 1
15. }
16. define service{
17. use generic-service
18. host_name 10.130.2.22
19. service_description check_ssh
20. check_command check_ssh
21. max_check_attempts 5
22. normal_check_interval 1
23. }

3.reload nagios服务端使配置生效

1. service nagios reload

重新加载nagios后就可以在nagios的界面上看到新的被监控的机器了
4.添加使用nrpe的监控

1. 在/opt/hadoop/nagios/etc/objects/commands.cfg里增加如下行
2. define command{
3. command_name check_nrpe
4. command_line $USER1$/check_nrpe -H $HOSTADDRESS$ -c $ARG1$
5. }

在服务器监控配置文件中加入如下行，确保被监控机的nrpe服务是开的

1. define service{
2. use generic-service
3. host_name 10.130.2.22
4. service_description check_load
5. check_command check_nrpe!check_load
6. max_check_attempts 5
7. normal_check_interval 1
8. }

重新加载nagios使配置生效。

1. service nagios reload

5.自定义监控脚本
编写脚本check_diskmount.sh

1. vim /opt/hadoop/nagios/libexec/check_diskmount.sh
2. #!/bin/bash
3. num=`cat /proc/mounts | grep '/disk' | wc -l`
4. if[ $num -eq 12];then
5. echo "OK - mount disk is $num"
6. exit 0
7. else
8. echo "Critical - mount disk is $num"
9. exit 1
10. fi

加上可执行权限

1. chmod +x /opt/hadoop/nagios/libexec/check_diskmount.sh

在被监控机的nrpe里加入自定义脚本路径

1. vim /opt/hadoop/nagios/etc/nrpe.cfg
2. command[check_diskmount]=/opt/hadoop/nagios/libexec/check_diskmount.sh

重启nrpe

1. /etc/init.d/nrpe restart

在nagios服务端加入配置

1. vim /opt/hadoop/nagios/etc/servers/10.130.2.22.cfg
2. define service{
3. use generic-service
4. host_name s9xplan2.isv.cm6
5. service_description check_diskmount
6. check_command check_nrpe!check_diskmount
7. max_check_attempts 3
8. normal_check_interval 1
9. }

重新加载nagios，使得配置生效

1. service nagios reload

摘自：http://www.opstool.com/article/236