3 编写脚本来监控nginx服务
3.1 调试详细经过
[root@lb-net-2 run]# find / -name nginx.pid
/usr/local/nginx/logs/nginx.pid
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -n /usr/local/nginx/logs/nginx.pid -s nginx_status -o /tmp/ -w 1500 -c 2000
expr: 参数数目错误
expr: 语法错误
(standard_in) 1: syntax error
/usr/lib/nagios/plugins/check_nginxstatus: line 258: [: : integer expression expected
/usr/lib/nagios/plugins/check_nginxstatus: line 262: [: : integer expression expected
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
去查看262行，将逻辑运算符 "-a" 改成 "&&"
[root@lb-net-2 run]# vim /usr/lib/nagios/plugins/check_nginxstatus
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -n /usr/local/nginx/logs/nginx.pid -s nginx_status -o /tmp/ -w 1500 -c 2000
expr: 参数数目错误
expr: 语法错误
(standard_in) 1: syntax error
/usr/lib/nagios/plugins/check_nginxstatus: line 258: [: missing `]'
/usr/lib/nagios/plugins/check_nginxstatus: line 262: [: : integer expression expected
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
[root@lb-net-2 run]#
看到已经OK了，再修改文件。
[root@lb-net-2 run]# vim /usr/lib/nagios/plugins/check_nginxstatus
[root@lb-net-2 run]#
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -n /usr/local/nginx/logs/nginx.pid -s nginx_status -o /tmp/ -w 1500 -c 2000
expr: 参数数目错误
expr: 语法错误
(standard_in) 1: syntax error
/usr/lib/nagios/plugins/check_nginxstatus: line 258: [: missing `]'
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
[root@lb-net-2 run]#
将[]改成使用"[[]]"， 即可！
[root@lb-net-2 run]# vim /usr/lib/nagios/plugins/check_nginxstatus
[root@lb-net-2 run]#
[root@lb-net-2 run]#
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -n /usr/local/nginx/logs/nginx.pid -s nginx_status -o /tmp/ -w 1500 -c 2000
expr: 参数数目错误
expr: 语法错误
(standard_in) 1: syntax error
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
[root@lb-net-2 run]#
注释掉#reqpcon=`echo "scale=2; $reqpsec / $conpsec" | bc -l`之后，就不会报(standard_in) 1: syntax error错误，如下所示：
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -s nginx_status -n nginx.pid -w 15000 -c 20000
expr: 参数数目错误
expr: 语法错误
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
[root@lb-net-2 run]#
注释掉# reqpsec=`expr $tmp2_reqpsec - $tmp1_reqpsec` 就不会再报 expr: 参数数目错误，如下所示：
报错：
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -s nginx_status -n nginx.pid -w 15000 -c 20000
expr: 语法错误
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
再次注释掉 #reqpcon=`echo "scale=2; $reqpsec / $conpsec" | bc -l` 后，运行不会报expr: 语法错误，如下所示：
[root@lb-net-2 run]# /usr/lib/nagios/plugins/check_nginxstatus -H localhost -P 80 -p /usr/local/nginx/logs/ -s nginx_status -n nginx.pid -w 15000 -c 20000
OK - nginx is running. requests per second, connections per second ( requests per connection) | 'reqpsec'= 'conpsec'= 'conpreq'= ]
[root@lb-net-2 run]#
看到这里发现'reqpsec'= 'conpsec'= 'conpreq'=都没有值，但是nginx又是在启动运行着，问题出在哪里？经过排查，原来是nginx_status服务没有启动，需要在/usr/local/nginx/conf/nginx.conf配置文件里面添加如下配置：
# 添加pid参数
pid logs/nginx.pid;
#charset koi8-r;
access_log logs/host.access.log main;
location /nginx_status {
stub_status on;
access_log off;
deny all;
}
然后重新加载nginx，看到新的nginx-status文件是生成了，但是文件内容为空，如下所示：
[root@lb-net-2 logs]# ll /tmp/nginx*
-rw-r--r--. 1 root root 0 7月 3 15:06 /tmp/nginx-status.1
[root@lb-net-2 logs]#
去查看ngins后台日志
[root@lb-net-2 logs]# cd /usr/local/nginx/
[root@lb-net-2 logs]# tail -n 300 error.log
……
2014/07/03 15:05:47 [error] 4285#0: *1851293 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
2014/07/03 15:05:48 [error] 4285#0: *1851294 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
2014/07/03 15:06:12 [error] 4282#0: *1851362 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
2014/07/03 15:06:13 [error] 4282#0: *1851363 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
2014/07/03 15:06:55 [error] 4285#0: *1851509 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
2014/07/03 15:06:56 [error] 4285#0: *1851519 access forbidden by rule, client: 127.0.0.1, server: localhost, request: "GET /nginx_status HTTP/1.0", host: "localhost"
查看nginx编译参数
[root@lb-net-2 logs]# /usr/local/nginx/sbin/nginx -V
nginx version: nginx/1.4.2
built by gcc 4.4.7 20120313 (Red Hat 4.4.7-3) (GCC)
configure arguments: --prefix=/usr/local/nginx --with-http_stub_status_module --with-http_realip_module
证明确实是加载了stub_status插件，之后去修改配置文件，注释掉deny all;重新加载nginx。
[root@lb-net-2 logs]# vim /usr/local/nginx/conf/nginx.conf
#deny all;
[root@lb-net-2 logs]# service nginx reload
reload nginx
[root@lb-net-2 logs]#
[root@lb-net-2 logs]# ll /tmp/nginx*
ls: 无法访问/tmp/nginx*: 没有那个文件或目录
[root@lb-net-2 logs]#
还是没有看到/tmp/nginx-status.1状态文件生成，因为nagios下监控nginx的脚本是从nginx-status.1获取数据，如果没有这个文件，没有办法获取数据。
继续google，”nginx stub_status没有生成nginx-status.1”文件，看到有人说只要配置好了这个状态文件有没有无所谓，我就试着直接运行脚本看看能否生效。
[root@lb-net-2 logs]# ll /tmp/nginx*
ls: 无法访问/tmp/nginx*: 没有那个文件或目录
[root@lb-net-2 logs]# /root/check_nginx2.sh -H localhost -P 80 -p /usr/local/nginx/logs/ -n nginx.pid -s nginx_status -w 15000 -c 20000
OK - nginx is running. 1 requests per second, 2 connections per second (.50 requests per connection) | 'reqpsec'=1 'conpsec'=2 'conpreq'=.50 ]
[root@lb-net-2 logs]#
看到'reqpsec'=1 'conpsec'=2 'conpreq'=.50里面有数据了，再去check下文件有没有生成，如下所示：
[root@lb-net-2 logs]# ll /tmp/nginx*
ls: 无法访问/tmp/nginx*: 没有那个文件或目录
[root@lb-net-2 logs]#
还是没有文件生成，但是check已经有数据了，证明不一定要拘泥于是否在/tmp/目录下是否有nginx-status.1文件。通过脚本分析如下：
[root@lb-net-2 logs]# vim /usr/lib/nagios/plugins/check_nginxstatus
180 get_status() {
181 if [ "$secure" = 1 ]
182 then
183 wget_opts="-O- -q -t 3 -T 3 --no-check-certificate"
184 out1=`wget ${wget_opts} ${hostname}:${port}/${status_page}`
185 sleep 1
186 out2=`wget ${wget_opts} ${hostname}:${port}/${status_page}`
187 else
188 wget_opts="-O- -q -t 3 -T 3"
189 out1=`wget ${wget_opts} ${hostname}:${port}/${status_page}`
190 sleep 1
191 out2=`wget ${wget_opts} ${hostname}:${port}/${status_page}`
192 fi
193
194 if [ -z "$out1" -o -z "$out2" ]
195 then
196 echo "UNKNOWN - Local copy/copies of $status_page is empty."
197 exit $ST_UK
198 fi
199 }
是通过访问`wget -O- -q -t 3 -T 3 --no-check-certificate :80/nginx_status`这个链接来获取status的数据记录的，而不是去加载/tmp/nginx-status.1文件来获取数据的。直接访问 :80/nginx_status 地址就能获取nginx运行数据，如下图所示：
在nagios服务器上check下，报错：
[root@cache-2 ~]# /usr/local/nagios/libexec/check_nrpe -H10.xx.xx.xx -c check_nginx_status
UNKNOWN - Local copy/copies of nginx_status is empty.
[root@cache-2 ~]#
检查监控脚本，搜索 ‘Local copy/copies of nginx_status is empty.’在第197行，有如下代码：
195 if [ -z "$out1" -o -z "$out2" ]
196 then
197 echo "UNKNOWN - Local copy/copies of $status_page is empty."
198 exit $ST_UK
199 fi
看出是由于if [ -z "$out1" -o -z "$out2" ]这个判断生效，导致监控脚本运行到这里就exit了。继续调试，发现用nagios服务器调用脚本的时候，执行到以下第190行到第192行
out1=`/usr/bin/wget ${wget_opts} ${hostname}:${port}/${status_page}`
sleep 1
out2=`/usr/bin/wget ${wget_opts} ${hostname}:${port}/${status_page}`
的时候，out1为空，out2也为空，所以在后面的if [ -z "$out1" -o -z "$out2" ]判断通过报出信息为：UNKNOWN - Local copy/copies of $status_page is empty. 然后直接exit。
说明：由于nginx是要调用wget命令来获取nginx_status状态的，而wget命令是只能以root用户来运行的, 所以需要将nagios用户设置成可以无需密码直接su成root，这样就能以nagios用户运行命令sudo /usr/lib/nagios/plugins/check_nginxstatus 。在CentOS系统中，无法直接调用sudo命令，需要修改/etc/sudoers, 找到 #Defaults requiretty 并取消注释，另外新增一行。表示nagios用户不需要登陆终端就可以调用命令，如下所示：
Defaults requiretty
Defaults:nagios !requiretty
#添加nagios 请求sudo，允许特定指令时（可跟参数），不需要密码（如）。
nagios ALL=(ALL) NOPASSWD: ALL
修改完后，再check，数据出来了：
[root@cache-2 ~]# /usr/local/nagios/libexec/check_nrpe -H10.xx.xx.xx -c check_nginx_status
OK - nginx is running. 1 requests per second, 1 connections per second (1.00 requests per connection) | 'reqpsec'=1 'conpsec'=1 'conpreq'=1.00 ]
[root@cache-2 ~]#
搭建基于Linux6.3+Nginx1.2+PHP5+MySQL5.5的Web服务器全过程
CentOS 6.3下配置Nginx加载ngx_pagespeed模块
CentOS 6.4安装配置Nginx+Pcre+php-fpm