nginx perl fastcgi

首先：有三种方式（三个不同的脚本）搭建fastcgi环境，第二种，在运行时，nginx的error里会报错，但是不影响使用FastCGI sent in stderr: "Bad file descriptor at /EBS/fcgi.pl line 125",第二种脚本的优点是可以起多个进程（process），第三个脚本的优点是配置简单。

第二种方法的官方文档：

方法一

参考：

1、安装FCGI模块：

到网站上下载FCGI模块：

wget CPAN/authors/id/S/SK/SKIMO/FCGI-0.67.tar.gz

 tar -zxvf FCGI-0.67.tar.gz. 

 cd FCGI-0.67 

 perl Makefile.PL

 make 

 make install

2、安装 IO 和 IO::ALL模块

下载：

wget CPAN/authors/id/G/GB/GBARR/IO-1.25.tar.gz

CPAN/authors/id/I/IN/INGY/IO-All-0.39.tar.gz

安装：

tar -xvzf IO-1.25.tar.gz

cd IO-1.25

 perl Makefile.PL

 make 

 make install

tar -xvzf IO-All-0.39.tar.gz

cd IO-ALL-0.39

 perl Makefile.PL

 make 

 make install

3、如果需要，安装perl-Getopt和 perl-Socket

4、perl cgi脚本

文件:	nginx-fcgi.rar
大小:	2KB
下载:	下载

或从下面的地址下载

5、启动cgi脚本 （nobody为nginx的运行用户）

vi start_perl_cgi.sh

#!/bin/bash

#set -x

if [[ $# != 1 ]];then

echo "usage $0 start|stop|restart"

exit 1

fi

dir=/usr/local/nginx

stop ()

{

#pkill  -f  $dir/perl-fcgi.pl

kill $(cat $dir/logs/perl-fcgi.pid)

rm $dir/logs/perl-fcgi.pid 2>/dev/null

rm $dir/logs/perl-fcgi.sock 2>/dev/null

echo "stop perl-fcgi done"

}

start ()

{

rm $dir/now_start_perl_fcgi.sh 2>/dev/null

chown nobody.root $dir/logs

echo "$dir/perl-fcgi.pl -l $dir/logs/perl-fcgi.log -pid $dir/logs/perl-fcgi.pid -S $dir/logs/perl-fcgi.sock" >>$dir/now_start_perl_fcgi.sh

chown nobody.nobody $dir/now_start_perl_fcgi.sh

chmod u+x $dir/now_start_perl_fcgi.sh

sudo -u nobody $dir/now_start_perl_fcgi.sh

echo "start perl-fcgi done"

}

case $1 in

stop)

stop

;;

start)

start

;;

restart)

stop

start

;;

esac

启动

./start_perl_cgi.sh start

6、为Nginx添加FCGI支持(以安装nagios为例）

        location ~* .*\.cgi$

                {

                     rewrite ^/nagios/cgi-bin/(.*)\.cgi /$1.cgi break;

                      include perl_fcgi.conf;

        }

重写请求cgi的url为cgi的文件名，组成下面的CRIPT_FILENAMEvi perl_fcgi.conf

gzip off; #gzip makes scripts feel slower since they have to complete before getting gzipped
fastcgi_pass unix:/usr/local/nginx/logs/perl-fcgi.sock;
fastcgi_index index.cgi;

fastcgi_param SCRIPT_FILENAME /usr/local/nagios/sbin$fastcgi_script_name;

fastcgi_param QUERY_STRING $query_string;
fastcgi_param REQUEST_METHOD $request_method;
fastcgi_param CONTENT_TYPE $content_type;
fastcgi_param CONTENT_LENGTH $content_length;
fastcgi_param GATEWAY_INTERFACE CGI/1.1;
fastcgi_param SERVER_SOFTWARE nginx;
fastcgi_param SCRIPT_NAME $fastcgi_script_name;
fastcgi_param REQUEST_URI $request_uri;
fastcgi_param DOCUMENT_URI $document_uri;
fastcgi_param DOCUMENT_ROOT $document_root;
fastcgi_param SERVER_PROTOCOL $server_protocol;
fastcgi_param REMOTE_ADDR $remote_addr;
fastcgi_param REMOTE_PORT $remote_port;
fastcgi_param SERVER_ADDR $server_addr;
fastcgi_param SERVER_PORT $server_port;
fastcgi_param SERVER_NAME $server_name

其中：$fastcgi_script_name就是访问的URL中的URI

测试页

#!/usr/bin/perl -w
print "Content-type: text/plain\r\n\r\n";#发送头信息
print "test";

记得一定要给nginx发送头信息 不然会报504错误的。

另：

cgi-bin存在于html目录下时（不是为了部署nagios）

nginx.conf更改下面的部分:

location ~* .*\.cgi$
                {
                        include perl_fcgi.conf;
                }

perl_fcgi.conf更改下面的部分:

gzip off; #gzip makes scripts feel slower since they have to complete before getting gzipped
fastcgi_pass unix:/usr/local/nginx/logs/perl-fcgi.sock;
fastcgi_index index.cgi;

fastcgi_param SCRIPT_FILENAME /usr/local/nginx/html$fastcgi_script_name;

方法二

1、安装FCGI模块：

到网站上下载FCGI模块：

wget CPAN/authors/id/S/SK/SKIMO/FCGI-0.67.tar.gz

 tar -zxvf FCGI-0.67.tar.gz. 

 cd FCGI-0.67 

 perl Makefile.PL

 make 

 make install

也可以这样安装：perl -MCPAN -e 'install FCGI'

（我在使用这种方法时，没有找到FCGI模块）

2、安装FCGI-ProcManager模块：

 wget CPAN/authors/id/G/GB/GBJK/FCGI-ProcManager-0.19.tar.gz

 tar -xvzf FCGI-ProcManager-0.19.tar.gz

 perl Makefile.PL

 make

 make install

3、Fastcgi 启动脚本

fcgi.pl

1. #!/usr/bin/perl -w
2. use FCGI;
3. use Socket;
4. use FCGI::ProcManager;
5. sub shutdown { FCGI::CloseSocket($socket); exit; }
6. sub restart { FCGI::CloseSocket($socket); &main; }
7. use sigtrap 'handler', \&shutdown, 'normal-signals';
8. use sigtrap 'handler', \&restart, 'HUP';
9. require 'syscall.ph';
10. use POSIX qw(setsid);
11.  
12. #&daemonize; we don't daemonize when running under runsv
13. #this keeps the program alive or something after exec'ing perl scripts
14. END() { }
15. BEGIN() { }
16. {
17. no warnings;
18. *CORE::GLOBAL::exit = sub { die "fakeexit\nrc=" . shift() . "\n"; };
19. };
20. eval q{exit};
21. if ($@) {
22. exit unless $@ =~ /^fakeexit/;
23. }
24. &main;
25.  
26. sub daemonize() {
27. chdir '/' or die "Can't chdir to /: $!";
28. defined( my $pid = fork ) or die "Can't fork: $!";
29. exit if $pid;
30. setsid() or die "Can't start a new session: $!";
31. umask 0;
32. }
33.  
34. sub main {
35. #如果使用 IP sockets
36. #$socket = FCGI::OpenSocket( "127.0.0.1:8999", 10 );
37. #如果使用 UNIX sockets
38. #$socket = FCGI::OpenSocket( "/var/run/perl_cgi-dispatch.sock", 10 );
39.  
40. #foreach $item (keys %ENV) { delete $ENV{$item}; }
41. #设置fastcgi进程数，默认四个
42. my $n_processes = $ENV{FCGI_NPROCESSES} || 4;
43. $proc_manager = FCGI::ProcManager->new( {n_processes => $n_processes} );
44. #使用unix socket
45. $socket = FCGI::OpenSocket( "$ENV{FCGI_SOCKET_PATH}", 10 );
46. #设置Socket权限
47. chmod 0777, $ENV{FCGI_SOCKET_PATH};
48.  
49. ; #use UNIX sockets - user running this script must have w access to the 'nginx' folder!!
50. $request =
51. FCGI::Request( \*STDIN, \*STDOUT, \*STDERR, \%req_params, $socket,
52. &FCGI::FAIL_ACCEPT_ON_INTR );
53. $proc_manager->pm_manage();
54. if ($request) { request_loop() }
55. FCGI::CloseSocket($socket);
56. }
57.  
58. sub request_loop {
59. while ( $request->Accept() >= 0 ) {
60. $proc_manager->pm_pre_dispatch();
61.  
62. #processing any STDIN input from WebServer (for CGI-POST actions)
63. $stdin_passthrough = '';
64. { no warnings; $req_len = 0 + $req_params{'CONTENT_LENGTH'}; };
65. if ( ( $req_params{'REQUEST_METHOD'} eq 'POST' ) && ( $req_len != 0 ) )
66. {
67. my $bytes_read = 0;
68. while ( $bytes_read < $req_len ) {
69. my $data = '';
70. my $bytes = read( STDIN, $data, ( $req_len - $bytes_read ) );
71. last if ( $bytes == 0 || !defined($bytes) );
72. $stdin_passthrough .= $data;
73. $bytes_read += $bytes;
74. }
75. }
76.  
77. #running the cgi app
78. if (
79. ( -x $req_params{SCRIPT_FILENAME} ) && #can I execute this?
80. ( -s $req_params{SCRIPT_FILENAME} ) && #Is this file empty?
81. ( -r $req_params{SCRIPT_FILENAME} ) #can I read this file?
82. )
83. {
84. pipe( CHILD_RD, PARENT_WR );
85. pipe( PARENT_ERR, CHILD_ERR );
86. my $pid = open( CHILD_O, "-|" );
87. unless ( defined($pid) ) {
88. print("Content-type: text/plain\r\n\r\n");
89. print
90. "Error: CGI app returned no output - Executing $req_params{SCRIPT_FILENAME} failed !\n";
91. next;
92. }
93. $oldfh = select(PARENT_ERR);
94. $| = 1;
95. select(CHILD_O);
96. $| = 1;
97. select($oldfh);
98. if ( $pid > 0 ) {
99. close(CHILD_RD);
100. close(CHILD_ERR);
101. print PARENT_WR $stdin_passthrough;
102. close(PARENT_WR);
103. $rin = $rout = $ein = $eout = '';
104. vec( $rin, fileno(CHILD_O), 1 ) = 1;
105. vec( $rin, fileno(PARENT_ERR), 1 ) = 1;
106. $ein = $rin;
107. $nfound = 0;
108.  
109. while ( $nfound =
110. select( $rout = $rin, undef, $ein = $eout, 10 ) )
111. {
112. die "$!" unless $nfound != -1;
113. $r1 = vec( $rout, fileno(PARENT_ERR), 1 ) == 1;
114. $r2 = vec( $rout, fileno(CHILD_O), 1 ) == 1;
115. $e1 = vec( $eout, fileno(PARENT_ERR), 1 ) == 1;
116. $e2 = vec( $eout, fileno(CHILD_O), 1 ) == 1;
117.  
118. if ($r1) {
119. while ( $bytes = read( PARENT_ERR, $errbytes, 4096 ) ) {
120. print STDERR $errbytes;
121. }
122.  
123. if ($!) {
124. $err = $!;
125. die $!;
126. vec( $rin, fileno(PARENT_ERR), 1 ) = 0
127. unless ( $err == EINTR or $err == EAGAIN );
128. }
129. }
130. if ($r2) {
131. while ( $bytes = read( CHILD_O, $s, 4096 ) ) {
132. print $s;
133. }
134. if ( !defined($bytes) ) {
135. $err = $!;
136. die $!;
137. vec( $rin, fileno(CHILD_O), 1 ) = 0
138. unless ( $err == EINTR or $err == EAGAIN );
139. }
140. }
141. last if ( $e1 || $e2 );
142. }
143. close CHILD_RD;
144. close PARENT_ERR;
145. waitpid( $pid, 0 );
146. } else {
147. foreach $key ( keys %req_params ) {
148. $ENV{$key} = $req_params{$key};      
149. }
150.  
151. # cd to the script's local directory
152. if ( $req_params{SCRIPT_FILENAME} =~ /^(.*)\/[^\/]+$/ ) {
153. chdir $1;
154. }
155. close(PARENT_WR);
156.  
157. #close(PARENT_ERR);
158. close(STDIN);
159. close(STDERR);
160.  
161. #fcntl(CHILD_RD, F_DUPFD, 0);
162. syscall( &SYS_dup2, fileno(CHILD_RD), 0 );
163. syscall( &SYS_dup2, fileno(CHILD_ERR), 2 );
164.  
165. #open(STDIN, "<&CHILD_RD");
166. exec( $req_params{SCRIPT_FILENAME} );
167. die("exec failed");
168. }
169. } else {
170. print("Content-type: text/plain\r\n\r\n");
171. print
172. "Error: No such CGI app - $req_params{SCRIPT_FILENAME} may not exist or is not executable by this process.\n";
173. }
174. }
175. }

使用下面的脚本启动

vi start_fcgi.sh

1. export FCGI_SOCKET_PATH="/tmp/perl_fcgi.socket"
2. export FCGI_NPROCESSES=4
3. ./fcgi.pl &

4、为Nginx添加FCGI支持(以安装nagios为例）

location ~* .*\.cgi$
                {
                      rewrite ^/nagios/cgi-bin/(.*)\.cgi /$1.cgi break;
                        include perl_fcgi.conf;
                }

将请求cgi的url由/nagios/cgi-bin/*.cgi 重写为/*.cgi，这样，即下面的fastcgi_param中的$fastcgi_script_name，这样，可以正确的拼出cgi文件的路径。vi perl_fcgi.conf

gzip off; #gzip makes scripts feel slower since they have to complete before getting gzipped
fastcgi_pass unix:/tmp/perl_fcgi.socket;
fastcgi_index index.cgi;

fastcgi_param SCRIPT_FILENAME /usr/local/nagios/sbin$fastcgi_script_name;

fastcgi_param QUERY_STRING $query_string;
fastcgi_param REQUEST_METHOD $request_method;
fastcgi_param CONTENT_TYPE $content_type;
fastcgi_param CONTENT_LENGTH $content_length;
fastcgi_param GATEWAY_INTERFACE CGI/1.1;
fastcgi_param SERVER_SOFTWARE nginx;
fastcgi_param SCRIPT_NAME $fastcgi_script_name;
fastcgi_param REQUEST_URI $request_uri;
fastcgi_param DOCUMENT_URI $document_uri;
fastcgi_param DOCUMENT_ROOT $document_root;
fastcgi_param SERVER_PROTOCOL $server_protocol;
fastcgi_param REMOTE_ADDR $remote_addr;
fastcgi_param REMOTE_PORT $remote_port;
fastcgi_param SERVER_ADDR $server_addr;
fastcgi_param SERVER_PORT $server_port;
fastcgi_param SERVER_NAME $server_name

方法三

fastcgi-wrapper.pl脚本

下载地址：

或从下面下载，记得修改扩展名。

 fastcgi-wrapper.txt   

添加可执行权限，启动：

sudo -u nobody ./fastcgi-wrapper.pl >>log.cgi 2>&1 &

log.cgi中记录错误信息。

关于：bind/listen: Permission denied

检查sock文件所在目录对于nobody是否有写的权限

添加nginx配置：

location ~ \.cgi$ {
fastcgi_pass    unix:/var/run/nginx/perl_cgi-dispatch.sock;
fastcgi_index   index.cgi;
fastcgi_param  SCRIPT_FILENAME  $document_root$fastcgi_script_name;
include         fastcgi_params;
}

测试页面：

#!/usr/bin/perl -w
print "Content-type: text/plain\r\n\r\n";#发送头信息
print "test";

记得一定要给nginx发送头信息 不然会报504错误的。