Project

General

Profile

Mod access » History » Revision 28

Revision 27 (gstrauss, 2021-03-18 17:13) → Revision 28/29 (gstrauss, 2021-03-18 18:13)

h1. Module mod_access 

 {{>toc}} 

 h2. Description 

 The access module is used to deny access to files. 

 h2. Options 

 *url.access-allow* 
 Allow access only to files with any of given trailing path names. (since 1.4.40) 
 Default value:empty 

 *url.access-deny* 
 Denies access to all files with any of given trailing path names.  
 Default value:empty 

 h2. Usage examples 

 *url.access-allow* 
 <pre> 
     url.access-allow = ( ".jpg", ".gif") 
 </pre> 

 You might want to deny access to all files ending with a tilde (~) or .inc because of:     

 # Text editors often use a trailing tilde for backup files. 
 # And the .inc extension is often used for include files with code. 

 *url.access-deny* 
 <pre> 
     url.access-deny = ( "~", ".inc") 
 </pre> 

 *Directory deny access* 
 An empty string in @url.access-deny@ matches all requests 
 <pre> 
     $HTTP["url"] =~ "^/libraries" { 
         url.access-deny = ("") 
     } 
 </pre> 

 Note: Creating a very, very large list of conditions is inefficient.    If creating conditions from a list, consider batching them with regex alternations into many fewer conditions, each with a large regex.    See #3074 

 *Deny bots* if User-Agent matches robots.txt 
 Sample one-liner to generate config: 
 @curl -s https://raw.githubusercontent.com/mitchellkrogza/nginx-ultimate-bad-bot-blocker/master/robots.txt/robots.txt | perl -e 'while (<>) { /User-agent:\s*(.+)/ && push @x, quotemeta($1); } print "\$HTTP[\"user-agent\"] =~ \"\\b(?i:", join("|",@x), ")\\b\" { url.access-deny = (\"\") }\n"'@ 

 *Deny lots* 
 https://github.com/mitchellkrogza 
 https://github.com/mitchellkrogza/nginx-ultimate-bad-bot-blocker 
 The following converts some of the rejections from @globalblacklist.conf@ into lighttpd syntax.    You should review the output file @rejections@ before adding @include "/path/to/rejections"@ in @lighttpd.conf@.    The IPs might be better off in firewall rules.    Yes, this could be cleaned up rather than a huge perl one-liner, but this is a quick response to #3074 that others might find useful, so here is a starting point. 

 @curl -s https://raw.githubusercontent.com/mitchellkrogza/nginx-ultimate-bad-bot-blocker/master/conf.d/globalblacklist.conf | perl -e '$/=undef; $list=<>; while ($list =~ /^# START ([^#]+) ### DO NOT EDIT THIS LINE AT ALL ###$(.*?)# END \1 ### DO NOT EDIT THIS LINE AT ALL ###/gms) { $x{$1} = $2; } push @bad_bots, map { /\(\?:\\b\)(.+?)\(\?:\\b\)/ && $1; } split("\n",$x{"BAD BOTS"}); push @bad_referrers, map { /\(\?:\\b\)(.+?)\(\?:\\b\)/ && $1; } split("\n",$x{"BAD REFERRERS"}); push @bad_ips, map { /\s*(\S+)\s*1;/ && $1; } split("\n",$x{"KNOWN BAD IP ADDRESSES"}); undef @strs; $str=""; foreach (@bad_bots) { $str .= $_."|"; if (length($str) > 32000) { substr($str,-1,1,""); push @strs, $str; $str=""; } } substr($str,-1,1,""); push @strs, $str; print "\$HTTP[\"user-agent\"] =~ \"\\b(?i:$_)\\b\" { url.access-deny = ( \"\" ) }\n\n" foreach (@strs); undef @strs; $str=""; foreach (@bad_referrers) { $str .= $_."|"; if (length($str) > 32000) { substr($str,-1,1,""); push @strs, $str; $str=""; } } substr($str,-1,1,""); push @strs, $str; print "\$HTTP[\"referer\"] =~ \"(?:\\.|^)(?i:$_)\$\" { url.access-deny = ( \"\" ) }\n\n" foreach (@strs); undef @strs; $str=""; foreach (@bad_ips) { $str .= quotemeta($_)."|"; if (length($str) > 32000) { substr($str,-1,1,""); push @strs, $str; $str=""; } } substr($str,-1,1,""); push @strs, $str; print "\$HTTP[\"remote-ip\"] =~ \"^(?:$_)\$\" { url.access-deny = ( \"\" ) }\n\n" foreach (@strs);' > rejections@