Hi Erik,
Am 19.10.2008 3:13 Uhr, Eric Shubert schrieb:
> Felix Buenemann wrote:
>> Hi Erik,
>>
>> Am 18.10.2008 20:39 Uhr, Eric Shubert schrieb:
>>> Sergio Minini wrote:
>>>> Erik, try:
>>>> # cat /var/log/maillog | ./spamdyke_stats.pl
>> [...]
>>>> % Valid: 29.11%
>>>> % Spam : 69.42%
>>>> % Error: 1.47%
>>>>
>>>> Good luck/
>>>> Sergio
>>> Thanks, but that doesn't work for me. My spamdyke log messages are going
>>> to the qmail smtp log (log-target=0|stderr). Perhaps that's where the
>>> problem lies, as the formatting would be slightly different.
>>>
>> Yes, the script currently expects syslog syntax. If you can provide a
>> few sample lines from the log and I can modify the script to work with it.
>>
>> -- Felix
>
> Thanks, Felix. Here's a sample rejection (each line begins with @40000):
> @4000000048fa5df51149c60c tcpserver: status: 1/100
> @4000000048fa5df51149d5ac tcpserver: pid 22865 from 209.133.101.250
> @4000000048fa5df51149e164 tcpserver: ok 22865
> doris.shubes:192.168.171.11:25 :209.133.101.250::58673
> @4000000048fa5dfc34b1ebec DENIED_SENDER_NO_MX from:
> [EMAIL PROTECTED]
> to: [EMAIL PROTECTED] origin_ip: 209.1
> 33.101.250 origin_rdns: broadcaster.eonline.com auth: (unknown)
> @4000000048fa5dfd01593124 tcpserver: end 22865 status 0
> @4000000048fa5dfd015a7d2c tcpserver: status: 0/100
>
> In case you don't know, the first (@4000...) field is a date/time stamp,
> and can be converted to local time with the tai64nlocal program. I don't
> know if that'll be necessary or not.
>
> Thanks again. I'm eager to see the results.
Please test the attached version. You have to pass --nosyslog parameter
to the script. Also please check which of the $linematch variants
beginning at line 33 is the fastest (run time cat /your/log |
spamdyke-stats.pl --nosyslog to find out) and report it back to me.
Btw. the new version is about 25% faster than the last one, due to
optimizations of the regular expression matching, so updating might be
useful to others aswell.
-- Felix
#!/usr/bin/perl -w
# build 2008101905
use diagnostics;
use strict;
use Getopt::Long;
my $tldtop = 0;
my $detailed = 1;
my $syslog = 1;
GetOptions (
"tld=i" => \$tldtop,
"detail!" => \$detailed,
"syslog!" => \$syslog
) or exit 1;
# Usage: # cat /var/log/qmail/smtpd/current | ./this_file
my %status = (); # hash of status connections
my %origin = (); # hash of tld per status code
my %originsum = (); # hash of tld per status code sums
my %rblstat = (); # hash of DNSBL lists matched
my %rhsblstat = (); # hash of RHSBL lists matched
my %rdnsblstat = (); # hash of patterns in IP_IN_RDNS_BLACKLIST matched
my ($allow, $deny, $error, $allowpercentage, $errorpercentage, $spampercentage,
$sum, $rblsum, $rhsblsum, $rdnsblsum);
$allow = 0;
$deny = 0;
$error = 0;
my $linematch = "";
if($syslog) {
$linematch = 'spamdyke\[\d+\]:';
} else {
$linematch = '[EMAIL PROTECTED],24}';
#$linematch = '[EMAIL PROTECTED],24}';
#$linematch = '[EMAIL PROTECTED] ]+';
}
while(<>){
# if( m/(spamdyke\[\d+\]:|[EMAIL PROTECTED],24})
(ALLOWED|ERROR|TIMEOUT|((DENIED|FILTER)_[^ ]+))(.*)/ ){
if( m/$linematch (.*)/ ){
my $line = $1;
if( $line =~ m/^(ALLOWED|ERROR|TIMEOUT|((DENIED|FILTER)_[^
]+))/ ) {
my $sdstatus = $1;
if( $sdstatus =~ m/FILTER_RBL_MATCH/ ){
$line =~ m/rbl: (\S+)/;
$rblstat{$1}++;
$rblsum++;
}
elsif( $sdstatus =~ m/FILTER_RHSBL_MATCH/ ){
$line =~ m/rhsbl: (\S+)/;
$rhsblstat{$1}++;
$rhsblsum++;
}
elsif( $sdstatus =~ m/FILTER_IP_IN_RDNS_BLACKLIST/ ){
$line =~ m/keyword: (\S+)/;
$rdnsblstat{$1}++;
$rdnsblsum++;
}
next if $sdstatus =~ m/CHKUSER|FILTER_.*/;
$status{$sdstatus}++;
if($tldtop and $line =~ m/ origin_rdns: ([^ ]+)/) {
my $rdns = $1;
$originsum{$sdstatus}++;
if($rdns =~ m/^\(unknown\)$/){
#$origin{$sdstatus}{'unknown'}++;
next;
} elsif($rdns =~ m/\.(com|net)$/){
$origin{$sdstatus}{$1}++;
} elsif($rdns =~
m/\.([a-z]{2,2}\.[a-z]{2,2})$/){ # co.uk
$origin{$sdstatus}{$1}++;
} elsif($rdns =~ m/\.([a-z]{2,})$/){ # de, ru,
...
$origin{$sdstatus}{$1}++
} else {
#$origin{$sdstatus}{'unknown'}++;
next;
}
}
}
}
}
foreach my $stat (sort keys %status){
if( $stat =~ m/ALLOWED/ ){
$allow = $status{$stat};
}
elsif( $stat =~ m/TIMEOUT|ERROR/ ){
$error += $status{$stat};
}
else{
$deny += $status{$stat};
}
}
my $aed_sum = $allow+$error+$deny;
if($aed_sum > 0) {
$spampercentage = sprintf("%2.2f", ($deny/($aed_sum)*100) );
$errorpercentage = sprintf("%2.2f", ($error/($aed_sum)*100) );
$allowpercentage = sprintf("%2.2f", ($allow/($aed_sum)*100) );
} else {
$spampercentage = $errorpercentage = $allowpercentage =
sprintf("%2.2f", 0);
}
foreach my $key (sort { $status{$b} <=> $status{$a} || $a cmp $b; } keys
%status){
print "$status{$key}\t$key\n";
if($detailed and $key eq "DENIED_RBL_MATCH" ){
print "-- Breakdown --\n";
foreach my $key (sort { $rblstat{$b} <=> $rblstat{$a} || $a cmp
$b; } keys %rblstat){
printf "%2.2f%%\t$key\n", ($rblstat{$key}/$rblsum*100);
}
print "---------------\n";
}
elsif($detailed and $key eq "DENIED_RHSBL_MATCH" ){
print "-- Breakdown --\n";
foreach my $key (sort { $rhsblstat{$b} <=> $rblstat{$a} || $a
cmp $b; } keys %rhsblstat){
printf "%2.2f%%\t$key\n",
($rhsblstat{$key}/$rhsblsum*100);
}
print "---------------\n";
}
elsif($detailed and $key eq "DENIED_IP_IN_RDNS" ){
print "-- Breakdown --\n";
foreach my $key (sort { $rdnsblstat{$b} <=> $rdnsblstat{$a} ||
$a cmp $b; } keys %rdnsblstat){
printf "%2.2f%%\t$key\n",
($rdnsblstat{$key}/$rdnsblsum*100);
}
print "---------------\n";
}
if($tldtop && $origin{$key}) {
my $top = $tldtop;
print "-- Top $top TLD --\n";
my $tldsum = 0;
my $lastsum = 0;
my @tldgroup = ();
my %neworigin = ();
foreach my $tld (sort { $origin{$key}{$a} <=> $origin{$key}{$b}
} keys %{$origin{$key}}){
if(($origin{$key}{$tld}/$originsum{$key}*100) ==
$lastsum) {
#print "push tldgroup, $tld
($origin{$key}{$tld})\n";
push(@tldgroup, $tld);
} else {
if(scalar @tldgroup) {
$neworigin{join(', ', @tldgroup)} =
$lastsum;
#print "tldgroup=". join(', ',
@tldgroup) ." ($lastsum)\n";
@tldgroup = ();
}
#print "push tldgroup, $tld
($origin{$key}{$tld})\n";
push(@tldgroup, $tld);
}
$lastsum = $origin{$key}{$tld}/$originsum{$key}*100;
$tldsum += $origin{$key}{$tld};
}
if(scalar @tldgroup) {
$neworigin{join(', ', @tldgroup)} = $lastsum * length
@tldgroup;
#print "tldgroup=". join(', ', @tldgroup) ."
($lastsum)\n";
}
foreach my $tld (sort { $neworigin{$b} <=> $neworigin{$a} }
keys %neworigin){
printf "%2.2f%%\t$tld\n", $neworigin{$tld};
last unless --$top;
}
#printf "%2.2f%%\t(unknown/illegal)\n",
(($originsum{$key}-$tldsum)/$originsum{$key}*100) if $tldsum &&
($originsum{$key}-$tldsum);
print "------------\n";
}
}
$sum = ($deny + $error + $allow);
print "\n";
print "Allowed: $allow \n";
print "Denied : $deny \n";
print "Errors : $error \n";
print "Total : $sum \n";
print "% Valid: $allowpercentage% \n";
print "% Spam : $spampercentage% \n";
print "% Error: $errorpercentage% \n";
_______________________________________________
spamdyke-users mailing list
spamdyke-users@spamdyke.org
http://www.spamdyke.org/mailman/listinfo/spamdyke-users
