#! @PERL@ -w # vim:syntax=perl use strict; use DB_File; (my $program = $0) =~ s%.*/%%; my $stem = shift or die "$program: give dumpfilestem as arg\n"; my $max = 99999; # max number of objects my $maxemail = $max; # max number of email addresses of form # john.doe.#@example.com to generate. my $maxdomain = $max; my $maxmaildomain = $max; my $inaddr = '1.0.0.10.in-addr.arpa'; my $mailuser = 'john.doe.1'; my $maildomain = '1.mail.example.com'; my $domain = '1.example.com'; my $ip = '10.0.0.1'; my %h; # hash to store dump in my @dbs = ('inaddr', 'maildomain', 'email', 'ip', 'domain'); my $tmpdir = defined $ENV{'TMPDIR'} ? $ENV{'TMPDIR'} : '/tmp'; my $uniq = "$program.$$"; for my $name (@dbs) { my $dbfile = "$tmpdir/$uniq.$name"; -e $dbfile and ( unlink $dbfile or die "$program: cannot unlink $dbfile\n" ); tie %{ $h{$name} }, "DB_File", "$dbfile" or die "$program: cannot tie to $dbfile\n"; } sub up { my $type = shift; my $what = shift; if ($type eq 'inaddr') { my ($a, $b, $c, $r) = split /\./, $what; if ($a < 254) { $a++; } elsif ($b < 254) { $b++; $a = 1; } elsif ($c < 254) { $c++; $b = 0; $a = 1; } else { warn "$program: up: cannot up inaddr $what. out of ip addresses.\n"; return 0; } return "$a.$b.$c.10.in-addr.arpa"; } elsif ($type eq 'mailuser') { # john.doe.#@example.com my ($n, $t); ($t, $t, $n) = split /\./, $what; if ($n < $maxemail) { $n++; } else { warn "$program: up: cannot up mailuser '$what'. used $maxemail " . "emailaddresses\n"; return 0; } return 'john.doe.' . "$n"; } elsif ($type eq 'maildomain') { my ($n, $d) = split /\./, $what; if ($n < $maxmaildomain) { $n++; } else { warn "$program: up: cannot up maildomain '$what'. used " . "$maxmaildomain maildomains\n"; return 0; } return "$n.mail.example.com"; } elsif ($type eq 'ip') { my ($r, $a, $b, $c) = split /\./, $what; if ($c < 254) { $c++; } elsif ($b < 254) { $b++; $c = 1; } elsif ($a < 254) { $a++; $b = 0; $c = 1; } else { warn "$program: up: cannot up ip $what. out of ip addresses.\n"; return 0; } return "10.$a.$b.$c"; } elsif ($type eq 'domain') { # #.example.com my ($n, $r) = split /\./, $what; if ($n < $maxdomain) { $n++; } else { warn "$program: up: cannot up domain '$what'. used $maxdomain " . "domains\n"; } return "$n" . '.example.com'; } else { warn "$program: up: unsupported type '$type'\n"; return 0; } } while (<>) { chomp; my $r = ''; while (length) { if (/^(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}\.in-addr\.arpa)/i) { unless (defined $h{'inaddr'}{$1}) { $h{'inaddr'}{$1} = $inaddr; $inaddr = &up('inaddr', $inaddr) or die "$program: cannot up $inaddr\n"; } $r .= $h{'inaddr'}{$1}; $_ =~ s/^$1//; } elsif (/^([^ =\(\)<>@,;:\\"\[\]]+@([-a-z\d\.]+))/i) { # we match an emailaddress. we do not support user@[ipaddress]. # NOTE: we do not allow '=' in username part, since = is used # as a separator in sendmail logs. # we store lowercase only my $d = lc $2; my $e = lc $1; if (defined $h{'maildomain'}{$d}) { unless (defined $h{'email'}{$e}) { $h{'email'}{$e} = $mailuser . "@" . $h{'maildomain'}{$d}; $mailuser = &up('mailuser', $mailuser) or die "$program: cannot up mailuser '$mailuser'\n"; } } else { $h{'email'}{$e} = $mailuser . "@" . $maildomain; $mailuser = &up('mailuser', $mailuser) or die "$program: cannot up mailuser '$mailuser'\n"; $h{'maildomain'}{$d} = $maildomain; $maildomain = &up('maildomain', $maildomain) or die "$program: cannot up maildomain '$maildomain'\n"; } $r .= $h{'email'}{$e}; $_ =~ s/^\Q$1//; } elsif (/^(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3})([^\d]|$)/) { # we match an ip adress unless (defined $h{'ip'}{$1}) { $h{'ip'}{$1} = $ip; $ip = &up('ip', $ip) or die "$program: cannot up ip '$ip'\n"; } $r .= $h{'ip'}{$1}; $_ =~ s/^$1//; } elsif (/^\/(([-_a-z\d]+\.)+[-_a-z\d]+[a-z]{2})/i) { # this is a workaround to make sure file names are no # recognized as domain names. File names differ in that # that they start with a "/", domain names do not $r .= "/$1"; $_ =~ s/^\/$1//; } elsif (/^(([-_a-z\d]+\.)+[-_a-z\d]+[a-z]{2})/i) { # one or more non-dot's dot, some non-dot's, two letters # note: we accept _'s in domainnames, since this is a very # often made error. we are especially interested in these in bind # logs. my $d = lc $1; unless (defined $h{'domain'}{$d}) { $h{'domain'}{$d} = $domain; $domain = &up('domain', $domain) or die "$program: cannot up domain '$domain'\n"; } $r .= $h{'domain'}{$d}; $_ =~ s/^\Q$1//; } else { $r .= substr($_, 0, 1); $_ =~ s/^.//; } } print "$r\n" or die "$program: cannot print\n"; } # now build inverse dumps # tied hashes to store inverse-d %h{$name}'s in my %inv; for my $name (@dbs) { my $dbfile = "$stem.$name"; -e $dbfile and ( unlink $dbfile or die "$program: cannot unlink $dbfile\n" ); tie %{ $inv{$name} }, "DB_File", "$dbfile" or die "$program: cannot tie to $dbfile\n"; while (my ($k, $v) = each %{ $h{$name} }) { $inv{$name}{$v} = $k; } untie %{ $inv{$name} } or die "$program: cannot untie inverse hash key $name from file " . "$dbfile\n"; $dbfile = "$tmpdir/$uniq.$name"; untie %{ $h{$name} } or die "$program: cannot untie hash key $name from file $dbfile\n"; unlink $dbfile or die "$program: cannot unlink $dbfile\n"; } __END__ =pod =head1 NAME lr_anonymize - anonymize a logfile =head1 SYNOPSIS B I =head1 DESCRIPTION B is typically used when sending logs to a responder. See the section on "Sending Anonymized Log Files To A Responder" in the chapter on "Using A Responder" in the Lire User Manual for usage examples. B reads a file containing emailaddresses, ipnumbers, and hostnames (typically a logfile from an internet service) from stdin, and prints an "anonymized" version of this file to stdout. It dumps the information to "deanonymize" the file, using lr_deanonymize(1), to 5 Berkeley DB databases; the names of the files holding these databases is contructed by concatenating the strings 'inaddr', 'maildomain', 'email', 'ip' and 'domain' to I. I can be e.g. /tmp/dump, ../../var/dump or dump. The db files created silently overwrite stale ones, if found. The script builds temporary files. By default, these are created in /tmp. One can override this by setting the TMPDIR environment variable. =head1 CONNECTION WITH lr_deanonymize When running $ lr_anonymize dump < log > log.anon $ lr_deanonymize dump < log.anon > log.new then log and log.new have the same content (except for case, check it with B). =head1 NOTES We tried to optimize this script for memoryusage. This has the drawback the script will run for quite some time when anonymizing a big logfile. We've run the script on a 25 M sendmail logfile. Typical values in such a case are 2500 K for total amount of physical memory used, and 15m real, 8m user and 22s systime spent on a 64 MB system with a 300 MHz Pentium II processor. de_anonymizing this file took 11m real, 8m user and 17s system time. We store maildomains in the dumpfile. These are used by lr_deanonymize(1), in case email addresses in our input file reoccur in the file read by lr_deanonymize(1) in split form, i.e. joe.user@example.com occurs as both joe.user@example.com and example.com in the to be deanonymized file. All dumped objects are casted to lowercase. =head1 BACKGROUND For your convenience, we quote a bit of rfc822: SPACE = ; ( 40, 32.) CTL = ; ( 177, 127.) specials = "(" / ")" / "<" / ">" / "@" ; Must be in quoted- / "," / ";" / ":" / "\" / <"> ; string, to use / "." / "[" / "]" ; within a word. atom = 1* quoted-string = <"> *(qtext/quoted-pair) <">; Regular qtext or ; quoted chars. word = atom / quoted-string domain-ref = atom ; symbolic reference domain-literal = "[" *(dtext / quoted-pair) "]" sub-domain = domain-ref / domain-literal domain = sub-domain *("." sub-domain) local-part = word *("." word) ; uninterpreted ; case-preserved addr-spec = local-part "@" domain ; global address and of rfc 2181 The DNS itself places only one restriction on the particular labels that can be used to identify resource records. That one restriction relates to the length of the label and the full name. The length of any one label is limited to between 1 and 63 octets. A full domain name is limited to 255 octets (including the separators). rfc1123 However, a valid host name can never have the dotted-decimal form #.#.#.#, since at least the highest-level component label will be alphabetic. rfc819 ::= | "." ::= |
::= ::= | ::= | ::= | | "-" ::= any one of the 52 alphabetic characters A through Z in upper case and a through z in lower case ::= any one of the ten digits 0 through 9 =head1 EXAMPLE A 'logfile' like e.g. blaat fkrf 1.2.3.4.in-addr.arpa] pietje@bigcompany.com bla 1 2 3 lj;agas;gag blaat 1.2.3.4 fkrf 3.2.3.4.in-addr.arpa] bla 1 www.hotsex.com 2 3 lj;agas;gag jan@blaat.frut.com agagag blaat fkrf 4.2.3.4.in-addr.arpa] bla pietje@bigcompany.com www.hotsex.com 234.34.2.0 jan@blaat.frut.com 4.2.3.4.in-addr.arpa1 2 3 lj;agas;gag blaat fkrf tweede 3.2.3.4.in-addr.arpa] bla 1.2.3.4 1 blablabla.com 2 mdcc.cx 3 lj;agas;gag wil get anonymized to blaat fkrf 1.0.0.10.in-addr.arpa] john.doe.1@example.com bla 1 2 3 lj;agas;gag blaat 10.0.0.1 fkrf 2.0.0.10.in-addr.arpa] bla 1 1.example.com 2 3 lj;agas;gag john.doe.2@example.com agagag blaat fkrf 3.0.0.10.in-addr.arpa] bla john.doe.1@example.com 1.example.com 10.0.0.2 john.doe.2@example.com 3.0.0.10.in-addr.arpa1 2 3 lj;agas;gag blaat fkrf tweede 2.0.0.10.in-addr.arpa] bla 10.0.0.1 1 2.example.com 2 3.example.com 3 lj;agas;gag The dumps will represent something like ip 234.34.2.0 10.0.0.2 ip 1.2.3.4 10.0.0.1 inaddr 3.2.3.4.in-addr.arpa 2.0.0.10.in-addr.arpa inaddr 1.2.3.4.in-addr.arpa 1.0.0.10.in-addr.arpa inaddr 4.2.3.4.in-addr.arpa 3.0.0.10.in-addr.arpa domain mdcc.cx 3.example.com domain blablabla.com 2.example.com domain www.hotsex.com 1.example.com email jan@blaat.frut.com john.doe.2@example.com email pietje@bigcompany.com john.doe.1@example.com =head1 BUGS We can't handle files containing hostnames or email addresses in the example.com domain, usernames of the form john.doe. or ipnumbers in the rfc 1918 private network 10.0.0.0 - 10.255.255.255 (10/8 prefix). We don't handle quoted-string's in email addresses. We don't handle domain-literals in email addresses' domain. We regard 999.999.999.999 as an IP address: we don't mind the 255 limit. We don't treat network ipaddresses like 100.10.3 as ipaddresses. These will I get anonymized. =head1 SEE ALSO lr_deanonymize(1), the README.lire-client file in the Lire distribution. =head1 VERSION $Id: lr_anonymize.in,v 1.5 2006/07/23 13:16:32 vanbaal Exp $ =head1 COPYRIGHT Copyright (C) 2000-2001 Stichting LogReport Foundation LogReport@LogReport.org This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program (see COPYING); if not, check with http://www.gnu.org/copyleft/gpl.html. =head1 AUTHOR Joost van Baal =cut # Local Variables: # mode: cperl # End: