破解故事——我是如何破解1亿2200万SHA1和Md5哈希密码
原文：http://blog.thireus.com/cracking-story-how-i-cracked-over-122-million-sha1-and-md5-hashed-passwords/
这是关于我如何破解密码的哈希值1亿2200万* John the Ripper和oclHashcat-plus故事。
这是几个月前，当我看到一条推特（https://twitter.com/m3g9tr0n）从korelogic约含共1亿4600万个密码的密码哈希表的torrent文件（https://www.korelogic.com/InfoSecSouthwest2012_Ripe_Hashes.html）。这个非常大的数量的密码哈希起初让我很沮丧，因为我自己只有一台经典配置的AMD Phenom II 4核3.2 MHz的电脑,外加一个ATI/AMD 5770显卡。但我真的很想给尝试一下，因为该领域的密码破解让我着迷。
在这漫长的旅程中我用的密码破解工具是John the Ripper和oclHashcat-puls。这篇文章是关于破解korelogic提供的MD5哈希值，但同样的策略也适用于SHA1哈希破解。
更新：
08/29/2012 –新的例子在John the Ripper部分: "在dict2hash.pl脚本的帮助下破解双MD5密码值"
08/29/2012 – 所有的都进行过排序和整理新的版本下载。
处理哈希值：
首先，所有的korelogic torrent文件必须被解压缩，它包含一个命名为“校验”文件夹。让我们检查这个文件夹的内容…
root@m3g9tr0n:~/hashes$ ls
longer_salts raw-md5.hashes.txt salted_with_md5 SHA1 vBulletin-v3.8.4
从现在起我们集中在raw-md5.hashes.txt文件列表，这个文件是4.3GB，通过wc工具可以看出它包括139444502行。
root@m3g9tr0n:~/hashes$ wc -lraw-md5.hashes.txt
139444502 raw-md5.hashes.txt
正如你可以假设，无论是John the Ripper和oclHashcat-plus都无法加载该文件，因为它太大了。出于这个原因，我们需要分割该文件。在Linux下，我们有一个很好的工具称为split 做这件工作得很好：
root@m3g9tr0n:~$ split --help
Usage: split [OPTION]... [INPUT[PREFIX]]
Output fixed-size pieces of INPUT toPREFIXaa, PREFIXab, ...; default
size is 1000 lines, and default PREFIXis `x'. With no INPUT, or when INPUT
is -, read standard input.
Mandatory arguments to long options aremandatory for short options too.
-a, --suffix-length=N usesuffixes of length N (default 2)
-b, --bytes=SIZE put SIZEbytes per output file
-C, --line-bytes=SIZE put atmost SIZE bytes of lines per output file
-d, --numeric-suffixes usenumeric suffixes instead of alphabetic
-l, --lines=NUMBER put NUMBERlines per output file
--verbose print adiagnostic just before each
output file isopened
--help display this help andexit
--version output versioninformation and exit
SIZE may be (or may be an integeroptionally followed by) one of following:
KB 1000, K 1024, MB 1000*1000, M1024*1024, and so on for G, T, P, E, Z, Y.
We can use the --lines=NUMBER parameter tosplit our raw-md5.hashes.txt file.
我们可以使用“--lines=NUMBER”参数来分割我们的原始md5.hashes.txt文件。
root@m3g9tr0n:~/hashes$ split -l 3000000raw-md5.hashes.txt part
请注意，考虑到每个MD5哈希是32字节长，我们还可以基于MB来进行文件分割。
使用oclHashcat-plus来破解密码
我开始玩oclHashcat-plus，因为它包含了-remove选项，一旦被破解它从hashfile删除破解成功的哈希值，这是真的很方便。oclHashcat-plus唯一的限制，是它对密码长度的限制。换句话说，它仅能够破解密码最多15个字符。我使用oclHashcat-plus规则是base64.rule，passwordspro.rule，T0XlC.rule在某些情况下d3ad0ne.rule。这些规则可以直接从oclHashcat-plus中找到。
暴力破解技术是不是我的第一选择。我用的字典是从g0tm1lk的BlogSpot下载的（http://blog.g0tmi1k.com/2011/06/dictionaries-wordlists/?redirect）。在g0tmi1k的文章其他外部链接中你会发现更多的字典。破解过程中最大的一部分是通过使用这些字典与上述规则进行。让我们来看看一些例子
使用单一规则：
./oclHashcat-plus64.bin -m 0 ~/hashes/md5_1~/Wordlists/d3ad0ne.dic -r rules/best64.rule -o Ultimate_Crack/eNtr0pY_1--remove
使用组合规则：
./oclHashcat-plus64.bin -m 0 ~/hashes/md5_1~/Wordlists/d3ad0ne.dic -r rules/best64.rule r rules/passwordspro.rule -oUltimate_Crack/eNtr0pY_1 --remove
Bruteforce attack with mask (you canspecify whichever charset you want):
使用掩码进行暴力破解***（可以指定任何你想要的字符集）：
./oclHashcat-plus64.bin -a 3 -1 ?l?d?u?s -m0 ~/hashes/md5_1 ?1?1?1?1?1?1?1?1 -o Ultimate_Crack/eNtr0pY_1 --remove
组合破解：
./oclHashcat-plus64.bin -a 1 -m 0~/hashes/md5_1 ~/Wordlists/d3ad0ne.dic ~/Wordlists/list -oUltimate_Crack/eNtr0pY_1 --remove
使用规则进行组合破解：
./oclHashcat-plus64.bin -a 1 -m 0~/hashes/md5_1 ~/Wordlists/d3ad0ne.dic ~/Wordlists/list -r rules/passwordspro.rule-o Ultimate_Crack/eNtr0pY_1 --remove
排列组合破解：
./oclHashcat-plus64.bin -a 4 -m 0~/hashes/md5_1 ~/Wordlists/d3ad0ne.dic -o Ultimate_Crack/eNtr0pY_1 --remove
排列组合规则破解：
./oclHashcat-plus64.bin -a 4 -m 0~/hashes/md5_1 ~/Wordlists/d3ad0ne.dic -r rules/best64.rule -oUltimate_Crack/eNtr0pY_1 --remove
In some cases, I used the hybrid + maskattack technique:
在某些情况下，我使用了混合 + 掩码***技术：
./oclHashcat-plus64.bin -a 6 -1 ?l?d -m 0 ~/hashes/md5_1~/Wordlists/d3ad0ne.dic ?1?1 -o Ultimate_Crack/eNtr0pY_1 --remove
混合掩码规则破解：
./oclHashcat-plus64.bin -a 6 -1 ?l?d -m 0~/hashes/md5_1 ~/Wordlists/d3ad0ne.dic ?1?1 -r rules/best64.rule -oUltimate_Crack/eNtr0pY_1 --remove
在这一点上，我没有用最后两个方法，因为他们是非常耗时的。我宁愿为John the Ripper 找到一个更好的KoreLogic的规则，将John the Ripper 的输出管道到oclHashcat-plus。正如我所说，oclHashcat-plus能够破解的密码最多15个字符。出于这个原因，我不得不通过“--stdout”选项限定每一次所产生的单词长度。如果你拥有一个非常快速的GPU可以跳过下面的例子。
./john --wordlist=~/Wordlists/all.lst-rules:KoreLogicRulesPrependYears --stdout=10 | ./oclHashcat-plus64.bin -m 0~/hashes/md5_1 -o Ultimate_Crack/eNtr0pY_1 --remove
当然你也可以使用来自Korelogic创建的其他前置的规则，像KoreLogicRulesPrependNumNum，甚至更好的创建自己的规则！是时候从破解密码的单词表中生成字典来破解剩余哈希值了。从eNtr0pY_1，使用以下命令我删除了MD5哈希值。
cut -b34- eNtr0pY_1 >eNtr0pY_1.dic
通过使用上述制造的字典，大量MD5哈希值被这种指纹***破解。你可以阅读更多关于马丁博斯@purehate这种***，我向你保证，这个技术是非常成功的！
当然，你也可以使用hashcat-utils中的二进制文件以及每一个管道到oclHashcat-plus的输出。
root@m3g9tr0n:~/oclHashcat-plus-0.08/hashcat-utils$ls
combinator.bin expander.bin gate.bin len.bin mp32.bin permute.bin prepare.bin req.bin splitlen.bin
使用John the Ripper破解密码
测试完所有我收集的字典以及几天后，是时候转移到john the Ripper来破解密码哈希值的剩余部分...
我用的Magnum-ripper编译OpenCL支持ATI / AMD显卡，因为我想用“--format=raw-md5-opencl参数”。相比“--format=raw-md5”，这是更快的方式为它使用你的CPU和GPU。
与john the Ripper中使用的规则是：
wordlist
Single
NT
Extra
KoreLogicRulesAppendNumbersandSpecials_Simple
KoreLogicRulesAppend6Num
KoreLogicRulesPrependAndAppendSpecial
KoreLogicRulesAppendNumNum_AddSpecialEverywhere
KoreLogicRulesAppendNumNumNum_AddSpecialEverywhere
KoreLogicRulesL33t.
您可以下载这些规则，并将其添加到您的john.conf文件中：
http://paste2.org/p/2048517
http://paste2.org/p/2048512
现在，让我们看到john the Ripper的一些例子...
使用--rules=Single：
./john --format=raw-md5-opencl--wordlist=../../Wordlists/all.lst --rules:Single ~/hashes/md5_1
哈希破解结果存储在默认文件john.pot中。您可以使用cat，more，head和tail检查它的内容。
root@m3g9tr0n:~/Tools/Password_Cracking/magnum-jumbo-OpenCL/run$tail -n 9 john.pot
$MD5$0fad81e7a61b47d387dde893fcf8e88a:anacarolinagu
$MD5$0f82fc9a81f5db07eb9289767390fd2b:fabulousfoodsu
$MD5$0e22933267b2e7df062703c4e5842029:fabuloustravelu
$MD5$0d40086a54fefe993c9816d1441672ac:modularhomeu
$MD5$0ed8181fc4d18e260dd8e36633124bfd:greenshoppingu
$MD5$0d6e8da4017ec5c384ac5536087da44d:lawofattractionu
$MD5$0eb916d3c6a66a32cedd4acc6edb1dbb:hotreportu
$MD5$0e241f99b5c13d56686ec618ab54d5fa:flightsandholidaysu
$MD5$0f3c99478362aae389d2cbf716394269:stthomasmoresu
To generate a wordlist from the john.pot file,you can use the following command.
要从john.pot文件生成从一个字典，可以使用下面的命令。
cut -d: -f 2- john.pot | sort -u >cracked.dic
生成的字典可以与上述规则相结合来破解哈希值。
当我使用oclHashcat-plus破解MD5哈希值时，我观察到，一些生成的密码被拒绝。这是因为oclHashcat-plus有一个关于字符串长度的限制。出于这个原因，我管道hashcat的结果输出到john the Ripper，并且在john the Ripper中更好的使用hashcat规则。
./hashcat-cli64.bin --stdout~/Wordlists/d3ad0ne.dic -r rules/best64.rule | ./john --format=raw-md5-opencl--stdin ~/hashes/md5_1
尝试所有字典与上述规则相结合后，是时候使用john the Ripper进行暴力破解破解。不幸的是，john the Ripper暴力破解***时不使用掩码***以产生密码。我们有基于包含在破解的密码来创建我们自己的字符集john.pot。
./john --make-charset=eNtr0pY.chr
Loaded 7948325 plaintexts
Generating charsets... 1 2 3 4 5 6 78 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 DONE
Generating cracking order... DONE
Successfully written charset file:eNtr0pY.chr (95 characters)
你们中许多人会想知道“DONE 31”......这仅仅是因为我编译john the Ripper用31个字符长度。默认情况下，john the Ripper被编译为8个字符长度的支持，所以最好通过修改头文件params.h中的以下行来改变，它位于john the Ripper的scr文件夹中。
#define CHARSET_MIN ' '
#define CHARSET_MAX 0x7E
#define CHARSET_SIZE (CHARSET_MAX - CHARSET_MIN+ 1)
#define CHARSET_LENGTH 8 //更改为31或者你期待的值
最后，在john.conf中你必须包括您所创建的字符集，以如本例中规定：
# Incremental modes
[Incremental:eNtr0pY]
File = $JOHN/eNtr0pY.chr
MinLen = 0
MaxLen = 31
CharCount = 95
Now it is time to use bruteforce attackswith our own charstet!
现在是时候用我们自己的字符串来进行暴力破解***！
./john --format=raw-md5-opencl--incremental=eNtr0pY ~/hashes/md5_1
如果你查看john.conf，你会看到一些暴力破解***模式分为外部。这些都是双重的，带键盘（使用从键盘字符产生的相邻组合），KnownForce，日期时间，重复，顺序，亚群和DumbForce为疯狂的密码格式。
./john --format=raw-md5-opencl--external=DumbForce ~/hashes/md5_1
在这里提供的dict2hash.pl脚本帮助下，我们也喜欢破解双重MD5哈希值。
perl dict2hash.pl < rockyou.txt | ./john--format=raw-md5-opencl --stdin ~/md5_1
在这里，您可以看到与john the Ripper破获MD5值的一些示例：
http://pastebin.com/uaGFXkBA
http://pastebin.com/LT5Gda9L
Personally, I believe a password like"$MD5$0b26a0faf1344d6e772bf55628e10e29:n34=mn { .clipboard $me }" isimpossible to crack with bruteforce attacks.
Note: All the abovementioned techniques canbe used with oclHashcat-plus by defining -m 100 and with John the Ripper bydefining --format=raw-sha1-opencl for SHA1 cracking with OpenCL!
就个人而言，我相信密码像“ $ MD5 $ 0b26a0faf1344d6e772bf55628e10e29：N34 = {MN $.clipboard$me} ”是不可能使用暴力破解***来破解。
注：所有上述技术可以在oclHashcat-plus定义参数“-m 100”使用，john the Ripper定义“--format=raw-sha1-opencl”使用OpenCL破解SHA1
密码分析
最后，值得一看的分析使用pipal（密码分析者）从破解结果中收集的样本。
root@m3g9tr0n:~/pipal$ ruby1.9.1 pipal.rb \
-o eNtr0pY_1~/Wordlists/Ultimate/Part1/eNtr0pY_5.dic
Total entries = 759103
Total unique entries = 758299
Top10 passwords
niezgadniesz123 = 3 (0.0%)
ubqu = 3 (0.0%)
amonys = 3 (0.0%)
centralitie = 3 (0.0%)
bobydu = 3 (0.0%)
hanghuynh = 3 (0.0%)
hmadyousi = 3 (0.0%)
matthewperman = 3 (0.0%)
shadowninja2 = 3 (0.0%)
lhz4 = 3 (0.0%)
Top10 基本密码
august = 219 (0.03%)
july = 205 (0.03%)
april = 199 (0.03%)
june = 195 (0.03%)
march = 165 (0.02%)
alex = 161 (0.02%)
love = 132 (0.02%)
chris = 130 (0.02%)
daniel = 128 (0.02%)
dragon = 122 (0.02%)
密码长度(以长度进行排序)
1 = 13 (0.0%)
2 = 103 (0.01%)
3 = 1332 (0.18%)
4 = 16781 (2.21%)
5 = 19831 (2.61%)
6 = 95800 (12.62%)
7 = 202414 (26.66%)
8 = 158562 (20.89%)
9 = 103855 (13.68%)
10 = 75652 (9.97%)
11 = 46023 (6.06%)
12 = 24997 (3.29%)
13 = 8423 (1.11%)
14 = 3772 (0.5%)
15 = 1560 (0.21%)
密码长度(以计数排序)
7 = 202414 (26.66%)
8 = 158562 (20.89%)
9 = 103855 (13.68%)
6 = 95800 (12.62%)
10 = 75652 (9.97%)
11 = 46023 (6.06%)
12 = 24997 (3.29%)
5 = 19831 (2.61%)
4 = 16781 (2.21%)
13 = 8423 (1.11%)
14 = 3772 (0.5%)
15 = 1560 (0.21%)
3 = 1332 (0.18%)
2 = 103 (0.01%)
1 = 13 (0.0%)
|
|
|
||
||
||
||
|||
||||
||||
|||||
|||||
||||||
||||||
|||||||||
|||||||||||||||||
00000000001111111
01234567890123456
Oneto six characters = 133854 (17.63%)
One to eight characters = 494828 (65.19%)
More than eight characters = 264275(34.81%)
Onlylowercase alpha = 154996 (20.42%)
Only uppercase alpha = 14072 (1.85%)
Only alpha = 169068 (22.27%)
Only numeric = 119581 (15.75%)
First capital last symbol = 6088 (0.8%)
First capital last number = 73611 (9.7%)
Months
january = 109 (0.01%)
february = 45 (0.01%)
march = 247 (0.03%)
april = 251 (0.03%)
may = 850 (0.11%)
june = 246 (0.03%)
july = 223 (0.03%)
august = 300 (0.04%)
september = 80 (0.01%)
october = 134 (0.02%)
november = 113 (0.01%)
december = 115 (0.02%)
Days
monday = 59 (0.01%)
tuesday = 20 (0.0%)
wednesday = 7 (0.0%)
thursday = 38 (0.01%)
friday = 46 (0.01%)
saturday = 7 (0.0%)
sunday = 70 (0.01%)
Months (Abreviated)
jan = 1482 (0.2%)
feb = 249 (0.03%)
mar = 8397 (1.11%)
apr = 692 (0.09%)
may = 850 (0.11%)
jun = 889 (0.12%)
jul = 1051 (0.14%)
aug = 785 (0.1%)
sept = 215 (0.03%)
oct = 512 (0.07%)
nov = 821 (0.11%)
dec = 874 (0.12%)
Days (Abreviated)
mon = 4319 (0.57%)
tues = 28 (0.0%)
wed = 217 (0.03%)
thurs = 44 (0.01%)
fri = 758 (0.1%)
sat = 769 (0.1%)
sun = 1018 (0.13%)
Includes years
1975 = 411 (0.05%)
1976 = 388 (0.05%)
1977 = 446 (0.06%)
1978 = 432 (0.06%)
1979 = 441 (0.06%)
1980 = 541 (0.07%)
1981 = 453 (0.06%)
1982 = 519 (0.07%)
1983 = 533 (0.07%)
1984 = 603 (0.08%)
1985 = 585 (0.08%)
1986 = 616 (0.08%)
1987 = 710 (0.09%)
1988 = 641 (0.08%)
1989 = 941 (0.12%)
1990 = 931 (0.12%)
1991 = 995 (0.13%)
1992 = 935 (0.12%)
1993 = 905 (0.12%)
1994 = 907 (0.12%)
1995 = 4021 (0.53%)
1996 = 858 (0.11%)
1997 = 486 (0.06%)
1998 = 443 (0.06%)
1999 = 416 (0.05%)
2000 = 1024 (0.13%)
2001 = 643 (0.08%)
2002 = 586 (0.08%)
2003 = 1132 (0.15%)
2004 = 1254 (0.17%)
2005 = 796 (0.1%)
2006 = 818 (0.11%)
2007 = 1442 (0.19%)
2008 = 1019 (0.13%)
2009 = 742 (0.1%)
2010 = 767 (0.1%)
2011 = 516 (0.07%)
2012 = 925 (0.12%)
2013 = 165 (0.02%)
2014 = 142 (0.02%)
2015 = 146 (0.02%)
2016 = 118 (0.02%)
2017 = 139 (0.02%)
2018 = 131 (0.02%)
2019 = 172 (0.02%)
2020 = 179 (0.02%)
Years (Top 10)
1995 = 4021 (0.53%)
2007 = 1442 (0.19%)
2004 = 1254 (0.17%)
2003 = 1132 (0.15%)
2000 = 1024 (0.13%)
2008 = 1019 (0.13%)
1991 = 995 (0.13%)
1989 = 941 (0.12%)
1992 = 935 (0.12%)
1990 = 931 (0.12%)
Colours
black = 485 (0.06%)
blue = 549 (0.07%)
brown = 184 (0.02%)
gray = 89 (0.01%)
green = 348 (0.05%)
orange = 125 (0.02%)
pink = 262 (0.03%)
purple = 73 (0.01%)
red = 2974 (0.39%)
white = 179 (0.02%)
yellow = 85 (0.01%)
violet = 63 (0.01%)
indigo = 22 (0.0%)
Single digit on the end = 92080 (12.13%)
Two digits on the end = 87587 (11.54%)
Three digits on the end = 103715 (13.66%)
Last number
0 = 45407 (5.98%)
1 = 64764 (8.53%)
2 = 52570 (6.93%)
3 = 52890 (6.97%)
4 = 43719 (5.76%)
5 = 55185 (7.27%)
6 = 42826 (5.64%)
7 = 46169 (6.08%)
8 = 42475 (5.6%)
9 = 44930 (5.92%)
|
|
| ||
||||
|||| | | |
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
||||||||||
0123456789
Last digit
1 = 64764 (8.53%)
5 = 55185 (7.27%)
3 = 52890 (6.97%)
2 = 52570 (6.93%)
7 = 46169 (6.08%)
0 = 45407 (5.98%)
9 = 44930 (5.92%)
4 = 43719 (5.76%)
6 = 42826 (5.64%)
8 = 42475 (5.6%)
Last 2 digits (Top 10)
95 = 14675 (1.93%)
23 = 12192 (1.61%)
12 = 9230 (1.22%)
11 = 8214 (1.08%)
01 = 7606 (1.0%)
00 = 7131 (0.94%)
07 = 6295 (0.83%)
10 = 6182 (0.81%)
21 = 5881 (0.77%)
99 = 5868 (0.77%)
Last 3 digits (Top 10)
123 = 6857 (0.9%)
995 = 4122 (0.54%)
971 = 2916 (0.38%)
972 = 2850 (0.38%)
007 = 2514 (0.33%)
000 = 1868 (0.25%)
234 = 1725 (0.23%)
666 = 1465 (0.19%)
777 = 1389 (0.18%)
004 = 1347 (0.18%)
Last 4 digits (Top 10)
1995 = 3886 (0.51%)
1234 = 1379 (0.18%)
2007 = 1325 (0.17%)
2004 = 1121 (0.15%)
2003 = 1016 (0.13%)
2008 = 869 (0.11%)
2000 = 846 (0.11%)
1991 = 819 (0.11%)
2012 = 809 (0.11%)
1990 = 789 (0.1%)
Last 5 digits (Top 10)
12345 = 743 (0.1%)
23456 = 652 (0.09%)
54321 = 189 (0.02%)
23123 = 140 (0.02%)
56789 = 127 (0.02%)
34567 = 102 (0.01%)
11111 = 99 (0.01%)
45678 = 75 (0.01%)
00000 = 73 (0.01%)
88888 = 68 (0.01%)
US Area Codes
971 = Oregon: Metropolitan Portland,
Salem/Keizer area,
incl Cricket Wireless (OR)
972 = Texas: Dallas Metro (TX)
234 = NE Ohio: Canton, Akron (OH)
Character sets
loweralphanum: 330937 (43.6%)
loweralpha: 154996 (20.42%)
numeric: 119581 (15.75%)
mixedalphanum: 41121 (5.42%)
upperalphanum: 41078 (5.41%)
mixedalpha: 28464 (3.75%)
upperalpha: 14072 (1.85%)
loweralphaspecial: 10222 (1.35%)
loweralphaspecialnum: 5735 (0.76%)
mixedalphaspecial: 4724 (0.62%)
upperalphaspecial: 2939 (0.39%)
mixedalphaspecialnum: 2247 (0.3%)
specialnum: 648 (0.09%)
upperalphaspecialnum: 374 (0.05%)
special: 47 (0.01%)
Character set ordering
stringdigit: 349534 (46.05%)
allstring: 197532 (26.02%)
alldigit: 119581 (15.75%)
digitstring: 28873 (3.8%)
othermask: 18649 (2.46%)
stringdigitstring: 14577 (1.92%)
stringspecial: 10441 (1.38%)
digitstringdigit: 9981 (1.31%)
stringspecialstring: 5469 (0.72%)
stringspecialdigit: 3075 (0.41%)
specialstring: 834 (0.11%)
specialstringspecial: 510 (0.07%)
allspecial: 47 (0.01%)
Hashcat masks (Top 10)
?d?d?d?d?d?d?d: 85053 (11.2%)
?l?l?l?l?l?l: 38400 (5.06%)
?l?l?l?l?l?l?l?l: 36217 (4.77%)
?l?l?l?l?l?l?l: 35468 (4.67%)
?l?l?l?l?l?l?d?d?d: 24051 (3.17%)
?l?l?l?l?l?l?d?d: 18591 (2.45%)
?l?l?l?l?l?d?d?d: 18047 (2.38%)
?d?d?d?d?d?d: 16048 (2.11%)
?l?l?l?l?l?l?l?l?l: 14236 (1.88%)
?l?l?l?l?d?d?d: 13802 (1.82%)
结论
这是一个非常耗费时间和艰苦的工作，因为我不拥有最快的显卡。整个破解过程历时约5个月来完成，因为我必须完成我的CCNP认证学习。
从中学到的教训是，一个好的智能词典结合方便的规则，使用Hashcat或John the Ripper甚至可以破解强密码。基于以上情况，管理员应该使用更强的哈希算法（加盐）来存储你的密码，并在在定期的基础上更改自己的密码。
谢谢阅读
你可以在推特上找到我, @m3g9tr0n.
下载
你可以下载破解的哈希结果值
721.9 MB -m3g9tr0n_122Million_Passwords_WordLists.zip
所提供的KoreLogic torrent文件包含了各种独特的密码哈希值。出于这个原因，你可能会在这些字典中发现重复的密码，作为一个单独的密码，可以使用各种算法来计算散列值！这意味着1.22亿独特的哈希值（MD5，SHA1，双MD5等）被破解，结果中有8亿3600万个独特的密码。
您也可以下载“都在一个”版本，干净和分类：
270.2 MB -m3g9tr0n_Passwords_WordList_CLEANED.zip
export LC_ALL='C' && cat * | sort |uniq > eNtr0pY_ALL_sort_uniq.dic
参考：
hashcat Wiki [ http://hashcat.net/wiki/ ]
John the Ripper Wiki [http://openwall.info/wiki/john ]