Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I tried to change a Linux password to "sitonapotatopanotis" and got this error: BAD PASSWORD: is a palindrome

Why does this rule exist?

share|improve this question
    
Nobody but the developers of pam_cracklib could answer this. I tried looking at the manpage and did a quick web search but no luck. –  Belmin Fernandez Mar 2 '12 at 17:59
2  
It's nearly impossible to figure out what the person who blocked it was thinking. But when someone's entire job is to think up passwords we're not allowed to use, they eventually start reaching for more stuff to add. i think to answer your question: Why? - someone had too much time on their hands. –  Ian Boyd Mar 3 '12 at 3:38
    
That seems like a good password to me, I've seen much worse. –  nikhil Mar 9 '12 at 4:16

3 Answers 3

up vote 10 down vote accepted

The manpage for pam_cracklib (responsible for password strength checking) does not specify why this is done:

   The strength checks works in the following manner: at first the Cracklib routine is
   called to check if the password is part of a dictionary; if this is not the case an
   additional set of strength checks is done. These checks are:

   Palindrome
       Is the new password a palindrome?

However, it's not difficult to imagine that there are some password cracking softwares that try palindromes.

I would not recommend using such a password but it's up to you evaluate what security trade-offs you're comfortable with making (you could use sudo or root account to change the password and it will allow you to change it to whatever you want).

share|improve this answer
2  
So if he added/subtracted one character the password would be fine? –  Daniel R Hicks Mar 2 '12 at 19:11
10  
@DanH: Yes. If a cracking program is going to try "near palindromes", it pretty much has to try everything. –  David Schwartz Mar 2 '12 at 20:11
6  
It seems to me that a palindrome should be counted as valid if the first half counts as a valid password. (But that's nit-picking, of course). –  Daniel R Hicks Mar 3 '12 at 14:35

People are simply more likely to choose "racecar" as their password cause they like it. So those words are high up on all wordlists (which are used before any brute-forcing). And it's simpler to check against all palindromes than to maintain a list of palindromes in the password checking library.


Some passwords are great and some are really bad.
We use certain factors to judge the quality of a password. Like length or what different characters are used.

For some passwords, these factors become less relevant or not relevant at all.

Like, this is a great password:

v10H73nqMQPkbUvTLOPyKBg4KnkUjWgF

This one, not so much:

acbaacbacaabcabbbaaabcaccbbbaaac

Even though it has the same length, if the same password rules apply and you brute-force it, the second password will be tried a lot sooner than the first password.

Let's have a look at this one:

qwertyuiopasdfghjklzxcvbnm123456

Now, we're rollin' with a serious password! Only that it's almost the worst possible password ever because all letters are used in the same pattern as they appear on a very popular keyboard type.

Someone might look at that password and think it's super awesome because he chooses it under false assumptions (length being most important for a password).

The same could be said for palindromes. First of all, they give a false feeling of security (as Mike notes) because their length is increased by simply duplicating all letters. But the real problem with them is that they are easy to remember and somewhat of a commodity.

share|improve this answer
12  
-1 "v10H73nqMQPkbUvTLOPyKBg4KnkUjWgF" this is not a great password, it's an awful one, since you simply cannot remember it. –  Lohoris Mar 2 '12 at 22:48
2  
The only reason why that's a bad password is just because I mentioned it here and it is no longer secret. I only use randomly generated passwords combined with a single-sign-on solution. –  Oliver Salzburg Mar 3 '12 at 0:04
2  
Version 10 has 73 new quotes. More quotes per knowledge-base under varyingly terse, low orders, pays you kindly, big growth for keen novices’ knowledge. Useful jobs wait going forward. –  Synetech Mar 3 '12 at 5:12
1  
@OliverSalzburg: single sign-in are good methods, I use them myself, but here we are in the domain of plain old "remember them" passwords, so it's not a valid twist. Remember that due to other people's idiocy, sometimes you are forced to enter passwords yourself anyway. –  Lohoris Mar 3 '12 at 9:20
2  
@OliverSalzburg You don't have to remember the password; it's written on the Post-it attached to my monitor. –  Ian Boyd Apr 14 '12 at 18:20

Because a 20-character palindromic password is only as secure as a 10-character password -- there's essentially no extra entropy in the last 10 characters. So you're getting a false sense of security from having a long password.

share|improve this answer
10  
Might be wrong here, but the effective security still depends on how you attempt to crack such a password. Does the attacker know there is a limited character set being used? Do we know the password length? –  slhck Mar 2 '12 at 16:18
17  
Do password crackers typically try palindromes of each guess? –  Joseph Mornin Mar 2 '12 at 16:19
1  
Hm, it certainly adds to password entropy. However, I certainly wouldn't recommend it. It all depends on how the password cracking software generates permutations. –  Belmin Fernandez Mar 2 '12 at 17:19
11  
A palindromic password adds exactly one bit of entropy (is palindrome vs. is not palindrome). It is not "only as secure as a 10-character password", but it's much less secure than an 11 character one. –  user35138 Mar 2 '12 at 17:49
2  
I'd say sitonapotatopanotis probably significantly less entropic than a standard 10 letter password made out of English words. Grammatically correct palindormes are very rare. It would be about as secure if you made a palindrome out of 10 random characters mwiovqfbzczbfqvoiwm, or if you just took words and made the palindrome part nonsense doctorwormrowrotcod. Also it adds a little more than a bit of entropy (you could vary how you do the palindrome; e.g., doctorwormrowrotcod or doctorwormmrowrotcod or doctorotcodwormmrowr, etc. But in general palindromes are a bad idea in pw. –  dr jimbob Mar 2 '12 at 21:03

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.