Re: pam_unix.so and unix_chkpw setgid - does it work for regular users?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hello

On Thu, Aug 2, 2012 at 9:06 PM, Wolfgang Draxinger <Wolfgang.Draxinger@xxxxxxxxxxxxxxxxxxxxxx> wrote:
Hi,

I'm currently trying to configure user authentication on a webserver,
that shall use the normal system user names and passwords. I'm using
Nginx as webserver, together with the auth_pam module, as packages by
Debian wheezy.

I expected that since unix_chkpw is set setgid shadow I could use
pam_unix.so for the webserver service just as is. However it turned
out, that the user for the webserver process must be in the group
"shadow" for authentication to work. If the webserver can't read shadow
it doesn't work.

I was under the impression the idea of unix_chkpw was to have process
separation and by having a thoroughly audited helper program, that can
be setgid safely so that a regular user can perform pam_unix.so tests.

Did I miss something here?



 
Why are you using pam authentication for web server ?Using PAM authentication with apache/ngnix is a very bad idea. Here are some reasons :

* The Web technology provides no governors on how often or how rapidly password (authentication failure) retries can be made. That means that someone can hammer away at your system's root password using the Web, using a dictionary or similar mass attack, just as fast as the wire and your server can handle the requests. Most operating systems these days include attack detection (such as n failed passwords for the same account within m seconds) and evasion (breaking the connection, disabling the account under attack, disabling all logins from that source, et cetera), but the Web does not.

* An account under attack isn't notified (unless the server is heavily modified); there's no "You have 19483 login failures" message when the legitimate owner logs in.

* Without an exhaustive and error-prone examination of the server logs, you can't tell whether an account has been compromised. Detecting that an attack has occurred, or is in progress, is fairly obvious, though - if you look at the logs.

* Web authentication passwords (at least for Basic authentication) generally fly across the wire, and through intermediate proxy systems, in what amounts to plain text. "O'er the net we go/Caching all the way;/O what fun it is to surf/Giving my password away!"

* Since HTTP is stateless, information about the authentication is transmitted each and every time a request is made to the server. Essentially, the client caches it after the first successful access, and transmits it without asking for all subsequent requests to the same server.

* It's relatively trivial for someone on your system to put up a page that will steal the cached password from a client's cache without them knowing. Can you say "password grabber"?

Refer http://httpd.apache.org/docs/1.3/misc/FAQ.html#passwdauth

A possible solution for you will be to Add all your users in LDAP and use LDAP auth instead. Benefits of using LDAP auth are

- All apache servers can access LDAP server & create a centralized authentication setup.
- You can configure LDAP on secure port and all data transfer will be done on SSL

Regards
Arpit Tolani

_______________________________________________
Pam-list mailing list
Pam-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/pam-list

[Index of Archives]     [Fedora Users]     [Kernel]     [Red Hat Install]     [Linux for the blind]     [Gimp]

  Powered by Linux