-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At Mon, 25 Jun 2012 07:24:58 -0500,
Mark McCullough wrote:
> 
> I do audit response, trying to guess what the auditors will demand as a big
> part of my job today, so please don't take this personally, more as
> identifying areas auditors may need addressed.
> 

Not at all, Mark, I very much appreciate your feedback! 

> 
> You need to ensure you are maintaining at least four old passwords.
> Password changes may not be done for one reason or another when a new one is
> issued.  Nothing is worse than a root password management system that you
> can't use to get the root password.
> 

Agreed. This functionality is planned [0].

> 
> What guarantee do you have that the hashes are actually deployed rather than
> just leaving the root password on some insecure value?  Even if it is just a
> "Yes, I updated the root password and validated it" signoff for each system
> with a person's name who did the work, something would need to be in place
> to ensure it is the actual password.
> 

Personally, I can't see implementing this without a configuration management
system like puppet (or friends). It's not implemented yet, but what I plan to
do is to provide for live audit using puppet's reporting capability (whereby
nodes will report a hash of their root hash) so this can be compared with what
credmgr *thinks* is deployed on the system.

> 
> Password age will be key.  There should be some framework (either political
> or software) to ensure the password never actually gets beyond the maximum
> age as defined by your organization.
> 

Agreed, this functionality is planned [0].

> 
> What about people leaving the organization?  The obvious answer with your
> software is in such an event, issue a new password, but it is something that
> needs to be kept in mind.  Password change events aren't just the fixed
> schedule of age.
> 

Agreed, this functionality is planned [0].

> 
> An auditor will likely have a hard time with even encrypted email.  It's a
> reflex thing.  If you use this with the intention of presenting to an
> auditor, have an answer ready.
> 

Well, short answer: yes. Long answer, this is *precisely* how the DNSSEC
signing key for the root is protected. If it's good enough for ICANN, it
*should* be good enough for an auditor.

> 
> Outside your software, but be sure you have clearly defined properties on
> who has access to that password.  You could manage it from the encryption
> keys, but have something to say that yes, this person is authorized.
> 

I can give people a tool but I cannot write their corporate policies. I agree
that this is equally vital. Hopefully anyone sufficiently clueful to manage
root like this is aware that technical solutions *alone* aren't sufficient.

> 
> Written before first cup of tea, so take this for mainly worth the electrons
> allocated to send and display this.
> 

Thanks again for your helpful feedback, Mark! I'm in the middle of moving
house at the moment, so I can't promise when the next release will be...but
soon. 

[0]: https://github.com/treyka/credmgr#todo

Cheers,
Trey
++--------------------------------------------------------------------------++
Kingfisher Operations, Sprl
Trey Darley - Principal
gpg fingerprint: C2AD E2A8 440C 8785 1958 B6DA 4176 9233 8F6D 8AF0
++--------------------------------------------------------------------------++
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAk/oqZAACgkQQXaSM49tivDjhwCZAQL1aWB/Ge1dmZZ388Yr8vyq
juIAoJ9euUngLcLHDLqLppr9IDxSycCL
=H+fn
-----END PGP SIGNATURE-----
_______________________________________________
Discuss mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to