Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.os.linux.networking > #800 > unrolled thread

ssh security question

Started byGeneral Schvantzkoph <schvantzkoph@yahoo.com>
First post2011-11-14 19:15 +0000
Last post2011-12-01 14:37 -0800
Articles 4 — 4 participants

Back to article view | Back to comp.os.linux.networking


Contents

  ssh security question General Schvantzkoph <schvantzkoph@yahoo.com> - 2011-11-14 19:15 +0000
    Re: ssh security question Jorgen Grahn <grahn+nntp@snipabacken.se> - 2011-11-14 20:02 +0000
    Re: ssh security question Richard Kettlewell <rjk@greenend.org.uk> - 2011-11-14 21:05 +0000
    Re: ssh security question David Schwartz <davids@webmaster.com> - 2011-12-01 14:37 -0800

#800 — ssh security question

FromGeneral Schvantzkoph <schvantzkoph@yahoo.com>
Date2011-11-14 19:15 +0000
Subjectssh security question
Message-ID<9id7mmF8mjU1@mid.individual.net>
I just regenerated the keys on one of my F14 systems. I am still able to 
access systems which don't have the new public key in their 
authorized_keys file. The one thing I did differently this time was that I 
did an ssh-add after I regenerated the keys. I did the ssh-add because 
putting the new public key into the authorized_keys files of my other 
systems wasn't sufficient to give me access. After the ssh-add I could 
access the other systems, however I could also access systems that don't 
have the new authorized_keys file.

Does ssh-add keep the old keys in the authentication agent as well as the 
new key? This would negate the value of changing keys if true.

[toc] | [next] | [standalone]


#801

FromJorgen Grahn <grahn+nntp@snipabacken.se>
Date2011-11-14 20:02 +0000
Message-ID<slrnjc2su0.1bk.grahn+nntp@frailea.sa.invalid>
In reply to#800
On Mon, 2011-11-14, General Schvantzkoph wrote:
...
> Does ssh-add keep the old keys in the authentication agent as well as the 
> new key?

Ask your agent: ssh-add -l

> This would negate the value of changing keys if true.

If you delete your old key and plan to keep the agent running for a
long time, you should probably remove the old key from the agent too,
yes. Once again, use ssh-add to do that.

/Jorgen

-- 
  // Jorgen Grahn <grahn@  Oo  o.   .     .
\X/     snipabacken.se>   O  o   .

[toc] | [prev] | [next] | [standalone]


#802

FromRichard Kettlewell <rjk@greenend.org.uk>
Date2011-11-14 21:05 +0000
Message-ID<8739dqtnws.fsf@araminta.anjou.terraraq.org.uk>
In reply to#800
General Schvantzkoph <schvantzkoph@yahoo.com> writes:
> I just regenerated the keys on one of my F14 systems. I am still able to 
> access systems which don't have the new public key in their 
> authorized_keys file. The one thing I did differently this time was that I 
> did an ssh-add after I regenerated the keys. I did the ssh-add because 
> putting the new public key into the authorized_keys files of my other 
> systems wasn't sufficient to give me access. After the ssh-add I could 
> access the other systems, however I could also access systems that don't 
> have the new authorized_keys file.
>
> Does ssh-add keep the old keys in the authentication agent as well as the 
> new key? This would negate the value of changing keys if true.

I disagree.  If you want to revoke the old key, you need to remove its
public half from all the authorized_keys files that list it.  

-- 
http://www.greenend.org.uk/rjk/

[toc] | [prev] | [next] | [standalone]


#854

FromDavid Schwartz <davids@webmaster.com>
Date2011-12-01 14:37 -0800
Message-ID<ca6ed6d1-f19a-4a6e-8630-4b56b6587409@w3g2000vbw.googlegroups.com>
In reply to#800
On Nov 14, 11:15 am, General Schvantzkoph <schvantzk...@yahoo.com>
wrote:

> Does ssh-add keep the old keys in the authentication agent as well as the
> new key? This would negate the value of changing keys if true.

There's no point in changing keys. The point is in changing *locks*.
Sure, if you change the locks on your back door, there's no security
advantage to throwing away the key to your front door. But if you
change *locks* so that only a new key is accepted *by* *the* *lock*,
then there's a security advantage.

Throwing away a key conveys no security advantage. That's is correct.

DS

[toc] | [prev] | [standalone]


Back to top | Article view | comp.os.linux.networking


csiph-web