Mailing List Archive

Interesting incident involving Gentoo hardened linux
Hey guys,

Had an interesting moment today. Running a Gentoo Linux system on a Sun Fire
v120, uptime of about 13 days, around 10 or 12 active users at any time.

One of my users reported that he was logged in and unable to view his
processes. I was logged in as root and also was unable to view his active
processes. The PID of his shell was not listed in /proc, nor was the PID of
any processes owned by any subsequent user.

Took the system down to init 1 and checked it out for any signs of foul
play, found none. No anomalous behavior in the logs, nothing weird that
grsec reported. Nothing in the NIDS logs of the attached system..

At this point I'm thinking it's more of a strange sshd bug (it's tracking
~sparc) as I logged in from the root console and was granted a PID and
visibility in 'w' output as well as 'who' output.

But still... anyone else seen this behavior?

--Andrew Ruef

--
gentoo-sparc@gentoo.org mailing list
Re: Interesting incident involving Gentoo hardened linux [ In reply to ]
On Wed, 29 Jun 2005, Andrew Ruef wrote:
> Took the system down to init 1 and checked it out for any signs of foul
> play, found none. No anomalous behavior in the logs, nothing weird that
> grsec reported. Nothing in the NIDS logs of the attached system..

Did you do an MD5 comparison between the 'ps' command on your box and a
known good binary? That sounds like a trojaned ps binary or something
amiss in the kernel.

> But still... anyone else seen this behavior?

--
gentoo-sparc@gentoo.org mailing list
RE: Interesting incident involving Gentoo hardened linux [ In reply to ]
No... due to piss poor administration and that it's a Gentoo box those md5's
don't exist. Although the strange thing is, after sshd has been restarted
everything works fine...

I think I'm reaching for straws but it was as if sshd wasn't forking a bash
shell properly. Users could enter into their shells entry in /proc, it just
wasn't being displayed in 'w' or 'ps'....

This just sounds bad the more I think about it.

I'm going to try and reproduce the bug, if it can be...

--Andrew Ruef

-----Original Message-----
From: Gary [mailto:gary@linuxforce.org]
Sent: Wednesday, June 29, 2005 4:50 PM
To: gentoo-sparc@lists.gentoo.org
Subject: Re: [gentoo-sparc] Interesting incident involving Gentoo hardened
linux

On Wed, 29 Jun 2005, Andrew Ruef wrote:
> Took the system down to init 1 and checked it out for any signs of foul
> play, found none. No anomalous behavior in the logs, nothing weird that
> grsec reported. Nothing in the NIDS logs of the attached system..

Did you do an MD5 comparison between the 'ps' command on your box and a
known good binary? That sounds like a trojaned ps binary or something
amiss in the kernel.

> But still... anyone else seen this behavior?

--
gentoo-sparc@gentoo.org mailing list


--
gentoo-sparc@gentoo.org mailing list
RE: Interesting incident involving Gentoo hardened linux [ In reply to ]
Interestingly enough, there are entries in my syslog along the following:

sshd[253]: syslogin_perform_logout: logout() returned an error

There appear to be one of these for every logout action taken by a user...

This is strange. Could this maybe produce starvation of a resource
indicating when / which users are logged in if it creates a host of undead
not-quite-logged-in users?

(really sounds like I'm grasping for straws. Sigh)

--Andrew Ruef

-----Original Message-----
From: Andrew Ruef [mailto:munin@speakeasy.net]
Sent: Wednesday, June 29, 2005 5:20 PM
To: gentoo-sparc@lists.gentoo.org
Subject: RE: [gentoo-sparc] Interesting incident involving Gentoo hardened
linux

No... due to piss poor administration and that it's a Gentoo box those md5's
don't exist. Although the strange thing is, after sshd has been restarted
everything works fine...

I think I'm reaching for straws but it was as if sshd wasn't forking a bash
shell properly. Users could enter into their shells entry in /proc, it just
wasn't being displayed in 'w' or 'ps'....

This just sounds bad the more I think about it.

I'm going to try and reproduce the bug, if it can be...

--Andrew Ruef

-----Original Message-----
From: Gary [mailto:gary@linuxforce.org]
Sent: Wednesday, June 29, 2005 4:50 PM
To: gentoo-sparc@lists.gentoo.org
Subject: Re: [gentoo-sparc] Interesting incident involving Gentoo hardened
linux

On Wed, 29 Jun 2005, Andrew Ruef wrote:
> Took the system down to init 1 and checked it out for any signs of foul
> play, found none. No anomalous behavior in the logs, nothing weird that
> grsec reported. Nothing in the NIDS logs of the attached system..

Did you do an MD5 comparison between the 'ps' command on your box and a
known good binary? That sounds like a trojaned ps binary or something
amiss in the kernel.

> But still... anyone else seen this behavior?

--
gentoo-sparc@gentoo.org mailing list


--
gentoo-sparc@gentoo.org mailing list


--
gentoo-sparc@gentoo.org mailing list