Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 4 Mar 2001 04:51:53 -0800 (PST)
From:      "Michael A. Dickerson" <mikey@singingtree.com>
To:        freebsd-security@freebsd.org
Subject:   "Input/output error" on a variety of devices
Message-ID:  <Pine.BSF.4.21.0103040400280.55542-100000@redlance.singingtree.com>

next in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello -security, something peculiar happened to a machine I'm responsible
for today.  The information in the "daily run output" and "security check
output" email is all I have to go on:

> Subject: myhost security check output
>
> checking setuid files and devices:
> find: /dev/rda0: Input/output error
> find: /dev/da0: Input/output error
> find: /dev/rda0s1: Input/output error
> find: /dev/rda0s1c: Input/output error
> find: /dev/da0s1: Input/output error
> find: /dev/rda0s1a: Input/output error
> find: /dev/da0s1a: Input/output error
> find: /dev/bpf0: Input/output error
> find: /dev/card0: Input/output error
> find: /dev/card1: Input/output error
> find: /dev/card2: Input/output error
> find: /dev/card3: Input/output error
> find: /dev/kbd0: Input/output error
> find: /dev/kmem: Input/output error
> find: /dev/mem: Input/output error
> find: /dev/tty: Input/output error
> find: /dev/ugen0: Input/output error
> find: /dev/uhid0: Input/output error
> find: /dev/ulpt0: Input/output error
>
> checking for uids of 0:
> tee: /dev/stderr: Input/output error
>
> checking for passwordless accounts:
> tee: /dev/stderr: Input/output error
>
> sentry.cduniverse.com login failures:
> tee: /dev/stderr: Input/output error
>
> sentry.cduniverse.com refused connections:
> tee: /dev/stderr: Input/output error

Clearly the 'find' didn't break on all devices, but if there's a pattern
in the ones that failed, it eludes me.  I suppose find was just trying to
stat the nodes to get their permissions(?).  At this point I was
suspecting a full disk might be upsetting the kernel, since this machine
logs for others (and a DoS attack would not be terribly surprising in this
environment).  However, the daily script thinks the disks are OK:

> Subject: myhost daily run output
> 
> Removing stale files from /var/preserve:
> 
> Cleaning out old system announcements:
>
> Removing stale files from /var/rwho:
> 
> Backup passwd and group files:
>
> Verifying group file syntax:
> 
> Backing up mail aliases:
>
> Disk status:
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on
> /dev/da0s1a     49583    33204    12413    73%    /
> /dev/da0s1f   7956270   713531  6606238    10%    /usr
> /dev/da0s1e     99183     7607    83642     8%    /var
> procfs              4        4        0   100%    /proc
> 
> Last dump(s) done (Dump '>' file systems):
>
> UUCP status:
>
> Network interface status:
> netstat: kvm not available
> ifnet: symbol not defined
>
> Local system status:
>  1:59AM  up 10 days, 19:03, 0 users, load averages: 0.07, 0.02, 0.00
... and the rest (mailq and some local scripts) is normal.

It seems the 'kvm not available' is not surprising if /dev/mem is broken
somehow, and I'm guessing that ifnet's complaint was just spurious and
caused by the first.

With ssh failing to connect, there's not much more information I can get
from this machine.  It still responds to pings, but I've learned that the
most brain damaged of kernels can still usually manage that:

elsewhere# ssh -l mikey xx.yy.zz.ww
Connection closed by xx.yy.zz.ww
elsewhere# nmap -sS xx.yy.zz.ww

Starting nmap V. 2.53 by fyodor@insecure.org ( www.insecure.org/nmap/ )
Interesting ports on  (xx.yy.zz.ww):
(The 1522 ports scanned but not shown below are in state: closed)
Port       State       Service
22/tcp     open        ssh                     

Nmap run completed -- 1 IP address (1 host up) scanned in 30 seconds

elsewhere#

Anyway, my question is this: Has anybody ever seen anything resembling
this behavior?  Specifically, does it seem likely that this host was
rooted?  I'm thinking not, but I ask because this machines lives in a
hostile environment and I have to be suspicious of anything weird that
happens on that network.  In fact, this was the hopefully "secure" machine
which exists only to monitor and log for others (which may have been
recently rooted through bind; we're still investigating).

It seems that even if it was compromised, the attacker has probably locked
himself out as well as me.  Hopefully it was a hardware failure or pilot
error and I'll be off to -stable.  Looks like I'll be adding another log
host for the log host..

Thanks very much,
M. Dickerson


-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
Charset: noconv

iQA/AwUBOqI6bxvDsQU/S3JEEQJmogCgpnY61LjUTLDEvNeeqS3390DlXMYAoKfQ
ZjW6fiOnHDbb9m2dUct0GfdD
=tnp9
-----END PGP SIGNATURE-----


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-security" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0103040400280.55542-100000>