Security Incidents mailing list archives

Re: DNS zone transfer


From: James Hoagland <hoagland () SILICONDEFENSE COM>
Date: Fri, 1 Sep 2000 15:45:28 -0700

At 3:18 PM +0100 9/1/00, Fernando Cardoso wrote:
I guess you are used to see (as I am) lots of AXFRs from all places. Usually
they came along with bind.version queries since the named NXT bug scripts
are still hot 3lee7 stuff. They don't cause any problem except for a couple
lines in my logs and, sometimes, a message to the tech contact of a
compromised machine (hello .kr!!).

Yesterday, another AXFR try was made. This time from Canada:
ts1-193.mtrl.ca.ziplink.net

My IDS logged the try:

[**] IDS212/dns-zone-transfer [**]
08/31-17:19:10.789779 165.154.200.193:21368 -> my.name.server:53
TCP TTL:109 TOS:0x0 ID:44578  DF
*****PA* Seq: 0xB4A43A   Ack: 0xE367A43   Win: 0x2000
00 17 86 39 01 00 00 01 00 00 00 00 00 00 02 62  ...9...........b
6E 02 70 74 00 00 0F 00 01                       n.pt.....

Nothing new here... What is strange is that nothing was logged in the
nameserver!! I've tried zone transfers with dig, nslookup, host and even
with Sam Spade and all of them left a log entry in the nameserver (bind
8.2.2-P5).

Actually, looking at your packet dump, it is not a zone transfer.  It
is a query for MX of bn.pt.  You can read about the structure of DNS
packets in RFC 1035 (ftp://ftp.is.co.za/rfc/rfc1035.txt).

I recently posted to the Snort-users mailing list about possible
improvements to the arachNIDS signature for zone transfers.  In short
look for 0xFC past byte 13.  See
http://www.geocrawler.com/archives/3/4890/2000/8/0/4258922/ for my
post.

Kind regards,

  Jim
--
|*   Jim Hoagland, Associate Researcher, Silicon Defense    *|
|*               hoagland () SiliconDefense com                *|
|*  Voice: (707) 445-4355 x13          Fax: (707) 826-7571  *|


Current thread: