Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 14 May 2001 22:23:28 -0700
From:      "Ted Mittelstaedt" <tedm@toybox.placo.com>
To:        "Kris Kennaway" <kris@obsecurity.org>
Cc:        "John Baxter" <jbaxter@mmcable.com>, "Dan Mahoney, System Admin" <danm@prime.gushi.org>, <questions@FreeBSD.ORG>
Subject:   RE: onitoring named
Message-ID:  <006b01c0dcff$2c7dff80$1401a8c0@tedm.placo.com>
In-Reply-To: <20010514200140.A93481@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
>-----Original Message-----
>From: Kris Kennaway [mailto:kris@obsecurity.org]
>Sent: Monday, May 14, 2001 8:02 PM
>To: Ted Mittelstaedt
>Cc: Kris Kennaway; John Baxter; Dan Mahoney, System Admin;
>questions@FreeBSD.ORG
>
>Both: >95% of the reported problems with named crashes on FreeBSD
>lists in the past 4 months have been penetration attempts, or at least
>occurred to people running vulnerable versions of named with symptoms
>perfectly consistent to being attacked.  Therefore this is the best
>initial diagnosis for people reporting problems with their named,
>until they go further and rule it out by indicating that they're
>already running 8.2.3-REL or a version of 9.x.  At that point more
>detailed analysis is obviously required (which perhaps might be better
>carried out on the bind support mailing lists).
>

The only problem with this statistic (assuming the 95% is
accurate) is that for it to be a valid indicator, this would
require that all the people having problems with bind
did, in fact, query the FreeBSD lists first, instead of
posting in the newsgroups or mailing lists.

This is an interesting debate in and of itself, though.  For
most programs, (I'll use Sendmail for example) there is a
Sendmail-specific support channel in addition to the general
FreeBSD questions.  I wonder what the percentage of people are
who post their question on the general FreeBSD mailing list
is, compared to people that post their specific question on the
support venue for that specific program.

That is one thing that I would caution you about drawing
conclusions on - looking at just the things posted on freebsd-questions
doesen't really give a good cross section of the problems people are
having with FreeBSD, let alone programs on it.  There's a lot of people
that never post here but use Usenet, and there's a lot of folks that never
use either forum but use other forums specific to what they are doing.

I'd speculate with bind that only the most greenhorn of bind admins
would post in this mailing list, unless it was something obviously
directly caused by FreeBSD.   I would think that most of them would
have gone to the Usenet group comp.protocols.dns.bind first.

This also doesen't report on the number of people who queried DejaNews
or the search indexes on the FreeBSD mailing lists for answers to their
problems
first, then found things to try and as a result never posted here (or
anywhere) at all.

You are right though that anyone running bind on a production system
should be running the secured code, and if they are having problems then
upgrading to the secured code is a perfectly valid step to take.  Of
course, upgrading to the current release of software is ALSO the correct
step to take when you suspect a plain old software bug, too, so whether
they do it for fear of cracking or do it to test for a bug, either way
they are going to be upgrading.

>> Maybe we ought to tell the next person who complains that their
>> nameserver is crashing, that this means their ram is bad and to go
>> replace it all. ;-)
>
>Well, that's also a possible explanation, but not the most likely one.
>

The only problem with taking troubleshooting steps just for no other reason
that they are A Good Thing is that you can easily waste a lot of time
chasing
down a dry hole when the problem is really somewhere else.

I still say that a modicum of analysis and observation applied to
troubleshooting
a problem is better than a knee-jerk "upgrade it" response.  You also risk
fixing the problem inadvertently, by accident, and leaving in place the
systems
that broke it to start with.  (and that are just going to break it again in
the future)

Ted Mittelstaedt                      tedm@toybox.placo.com
Author of:          The FreeBSD Corporate Networker's Guide
Book website:         http://www.freebsd-corp-net-guide.com


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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?006b01c0dcff$2c7dff80$1401a8c0>