Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 19 Apr 2002 18:22:44 -0400
From:      "Peter C. Lai" <sirmoo@cowbert.2y.net>
To:        Greg Fortune <megatontech@pacbell.net>
Cc:        Brett Glass <brett@lariat.org>, Ken McGlothlen <mcglk@artlogix.com>, security@freebsd.org
Subject:   Re: FreeBSD Security Advisory FreeBSD-SA-02:21.tcpip
Message-ID:  <20020419182244.A27580@cowbert.2y.net>
In-Reply-To: <5.1.0.14.2.20020419101925.00ab2200@postoffice.pacbell.net>; from megatontech@pacbell.net on Fri, Apr 19, 2002 at 12:08:25PM -0700
References:  <878z7k4oz9.fsf@ralf.artlogix.com> <4.3.2.7.2.20020418143615.021a8460@nospam.lariat.org> <4.3.2.7.2.20020418095356.024354c0@nospam.lariat.org> <4.3.2.7.2.20020417230144.032ad390@nospam.lariat.org> <200204171923.g3HJNga58899@freefall.freebsd.org> <4.3.2.7.2.20020418095356.024354c0@nospam.lariat.org> <4.3.2.7.2.20020418143615.021a8460@nospam.lariat.org> <4.3.2.7.2.20020418202335.0229b540@nospam.lariat.org> <5.1.0.14.2.20020419101925.00ab2200@postoffice.pacbell.net>

next in thread | previous in thread | raw e-mail | index | archive | help

On Fri, Apr 19, 2002 at 12:08:25PM -0700, Greg Fortune wrote:
> At 08:30 PM 4/18/2002 -0600, Brett Glass wrote:
> 
> >Having a local build server is a nice idea, especially if you're
> >a large shop, but doesn't get newcomers a safe version to install
> >(important; if they're hacked they'll sour on FreeBSD) or give
> >an admin a build to which she can just upgrade quickly and know
> >that the latest holes are closed.
> >
> >--Brett
> 
> Brett,
> 
> I've been watching this thread quietly, as I am a "newcomer" to FreeBSD. 
> However your intimation that we'll run for the hills like children at the 
> first sign of difficult offends me.
> 
> First, anyone connected to the net who ever thinks that their box is ever 
> "safe" needs a reality check. Pretty good assumption for a newcomer, eh? I 
> came to FreeBSD because of its security and groups like this. If my site 
> gets hacked, I'm not going to "sour" on FreeBSD, I'm going to take 
> advantage of this group and all the other wonderful resources available to 
> this community and figure out what I need to learn to do better.
> 
> Just because we're new to FreeBSD doesn't mean we're sheep. We all know 
> where the sheep graze. Nobody ever told me that FreeBSD was easy. Nobody 
> ever told me it was secure "out of the box". What I heard was that if I was 
> willing to learn how to do it, FreeBSD has the potential to be one of the 
> most powerful and secure operating systems out there. I never thought that 
> all the work was going to be done for me, or that the process would be easy 
> of end. If technology was easy, sysadmins would get paid minimum wage and 
> have to wear polyester uniforms and funny little hats.
> 

It has been said (by various people, mostly those from the latter computing
age of PDP,VAX,and s/390) that a good sysadmin is one that should be able
to script (or otherwise automate/routin-ize) themselves out of a job.

Administration is just that. Read: management from the desk, planning,
communications, finding people and tasking them to deploy or implement.

Sysadmin-ship historically was maintaining system components that could not
maintain themselves. This included loading software from tape, backing up
to tape, providing user-requested features and fixing failures. With
modern systems, the OS is but one very small part of the whole equation.  
It is supposed to provide a user-computer interface to load and run
programs. It ought to be as automated and easy to implement as possible,
with high reliability and security.

There really is no reason why UNIX or FreeBSD should be harder to
deploy or implement than WinNT or Solaris.  A "solution" being the
buzzwords of these days, is exactly what it should mean. 
You are supposed to tell your boss "we need this functionality,
this vendor supplies something with that. It costs this much compared
to this other thing, and the implementation time is 1 day"
Unless you are truly masochistic, I'm pretty sure you don't want to spend your
nights trying learning the nuances of an OS that you picked because
you lost an OS flamewar with your favorite security mailing list ;)

In effect, the old saying  "Unix is userfriendly, it's just picky about its users"
should really ring less and less true as we develop more advanced
versions of it.

> Anyone who runs from an OS due to their own inability to learn how to 
> properly configure/maintain it can go run Windows and contribute to 
> Microsoft's ongoing track record for security and stability.
> 

It isn't running away. see above :) At a company, you don't *learn*
how to properly configure an OS, you do it. Years ago, I used to work at a place
where the motivational poster was "This is not a University".
Companies who hire administrators expect that their people know what's going
on and enough knowledge to run the systems. I suppose if someone
wants to migrate software platforms they should be educated to some
extent about the target platform, but how do we use this as a
FreeBSD selling point instead of hindering potential users
to begin using FreeBSD? (see comment below)

If it's going to take additional human resources to implement FreeBSD over
some other OS, with the same sort of stability and reliability, then
maybe it's not such a good idea. Sysadmins have better things to do
than maintain build servers and worry if the next patch breaks the OS.
They should be figuring out improvements in efficiency, user training,
uptime, infrastructure growth and assessing the needs of users or clients.

> You sound like you know exactly what you want. Why not put it together? 
> Hey, if you build it, it'll be done exactly the way you want it done, won't 
> it? Don't let this opportunity pass you up! Here's your chance to have a 
> piece of FreeBSD work perfectly for you! I'd code it, but my skills aren't 
> up to snuff (yet) and I don't figure that any of these kind people should 
> have to bear the burden of holding my hand. So I send my money to O'Reilly 
> and I spend my time learning how to do new things. One of these days I will 
> contribute to this body of work, but not until I've got the chops (I'd like 
> to fix bugs, not introduce them ;-) ).
> 
> If you aren't careful, one of these days you'll be griping about the update 
> mechanism I wrote, because I won't code it the way you want, I'll code it 
> the way I want.
> 
> Life is wonderful when you just deal with what IS. I read this list to 
> learn how to use the tools I currently have to do the best job I can, not 
> to watch theory wars via email. If you don't like things the way they are, 
> step up to the plate and do something about it. Otherwise, we all heard 
> what you said, so please remain in the audience and take your seat.
> 
> Personally, my hat's off to the fine folks who post the security 
> notices,  analyze the bugs, write the code, debug the code, and maintain 
> the source tree, all for a FREE OS! Without the people who actually do all 
> the work that you're complaining about, you'd have to do all that work 
> yourself (or "sour" on FreeBSD, as you put it). Try applying THAT across 
> 1000 servers sometime.

But then again, the objective of FreeBSD advocacy is to say that we
provide a suitable replacement enterprise level OS in a production
environment on mission critical systems.  The main argument
would favor improved binary patch system with minimal downtime and
maximum stability. If more people are to expected to adopt open source
operating systems, then Brett's point is that a successful binary
patch system is also an important marketing feature.

Normally, with commercial vendors, the sysadmin will consult those
technicians to result in a working solution to a patch.  That's the
price of a support contract. You are walked through the upgrade
process, and if something breaks, the vendor is responsible for fixing
it. (I'm talking about large implementations here, such as our S/390
support contracts. Downtime of over an hour is unacceptable, so
the protocols for microcode updates have been written by IBM
for our customized systems, and in the case they failed to forsee
an event, they have a tech on hand. Similarly, I've never seen
any particularly involved AIX patch because we needed to reinstall all
the core binaries for an update - we just install the binaries on
the patch CD, and half the time don't even need to reboot.)
With open source, mailing lists such as these are typically
your main source of support.  However, utilities facilitating
easy system upgrades such as a reliable binary patch system
would again be beneficial not only to existing users, but
also to potential users.

As a sidenote, linux operators commonly exclaim why I have to spend
hours compiling all of my core software, and then take down the system
to patch a system when all they do to fix vulns is to download the latest
rpm or deb. Similarly, microsofties download the latest SP (even though
it's usually 5 months later :) and reboot.

> 
> -Greg
> 
> P.S. If you really must respond to this, please email me directly. No need 
> to clutter the group with more witty banter or high drama.
> 
> Greg Fortune
> Megaton Technologies
> megatontech@pacbell.net
> ------------------------------------------
> "Those who say it can't be done should
> get out of the way of those who are doing it."

-- 
Peter C. Lai
University of Connecticut
Dept. of Residential Life | Programmer
Dept. of Molecular and Cell Biology | Undergraduate Research Assistant
http://cowbert.2y.net/
860.427.4542 (Room)
860.486.1899 (Lab)
203.206.3784 (Cellphone)

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?20020419182244.A27580>