On 29/11/2017 11:06, Ingo Wichmann wrote:
> Hi there,
> 
> Am 29.11.2017 um 06:34 schrieb Bryan Smith:
> 
>> One thing I love about the LPIC-1 program is that it exposes
>> candidates to _all_ common technologies on a system ... especially
>> useful for troubleshooting.  I.e., if one can't figure out why a
>> service isn't accessible, knowing _all_ the common places to look is
>> very useful.
> 
> I question, whether tcp wrappers are really commonly used today.
> I do not question, that they are available for some services.
> I do not question, that they are useful in some corner cases.

I'm probably going to agree in large part with Ingo here. I have not
myself used tcpd for anything in over 5 years. Like all tools that
originate from the 90s era, the original reason for tcpd seems to have
gone, way back then we seldom bothered writing access control into the
many daemons around and preferred to ship it off to something else.

But the software landscape has changed dramatically since and every
daemon I care about does it's own access control in some way. Even nrpe
has allowed_hosts, and were I to pick one app where offloading access
control elsewhere was suitable, it would be that. It's also a prime
candidate for running under xinetd but that's another story.

Sooner or later, tcpd will go the way of all old software, to my mind
it's now time to put it on the sunset path

> 
> I'd say, iptables, ACL's, capabilities, AppArmor, SELinux, cgroups are
> much more commonly used than TCP wrappers. So candidates are much more
> likely to be hit by one of these.
> 
> The most common service mentioned in this thread was ssh. But which
> admin still implements TCP wrappers on sshd, if he knows the Match
> keyword in sshd_config?
> 
> There is only one reason to me: because he did so 15 years ago.

hehe. That's the truth

> 
>> Alessandro Selli <[email protected]> wrote:
>>> Say you change in.ftpd service
>>> port to some non-standard one in it's configuration file.
>>> in.ftpd: 192.168.0.0/255.255.0.0 EXCEPT 192.168.1.0/255.255.255.0
>>> in /etc/hosts.allow is still going to work,
>>> iptables -I INPUT -i eth0 -p tcp --dport ftp -m conntrack --ctstate NEW -j
>>> REJECT
>>> no longer will.
> 
> FTP is an example of a protocol, that has lost importance. If this is
> the best use case for TCP wrappers we come up with, we should remove it
> from LPIC 1.

Don't deal with corporate billing departments much? When these guys
start moving files with 100 million CDRs around, everything except ftp
falls over and dies. But I get your point, the days when everyone had an
ftp.mycompany.com are long gone, and those users that do use ftp likely
have corporate-quality network firewalls doing what tcpd would do

> 
> Here's some indication, that on servers iptables is more commonly used
> than TCP wrappers:
> 
> * No distro comes with TCP wrappers blocking some service by default,
> other than iptables
> * Ansible has an iptables module by default, TCP wrappers are in galaxy
> * PuppetForge finds 4 modules tagged with 'tcpwrappers' - none of them
> is supported by puppetlabs. And 22 modules tagged with 'iptables', one
> of them the "official" puppetlabs module
> 
> So please TCP wrappers users: where is the indication, that TCP wrappers
> kept their importance since they have been introduced in LPIC 1 in 2001?
> Linux has changed, since!

I still like good old tcpd, I find it's a very valuable tool to
demonstrate access control concepts to bright-eyed shiny uni grads. But
I don't let them actually *use* it, mostly because none of the sysadmins
team ever think to look in hosts.{allow,deny} (unless they happen to
grep -r /etc)

I don't say tcpd is useless, it obviously has it's uses. But we can't
included everything in the exam, and LPI is not in the business of
certifying that someone knows hwo to use tcpd/httpd/sshd correctly. LPI
is in the business of giving employers a reasonable assurance that the
new hire probably won't destroy the servers while trying to do routine
work. I also like rcs but would never think of putting it on any exam.


-- 
Alan McKinnon
[email protected]

_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev

Reply via email to