1600 lines?  That's nothing.  If a solution was good yesterday, it is
good today.
Yesterday we fixed a problem by adding code to the kickstart %post section.
Today we solved another problem the same way.  Tomorrow we'll add more
code to %post.  Then more code.  We're solving problems, buddy. Don't
get in our way.

What about legacy machines?  Let them wipe and reload.


Of course, that speech was a "re-creation" of the "Evil Captain Kirk"
from um... Episode, um... [ Emcee whispers ] THIRTY-SEVEN... uhh...
called... [ another whisper ] "The Enemy Within."
 http://snltranscripts.jt.org/86/86hgetalife.phtml

But seriously...

You know the answer or you wouldn't have been asking about it.  Yes,
putting that much in %post is bad.  Legacy systems aren't going to get
updates.

A better question is, "How do I convince this other coworker that I'm right?"

The answer is: You can't.  Some people can't be talked into a better
solution.  However, if you provide an easier solution they'll switch
to it out of laziness.  Laziness is the strongest force in the
universe.

If you set up cfengine, puppet, or any of the fine Configuration
Management systems out there, and make it easier to add to that
configuration than to add to the %post stanza, he may start using it.
If you start porting things from the %post area, one at a time, to
your system, eventually he'll get the hint.  (Jedi mind trick: don't
say that you are back-porting his code. Say that you are creating the
equivalent in Puppet so that legacy machines receive the benefit too.
Eventually he may be the one to suggest removing the redundancy from
%post).

Putting things in cfegine or Puppet is, among other things, easier to
test that code in %post.  You can remove the barrier to entry by
writing a good "getting started" doc that includes a working example,
and recipes for getting it working, tested, etc.

Yes, you could try to change his mind. Or, you could just assume that
if you do a better job than him he will eventually go away.  That's
better than management, that's leadership.

And now... since no post of mine is complete with a boring story from
days of yore...

Back in the 1980s, before I was a sysadmin, Ethernet involved huge
thick cables that were literally drilled into ("tapped") to add new
connections.  It took ages and was very error prone.  One slip of the
drill and you ruined the hallway's ability to get on the network.
People at Bell Labs went on a quest to make Ethernet as easy to
install as telephones with modular plugs.  Everyone said that it
couldn't be done.  They said things like "Anyone that understands the
physics of how Ethernet works knows you can't possibly do it over
twisted pair."  So then Bell Labs invented 10base-T: Ethernet over
twisted pair, with telephone-like modular jacks.  The nay-sayers
response, "Oh, sure!  If you do it THAT way!"
http://www.instantrimshot.com/

Once 10base-T was invented the nay-sayers claimed it was an
abomination because "it wasn't Ethernet!".  It removed the "bus"
structure and turned it into a "star".  "That's not ethernet!  You
changed too much!"  Well, nothing changed on the software side.  There
was basically no barrier to entry other than buying a hub.  Those
nay-sayers used it anyway because it was so much easier to install.
No more drilling and hacking to add a new connection.  Whether or not
it was "real" Ethernet didn't matter because laziness is more powerful
than doctrine.

HTH.  Let us know how the project goes!

Tom
_______________________________________________
Tech mailing list
[email protected]
http://lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to