/tftpboot vs. /var/tftp vs. something else?

[Stephen John Smoogen]

On Nov 14, 2007 3:52 PM, Jon Masters <jonathan at jonmasters.org> wrote:
> On Tue, 2007-11-13 at 08:34 -0700, Stephen John Smoogen wrote:
> > On Nov 13, 2007 6:35 AM, Alan Cox <alan at redhat.com> wrote:
> > > On Mon, Nov 12, 2007 at 09:20:28PM -0700, Richi Plana wrote:
> > > > On Mon, 2007-11-12 at 22:45 -0500, Warren Togami wrote:jA
> > > > > Edubuntu is using /var/lib/tftp as their tftpdir.  Should we use it as well?
>
> > > /tftpboot is the historical tradition going back about thirty years. Why
> > > break every script, every book and every third party management tool ?
>
> > What would be the proper RFC process to go over such a change? As in
> > say creating a /srv/fedora/ and start populating it with data or some
> > such?
>
> Don't take this personally, but that is absolutely the worst single idea
> I have heard so far in this thread. So now we go from saying "ah, let's
> move /tftpboot because it's the in thing to move stuff" to "hey! We can
> make everything fedora-centric so sysadmins need to relearn everything".
>
> Woooo! :-)
>

You missed the first part of the sentance. What would be the proper
RFC process for doing something. It doesnt mean that the suggestion is
a good one.. just that instead of some packager doing willynilly what
they want.. how should they approach the problem that gets a proper
technical engineering response.


> > Or say putting in a notice that we would like to move /tftpboot
> > to something else and please give us feedback on how we can accomplish
> > this with the least amount of pain, and then putting it to a proper
> > engineering group to vote on it?
>
> I vote for just not moving it. It's been around many times the lifetime
> of FHS and other "standards", that themselves were based upon what
> existed at their inception. If anything these "standards" should be
> updated with occasional exceptions, based upon years of reality.
>
> Yeah, I know, this is a very conservative attitude for a change. I guess
> we could just throw out the Linux kernel because it's also based on "30
> year old technology", but it seems to work well for most of us.
>

I would probably vote for the same thing. However, there needs to be a
process for making changes or innovations and getting them advertised
in a way that meets the communities needs.

-- 
Stephen J Smoogen. -- CSIRT/Linux System Administrator
How far that little candle throws his beams! So shines a good deed
in a naughty world. = Shakespeare. "The Merchant of Venice"