A policy for removing named.conf options.

[Warren Kumari]

On Thu, Jun 13, 2019 at 2:43 PM Evan Hunt <each at isc.org> wrote:
>
> > > Is it really much of a hassle to leave the obsolete options in the
> > > parser, but just ignore them?
>
> IMHO, it depends on the option. For something like "managed-keys" and
> "trusted-keys", there are clear security implications.  Once those are no
> longer effective, it would be dangerous to have named ignore them - even
> with a logged warning. Operators who didn't notice the log message wouldn't
> realize they were running without the security they'd configured.
>
> For something like "cleaning-interval" or "max-acache-size", IMHO it would
> be safe to let it slide. With "dnssec-enable" or "queryport-pool-ports",
> maybe those fall somewhere in between, I could see arguments either way.

I personally think that while it may or may not be a hassle to have
the parser ignore them, it would be a significant operational risk /
annoyance.
Having knobs which you can twiddle which don't do anything leads to
all sorts of annoyance -- if I'm running low on space for cache, and
spend much time twiddling the "max-acache-size" knob before
discovering that someone has simply snipped the wires to it, I'd be
super-grumpy.

I'm expecting some issues when knobs get deprecated (and I'm likely to
run into a few lurking in old configs which have grown over time), but
I'd rather have named not start just after I've upgraded it than be
running in some partially undefined state.

W

>
> In any case, if we're going to make a policy that covers the whole range of
> possibilities, then it needs to address the case when an option must
> removed, and how to ensure operators aren't blindsided by that.
>
> --
> Evan Hunt -- each at isc.org
> Internet Systems Consortium, Inc.
> _______________________________________________
> Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list
>
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users



-- 
I don't think the execution is relevant when it was obviously a bad
idea in the first place.
This is like putting rabid weasels in your pants, and later expressing
regret at having chosen those particular rabid weasels and that pair
of pants.
   ---maf