[dns-operations] FreeBSD and the slaving of the root zone

[Michael Sinatra]

Roland Dobbins wrote:
> On Aug 2, 2007, at 10:33 AM, Michael Sinatra wrote:
> 
>> So I agree that the change should be backed out, but not because it  
>> was thrust upon sysadmins.
> 
> Anyone who runs a box is a sysadmin, even for his own personal  
> workstation.  The caching-only thing a lot of users do is key.

And each of those sysadmins had to do something before this change took 
effect on their systems.  I don't see how that changes my point.  There 
*is* a difference between the way FreeBSD does updates and, say, Windows 
Update.  (Okay, that's a bit extreme...)

The only people this change was thrust upon were people who would have 
installed a release that hasn't been released yet--which is why it 
needed to be backed out.  Now that the change is backed out, that won't 
happen.

Paul Vixie wrote:

> also, while this isn't a freebsd mailing list, i want to say that saying
> "no" to mergemaster when it wants to import some change from the mothership
> is a wizardly activity, 

I'd actually argue that running mergemaster is itself a wizardly 
activity, which is why I made the point that I did.

> and commits a boxowner (sysadmin) to saying no to
> this and related/dependent changes throughout the future.  every time i say
> "no" i kick myself for the future workload i just bought myself.

It's not clear how this is true with named.conf, but in general, I think 
you're right.

Randy Bush wrote:

> in the workshops, we have been teaching, "it is easy and simple to run a
> caching server, and often what you should do for an server which is not
> undernourished."  we will no longer be teaching this; life in the
> classroom just got a lot harder.

Exactly.

michael