dhcp and dynamic dns updates

[Jiann-Ming Su]

On Tue, Mar 3, 2009 at 3:32 PM, David W. Hankins <David_Hankins at isc.org> wrote:
> On Tue, Mar 03, 2009 at 11:15:17AM -0500, Jiann-Ming Su wrote:
>> the same way.  Is there code in the dhcp server that would cause this
>> type of blocking?  Thanks for any insight.
>
> DDNS updates are done with a stripped-down BIND 8 libres we call
> 'minires'.  It is synchronous, which means while updates or
> resolutions are pending, the server waits.
>
> In some versions of 3.0.x, it is even worse, because we would
> initialize the resolver library in two different places (depending
> on wether a ddns update or removal was first), and the variable
> assignments were inconsistent.  The difference is 1 1-second timeout
> versus something like 3 timeouts over 15 seconds total.  I don't
> remember what version we fixed that in.  But it's always a good idea
> to track maintenance, and 3.0.7 is the latest 3.0.x track.
>
> We're working on an asynchronous DDNS update method, which is the only
> real solution to this problem.
>

Turns out we were able to reproduce this problem in the lab with a
minimal number of hosts.  Just had to use the right tools.  Thanks for
clarifying.  Turned out to be a good exercise for us.  We just need to
make sure our DNS server's network configs are correct.


-- 
Jiann-Ming Su
"I have to decide between two equally frightening options.
 If I wanted to do that, I'd vote." --Duckman
"The system's broke, Hank.  The election baby has peed in
the bath water.  You got to throw 'em both out."  --Dale Gribble
"Those who vote decide nothing.
Those who count the votes decide everything.”  --Joseph Stalin