bug in reload single zone feature

[John N Dvorak]

version: BIND 8.2.1

I could not find a reference to this in any previous posts.  I apologize
if this is a duplicate.
This may or may not be a bug--I may just be misinterpreting the feature.

Using the single domain reload (ndc reload domain.com), you can
instruction your master to reload a single domain.  However, there seems
to be a problem if you request multiple reload requests within a certain
period of time, even after the zones have transferred to the slaves.

For example:

1. make change to domain.com zone file, update serial
2. ndc reload domain.com
3. "Zone scheduled for reloading"
4. master updates its own record.  master sends NOTIFY to slaves.
5. slaves respond NOTIFY answer
6. zones transfer to update slaves
7. make another change to domain.com zone file, update serial
8. ndc reload domain.com
9. receive "Zone is already scheduled for reloading"

After receiving this message, it seems to be that the master will not
reload the zone, nor send requests--it basically junks the request (more
accurately, perhaps: ndc junks the request).

I tried again in 20-30 minutes with the same result.
When I tried again in a bit over an hour, the process was handled
properly.

Here's the big question: how does BIND handle its request TTL for
single-zone reloads?  Or, maybe the question should be: what is BIND
using as a criteria for a successful reload of the zone record?  Or is
there some general TTL for the reload process that basically locks out
further requests to reload the some zone for X seconds?

JD

===========================================
John N Dvorak | dvorak at capu.net
Director of Technology
CapuNet, LLC - Corporate Internet Solutions
(301) 881-4900 x8018
===========================================