Can some one help me in plain English

[Simon Waters]

Jeanne Ilchuk wrote:
> 
> When we set this up a few years
> ago, I thought the secondary would be there to answer
> queries in case the primary failed.

As far as I can see it is set up to act as a secondary
correctly.

The raid02 server isn't listed in the delegation data from the
GTLD-SERVERS but that isn't an unusual sin, and isn't the cause
of your problems.

Since RAID02 is apparently working as a secondary okay, I can
only assume you are using ns.wrlc.org to resolve queries from
local PC's. And that when you stop ns.wrlc.org this behaviour
stops ?

If so you need to make sure the PC's list both servers in their
DNS settings, either by updating the DHCP settings, or visiting
each client.

Mixing the resolution of local requests, and serving
authoritative data to the Internet from the same BIND instance
is considered less than best practise, and when upgraded I
suggest as a minimum you deny recursion to queries from outside
your own networks.


Switching RAID02 to be master is required if you expect to spend
more than a week upgrading, or if you need to make changes in
the meantime. Alternatively you could build a PC with a later
Linux and BIND versions to supply just the BIND functions and
drop that in as a replacement for ns.wrlc.org while you upgrade
it. Which is probably easier and safer than fiddling around with
the configuration on live servers.

-- 
Open Source Software Day at Exeter University Saturday 27th
April 2002
	http://www.dclug.org.uk/ossday/