Ryan.McCain at dss.state.la.us
Fri Sep 28 17:01:17 UTC 2007
That seems to be the best approach. zone1 is a bit different as its acting as my "master", while zone2 and zone3 are the same. I might be able to have zone1 look similar. I'll need to test around.
>>> On Thu, Sep 27, 2007 at 7:30 PM, in message
<barmar-4352F6.20305427092007 at comcast.dca.giganews.com>, Barry Margolin
<barmar at alum.mit.edu> wrote:
> In article <fdh9mk$2ek0$1 at sf1.isc.org>,
> Christian Robottom Reis <kiko at async.com.br> wrote:
>> On Thu, Sep 27, 2007 at 04:56:58PM -0500, Ryan McCain wrote:
>> > I have 3 zones that all point to a shared.zone file via $INCLUDE. When I
>> > add a new DNS record to shared.zone I now have to:
>> > 1) Update the reverse zone
>> > 2) Update the serial # in zone1
>> > 3) Update the serial # in zone2
>> > 4) Update the serial # in zone3
>> For the case of 2, 3 and 4, if they are truly identical you could do as
>> I do and just have a master file symlinked between them.
> But maybe they're not completely identical, they have some unique
> entries and then a bunch of shared entries in the shared.zone file.
> In this case, what you could do is put the SOA record in the shared.zone
> file as well. You'll need to update this serial number whenever you
> update shared.zone or any of the individual zone files, so at most
> you'll ever have to update two files when you add a new DNS record.
> This has one misfeature -- if you update zone1, the serial numbers of
> zone2 and zone3 will also be updated, and the slaves will transfer all
> of them. But unless your zones are huge I wouldn't worry too much about
More information about the bind-users