bind replication

[Gary Wallis]

pyh at mail.nsbeta.info wrote:
> Torinthiel writes:
> 
>>
>> If you know which zone has changed, than you can do "rndc reload 
>> zonename".
>> If you don't, than "rndc reload" reloads all zones.
>> You could also try "rndc reconfig", but I think it will only load new 
>> zonesm the ones just added in configuration, not never wersions of old 
>> zones).
> 
> What I'm not sure is, given I have two hosts A and B, A is master, B is 
> slave.
> B fetches the zone files from A via rsync.
> But, how B knows that the zone files have been changed and then run 
> "rndc reload" to tell bind reload the zones?
> Thanks & Happy New Year!
> Regards.
> _______________________________________________
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users
> 
You need to create your own replication scripts as part of your dns 
management process. Simple Perl scripts even BASH works fine. But it may 
be a lot of work. And as it has been pointed out, you need to mesh with 
the great ISC/BIND software:

rndc

And I would go further pre processing everything using:

named-checkzone
named-checkconf

Just as some have mentioned, when you need to manage many NSs with views 
note that the perfect primary/secondary transfer system built in to BIND 
(that is really a great distributed responsibility solution and even a 
true DNS cornerstone) breaks. It breaks by adding complexity.

If a single entity is responsible for many DNS servers and zones and the 
entity has complete control over all the systems, then, I personally 
think that managed replication is much better.

See http://en.wikipedia.org/wiki/DNS_management_software for some free 
GPL systems that do work.

Disclosure: I wrote one of them.

Happy new years,
Cheers!
Gary