Webmin BIND failover configuration

I am just replacing our two aging BIND DNS servers used by our hosting department.  I am starting from scratch and installing the latest build of Centos, with Webmin and BIND installed.  I have configured two physically separate servers and configured one as a mast and one as a slave.  However, i cannot work out how to do a couple of things!

Firstly - you are still able to delete the slave zone, at which point i cannot force the slave to re-sync with the master, other than by re-creating a salve zone  and then it eventually talks.

Secondly - The previous systems experienced problems whereby if the mast server failed, after about an hour, the slave server would also stop resolving details and then required a reboot after the master was brought back online!  I cannot find any information on how to stop this behaviour, or indeed if it was just due to the state of the configuration on these servers!.  I have also tried testing the "convert to Slave/Master" options.  When i convert the Slave to a Master, i can edit the zone, but the Master still stays as a master, and does not update with the changes from the newly converted zone.  If i then convert back to a Slave zone, I get errors regarding not being a master - the only way i can fix this is to delete the zone completely and recreate a slave zone.  I have tried configuring them both to replicate each other, but that seems to just confuse them further!

Can anyone help?
Who is Participating?
Jan SpringerConnect With a Mentor Commented:
When using two DNS servers as you are, I would recommend a master/master configuration.

Do all of your changes on one server and create a script/cron job to send the data to the other server every X (15?) minutes.

With a script/cron job on the other [master] server, every X minutes (but not the same X minutes as the sending server), look for the completed tar transfer, extract the data into the production directory and do a server reload.
Amaze_ITAuthor Commented:
Thanks for the reply - but will the master/slave configuration we originally used be sufficient for this sort of failover?  or is BIND through webmin not capable of doing this?
Jan SpringerCommented:
If you have really short SOA data, when the master fails, the slave will expire any data it knows about if the master is down too long.

I don't use webmin but I don't think that it can sync the data between the servers.
All Courses

From novice to tech pro — start learning today.