### NS record problem ###

A.R. abdul_rahman at sunbursthospitality.com
Wed Sep 15 15:42:40 UTC 1999


This is a somewhat longwinded email.  So please bear with me
as i explain what my problem is.

Here is a description of the servers i have configured:

Server 1: Red Hat Linux running Bind 8.2, one interface listening
on one ip address

Server 2: Sun Solaris 2.6 running Bind 8.2, one interface listening
on two ip addresses.  There are two named.conf files that each have an
ip address, one for internal dns and one for external dns, listening
to the different ip addresses respectively.  They both service the same
domains, ie, the internal ip services internal addresses for abc.com.
And
the external ip hosts external addresses for abc.com.  This is a split
brain configuration.

The records in question are as follows:


@ IN SOA ns1.abc.com. root.ns1.abc.com. (
        10
        3600
        3600
        604800
        86400   )

abc.com.        IN      NS      ns1.abc.com.
abc.com.        IN      NS      ns2.abc.com.

sap-server    IN    NS     serv1.abc.com.
sap-server    IN    NS     serv2.abc.com.
serv1              IN     A        10.0.0.10
serv2              IN     A        10.0.0.11

The goal of doing the double NS record configuration is
to be able to have 'auto' failover for dns to be run
by an auto sensing script feature on another set of network
boxes.  The bottom line is that if serv1 does not respond, serv2
can be brought up, transparent to users.

I will refer back to these records shortly.  The /etc/resolv.conf files
in server1 and 2 point to their ip addresses as name servers.



More information about the bind-users mailing list