Linux standard limitation is 3 nameservers.
Why do you want to use different nameservers on different members of the cluster? What do you expect to gain by doing this?
Cluster members must see identical things on the network. If you start abusing that then you will get unexpected behaviour.
You are right , I know the limitation of Linux OS....
Problem is that we have major 7 geographical regions covered by these Clusters. For the region's NS which is not listed in ADDM having name resolution problem.
Doing as you suggest is not going to help because you cannot control which machines in a cluster will discover which IP address.
If you must have the names you either need to use something like a DNS cache which can combine information from different naming servers together or to only scan those IPs which have the appropriate naming servers and use consolidation to combine ADDM information.
I've observed the same thing,, that we can't define scope for cluster member to scan bunch of specific IP's individually.