I have a NW6.5SP5 cluster attached to two lan segments (so each server
has two NICs).

I've added an additional secondary IP address for a cluster pool (in the
load script), so that an IP address of both lan segments is bound to
this volume/pool. Mapping directly to this IP addresses works fine.

Inside the login scripts I like to use the NDS way to map the volumes
for the users (e.q. "map x := volume_server.context"), instead of
mapping to the secondary IPs of that pool/volume.

But on doing this, all clients use the pool IP address of the *first*
lan segment, because this is the IP which is listed as cluster resource
address at ConsoleOne.

Is it possible to configure that the map command should use the
secondary IP address which lies within the workstations subnet? This
should prevent unnecessary routing on the servers.

Thanks in advance.