Hi,
This seems to work. The issue is scale. I have would conceivably have to have a huntgroup definition in the huntgroups file for each NAS. And if I wanted 30 vlans, I'd have to have 30 definitions like the ones above in my users file for EACH one of my NAS's.
that would depend on what scale this would have to go to. certainly if each switch were to hav different VLANs for each of the types of users eg switch 1 vlan 200 for staff, vlan 201 for researchers switch 2 vlan 300 for staff, vlan 301 for researchers this woul dget very big very quickly. however, if each switch only needs to feed the same VLAN depending on the class of user - ie those 30 VLANs are are the same on each switch, then you can simply define a normal huntgroup for the switch eg in $place/raddb/huntgroup my-switches NAS-IP-Address == 231.123.241.123 my-switches NAS-IP-Address == 231.123.241.124 my-switches NAS-IP-Address == 231.123.241.125 my-switches NAS-IP-Address == 231.123.241.126 etc etc. then, in your example , the entry looks like DEFAULT Huntgroup-Name == my-switches, Ldap-Group == student User-Name=`%{User-Name}`, Tunnel-Private-Group-Id=177, Tunnel-Type=VLAN, Fall-Through = no (plus the others for each class of user) a 'clear scale' way would otherwise to be having an SQL table which defines each VLAN for each Ldap-group for each switch (or NAS) and use Perl or python to extract that info and return the attributes based on the request. alan