On 24/02/16 13:49, A.L.M.Buxey@lboro.ac.uk wrote:
Hi,
update { &outer.session-state: += &reply: &outer.session-state:User-Name = &User-Name }
update outer.session-state { User-Name = &User-Name }
?
what does debug show?
With your suggested change, for some reason it does a noop (8) update outer.session-state { (8) &outer.session-state:User-Name = &User-Name -> "iser-linauth@bris.ac.uk" (8) } # update outer.session-state (noop) The outer User-Name should at this point be anonymous@bris.ac.uk so I would expect this update operation to make a change and set &outer.session-state:User-Name to iser-linauth etc. I'm not sure if I'm tying myself in knots here. Basically, in the past we've decided on the user's VLAN in outer post-auth based on their inner username, which we access like %{reply:User-Name} with use_tunneled_reply=yes. This doesn't work with resumed sessions in FR3 like it did on FR2 and we haven't been able to figure out why. We don't want to cache the VLAN number itself, i.e. don't want to make the VLAN decision in the inner post-auth but we do want to cache the inner User-Name so it can be easily accessed later to do the VLAN calculation on resumed sessions. Cheers, Jonathan