Still a newbie, still no joy. New issues...

There is one server.   I have been able to configure  it in a way usable 
from a local venus client and a laptop. The latter works both on the 
local network and on a remote one, using a openvpn VPN to access the 
server network.

On the remote network, there are two clients. Both are Fedora 13 boxes, 
configured identically (one being the laptop mentioned above). Still, 
one client can clog and use the /coda directory, the other one can't. 
 From a DNS perspective, the clients look the same from server side. 
Both clients sees the same DNS server.

The failing client gets no error message after the clog command. 
However, ctokens doesn't show any tokens at all, and /coda is thus 
empty. The one that works of course lists its token OK and /coda looks fine.

Both clients can connect to testserver.coda.cs.cmu.edu without problems(!)

Raising the log level I can see a difference in the communication 
pattern between client-server in the two clog cases (one more 
request-reply combo?), but the log does not reveal anything about the 
actual message contents.

On the failing client, venus.err tells me that cunlog actually discards 
the token it got through clog, but doesn't show using ctokens. This only 
happens after a seemingly succesful clog. So it seems that the client 
somehow doesn't want to use its token, nor want's to display it with 
ctokens.

Still confused, but at a higher level... any hint out there?

--alec