Harald Barth wrote: > Fortunately no NAT involved. It only changed IP. So yes, it is not > there any more but will be again tonight. Same situation as my > previous capture file.
In which case, the IP address is still associated with the same client and when you move from the office to home the AFS server will be able to notify you of callbacks that are broken. If that address was re-used by another client that either reported a different UUID or no UUID, the address would be removed from that host entry. I am not opposed to making changes in this area. For example, implementing a algorithm such as: * new connection from IP address X * query X for UUID * if UUID(X) is found in an existing host entry: - add X to host entry - Ping all IP addresses other than X stored in host entry - If UUID(Y) != UUID(X), remove Y from host entry [we do this now] - If no response, remove Y from host entry [this would be new] However, I do not believe that it is going to solve your problem because your delay is caused by the fact that UUID(X) == UUID(Y) and UUID(Y) is not responding. Hence, there is a delay. Jeffrey Altman
smime.p7s
Description: S/MIME Cryptographic Signature
