Because arp manages the communications between the physical hardware address and the "Internet address", dumping the tables is probably blowing away all active communications between the physical layer and the network layer.
No, this can't be the explanation. It's perfectly legitimate for arp entries to be randomly deleted (due to aging or being superceded by new mappings.) More likely the command in the UI to flush the arp table is either explicitly flushing the translation table, or something else is doing so as a side-effect... Hmmm, I tried an experiment. Went to a 3.5.3 gnatbox, looked at the active connections, flushed the arp table, looked at the connections again. My ssh session through the gnatbox was not interrupted and the connections already active were not disturbed. What release do you see this behavior in?
If so, this may be allowing the filters enough time to become active and block the renegotiation attempt of the traffic in question.
Just a thought, although not well expressed...
------------------------------------------------------ To unsubscribe: [EMAIL PROTECTED] For additional commands: [EMAIL PROTECTED] Archive: http://archives.gnatbox.com/gb-users/
