On Mon, 2016-11-28 at 20:44 +0000, David Woodhouse wrote: > On Mon, 2016-11-28 at 15:38 -0500, Stuart Luppescu wrote: > > > > > > I will do this. This seemed to coincide with upgrades to our > > analysis > > servers, but that should not have anything to do with the VPN > > gateway, > > should it? > > Maybe if they caused packets to be seen out of order, which would be > odd. And would have to be coupled with a strange bug like the > historical OpenSSL one, which is also unlikely. > > If you could ever reproduce that bug, I'd be very interested to see a > packet capture of the UDP frames on the public network, along with > the > master-secret and session-id headers from the VPN negotiation. > > You would be giving me *all* the traffic from that network connection > though, if you did that. And you'd want to sign off that session > before > sending them too, or I could use the secret/session-id to connect for > myself. So maybe just capture them to a file, and I can talk you > through interpreting them with wireshark. > Yeah, if only I could reproduce it. It generally only happens at night when I'm not using the computer, but it's still connected. (Last week I was doing some very, very long analyses that had to run over night, and I had to leave the VPN connected. I lost all that work when the VPN disconnected and had to do it all over again. I should have done it in a screen. Grrr.) I generally leave it running all the time, but I don't think I've seen it cut off during the day when I'm using it. I could set up Wireshark but I wonder if I have enough disk space to capture all the packets until it fails. -- Stuart Luppescu Chief Psychometrician (ret.) UChicago Consortium on School Research http://consortium.uchicago.edu
