[mosh-devel] Mosh connections didn't come back after ca. 18000 sec over 2x NAT

[Keith Winstein]

Do you think you could reproduce this? It would be wonderful to
capture what the server thinks is happening (i.e. the debugging output
of mosh-server new -v).

On Mon, Dec 31, 2012 at 12:47 AM, Axel Beckert <abe at deuxchevaux.org> wrote:
> Hi,
>
> On Mon, Dec 31, 2012 at 01:44:34AM +0100, Christoph von Stuckrad wrote:
>> On 30.12.2012 23:55, Axel Beckert wrote:
>> ...
>> > And now that you mention it: That clock still works, i.e. it's updated
>> > on the client. It updates once a second, but the blue bar on top is
>> > still there and says "mosh: Last reply 45804 seconds ago. [To quit:
>> > Ctrl-^ .]". But nothing I type is shown on the terminal while the
>> > clock still continues to update the time every second.
>>
>> This really sounds strange, so the client is updating received data
>> to its display, but it 'does not really know' that it got the connection
>> back and so does not send?
>
> It looks like the data sent from the client is not (correctly)
> received on the server side.
>
>> You can check, whether your 'screen' still is accepting input, by
>> accessing the same 'screen-window' from another session by
>> connecting to it via "screen -S ...name... -x"
>
> Works as expected: I can type stuff and it shows up in the mosh
> session, too.
>
> But not all my sessions that hung that way had a screen session
> behind. It was just the one I've chosen for tcpdumping and one of
> those which updates each second. There were at least four sessions
> to two hosts which had no screen session attached and still behaved
> that way.
>
>> I think I might have had a situation as the above, but could not
>> find out, what was happening (I was riding a car and used a slow
>> lossy link, so I assumed that to be lost).
>
> Interesting. So a situation to cause this may be more common than I
> thought.
>
> Here it's a stationary link, but just over EDGE and still with varying
> signal quality (at least according to the APs signal strength
> indicator).
>
>                 Kind regards, Axel
> --
> /~\  Plain Text Ribbon Campaign                   | Axel Beckert
> \ /  Say No to HTML in E-Mail and News            | abe at deuxchevaux.org  (Mail)
>  X   See http://www.asciiribbon.org/              | abe at noone.org (Mail+Jabber)
> / \  I love long mails: http://email.is-not-s.ms/ | http://noone.org/abe/ (Web)
> _______________________________________________
> mosh-devel mailing list
> mosh-devel at mit.edu
> http://mailman.mit.edu/mailman/listinfo/mosh-devel