linphone-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Linphone-developers] wrong NAT keep alive method over udp


From: Philip Morant
Subject: Re: [Linphone-developers] wrong NAT keep alive method over udp
Date: Tue, 20 Jun 2017 11:07:54 +0100

Hi Jehan,

The server is sending empty udp packets, ie length 0.  I'm guessing
that these empty packets are a response to the udp double CRLF ping.
linphone is interpreting those empty packets as an error, and then
starting more REGISTER transactions.  The REGISTERs are happening
approx one every minute instead of the once an hour that I would
expect.  My ISP confirmed that I have multiple concurrent
registrations on the server.

As for switching keep alive off, that would presumably just trade one
problem for another (ie lost NAT mappings).  Also, it sounds like a
recompile.

So can you comment on those empty reply packets ?

Thanks
Philip

On 20 June 2017 at 10:51,  <address@hidden> wrote:
> Hi Philip,
>
> You are right, CRLF are supposed to be for TCP only.
> However CRLF keep alives are pretty simple and works with most SIP server.
> If it’s not the case, you can still disabled keep alives using
> linphone_core_keep_alive_enabled(LinphoneCore* lc)
>
> Best regards
>
> Jehan
> www.linphone.org
>
>
>
> Le 19 juin 2017 à 20:57, Philip Morant <address@hidden> a écrit :
>
> Hi,
>
> I'm running linphone version 3.9.1-1231-gecda7e6 on Ubuntu, configured
> to use UDP (through NAT).
>
> I read in https://tools.ietf.org/html/rfc5626#section-4.4.1 that keep
> alive with CRLF
> " MUST only be used with connection oriented transports
>   such as TCP or SCTP; it MUST NOT be used with connection-less
>   transports such as UDP "
> yet when I wireshark my calls (which abort roughly one in every 4
> calls) I find that linphone is sending double CRLF keepalives over
> UDP.
>
> The code in static int send_keep_alive(belle_sip_channel_t* obj) in
> listeningpoint.c appears to confirm this, since it sends a double CRLF
> regardless of what transport is in use.
>
> Please confirm / deny this analysis.
>
> Thanks !
>
> Philip
>
>
> _______________________________________________
> Linphone-developers mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/linphone-developers
>
>
>
> _______________________________________________
> Linphone-developers mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/linphone-developers
>



reply via email to

[Prev in Thread] Current Thread [Next in Thread]