On 23 Jul 2014, at 14:38, Jeff Pyle <[email protected]> wrote:

> HI Saul,
> 
> The ports in use are 16384-32768.  That provides for ~8000 session, yes?  We 
> were nowhere near that many calls.  So, it's much more likely that ports were 
> not being released for some reason.
> 
> There is a limited amount of T.38 traffic through these relays.  The majority 
> of the T.38-speaking devices are Adtran TA900-series, which, in general, 
> behave quite well.  That's on our side.  On the partner side, usually Sonus 
> GSX 9000-series.  I bet it's the unusual client that is the problem.
> 
> A SIP capture isn't practical here.  :(   I'll try to correlate the 
> exceptions with CDR records to determine if there is a particular trunk group 
> on my network that needs its attitude adjusted.
> 
> Until we can update the entire call flow (Opensips, Mediaproxy, all deps, 
> ...), or find a definitive misbehaving client, I'll restart the relays on a 
> schedule.  If I am able to reproduce the exceptions I'll provide debug data.
> 

Ok, thanks! If you manage to get some SIP captures feel free to email them to 
me privately.

--
Saúl Ibarra Corretgé
AG Projects



Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

Reply via email to