[asterisk-users] Upgrade from version 1.6.24 to 1.8.12 - Retransmission timeout error

[Administrator TOOTAI]

Le 30/05/2012 14:44, Matthew J. Roth a écrit :
> Considering that you made progress on your initial problem by setting 
> "nat=force_rport" (resulting in connected calls with no audio) and now 
> you're mentioning the use of "externaddr", I'd recommend a very 
> careful reading of the "NAT SUPPORT" section of sip.conf.sample in the 
> configs directory of the Asterisk source tree.

I did read all those documentation, belive me. Also keep in mind that I 
*ONLY* face this problem with this provider, people using voipbuster or 
sipdiscount should have the same problem.

Concerning externaddr, this test server -dedicated to asterisk- being 
running in VM since ages, I never would suspect a NAT issue! Especially 
if previous 1.4 and 1.6 version are running smoothly ...

> In Asterisk 1.8, there is a new configuration option named 
> "media_address" which may be of particular interest.

media_address seems not an option, can be set only in general not per peer.

> This is confusing because your first email said you had "nat=no" in 
> your working 1.6.24 setup, but everything you're saying indicates a 
> NAT problem to me.

Again, 1.6 version is perfectly working with this setup and conf files, 
and before 1.4 was too. And those both asterisk versions with *this* 
provider.

> . A diagram showing all network elements between your Asterisk server 
> and the remote host would be helpful.

Phone registration:

phone (Snom320 and GS GXV3175) -> firewall1 (linux router) -> Internet 
-> firewall2 (linux router) -> VM -> phone account

Call:

phone account -> Out of VM -> firewall2 (linux router) -> Internet -> 
Peer IP -> ???

> To avoid further confusion, please include full and unaltered logs, 
> SIP traces, and configurations in future posts.

During the time you and Andres replied to my post ;-) I got the same 
idea then him; and guess what, it's working! So problem is Asterisk 
1.8/10 in VM _only_ this provider(s) which are all Dellmont services.

Can someone confirm the problem?

Question is now, who is faulty? Should I open a bug?

Thanks for your time and support.
-- 
Daniel