Transfer failed when calls on different nodes

[Alexander Mustafin]

Hello.

I have issue with transfer when calls use different FS nodes.

User1 --> User2 (FS1)

User2 --> User3 (FS2)

User2 transfer User1 to User3 = Transfer failed.

I've seen INVITE frome Node1 to Node2 which ends by 407 Authentication required.

kazoo.log

=======

2aad95de-a586-11e5-9cce-59492c4658b0|reg_authn_req:45 (<0.30329.952>) auth failure for Us...@test.example.com: not_found

Node1 freeswitch debug.log

+++++++++++++++++++++

2015-12-18 12:52:18.547821 [ERR] sofia_reg.c:2536 Cannot locate any authentication credentials to complete an authentication request for realm "54.54.214.231" (Public address of Node1)

Both nodes and kamailio placed on Amazon EC2, so I suspect problem related to NAT, because every server has internal IP and public IP.

If someone can help it would be great.

Best regards!

[James Aimonetti]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

If you can create a pastie/gist with the FreeSWITCH servers' logs for
the legs involved, it could help. The feature is nightmare transfer if
you want to dig in deeper yourself.

- --
James Aimonetti
Lead Systems Architect / Impressionable Scallywag
"If Dialyzer doesn't care, I don't care"

2600Hz | http://2600hz.com
sip:ja...@2600hz.com
tel:415.886.7905
irc:mc_ @ freenode
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBCAAGBQJWdHLYAAoJENTKa+JPXCVgLfMH/3vEC+PV9mtzN03l0M+JqLEc
hPM+2FzsiJaKMfsFEmyqAZ4JO2M3PEL6XdXf+0yA9SrPOcpTiE88Vy7SOt9GwNMM
M+fnvsDPyOt91GvoIFlDxv+hPRtpdNR1W4OUuOC6kpwvPR9GCNG9nCwq7049Kwm6
rFf4NlzM3ZqRwlkLZfdHo9fNPIUlyQlefcdex7jwGCyRY42IhIj5BWLFfu1ddibo
9RLlce2fWVy0kmMkqUkPiQg3L3L9Q/bqs+lCRtcrPBqNsKFLm0bukKVZpWk+B41P
NQ+rR2ho4y+RIILdsQ7Js6n3ARS4AD1czM9fG+PtseI7itERCR1/Y5+v7KQWi8w=
=3ejX
-----END PGP SIGNATURE-----

[Kirill Sysoev]

Hi Alexander,

Sorry for may be stupid question, 'cause it feels like it was discussed
several times in list,
but does your freeswitch nodes added to alc as carriers?

sup -necallmgr ecallmgr_maintenance allow_carrier FSNODE1 x.x.x.x/32
sup -necallmgr ecallmgr_maintenance allow_carrier FSNODE1 y.y.y.y/32

In this case no Auth would be required.

If I am not wrong it is a must for nightmare transfer to function properly.

Regards,
Kirill

[Darren Schreiber]

Hi there,
I agree with this as well. FS will reject nightmare transfer from an
unknown server I think.

Although, nightmare transfer should only be happening when your phone is
misconfigured :-) It may be using both proxies when it should be
preferring one or the other.

Either way, it should work (even when misconfigured) but Kirill is right
that the FS nodes themselves must be listed in the ACLs.

- Darren


On 12/18/15, 1:15 PM, "2600h...@googlegroups.com on behalf of Kirill
Sysoev" <2600h...@googlegroups.com on behalf of kirill...@gmail.com>
wrote:

>--
>You received this message because you are subscribed to the Google Groups
>"2600hz-dev" group.
>To unsubscribe from this group and stop receiving emails from it, send an
>email to 2600hz-dev+...@googlegroups.com.
>For more options, visit https://groups.google.com/d/optout.

[Alexander Mustafin]

Thanks for answers, guys. 

We use Cisco 7940/7960 phones, so this kind of transfer is default method. Blind transfer (REFER method works well).

I thought about that acl 'trusted', but I didn't see any points in documentation.

I'll add those IPs in acl and write back.

Thank you!

суббота, 19 декабря 2015 г., 2:33:33 UTC+5 пользователь kirill.sysoev написал:

[Alexander Mustafin]

Thanks for answers.

My tests show that all works fine, after I'd added IP in acl as suggested Kirill

[Kirill Sysoev]

Great to know, Alexander.

Btw, since it works, I am sure you've found my typo.
But to not to let others get cught by it:

 FSNODE1  should be changed to something like FSNODE2 in the second sup command :)

Regards,