[Spce-user] [EXTERNAL] BUG nathelper

[Marco Capetta]

Hi Danny,

It is difficult to do a full analysis of your issue without a PCAP trace
and kamailio proxy logs in verbose mode of both the REGISTER message and
the following natping.

When NGCP receives a REGISTER message it should store in the location the
the "path" parameter that contains: the LB internal sip interface and the
LB external sip interface (socket). These information should be used by the
natping process to sent the OPTION messages to the devices.

Please check the information stored in the location to see if everything is
stored correctly. You can use the command:
     *ngcp-usr-location **--subscriber username at domain --verbose*


Thank you
Marco


On Sat, Jun 27, 2026 at 12:12 PM Danny Cappellari | ConnectLife S.p.A <
cappellari at connectlife.it> wrote:

> I would like to report several critical errors in
> /var/log/ngcp/kamailio-proxy.log
>
> ERROR: <core> [core/udp_server.c:821]: udp_send(): sendto(... dst:(
> 1.2.3.5:5060) ...) - err: Invalid argument (22)
> CRITICAL: <core> [core/udp_server.c:828]: udp_send(): invalid sendto
> parameters ... server is bound to localhost and attempts to send to the net
> ERROR: nathelper [nathelper.c:2458]: nh_timer(): sip udp_send failed
>
> *Cause*
>
> This is the nathelper / NAT keepalive (natping) proxy module. From the
> active config:
> - natping_processes = 7, natping_interval = 30
> - no natping_socket defined.
>
> *The proxy sockets are 127.0.0.1:5060 <http://127.0.0.1:5060>,
> 1.2.3.4:5060 <http://1.2.3.4:5060>, 127.0.0.1:5062 <http://127.0.0.1:5062>.*
>
> REGISTERS arrive at the proxy from the load balancer via loopback, so the
> socket stored for contacts is 127.0.0.1. When natping tries to send a
> keepalive to the contact's public IP using the loopback socket as the
> source, the kernel rejects sendto() with EINVAL(22) → "server is bound to
> localhost". With ping_nated_only=0, it also tries non-NAT peers.
>
> *Impact*
>
> - CPU waste on the 7 natping processes and hundreds of MB of logs per day
> (risk of filling up the /var disk).
> - NAT keepalives don't work for anyone: the NAT pinholes of phones
> actually behind the NAT are no longer refreshed → after the NAT timeout,
> incoming calls to those phones may no longer arrive.
>
> Do you have any solutions?
> Version mr-13.5.1
>
>
> *----------*
> *Cappellari Danny*
> CEO | *ConnectLife S.p.A*
>
> Tel. +39 045 2081019 | Mobile: +39 392 4206436
> cappellari at connectlife.it
>
> [image: 1770113372656]
>
>
> *Informativa Privacy (art. 13 GDPR):* I dati personali trasmessi con la
> presente comunicazione saranno trattati da Connectlife S.p.A. nel rispetto
> del Regolamento (UE) 2016/679. Sul nostro sito sono disponibili informative
> specifiche per le diverse categorie di interessati. La invitiamo a
> consultare quella di Suo interesse alla pagina seguente per conoscere le
> finalità del trattamento ed i Suoi diritti:
> https://www.connectlife.it/documentazione/
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sipwise.com/pipermail/spce-user_lists.sipwise.com/attachments/20260629/f320096c/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OutlookEmoji-177011337265666a6b3ac-06f2-4f65-880d-b5fc95633729.png
Type: image/png
Size: 34834 bytes
Desc: not available
URL: <http://lists.sipwise.com/pipermail/spce-user_lists.sipwise.com/attachments/20260629/f320096c/attachment-0001.png>