Hi,
I’m experiencing an issue with a PJSIP trunk in NethVoice registered to our local provider (fixnet.cz). Initially, the trunk works fine — both incoming and outgoing calls function as expected. However, after some time (typically days), and possibly not yet discovered circumstances, incoming calls stop to work:
When calling the number from an external line, there is no ringtone, and the call does not reach any internal extension.
The inbound route has not been changed since it was working.
The NethVoice dashboard still shows the trunk as “registered.”
Outgoing calls continue to work without issue.
This SIP account worked reliably with a Grandstream VoIP gateway before switching to NethVoice, so I suspect the issue lies within NethVoice itself. I’ve noticed the problem may occur after unrelated configuration changes in NethVoice.
Restartiing the NethVoice app does not resolve the issue.
The only workaround I found is to delete and recreate the trunk in NethVoice / FreePBX, which restores functionality, but this is not a sustainable solution, regardless of that the problem is not indicated anywhere.
Could you help identify the root cause or suggest a more permanent fix?
Thanks in advance!
Thank you @mrmarkuz.
The firewall ports were not open, so I opened them now (I’m just wondering how it’s possible that it worked before and only failed after some time)
Browsing the logs perhaps following message, which occurs repeatedly during problematic periods, caught my attention:
2025-06-23T09:23:21+02:00 [1:nethvoice-proxy2:rtpengine] [1750663401.009178] WARNING: [80648ae4-2d68-4650-a515-5949c0037299 port 10088]: [core] No support for kernel packet forwarding available (interface to kernel module not open)
I have just updated nethvoice and nethvoice-proxy to the newest versions.After it, unfortunately, incoming calls not working again. After dialing from external phone: unusually long silence (15 sec), then ring tone, but no ringing withing NethVoice system. Trunk showing as registered, outbound calls working normally. No activity seen in nethvoice / asterisk / nethvoice-proxy logs during the call attempts. Nothing helped, just deleting and recreating the trunk again.
and launch it, check what the INVITE from the trunk does when it arrives.
A possible cause could be that OPTIONS are rejected and the INVITES find theports closed when it arrives. But also other bad things could happen.
The “No support for kernel packet forwarding available (interface to kernel module not open)” isn’t an issue: rtpengine would like to have this kernel module to forward RTP packets more efficently, but it’s not really necessary for the Nethvoice internal network configuration
select the INVITE and hit enter, you should se the details of how where the package goes. Take a screenshot and we’ll try to figure it out together. Also do the same with the OPTION as it could be the cause of the problems with the INVITE
Thans for your advices.
Seems the problem was caused by ports 5060-5061 not opened / forwarded on Firewall / NAT.
I was mislead by the fact it always worked first after creating the trunk so I have not put my attention this way.
Nevertheless I have learned about the sngrep tool in the process, so thanks again.
Unfortunately, the joy was too early… not working again. Here is the sngrep output:
UDP ports 5060-5061 and 10000 - 30000 opened and forwarded on firewall.
Problem seems to be resolved by switching the trunk transport from default udp protocol to tcp, which keeps the required firewall / nat connection open reliably.
Thanks all for your inputs.
