Phone Registers and then shows red under pjsip report.

I have a couple clients where the end point registers and shows green, and then turns red after a few seconds. trying to trouble shoot the cause, ANYONE?

Kinda Frustrated,
Jeff

Is this a self-hosted server?

Have you disabled the parameter “Allow transport reload” from the PJSIP settings?

Have you altered the “Default PJSIP Profile”?

Do you have enabled the parameter “Block ICMP Requests” from the Firewall Settings?

I’m having the same issue. Server is hosted in Google cloud and PJSIP settings have the external IP set in both the media and signal address. I’m running on Debian 11 with Vpbx 4.0.0-2.

All other settings (mentioned above) have been left default and no changes have been made to the default PJSIP profile.

The endpoint will register for exactly 3 seconds and then go unreachable. This is consistent using both TCP and TLS connections. I have tried multiple endpoints (softphone/phone) and they are behind a NAT router.

Outbound calling actually works fine but with the endpoint unreachable, incoming calls go to voicemail.

Any suggestions?

When an endpoint goes unreachable, it means that the server and client have not had a healthy OPTIONS conversation.

Usually, OPTIONS will get lost if there’s a misconfigured network, NAT Router, or a Firewall that hates UDP sessions.

The key is, to:

  1. Make sure your Networks and Transports are properly configured on the PBX.
  2. Make sure what’s mentioned above are properly configured.

Here’s a simple guide from VitalPBX: https://vitalpbx.com/blog/how-to-configure-nat-for-pjsip-endpoints/

With that being said, there are some crappy routers/firewalls that you cannot change the UDP session timeout settings, then you will have to lower the registration expiration times.

Hi Pitzkey,
My sip packet captures show exactly that… the server is requesting OPTIONS but getting nothing. The Default PJSIP profile on vpbx4 already has all of the recommended optioned enabled and my config for the network was set.

I have done some further testing and the connection works fine on UDP but not on TCP or TLS.

  • UDP - registration works, contact becomes available, incoming call works fine
  • TCP - registration ok, contact unavailable, incoming call fails
  • TLS - same as TCP

The router/firewall local to the softphone does not block outbound and the server firewall is set to allow both TCP and UDP to 5060 and 5061.

I’m not an expert on SIP but what could cause the connection to work fine over UDP but not TCP/TLS?

Can you please elaborate a bit more? Are the PBX and endpoint behind individual NAT routers?

It doesn’t necessarily block it, but rather “chocks” the connection.

However, when UDP sessions are getting blocked/chocked, TCP and TLS usually work fine. TLS is a bit more of a challenge to get working.

I speculate that something is misconfigured on the client side network, being that the PBX actually sends the OPTIONS but is not getting any response.

I would start with a packet capture on the firewall at the endpoint’s side and see what happens there…

You were correct, however, it’s an issue that something on the client side is stopping the OPTIONS message from reaching the endpoint, not stopping the endpoint from reaching the server.

When using UDP, the OPTIONS and NOTIFY messages reach the endpoint and are acknowledged but on TCP and TLS they don’t.

I’ll dig deeper on the client-side firewall to work out why.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.