Incorrect call routing

I have a Freepbx version 16.0.40.11 Asterisk version 18.20.0. It has about 280 yealink phones, all setup as pjsip. Intermittantly the phones (not all) will drop off and reregister. After that a call that is for one of the phones that went offline will be routed to a different phone. The original phone is able to place a call but not recieve a call. The really interesting part is that if there is a Clearly Anywhere app paired with this phone the app will get the incoming call. the desk phone does not react at all to an incoming call. After some indeterminate time the deskphone will start to work again. This issue is totally random throughout the system. We have been unable to catch this in a pcap. CDR shows the call as normal. Has anyone seen anything like this. The AOR appears normal as well.

Thanks

Check your registration timeout on the yealinks and make sure it is less than the max timeout for the pbx. By default I think they may match which means no time for a delayed or sluggish registration. Ill verify the setting when I am at my desk sometime but just trim the reg time on the yealink so it happens well before it actually expires on obx side .

Are the phones remote to the PBX?

Yes the system is in a cloud.

Sounds like NAT/firewall is causing the issues.

1 Like

are you using TLS for connections? If so the local firewall is probably cutting off the TLS Session pre-maturely.

This definitely sounds like a NAT issue

I had this exact situation at a client. The desk phones appeared (to the client) to be online, intermittently they would not ring when a call came in but the ClearlyAnywhere app (PC or Mobile) always would.

The issue was their sonicwall firewall interupting the TLS Sessions. I ran packet captures on the PBX side as I had no control or access to the clients network (one of the few we don’t control directly). After analyzing the captures everything pointed to the local firewall interupting the TLS Sessions. Because neither end (Phone or PBX) knew about this until the session naturally timed out they both appeared to show the endpoints online.

The Clearly Anywhere app works because it doesn’t go throught the firewall with a TLS session, it connects to the ClearlyIP proxies and they connect to the PBX…

I finally convinced the local IT staff to call their network vendor and have them look at their sonicwall. After that the issue “magically” went away… No idea what changes were made…

Apparently there were ego’s involved who wanted the problem to be the phones…