Yealink phones don't come back after internet outage

[Peter Richardson]

Hi All,

Still having this issue, posted twice previously here:

https://www.3cx.com/community/threads/yealink-phones-dont-come-back-after-internet-outage.63485/
https://www.3cx.com/community/threads/yealink-phones-dont-come-back-after-internet-outage.66046/

I assume @Mark Phillips is still having this issue?

• 3CX Version, Pro Annual 15.5.15502
• Server OS Debian 9
• Is the 3CX Server Hosted and where? - Our own data centre, located in the same city with 4 other 3CX VMs and 1 Asterisk VM
• IP Phone Make/Model/Firmware: Yealink T48S 66.83.0.55
• Provisioning Method: VPN
• Trunk Provider: DID Logic
• Has the Firewall Checker passed: YES
• Are custom Phone Templates being used: NO

I tested in our office by plugging a phone into the switch, then unplugging the router from the switch. Phone was set to DHCP (same as the problem site). I re-plugged the router after 6 minutes and the phones reconnected within 30 seconds. I tried again and re-plugged the router after 15 minutes. Same result. Tried again after 3 hours, same result.

Our data centre and all of our clients are using pfSense routers. They are very stable, we have very long uptimes with them and generally only see them reboot on power failure or if we are updating pfSense.

I still don't know why the Yealink template and/or web gui shows "server retry counts" at 3. @YiannisH_3CX said (in the first version of this thread, see link above) that it didn't affect testing. I can confirm that it didn't affect my testing here in the office either. But either the phone is ignoring that setting (why?) or........(any other ideas?)

Client called me today saying that the phones were down. I immediately logged into their router via the VPN, so their internet connection and their VPN connection was up, but the phones were not.

In the 3CX Event Log:
13/09/2019 18:18:51 - all trunks went down. I assume an internet outage at the data centre. BUT all I can see in the logs is that the trunks are down - no mention of extensions being down until...
14/09/2019 00:01:39 - all phones went unregistered
14/09/2019 00:14:02 - all phones re-registered
then there's a big gap in the logs - a few hacking attempts, database cleanup, etc, but nothing about phones going unregistered
16/09/2019 09:14:46 - 2 days later all the phones re-registered after I instructed the client to reboot the PoE switch. At this point I could log into their router via the VPN but prior to this, 3CX was showing all phones

Any ideas would be greatly appreciated!

 

[JohnS_3CX]

Hi Peter,

Under V16 SP2 Debian my T4x-series yealink phones reconnect with no issues whether the server was simply rebooted (short downtime) or shut down for a couple of days and then brought back online (long downtime). This was a local scenario with T46 and T48 and T32 on the same PBX so the provisioning method is the same as yours in principle.

Currently you are describing a system that is on V15.5 and on a phone on FW 66.83.xxx instead of the 66.84.xxx generation of firmware.

Would it be possible to consider bringing the PBX and firmware up to date before troubleshooting further?

 

[Peter Richardson]

Thanks very much John. Good point. It is still strange behaviour, and nobody had any answers to this problem while 15.5 was still in vogue, but we may as well rule out the easy ones first - I'll update 3CX and firmwares and report back.

 

[JohnS_3CX]

Hi Peter,

Great, let's see how it goes then. I would recommend resetting and reprovisioning the phones after a PBX and firmware update to ensure you are ready to go immediately and then cause a VPN dropout for a bit during a time that will not affect the customer (so we can replicate the problematic conditions)

 

[Mark Phillips]

I haven't noted any recurrence of the issue I originally reported with V15.5. However, we did upgrade to V16 a while back.

 

[Peter Richardson]

Thanks Mark. I have run all the updates (there were a few!) and updated phone firmwares. Now we wait.

I do have a Yealink W52 that is registered to a different 3CX running latest version that experiences the same drop outs, but I am yet to visit site to see what the issue here is.

• It is registered via remote SIP / STUN.
• The site has a basic home modem/router (Technicolour TG-789)
• No VLANs, no NAT or firewall trickery
• The site has 2 phones - Yealink T21P E2 and Yealink W52 cordless
• T21 is always up, no worries with this one.
• W52 is often offline. Client is usually unaware of this but I notice when checking extensions. I assume the only time it comes back online is when they have a power outage and it has a chance to reboot.
• W52 is plugged into another Technicolour TG-789 - this one is only being used for extending WiFi to the other side of the building (not exactly an amazing WiFi solution, but they couldn't afford Ubiquiti)

I can just go to site, reboot it and it'll come back, but I'm more interested in finding out why it is happening. On reboot, it does seem to stay connected for quite some time - I'm not sure exactly as I haven't timed it, but I'd guess 1 - 3 months.

Any ideas?

 

[JohnS_3CX]

This seems like a completely different situation unrelated to the above issues. Here your scenario is a W52P base behind a router, behind a second router and then reaches the PBX via internet using STUN

As a first step, follow the link from here to update the W52P base to 25.81.0.30 https://www.3cx.com/sip-phones/firmware-upgrade-yealink-dect-w52p/

The home routers are likely not as predictable as an enterprise grade firewall, you will have to ensure that no ALG is running on them, that the ports are forwarded as per 3CX recommendations (different sip and media ports per each phone), and that the VoIP profile of the provider is not interfering with the 3CX phones so I would avoid port 5060 in this case and do something like 5065 for one phone and 5066 for the other as an example.

Also note that these devices are usually provisioned via TR-069 by the provider so they may overwrite some of your settings, or in some cases not even give you access to certain settings that can be hidden from the UI of the router.

 

[Peter Richardson]

John, I really do appreciate your detailed and very helpful replies.

I will...

• remove the second router and replace with a Ubiquiti In-Wall AP that has an extra PoE pass through Ethernet port for the phone
• use a different sip port
• ensure sip and rtp ports are correctly forwarded through the firewall
• Update firmware on phone if required
• I might even try setting up the W52 with OpenVPN (which we use on most other sites), which would probably solve all my problems in one fell swoop
• Report back with the good news!

Thanks again!