Solved Phones cycling between provisioned(ext online) and unprovisioned(ext offline) every minute

[jcraig_techsolutions]

Hi everyone!

3cx Version 16.0.493
Sonicwall TZ300 Firewall
Unifi POE Switch
Linux installation on a Hyper-v VM
Fanvil x4g Phones

All firmware on all devices is up to date.

I have the PBX and phones on a flat network now.. I did have VLAN's configured ad was experiencing this issue so I flattened it to rule out any vlan configuration in the sonicwall or switches.

The phones will discover and be listed as NEW. I create an extension and associate it with a phone. The phone behaves properly with a reboot and it successfully downloads and applies the config file from the PBX. The phone will show online and registered to the proper ext when you are looking at the phone..

In the PBX. the phone will constantly cycle from assigned to the ext, to listed as NEW again.. back and forth. The extension will cycle from red to green.. back and forth, at least once per minute.

I did a packet capture and all I can decipher from that is that the phone will attempt to get the provisioning file and get an OK response, then it will try again and get a 404 Not Found response.. over and over..



Any ideas??

Thanks!!

Joe

 

[JohnS_3CX]

Hi Joe,

Start by checking the power on the phones. They seem to be rebooting and trying to reprovision themselves.

• Is the PoE switch able to provide them with reliable power?
• What about the cabling between the phones and the switch?
• Can you hook up a few phones directly to the switch with short ethernet cables and see if this behavior stops on those specific phones?
• Do you have power adapters to eliminate PoE issues?
• Have you checked the switch GUI to see if it reports the power drain?

An example from a PoE switch reporting the power consumption:

 

[craigreilly]

I had a phone doing exactly this - and someone was using the line from the back of the phone for a PC connected to a USB-c to Ethernet adapter. When the adapter wasn't plugged into a computer it would send the phone offline. Plug the adapter in to the computer, and voila, back online. (Or unplug the PC ethernet from the phone and it came back up)
This was on a Grandstream however... but figured I would mention it as another possibility.

Plugging directly into the switch seems like a good start to me with 1 phone.

 

[jcraig_techsolutions]

Thanks for the replies.
I have confirmed poe is not an issue. Power consumtion is well under limits and the phones them selves don't actually power off.. if you are looking at the phone.. it appears online and working fine.. but the PBX tells a different story and when we try a call between phones.. it forwards to voicemail as if it's offline. We tried a power adapter, we plugged directly into the switch. The behavior is the same with all 15 of the phones. We tested the cables with our Cable tester and they all seem fine.

 

[JohnS_3CX]

Hi,

Can you please confirm if the following is all true in your installation?

• The X4 all have FW version 2.10.1.6836
• The phones have been factory reset again
• They are on the same network range as the PBX
• They are provisioned via PnP as "Local LAN" mode
• The default 3CX templates are being used (and no pending updates exist)
• The Phones section on your PBX shows them registering with the same IP each time (their IP doesn't change whenever they re-register and they are in the correct IP range that you expect them to be).
• The Hyper-V based VM has a static MAC assigned, and the NTP service installed. The network adapter must not be using the Legacy Network adapter https://www.3cx.com/docs/installing-microsoft-hyper-v/
• You do not use any special characters in the names, and the phone credentials are in good order (no red exclamation mark appears in the extensions list)

The above covers a range of things that can mess up registrations and cause the phenomenon you described, so try to make sure all those points are validated before moving forward. Do not use any custom settings while testing, to avoid introducing any additional variables while we narrow doown the issue.

 

[jcraig_techsolutions]

Hello JohnS_3CX. Thank you. I did confirm all those things.. I have found the solution.. Everything started working perfectly once I disabled Time Syncronization in Integration Services in the VM Settings on the HyperV. I guess that Time Syncronization was conflicting with the NTP service installed on the Linux OS in the VM..?
Anyway.. it's working. Thanks for the help everyone!
Joe

 

[BrenttG]

Hello JohnS_3CX. Thank you. I did confirm all those things.. I have found the solution.. Everything started working perfectly once I disabled Time Syncronization in Integration Services in the VM Settings on the HyperV. I guess that Time Syncronization was conflicting with the NTP service installed on the Linux OS in the VM..?
Anyway.. it's working. Thanks for the help everyone!
Joe

We will have to add this to the list we have of why its a bad idea to run 3CX on HyperV, we avoid it if at all possible, this is a new issue, but we have seen many, namely strange issues that almost seem firewall related, packets arriving out of order, and some not arriving at all, the few we have running on HyperV seem to have strange quirks crop up randomly more then the others, and need rebooted much more often, even if they are linux on HyperV. Then theres those occasions where HyperV has a Senior moment and looses all the VMs from the HyperV manager, but they are still live in the background, or better yet, when HyperV is clustered, and it has a similar Senior Moment and suddenly cannot find the VMs running on Any of the cores, so it fires them up on another core, so you end up with 2 instances of several VMs running, and everything goes absolutely haywire. vSphere/vCenter on the other hand has been rock solid.

 

[JohnS_3CX]

Hello JohnS_3CX. Thank you. I did confirm all those things.. I have found the solution.. Everything started working perfectly once I disabled Time Syncronization in Integration Services in the VM Settings on the HyperV. I guess that Time Syncronization was conflicting with the NTP service installed on the Linux OS in the VM..?
Anyway.. it's working. Thanks for the help everyone!
Joe

Hi Joe,

Glad to hear this was resolved.