Intermittent Event Makes Random Inbound Calls Unable to Hear Our Outbound Audio

[kl116004]

Hi, I'm running a 3cx install at a small business, our general setup here is:

• Linux 3cx Appliance
• Patton Smartnode Gateways
• Yealink IP phones

Seemingly without cause, random callers start being unable to hear our outbound audio. This might happen every couple months, and I'll reboot all the boxes in the system and that has usually fixed it. I have not determined any cause for why it starts, why it stops, or who it affects and when. One caller might be able to hear us one call, same call might not get our outbound audio.

This is the worst kind of problem, and I have no idea what is going on because I haven't determined ANY cause yet.

 

[eddv123]

What Patton models are we talking about here ? either:

ISDN BRI/PRI or Analogue (BT lines) for external calls.

FXS Gateways (for analogue endpoints and FAX) with a SIP trunk for external calls and DDI's.

Please define make, model and firmware if possible. A well as location of the devices and 3CX (on prem or hosted).

 

[kl116004]

What Patton models are we talking about here ? either:

ISDN BRI/PRI or Analogue (BT lines) for external calls.

FXS Gateways (for analogue endpoints and FAX) with a SIP trunk for external calls and DDI's.

Please define make, model and firmware if possible. A well as location of the devices and 3CX (on prem or hosted).

3cx is on prem (16.0.1581)

2 x Smartnode 4114 (4 FXO)
1 x Smartnode 4114 (2 FXS, 2 FXO)
Gateways are for our external analog phone service and a fax machine. Running on software R6.9 2017-03-13 H323 SIP FXS FXO

Thanks for the assistance.

 

[eddv123]

To troubleshoot this due to its intermittent behaviour will be a real nightmare, you can run a trace from the Patton gateway which usually gives us everything we need, however the issue we have is the amount of time that it takes for the issue to occur.

One thing I would check on the lines is the service that is running, here in the UK we have something that the provider adds sometimes called "Featureline" or "Redcare" this is where PBX-like services are added to the analogue line by the provider.

This can cause lots of issues with cross over, audio etc so it is worth checking with your provider what service you have. A bill from them may also highlight this.

 

[kl116004]

I'm embarrassed I didn't make this abundantly clear (I'm stressed out of my head), it's happening RIGHT NOW and I haven't been able to mitigate it at all. I'm happy to try running a trace if you give me the particulars.

 

[eddv123]

Here you go trace from the CLI of the Patton, you can use either telnet/putty or the console cable that comes with the device:

enable
show running‐config
debug call‐router
debug call‐control
debug ccfxo
debug fxo
debug context sip‐gateway signaling detail 5
debug context sip‐gateway transport detail 5
debug context sip‐gateway error

Then "no debug all" after to cease the log flow. Cut and paste from the debug into a txt file and send it over. You can also log from 3CX with PCAP as well but I think the Patton will tell us more.

https://www.3cx.com/docs/capture-network-traffic/

Remember to let me know the times and numbers involved.

 

[kl116004]

One thing I would check on the lines is the service that is running, here in the UK we have something that the provider adds sometimes called "Featureline" or "Redcare" this is where PBX-like services are added to the analogue line by the provider.

Furthermore, we do have some of these types of features enabled, mainly for hunt-group functionality, we use three main numbers, and each is actually a hunt group of several analog phone lines. All those lines come into the local PBX.

 

[kl116004]

OK, got a live trace with an example.

Phone number is redacted
2:39:00 is the rough log time where the call occurs.

I took this call myself, could hear audio from caller, but he clearly could not hear me.

 

[kl116004]

Another wrinkle to this that I have noticed is that when one of these is occurring, it appears to affect calls that come in while it is still connected, so calls exhibiting the problem can arrive in bursts.

 

[eddv123]

I cannot see anything that jumps out as being a major issue.

The only thing that I would say is that there are a lot of ReINVITE's that I can see in this trace. I wonder if disabling "Supports Re-Invite" which is under "Options" may make a difference.

Is your PSTN gateway options >> Hunting mode is normal or cyclic ?

Cyclic = In hunting mode enables specifying a timeout after which the next destination interface is tried when there is no answer at all from the destination interface. Perhaps this may help also.

 

[kl116004]

Thank you for looking.

1. If I turn off the reinvite option, do I have to generate a device config and push it to the gateway?
2. The smartnode in the trace was in cyclic, other two were in normal.

 

[eddv123]

Good question, If you check the config file from 3CX it will tell you if the setting is transferred (search for re-invite or something like that) It maybe quicker however to just re-provision.

 

[eddv123]

Any update on this one? Has the issue reoccurred at all?

 

[kl116004]

After updating the Patton software, the problem appears to be fixed.