All extensions unregister and re-register simultaneously

[Jagk23]

We're having an issue where all of our extensions unregister and re-register all at the same time.
We're using a variety of Yealink, Polycom and Grandstream phones with a variety of different connection methods including Direct SIP, Local LAN, etc.
Doesn't matter what phone or connection method.
Version is the latest version 16.0.4.493 - only started happening on Friday after this update was installed on Thursday evening.
Server is hosted in AWS - not had any previous issues and can see no connectivity issues between site and AWS server. No performance issues on the server either.
Obviously all calls drop when the extensions unregister. Seems that they unregister approximately every 90 minutes to two hours.

 

[AWS2P]

Hi, try to give us better overview starting here
https://www.3cx.com/community/threads/information-to-provide-when-requesting-help.67558/

You said your PBX is AWS hosted, so do you have multiple connection locations ?
As you said direct sip, lan ...whatelse ? SBC ?

 

[Jagk23]

3CX Version -16.0.4.493
Server OS - Debian 9 Stretch
3CX Hosted in AWS - multiple connection locations
Phones are just provisioned using Direct SIP and Local LAN (Sorry thought there was other options).
Yealink T21_E2's using up to date firmware, Polycom VVX 300's using up to date firmware, Grandstream GXP1610 and GXP2130's all using up to date firmware.
Firewall Check is fine
Customer templates are not used

 

[Jagk23]

Not sure if helpful but i've pulled this error from an example Polycom phone.
0127095244|app1 |4|00|Event: Unrecognizable key - appKeyFunc = 327, scan code 0
0127095323|cfg |5|00|Prm|Parameter voIpProt.SIP.allowTransferOnProceeding requested type 2 but is of type 0
0127095323|app1 |4|00|Event: Unrecognizable key - appKeyFunc = 327, scan code 0
0127095323|sip |5|00|CUser::CallNew Call SubscribeDialogBla absent
0127095323|cfg |5|00|Prm|Parameter voIpProt.SIP.allowTransferOnProceeding requested type 2 but is of type 0
0127095323|sip |5|00|CUser::CallNew Call SubscribeDialogBla absent
0127095333|so |4|00|[SoNcasC]: appncascontext termination:1
0127095333|so |4|00|[SoNcasC]: Case Handling termination:1
0127095333|sip |*|00|Sip UnRegister Usr:244 Dsp:XXXXXX Auth:'XXXXXXXInx:0
0127095333|sip |*|00|SipUserRemove: user 0 being removed.
0127095333|app1 |*|00|SoRegistrationEventLineChanged - success lineIndex 0 RegListSize 0
0127095334|app1 |*|00|SoRegistrationEventLast - new AppRegLineC, szUser = 244
0127095334|so |4|00|[SoNcasC]: appncascontext termination:1
0127095334|cfg |5|00|Prm|Parameter attendant.reg requested type 2 but is of type 0
0127095334|so |4|00|[SoNcasC]: Case Handling termination:1
0127095334|cfg |5|00|Prm|Parameter reg.x.outboundProxy.port requested type 0 but is of type 2
0127095334|sip |*|00|Sip Register Usr:244 Dsp:XXXXXX Auth:'XXXXXXX' Inx:0

 

[AWS2P]

which phones are using direct sip? how many phones on that location?

 

[Jagk23]

Sorry I forgot to include the DECT phones, we also have 5 of the Yealink W52P DECT sets.

The DECT phones all use Direct SIP and we have approximately 4 Grand Stream phones using Direct SIP as well as 2 of the Yealink devices at our branch locations.

 

[AWS2P]

was it previously working fine and not anymore since you update to v16 SP4 ?
You should use an SBC instead of direct sip, easy to set and reliable

 

[Jagk23]

Yes everything was working fine until the automatic update was installed on Friday at 2AM. I know a lot of people recommend SBC but the point is it was working prior to this and it isn't isolated to the phones/extensions using Direct SIP, all of the extensions drop at the same time.

 

[Alphabetic]

How are you phones LAN when your 3CX is hosted? Have you got a VPN?

 

[Jagk23]

Yes there is a VPN between AWS and our head office. I've checked the VPN and we're not getting any drops and we also use the VPN for other services which are working fine.
Also checked our SIP trunks with our supplier and these are coming back fine.

 

[Alphabetic]

Have you tried a full reboot of the OS since upgrade?

 

[Jagk23]

Yes full reboot as well as doing a full check for system updates on the OS has been completed.

 

[Alphabetic]

And firewall passes?

 

[Jagk23]

No...my firewall test does now seem to fail - previously this was not the case. But I now get failures on 5060, 5090 and all ports 9000+.
We just have any in/out rules configured for our site IP's and SIP providers in AWS

 

[Alphabetic]

Then I would start by getting that sorted.

 

[Jagk23]

I'm a little uncertain what is required exactly...

 

[Alphabetic]

You will need to ensure you have port forwarding to those ports.
https://www.3cx.com/docs/manual/firewall-router-configuration/

 

[Jagk23]

It just states the ports need to be open from our VOIP providers, which they are.

 

[AWS2P]

you need to go in aws console and check all ports to be opened as 3cx require and not only for your sip provider otherwise it can't work in your case, this could be ok if all was in local (pbx +phones)

 

[Jagk23]

I found this information elswhere;

Can you guys confirm I am right in saying these are the ports to open up?

Inbound
TCP 5060
TCP 5001
TCP 443
TCP 5090
TCP 5061
UDP 5060
UDP 5090
UDP 9000-9500
TCP 22
TCP 5015

Outbound:
All traffic on all ports.