100% CPU Load then 3CX Crashes

[FMRS_IT]

Hello,

Problem:
I am experiencing a weird problem. I've a 3CX instance (v15, up-to-dated) installed at a Windows 2012 R2 VM. At around 8:30 AM every morning, the 3CX console will reach 100% CPU usage and eventually crash the 3CX service.

After restarting the VM and/or mode switching between static/dynamic IP, the external-to-internal call would drop at around 32 seconds. Internal-to-External calls would last about 2:30 seconds.

Sample Log:
05/29/2018 9:31:05 AM - Leg L:17.4[Extn:201] is terminated: Cause: BYE from local
05/29/2018 9:31:05 AM - Leg L:17.1[Line:10000<<[External_PhoneNumber] is terminated: Cause: BYE from local
05/29/2018 9:31:05 AM - [CM503008]: Call(C:17): Call is terminated
05/29/2018 9:31:05 AM - [CM503023]: Call(C:17): Call recording is stopped, audio file: E:\Recordings\201/[Unknown%3AOffice Hours]_[External_PhoneNumber]_20180529133033(17).wav
05/29/2018 9:31:05 AM - [CM503021]: Call(C:17): ACK is not received from sip:[External_PhoneNumber]@208.89.104.241

Attemped:
To temporary solve the problem, we restarted the VM and switch between static IP and dynamic IP, sometimes require multiple switching modes and eventually will settle down and 3CX will work without problem for the day.

Observation:
During around the 100% CPU usage & crash time, I notice the following error message:

Registration at RingOffice has failed. Destination (sip:[Office_PhoneNumber]@[ServiceProvide_SIP]) is not reachable, DNS error resolving FQDN, or service is not available.

The router we are using is Cisco RV042G and we have SIP ALG disabled.

Any help would be appreciated. Thank you!

 

[YiannisH_3CX]

Hello @FMRS_IT

Please make sure that the machine is fully up to date with windows updates and it is dedicated to the PBX. Also what resources are allocated to the VM and how many extensions do you have on the PBX?

After restarting the VM and/or mode switching between static/dynamic IP, the external-to-internal call would drop at around 32 seconds. Internal-to-External calls would last about 2:30 seconds.

This means that there is a routing issue somewhere as the ACK message is not received by the system and the call is terminated. Make sure that the firewall checker passes all tests to avoid any issues.

 

[FMRS_IT]

Update:

So we observed that on certain PCs with "3CX for Windows" installed, once the user login to the domain, the 3CX VM will then reach 100% CPU usage. The service that is using the CPU is "3CX PhoneSystem 01 SIP Server".


Once the 3CX for Windows is uninstalled, we observed user login no longer causing the high CPU usage.


From the attached screenshot we can see that once the user login, the user PC (192.168.1.101) issued multiple 550 Request: SUBSCRIBE to a multcast address 224.0.1.75 and eventually flood the VM and caused the 3CX service crashes.

Do you guys think it's a network problem, 3CX bug or something else?

 

[FMRS_IT]

Hello @FMRS_IT

Please make sure that the machine is fully up to date with windows updates and it is dedicated to the PBX. Also what resources are allocated to the VM and how many extensions do you have on the PBX?

This means that there is a routing issue somewhere as the ACK message is not received by the system and the call is terminated. Make sure that the firewall checker passes all tests to avoid any issues.

The VM is up-to-dated on Windows Update. It's also a dedicated VM for the 3CX. There are 10 extensions, usually only 3-5 are actively being used.

The VM has 4GB ram and still have more than 50% desk space free, so it shouldn't be a problem. It doesn't have a problem a month ago anyway ...

 

[Chris-MBN]

FMRS_IT,

Is the System Interrupts always that high? On my Hyper-V 3CX VMs it is not normally above 1%.

 

[YiannisH_3CX]

From the attached screenshot we can see that once the user login, the user PC (192.168.1.101) issued multiple 550 Request: SUBSCRIBE to a multcast address 224.0.1.75 and eventually flood the VM and caused the 3CX service crashes.

Once the 3CX client initiates and if there is no configuration on it it will send a subscribe to the PnP multicast address (224.0.1.75). It will send one SUBSCRIBE message from each NIC installed on the PC. The only way to make the client re-send the PnP request is to reload the client without configuration.
Can you replicate the issue on a PC that is not part of the domain? Are these physical machines that the clients are installed on? What is the version of the 3CX windows client that the PC's are running?