Sudden reboot of system when using MS Teams

[niels@nooby.tech]

Hi all,

First time user of Opnsense. I've got a very simple setup: 1 WAN (Cable connection 300Mbs), 1 LAN on Intel Pro 1000, currently no Vlans. I'm running some tests as I want to use this box for a webfilter, OpenVPN access. Currently never more than 2 clients connected through a Unify AP and I have continious crashes on the system. It just keeps rebooting during Team calls. No straight away, but anywhere between 2-20 minutes in. A simultaneous speed test sometimes triggers it, but not always. I've tried everything I can think off.

The crashes are only during MS Teams videocalls. CPU/mem/disk usage are still very low during this point. No error appears just a sudden reboot like the power was cut. It doesn't seem to get too hot (as far as I can tell).
I tried switching off all hardware accelleration on the interfaces, I played with any setting I could find but nothing seems to help. Any advice on how to fix/diagnose the issue? Thanks in advance.

Current Config:
OPNsense 22.1.10-amd64
FreeBSD 13.0-STABLE
OpenSSL 1.1.1q 5 Jul 2022
On a fujitsu S920 (AMD GX-222GC SOC with Radeon R5E Graphics (2 cores, 2 threads), 4GB ram, 8GB SSD) with a PCI Epxress Intel PRO1000 dual ports.

[bartjsmit]

Start with a memory test to make sure you don't hit a duff row in one of your DIMMs
https://www.memtest86.com/

Bart...

[niels@nooby.tech]

Hi Bart,

Thanks for the reply. I've been looking into this further. I tried running the WAN over the build-in network interface (realtek) instead of the dual port Intel and that didn't make any difference. Tried a re-install as well, tried the legacy 21.7 opnsense,... still the same behaviour. I'm starting to wonder if your memory test might indeed reveal something. I'll try to run it as soon as I can. The only thing I don't get is why it would only happen during MS Teams calls?

Niels

[franco]

I would have thought maybe it's a pf-IPv6-refrag issue we are still chasing, but I think that wasn't present in 21.7...


Cheers,
Franco

[garichd]

I have exactly the same machine and just installed opnsense v22.7 on to it . I have noticed that it's running quite hot at 92C.
Can you please check the Temp of your machine - you may have to add Thermal Widget in the dashboard and select the correct thermal sensors in System --> Settings -- > Miscellaneous. I am just trying to figure out if high Temp is normal or if something is wrong with my machine.


 

Hi all,

First time user of Opnsense. I've got a very simple setup: 1 WAN (Cable connection 300Mbs), 1 LAN on Intel Pro 1000, currently no Vlans. I'm running some tests as I want to use this box for a webfilter, OpenVPN access. Currently never more than 2 clients connected through a Unify AP and I have continious crashes on the system. It just keeps rebooting during Team calls. No straight away, but anywhere between 2-20 minutes in. A simultaneous speed test sometimes triggers it, but not always. I've tried everything I can think off.

The crashes are only during MS Teams videocalls. CPU/mem/disk usage are still very low during this point. No error appears just a sudden reboot like the power was cut. It doesn't seem to get too hot (as far as I can tell).
I tried switching off all hardware accelleration on the interfaces, I played with any setting I could find but nothing seems to help. Any advice on how to fix/diagnose the issue? Thanks in advance.

Current Config:
OPNsense 22.1.10-amd64
FreeBSD 13.0-STABLE
OpenSSL 1.1.1q 5 Jul 2022
On a fujitsu S920 (AMD GX-222GC SOC with Radeon R5E Graphics (2 cores, 2 threads), 4GB ram, 8GB SSD) with a PCI Epxress Intel PRO1000 dual ports.

[jezza007]

It certainly sounds like high temp system protection cutoff kicking in, if kernel is cutting power to save frying the cpu

92c does sound high, that chips normal operating range is between 50-70c, 50c when idle passively cooled.

Teams video uses high network bandwidth, so there maybe expected additional cpu load when there is increased traffic throughput, hence varying time for symptom to manifest. Its just a question of how long before temp rises before cutoff kicks in, ive seen this lots of times for various reasons.

1.Make sure plenty of space and airflow around unit
2.Check the cpu and heatsink
3.Check latest bios
4. Perhaps upgrade to 22.7 as there seems to have been a few issues with 22.1 to 22.1.10 with network hardware offloading changes
5.Try disable any on chip gpu and run headless may assist with temp
6. Disable in bios anything that is not absolutly nessesary such as audio, unused usb ports etc, i have seen sone posts where chipset feature were causing kernel issues

Hope these suggestions help you figure it out :-)