opensense crash 9000MTU during transfer dropoff

[ninja0n3]

Hello,

I just experienced a strange issue today.
My opnsense froze, I noticed a message `ketn.ipc.nmbjumbop` limit reached on the screen.

At the time this happened, I was running a download at around 3.2Gbps (~400Mb/s) through opnsense on one of my servers (both with 10Gbe, 9000MTU, iperf3 clocking at 9.89Gbps between opnsense and the server).
Due to a completely unrelated issue (different problem altogether associated with the ISP - due to some throttling), the connection dropped a couple of times and then opnsense froze (no direct input on console and I noticed the message I mentioned above).
The ISP throttling/connection dropoff lasts for about 6min before restoring the internet. So it goes around 5min on, 6 min off.

I tried to change the tunables to:
- kern.ipc.nmbclusters = 1048576
- kern.ipc.nmbjumbop = 524288
- kern.ipc.nmbjumbo9 = 524288

This didn't have any effect, whenever the ISP connection would drop after a couple of drops the system would freeze, this time with no message on screen.

The opnsense hardware is as follows:
- CPU: i7-7700K
- RAM: 16Gb

This is the first time I've experienced with issue. It started happening with the optimized MTU for the NICs in order to reach 10Gbps on the network.

Any help with this would be greatly appreciated.