OpnSense keeps crashing(Version 23.1.11 f1305748e)

[HunterB007]

Hello, I just setup my OPNSense router yesterday, so a fresh install and only a theme plugin. For some reason, the router keeps crashing. Every time I check the crash reporter(its huge btw), I see something related to the attached messages on there. I am new to networking and this whole system, so any advice or anything else you guys might need to help me, feel free to ask and I will try my best to help

[HunterB007]

Code Select Expand

trap number = 12
panic: page fault
cpuid = 0
time = 1688812201
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe00c5cd1b20
vpanic() at vpanic+0x17f/frame 0xfffffe00c5cd1b70
panic() at panic+0x43/frame 0xfffffe00c5cd1bd0
trap_fatal() at trap_fatal+0x385/frame 0xfffffe00c5cd1c30
trap_pfault() at trap_pfault+0x4f/frame 0xfffffe00c5cd1c90
calltrap() at calltrap+0x8/frame 0xfffffe00c5cd1c90
--- trap 0xc, rip = 0xffffffff82f1f043, rsp = 0xfffffe00c5cd1d60, rbp = 0xfffffe00c5cd1da0 ---
ng_netflow_expire() at ng_netflow_expire+0x273/frame 0xfffffe00c5cd1da0
softclock_call_cc() at softclock_call_cc+0x13d/frame 0xfffffe00c5cd1e40
softclock() at softclock+0x79/frame 0xfffffe00c5cd1e60
ithread_loop() at ithread_loop+0x25a/frame 0xfffffe00c5cd1ef0
fork_exit() at fork_exit+0x7e/frame 0xfffffe00c5cd1f30
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00c5cd1f30
--- trap 0x80d08f20, rip = 0xffffffff80c30e5f, rsp = 0, rbp = 0x2b56000 ---
mi_startup() at mi_startup+0xdf/frame 0x2b56000
KDB: enter: panic

Heres the part of the code,so you dont have to download that file

[CJ]

It sounds like you have a hardware issue.  Try doing some burn in testing with things like memtest, etc.

[HunterB007]

Downloading memtest86+

just for specs, im running and N200 with a 1x 16gb ddr5 4800mhz with a kingston 500gb ssd.

To note for the install of opnsense: When I was installing, I was trying to use the zfs version, but it kept failing, so i tried a few times and each time it would progress farther and then it finally worked. At the time, I just assumed it fixed whatever the issue was, but could that be the root of my problem?

[HunterB007]

Ran memtest86+ for 41:09 minutes and it detected 19 errors and 1 pass. I assume that means the memory's bad. memtest86+ recommends swapping one of the modules, but this is a one dimm system. I just got the ram, so should I just return in and replace it with a new one to be safe?

[Stormscape]

Yeah, best to return/exchange the RAM as faulty and get a new stick.

[wizdude]

There is a long thread on the servethehome forums where user Becks0815 has been trying to get OPNsense running natively on an N100 without success. Proxmox runs fine on the device so they have currently deployed it as a VM. After much experimentation they went back to try and run it natively again, without success.

It appears that Linux runs fine on the N100 and similar devices, but FreeBSD isn't yet stable on this hardware.

I'm not excluding that you may have a hardware fault, but I don't think N100 or N200 is a suitable platform at this stage.

I'm quite disappointed about this myself as I wanted to get a new firewall appliance for home and also wanted to run IPS. The N and J series boxes aren't grunty enough (from my understanding) and I'm looking for an option better than repurposing an i5 desktop or similar.

Cheers, Wiz!!

[iMx]

Just throwing it out there, I've also had issues with recent CPUs that use Intel Speed Shift - seems the N200 also does - seems to be common on FreeBSD 13, there were reports on the FreeBSD lists that it was resolved in FreeBSD 14.

Adding the below to /boot/loader.conf.local and then rebooting

Code Select Expand

hint.hwpstate_intel.0.disabled=1

.. causes it to then use SpeedStep instead, which can then be used with PowerD if you wish.  This fixed my random crashing issues.

[CJ]

If memtest is throwing errors it's a hardware issue, not FreeBSD.

Ran memtest86+ for 41:09 minutes and it detected 19 errors and 1 pass. I assume that means the memory's bad. memtest86+ recommends swapping one of the modules, but this is a one dimm system. I just got the ram, so should I just return in and replace it with a new one to be safe?

I'm assuming the ram came in the unit you purchased, so yes, just return the whole unit.  If you bought the ram yourself, then just return it.

[HunterB007]

Thanks for the thoughts and recommendations everyone. I am returning the RAM, as its covered by amazon(thanks Bezos). In terms of the N200, that's disappointing to hear the stability problems. I will try that command that was recommended, and hope for the best. Once again, thanks everyone, and I will reply to this thread once I get my new RAM, with hopefully good news!

[iMx]

How did you get on? ;)

[franco]

If it's of any use the crash posted here is about Insight reporting / NetFlow generation. Turning it off would prevent this if it reoccurs. In that case it's also unlikely bad memory.


Cheers,
Franco

[HunterB007]

Was a RAM issue. Been running for a week without fail :). Thanks everyone!