pfSense keeps "crashing" at least one or two times a day

Wanna 0

Hi,

I have the problem, that pfSense keeps crashing at least one or two times a day. I have no clue where to search for the error. Or should I just start from ground zero becuase reinstalling it and importing the configuration hasn't helped. Here is the crash log which I can't do anything with. Maybe some of you can gibe me a hint?

crashlog.txt

Gertjan

@Wanna-0

Pages fault all over the place.

Try this : take out the network card, and use another device and try again.

Wanna 0

Hey @Gertjan,

thank you for the answer. I am using a Protectli FW4B - 4 Port Intel J3160 as a base where pfSense is running on.

What exactly do you mean with "pages fault". The admin GUI pages??

Gertjan

@Wanna-0 said in pfSense keeps "crashing" at least one or two times a day:

What exactly do you mean with "pages fault". The admin GUI pages??

You said that ;)
There are multiple crash reports.

Filename: /var/crash/info.0
Filename: /var/crash/info.1
etc.

They all have panic reasons like :
Panic String: bad pte va 15bc19d62000 pte 0
Panic String: page fault
Panic String: page fault
Panic String: page fault
Panic String: page fault
Etc.

which means to me : "memory management errors".

I would go to https://www.freebsd.org/, and get a bare bone CLI only FreeBSD ISO and install that one as a test - or any other OS your device supports.
You've installed pfSense yourself ?

Wanna 0

@Gertjan Hey, thanks again.

Yes I've installed it myself and used the official ISO from here: https://www.pfsense.org/download/ to do it. I used the AMD64 bit version which should be the correct one for the protectli device.
I've also upgraded to the pfSense Plus version, but it also kept crashing.

crashdumps.zip

stephenw10

All those panics are different. That generally implies not a software issue. I'd suspect bad ram, try running memtest through a few cycles.

Wanna 0

@stephenw10 said in pfSense keeps "crashing" at least one or two times a day:

All those panics are different. That generally implies not a software issue. I'd suspect bad ram, try running memtest through a few cycles.

Hey, A first "quick" ram test showed zero errors.

Gertjan

@Wanna-0
It's 'just' a quick test, and notably with SMP (multi processor mode) disabled, so just processor core.
Keep in mind : today's machines are not : a processor, and bunch of address lines, data lines, read write line, and a RAM chip. Don't forget about the entire MMU scheme that provides virtual/paged/protected memory.

It's more work, but : install another OS, like "FreebSD 14 stock" and see how for you come.

Wanna 0

@Gertjan okay, I will try to re run the test with multicore. Otherwise I am considering switching to a official netgate device and keep the protectli as a backup if the netgate device fails someday.