pfSense 2.6.0 stable crashing with panic "Unknown caching mode 23"

ThatPickleDude

Running through xcp-ng on an i9-10900. Fresh install as of last night. Crash has happened twice in the past hour without touching anything. Not seeing anything about this on the forums and Google isn't showing any relevant results for me. Any ideas?

textdump.tar (1).0
info (1).0

stephenw10

Is it the same crash each time?

The message buffer shows the previous panic was different:

<118>savecore 106 - - reboot after panic: general protection fault
<118>savecore 106 - - writing core to /var/crash/textdump.tar.3

The backtrace on that one is pretty generic, it doesn't point to anything specific.

Steve

ThatPickleDude

@stephenw10 Letting it run overnight, I had three unique crashes. Another like the original I posted, one as you pointed, and a page fault.

Filename: /var/crash/info.2
Dump header from device: /dev/label/swap0
  Architecture: amd64
  Architecture Version: 4
  Dump Length: 74752
  Blocksize: 512
  Compression: none
  Dumptime: Thu Dec  1 07:01:46 2022
  Hostname: pfSense
  Magic: FreeBSD Text Dump
  Version String: FreeBSD 12.3-STABLE RELENG_2_6_0-n226742-1285d6d205f pfSense
  Panic String: page fault
  Dump Parity: 2066528263
  Bounds: 2
  Dump Status: good

SteveITS

@thatpickledude In general, with PC hardware I find random/changing crash errors are usually bad memory. You might run a memory test just to rule that out.

ThatPickleDude

@steveits Thanks! Ram tests look good, unfortunately

Strangely, I can't replicate the issue again. Wish I knew what happened, but this appears to be intermittent.

stephenw10

Mmm, hard to imagine it could be bad ram in VM. But it sure behaves like it I agree.

SteveITS

@stephenw10 said in pfSense 2.6.0 stable crashing with panic "Unknown caching mode 23":

Mmm, hard to imagine it could be bad ram in VM.

Well now that you say that…. Missed that tidbit. Host, maybe. Worth a shot to rule it out.

Recently had a client’s year old PC start throwing STOP errors every few days. Windows memory test passed. Finally a tech took memtest or something there and it flagged the RAM. No issues after replacing it.

ThatPickleDude

Okay... It's still happening. Getting a new unique crash though, so that may be progress. Panic string is: vm_fault: fault on nofault entry, addr: 0xffffffff83d92000. Also getting more page fault crashes. RAM checks out, and xcp-ng is updated with the latest patches.

vm_fault crash:
info.0
textdump.tar.0

Page fault crash:
info.0
textdump.tar.0

stephenw10

Hmm, they're all different. This 'feels' like some sort of hypervisor feature incompatibility.

Like maybe flexible virtual memory or disk size which is known to cause problems with some other hypervisors.

Were you running pfSense in xcp before?

Steve

ThatPickleDude

@stephenw10 Different hardware (two R710 in an HA pair), but yes. No real issue before. Maybe I'll try a reinstall in BIOS mode instead of UEFI

ThatPickleDude

Updated BIOS (was one patch version behind) and am updating to 2.7 DEVEL hoping this is a diver issue causing some weird compatibility that can be fixed in FreeBSD 14. Fingers crossed

PabzRoz

@thatpickledude Did you ever figure this out? I'm having the exact same issue with the exact same messages...

stephenw10

What are you actually seeing? The backtraces above are different.

Steve