SOLVED Fatal Trap 12: Page fault while in kernel mode (CIFS Transfer)

[Revojames]

My System:

Mobo - ASRock FM2A85X-ITX
CPU - AMD A4 5300 2-Core Processor
RAM - GeIL 16GB Kit (2x8GB) DDR3 Evo Veloce C9 1333MHz
HDDs - 6x WD Red 3TB
Network card - Intel PCIe CT Gigabit Ethernet Network Card
Boot device - Imation Classic USB 2.0 Flash Drive (8GB)
OS Version - FreeNas 8.3.1 Release P2

Current Setup

The FreeNAS box is currently connected to a gigabit switch, along with 2 Windows PCs and a PoE device, which connects to the router in another room.

Background

FreeNAS is running live from the USB drive as mentioned above - I've created a RAID-Z2 volume with all 6 disks, and a CIFS share to mnt/volume1.
Both PCs have no issue accessing that share and authenticating as root to gain access (haven't set up non-standard user accounts as of yet).

Issue
When I start transferring data (starting a 1.3tb transfer, from an internal SATA HDD to the CIFS share), from either PC, after anywhere between ~2-10 minutes the transfer freezes, and the box itself is displaying "Fatal trap 12: page fault while in kernel mode" - much like the screenshot posted by "MotorSport" in this thread. Apologies I don't have my own screenshot right now - forgot to take one before leaving for work! It's a pretty similar message, just different values for virtual address etc, I'll upload one when possible.

I've tried enabling all of the CIFS Aux settings & the Tunables mentioned by Proto in this post, as well as enabling/disabling a few different combinations of those tunables + doubling the values of arc_max & write_limit_override in an attempt to take the extra RAM into account - No stability change yet.

I tried clearing the config and rebooting the system 2-3 times with Autotune enabled, no change.

I ran a short memtest (~1 hour) with no results - but plan to leave it on overnight tonight to see what comes up.
The hardware's all new so obviously I'm hoping there's nothing wrong with it... but it wouldn't be the first time I've had new system problems caused by bad hardware.
I had a go at running a SMART test, but all I could see how to do was schedule on to start every day \ every hour \ every day \ every month and just wait an hour for it to start - is this the best way of starting the test?

I've updated to the latest BIOS available for my motherboard, didn't seem to make any change.

I did notice late last night that the switch I've been using has had a few issues, causing slow transfer speeds. I'm not sure if that could also be causing this, but I plan to pick up a new Gigabit switch (Linksys SE2500) after work tonight - as well as trying the nightly build of 9.1.

Sorry if I've missed / forgotten anything - Does anyone have an idea of what could be causing the problem?

Cheers!

 

[ProtoSD]

WOW! Amazing and thorough first post! You didn't mention whether you're using the x64 or x86 installation. You should be using the x64 version. I'd double check that.

I wouldn't try a nightly build of 9.1 yet, its not anywhere near ready to be used as a troubleshooting method.

It seems like you've covered all the initial stuff to try. Running the wrong architecture is the only thing that comes to mind at the moment.

Welcome to the FreeNAS 8 Forums!

 

[Revojames]

Thanks!

Knew I'd forgotten to mention something - double checked in the hopes that I might've grabbed the wrong version, but it's definitely running x64 :(.

Thanks for the suggestion anyway, and the quick reply :)

 

[ProtoSD]

You could try running a Live PCBSD 9.1 or FreeBSD 9.1 CD/DVD and importing your pool and see if that helps. I'm leaning toward a driver or hardware problem at the moment.

 

[gpsguy]

WOW! Amazing and thorough first post!

+1 My guess, is that it's a mobo/cpu/ram issue. I know your hardware is new, but sometimes one gets a lemon.

Eleven months ago, I experienced similar problems - Fatal-trap-12-while-copying-data-from-server-to-workstation. My machine would pass diagnostic tests and yet I still had the kernel panics.

Due to time constraints, I ended up buying new hardware and installed the old drives (imported the volume) and was back in business. Haven't had any problems with the new system.

 

[Revojames]

You could try running a Live PCBSD 9.1 or FreeBSD 9.1 CD/DVD and importing your pool and see if that helps. I'm leaning toward a driver or hardware problem at the moment.

Thanks! I'll give that a go before taking the thing apart again and testing hardware - fingers crossed it's just a driver issue, I figure that'd just mean waiting until 9.1 beta to get the box working?

Eleven months ago, I experienced similar problems - Fatal-trap-12-while-copying-data-from-server-to-workstation. My machine would pass diagnostic tests and yet I still had the kernel panics.

Due to time constraints, I ended up buying new hardware and installed the old drives (imported the volume) and was back in business. Haven't had any problems with the new system.

Thanks for the input; glad to hear you got yours working with new hardware! Assuming the above doesn't alleviate the issue it'll definitely be time to pull the box apart and do some hardware elimination testing - seems like it's gonna be a fun weekend :D

 

[Revojames]

Hey guys,

Turns out it was the RAM - Took it back in and got it replaced this morning, have transferred 600gb to the CIFS share so far with no crashes :). Thanks for the advice!

 

[ProtoSD]

Thanks for the followup, glad to hear you figured it out! :)