Crash/Fatal trap 12 during rm -rf operation

Status
Not open for further replies.

jaydub

Cadet
Joined
Oct 3, 2014
Messages
3
Hi all, I did do a search prior to posting, and I found a thread that seems to be a similar situation where the user experienced a crash during a delete operation: https://forums.freenas.org/index.ph...-kernel-mode-error-while-mounting-pool.16926/

It's old however and not sure if still relevant. My crash occurs at a different location, as well.

I ran rm -rf on a fairly large directory tree and a few minutes later, lost all network connectivity to the box. I plugged a monitor into it and saw the crash message on screen. Unfortunately I didn't have a keyboard plugged in at the time, and plugging in one at that point got no results, so I had no choice but to reboot, at which point the error returned when the system attempted to mount.

edited.jpg

Sorry about taking a picture of a monitor. :p I do see reference to serial console in the manual, but unless it's turned on by default, I don't think I'll be able to use it.

Taking some time to transcribe some of the output:

Fatal trap 12: page fault while in kernel mode
fault code: supervisor read, page not present
current process: 321 (txg_thread_enter)
Stopped at: range_tree_vacate+0x24

I also snapped the bt output, if that's helpful here:

bt.jpg

It seems to me from searching that simply doing a fresh install (in my case, setting up a new USB stick) and then importing the pool should do it. If that's the case I'll gladly do so.

My main goal is to rescue the data. Ironically, I was deleting some backups from the NAS, but I still had some that hadn't been restored yet.
 
Last edited:

jaydub

Cadet
Joined
Oct 3, 2014
Messages
3
I imaged a new USB stick and whaddya know, it worked. Imported the volume without a problem. I don't understand how that works, if it was a problem with the volume that caused the failure in the first place, I wouldn't expect it to import. I guess ZFS is just weird like that.

I've also found in my searching that this seems to be an extremely common thing with 9.2.x, ZFS, and anything shy of 8GB RAM. I don't seem to have lost any data but nevertheless, I'm dumping everything off my FreeNAS. Not sure if I'll rebuild as UFS or go with just plain linux and use lvm to handle all the drives.

I thought that I would just see slower performance if I had less RAM. I was okay with that, since the NAS would see light use (streaming two files max at a time, very occasional large transfers). I didn't think it was actually risky in the data loss sense.

Suggestion for the manual: big, red letters: 8GB or risk your data, no exceptions! :eek:
 

cyberjock

Inactive Account
Joined
Mar 25, 2012
Messages
19,526
UFS support is gone from the 9.3 code, so you should either plan to stay with ZFS (and upgrade to 8GB of RAM) or give up on FreeNAS completely.
 

jaydub

Cadet
Joined
Oct 3, 2014
Messages
3
Hmm, thanks for the warning. I think I will abandon the FreeNAS idea. It was a mistake to use it with ZFS on hardware that I knew was inadequate. I just didn't realize how big a mistake it was. I got lucky.
 
Status
Not open for further replies.
Top