Unable to import zpool after crash and subsequent upgrade

[mboalt]

About 5 days ago my FreeNAS crashed during the night with a Fatal Trap 18, and after an upgrade to 9.2.1.5 a couple of Fatal Trap 12s, and from that point just refused to boot properly.

I assumed, based on googling while semi-panicking, that it was a hardware problem, and replaced the motherboard, CPU and RAM entirely. I am now running:

ASUS E2KM1I-DELUXE (mITX with integrated AMD APU E2-2000)
Corsair Vengeance 16GB DDR3 PC-12800

(This was what my local vendor had in stock with 5+1 SATA ports. Let's leave it at that.)

Now, the system boots just fine, but I'm unable to auto-import my zpool. When I try to, I get tons of errors and the system halts (I can post a log if that helps). Switching USB boot sticks has no effect.

There are some similarities to this guy:

http://forums.freenas.org/index.php?threads/zfs-import-crashes-freenas.18571/

I have two major differences, though: I've run 3 passes of Memtest with no errors (although I never checked the memory of my initial setup). Secondly, I can import the the zpool from the CLI with read-only on, and I'm in the process of backing up the system before doing anything else to it.

Status, import and gpart as follows:

Code:

[root@freenas] /# zpool status
no pools available

[root@freenas] /# zpool import
   pool: Z2-Volume
     id: 17476285307126055496
  state: ONLINE
 status: The pool is formatted using a legacy on-disk version.
 action: The pool can be imported using its name or numeric identifier, though
        some features will not be available without an explicit 'zpool upgrade'.
 config:

        Z2-Volume   ONLINE
          raidz2-0  ONLINE
            ada1p2  ONLINE
            ada0p2  ONLINE
            ada2p2  ONLINE
            ada3p2  ONLINE
            ada4p2  ONLINE
            ada5p2  ONLINE

[root@freenas] /# gpart status
  Name  Status  Components
ada0p1      OK  ada0
ada0p2      OK  ada0
ada1p1      OK  ada1
ada1p2      OK  ada1
ada2p1      OK  ada2
ada2p2      OK  ada2
ada3p1      OK  ada3
ada3p2      OK  ada3
ada4p1      OK  ada4
ada4p2      OK  ada4
ada5p1      OK  ada5
ada5p2      OK  ada5
 da0s1      OK  da0
 da0s2      OK  da0
 da0s3      OK  da0
 da0s4      OK  da0
da0s1a      OK  da0s1

Normally, I would've cut my losses here, but since I can still access all data and the zpool appears unharmed, it seems to me that it should be possible to simply restore the zpool. Any suggestions on what to try once I finish the backup? Any more data I need to supply?

 

[cyberjock]

Your old hardware, did it have non-ECC RAM? What were the specs?

This sounds like a case of hardware failure causing corruption of the zpool. :/

 

[mboalt]

It did not. It was a simple Atom D510 system with 8GB RAM if I recall correctly. I didn't read your guide until a few days ago, when this happened.

Hardware failure sounds likely, but I don't know enough about ZFS to understand why I can still import the pool in read-only and access everything on it (as far as I can tell), but not through the GUI. I haven't tried any other CLI-importing until backup is done, due to your warnings here.

 

[cyberjock]

read-only ignores some assumptions because the pool is read-only. My guess is your pool is damaged and when you don't mark it read-only its not able to be mounted due to the damage. You wouldn't be the first person to be forced to use read-only, and you certainly won't be the last.

You also may or may not have other problems related to your choice of motherboard, etc. But as you said "lets leave it at that".

 

[mboalt]

So your advice would be to simply delete the pool and start over once backup is done?

Right. This was a quick fix just to see if I could access the data. I'll make sure not to make the same mistake.

 

[cyberjock]

So your advice would be to simply delete the pool and start over once backup is done?

Yes. Backup your pool, destroy and recreate it again. More than likely the hardware failure corrupted the pool and the only way to be sure you don't get bit later is to recreate the pool.