Problem importing old 1 drive zpool

[Husky]

I just tried to grab some data from an old 1 drive zfs pool that was not in use for about a year. It however doesn't work. Does anybody have an idea how to access the pool? I think the device was named different in the old machine.

Code:

[root@freenas] ~# zpool import
   pool: harddisk
     id: 11245568988528280564
  state: UNAVAIL
status: The pool was last accessed by another system.
action: The pool cannot be imported due to damaged devices or data.
   see: http://illumos.org/msg/ZFS-8000-EY
config:

        harddisk     UNAVAIL  missing device
          ada0p2     ONLINE

Code:

[root@freenas] ~# zpool import harddisk
cannot import 'harddisk': pool may be in use from other system, it was last accessed by freenas.local (hostid: 0x47307aed) on Fri Mar 15 03:01:18 2013
use '-f' to import anyway

Code:

[root@freenas] ~# zpool import -f harddisk
cannot import 'harddisk': one or more devices is currently unavailable

I thought it might be a missing log device but providing "-m" doesn't help:

Code:

[root@freenas] ~# zpool import -f -m harddisk
cannot import 'harddisk': one or more devices is currently unavailable

Additional info that might help to track down the issue

Code:

[root@freenas] ~# zdb -l /dev/gptid/7df89392-fc09-11e0-a869-003005c4d551
--------------------------------------------
LABEL 0
--------------------------------------------
    version: 15
    name: 'harddisk'
    state: 0
    txg: 201310
    pool_guid: 11245568988528280564
    hostid: 1194359533
    hostname: 'freenas.local'
    top_guid: 117745139941842
    guid: 117745139941842
    vdev_tree:
        type: 'disk'
        id: 1
        guid: 117745139941842
        path: '/dev/da2p2'
        whole_disk: 0
        metaslab_array: 23
        metaslab_shift: 33
        ashift: 9
        asize: 1498149552128
        is_log: 0
        DTL: 427
--------------------------------------------
LABEL 1
--------------------------------------------
    version: 15
    name: 'harddisk'
    state: 0
    txg: 201310
    pool_guid: 11245568988528280564
    hostid: 1194359533
    hostname: 'freenas.local'
    top_guid: 117745139941842
    guid: 117745139941842
    vdev_tree:
        type: 'disk'
        id: 1
        guid: 117745139941842
        path: '/dev/da2p2'
        whole_disk: 0
        metaslab_array: 23
        metaslab_shift: 33
        ashift: 9
        asize: 1498149552128
        is_log: 0
        DTL: 427
--------------------------------------------
LABEL 2
--------------------------------------------
    version: 15
    name: 'harddisk'
    state: 0
    txg: 201310
    pool_guid: 11245568988528280564
    hostid: 1194359533
    hostname: 'freenas.local'
    top_guid: 117745139941842
    guid: 117745139941842
    vdev_tree:
        type: 'disk'
        id: 1
        guid: 117745139941842
        path: '/dev/da2p2'
        whole_disk: 0
        metaslab_array: 23
        metaslab_shift: 33
        ashift: 9
        asize: 1498149552128
        is_log: 0
        DTL: 427
--------------------------------------------
LABEL 3
--------------------------------------------
    version: 15
    name: 'harddisk'
    state: 0
    txg: 201310
    pool_guid: 11245568988528280564
    hostid: 1194359533
    hostname: 'freenas.local'
    top_guid: 117745139941842
    guid: 117745139941842
    vdev_tree:
        type: 'disk'
        id: 1
        guid: 117745139941842
        path: '/dev/da2p2'
        whole_disk: 0
        metaslab_array: 23
        metaslab_shift: 33
        ashift: 9
        asize: 1498149552128
        is_log: 0
        DTL: 427

 

[cyberjock]

What OS was the drive created under? Was it FreeNAS 8+?

 

[Husky]

I am not 100% sure but i think so (ZFS V15 as detected by the last commands also indicates that).

I already tried to import with 8.3 but it didn't work either.

Edit: Oh I just realised that 8.3 already had V28 - will try again with 8.2

 

[cyberjock]

How about the output of "zdb -eC <poolname>" .

The output will need to be in pastebin, the forums will garble the output.

 

[Husky]

it says "device not configured" when i run this command

Btw. zpool import on freenas 8.2 says "insufficient replicas" for the pool while the single disk is still shown "online".

 

[cyberjock]

Try the zdb command I listed above with 9.2.1.8. There really is no reason to use the old FreeNAS versions at this point. Anything the old versions can do the new versions should be able to do.

 

[Husky]

Roger that - reinstalling the new version - will report back in a few minutes

 

[Husky]

Code:

[root@freenas] ~# zdb -eC harddisk
zdb: can't open 'harddisk': No such file or directory

The following commands shows sth, tho

Code:

[root@freenas] ~# zdb -e harddisk

Configuration for import:
        version: 15
        pool_guid: 11245568988528280564
        name: 'harddisk'
        state: 0
        hostid: 1194359533
        hostname: 'freenas.local'
        vdev_tree:
            type: 'root'
            id: 0
            guid: 11245568988528280564
            children[0]:
                type: 'missing'
                id: 0
                guid: 0
            children[1]:
                type: 'disk'
                id: 1
                guid: 117745139941842
                whole_disk: 0
                metaslab_array: 23
                metaslab_shift: 33
                ashift: 9
                asize: 1498149552128
                is_log: 0
                DTL: 427
                path: '/dev/ada0p2'
zdb: can't open 'harddisk': No such file or directory

 

[cyberjock]

Yeah, sounds like one of two things has happened:

1. The pool got corrupted and is giving output for commands that are incorrect for conditions.
2. The pool was not a single disk pool.

I'm leaning *heavily* towards #2. In particular, because it says a device is missing. I'm going to guess that you had more disks and you don't remember having more than 1 disk.

In any case, without all of the children to make the pool whole you won't be able to get access to the pool again. It was clearly made on FreeNAS and was v15.

 

[Husky]

I am 100% sure that it was a single disk pool. I think it might be the slog device that is missing. I am running zpool import -m -fFX harddisk atm. and at least it's doing something - dunno if it will be successfull thi

 

[cyberjock]

If you had an slog, you are screwed. Versions of ZFS that are below v19 (which yours is) won't mount without the slog.

 

[Husky]

Is there a way to access the data without mounting the zpool? Like a forensics script or sth. The data is very important (it was a vmdk file).

 

[cyberjock]

Is there a way to access the data without mounting the zpool? Like a forensics script or sth. The data is very important (it was a vmdk file).

Nope. The way to access the data is to mount the pool. No recovery tools for ZFS exist, which is mentioned on slide 11 of my noobie presentation. This is why using ZFS with sufficient redundancy and performing scrubs at regular intervals is crucial. If your zpool gets corrupted you'll lose your data forever.

If recovery is worth a 5-digit cost, give a company like Ontrack a call. But I'm serious when I say 5-digit. They will change you the cost of a mid-sized sedan because they know they have you by the balls. The one person I know of that actually did recovery with a company like ontrack paid $17k for 450GB of data. ;)