On 05/06/16 17:32, Pawel Jakub Dawidek wrote:
Unfortunately tools like photorec won't work (I assume the tool scans
the disk and looks for some JPEG headers, etc.) as this was RAIDZ1 pool,
so none of the disks contains recognizable data.

Are you sure? Isn't only the parity information unusable? So 2/3 of the
information should be there in plain. It would be good to revive
the broken disk, though.


Two questions come to my mind:
1. Is it possible to recover the data now?
2. Is it possible to recover the data in theory?

I'm afraid the answer to the first question is 'no', because we are not
aware of any existing tools that could help you with the process and to
do it now you would need to find ZFS expert willing to spend some
serious amount of time, which I'm sure will cost a lot.

But if I understand your description correctly, the data is still on the
disks, so in theory it should be possible to recover it. Maybe not now,
but in 5 years? 10 years? Family memories is probably not something you
need immediately and after some time someone may develop a tool to help
recover data from the pools like yours. Even if noone will, maybe you
will become rich enough (or maybe you are already, no idea) to sponsor
development of such a tool.

Writing such a tool would be fun, but it would probably take weeks to
months to write it. The question is if there are enough broken pools to
make it worth investing the time.

-Arne


My advise? Hold on to your data and don't lose hope!

On Sun, Jun 05, 2016 at 05:02:39PM +0200, Arne Jansen wrote:
Hello Evgeny,

without trying to understand what exactly you've done to the pool:
If you think that in theory that data should still be there you can
try to use tools like photorec. There's a chance it can recover most
of your data, but probably without filenames. Tons of work awaiting
you there...
On the other hand, to truly recover the data it might need an expert
dedicating hours and hours to the task. I don't know if anyone has
written a general-purpose recovery tool for messed-up zfs pools.
It's a good thing you've created bit-copies of the 2 drives. What about
the third? Is it truly gone? Maybe it is possible to scrape some data
off it, too. There are companies specialized in this.

I'd first give the first variant a try...

-Arne

On 03/06/16 19:56, esamorokov wrote:
Hello All,

    My name is Evgeny and I have 3 x 3TB in RAIDZ1, where one drive is
gone and I accidentally
    screwed the other two. The data should be fine, just need to revert
    uberblock in point of time, where i started doing changes.

    I AM KINDLY ASKING FOR HELP! The pool had all of the family memories
for many years Thanks in advance!

    I am not a FreeBSD guru and have been using ZFS for a couple of
years, but I know Linux and do some programming/scripting.
    Since I got that incident I started learning the depth of the ZFS,
but I definitely need help on it at this point.
    Please don't ask me why I did not have backups, I was building
backup server in my garage, when it happened

History:
    I was using WEB GUI of FreeNas and it reported a failed drive
    I shutdown the computer and replaced the drive, but I did not
noticed that I accidentally disconnected power of another drive
    I powered on the server and expanded the pool where there only one
drive of the pool was active
    Then I began to really learn ZFS and messing up with bits
    At some point I created a backup bit-to-bit images of the two drives
from the pool (using R-Studio)


Specs(ORIGINAL):
    OS: FreeBSD 9.2-RELEASE (FREENAS.amd64) #0 r+2315ea3: Fri Dec 20
12:48:50 PST 2013
    RAID:   [root@juicy] ~# camcontrol devlist
    <ST3000DM001-1CH166 CC29>          at scbus1 target 0 lun 0
(pass1,ada1)
    <ST3000DM001-1CH166 CC29>          at scbus2 target 0 lun 0
(ada2,pass2)
    <ST3000DM001-9YN166 CC4H>          at scbus3 target 0 lun 0
(pass3,ada3)
    [root@juicy] ~# zdb
zh_vol:
    version: 5000
    name: 'zh_vol'
    state: 0
    txg: 14106447
    pool_guid: 2918670121059000644
    hostid: 1802987710
    hostname: ''
    vdev_children: 1
    vdev_tree:
        type: 'root'
        id: 0
        guid: 2918670121059000644
        create_txg: 4
        children[0]:
            type: 'raidz'
            id: 0
            guid: 14123440993587991088
            nparity: 1
            metaslab_array: 34
            metaslab_shift: 36
            ashift: 12
            asize: 8995321675776
            is_log: 0
            create_txg: 4
            children[0]:
                type: 'disk'
                id: 0
                guid: 17624020450804741401
                path: '/dev/gptid/6e5cea27-7f52-11e3-9cd8-d43d7ed5b587'
                whole_disk: 1
                DTL: 137
                create_txg: 4
            children[1]:
                type: 'disk'
                id: 1
                guid: 3253299067537287428
                path: '/dev/gptid/2b70d9c0-8e40-11e3-aa1c-d43d7ed5b587'
                whole_disk: 1
                DTL: 133
                create_txg: 4
            children[2]:
                type: 'disk'
                id: 2
                guid: 17999524418015963258
                path: '/dev/gptid/1e898758-9488-11e3-a86e-d43d7ed5b587'
                whole_disk: 1
                DTL: 134
                create_txg: 4

State summary of the drives/pool:

    *State #1*      *State #2*      *State #3*      *State #4*      *State #5*
*ada1 (ada1p2)*     OK      OK      DISCONNECTED    CONNECTED (OLD POOL)    
CLONED to a new
drive, Prev. state
*ada2 (ada2p2)*     OK      OK      OK      OK (new Pool)   CLONED to a new 
drive, Prev. state
*ada3 (ada3p2)*     OK      FAILED  REPLACED        CONNECTED FAILED DRIVE (OLD 
POOL)
CLONED to ada4, Prev. state
*ZH_VOL (pool)*     OK      DEGRADED        DESTROYED       RECREATED       
Previous State



Specs (CURRENT):

    root@juicy] ~# camcontrol devlist
    <Patriot Pyro SE 332ABBF0>         at scbus0 target 0 lun 0 (ada0,pass0)
    <ST3000DM001-1ER166 CC25>          at scbus1 target 0 lun 0 (ada1,pass1)
    <ST3000DM001-1ER166 CC25>          at scbus2 target 0 lun 0 (ada2,pass2)
    <ST3000DM001-9YN166 CC4H>          at scbus3 target 0 lun 0 (ada3,pass3)
    <ST3000DM001-1ER166 CC26>          at scbus5 target 0 lun 0 (ada4,pass4)
    <Marvell 91xx Config 1.01>         at scbus11 target 0 lun 0 (pass5)


    [root@juicy] ~# zdb
    zh_vol:
        version: 5000
        name: 'zh_vol'
        state: 0
        txg: 1491
        pool_guid: 10149654347507244742
        hostid: 1802987710
        hostname: 'juicy.zhelana.local'
        vdev_children: 2
        vdev_tree:
            type: 'root'
            id: 0
            guid: 10149654347507244742
            create_txg: 4
            children[0]:
                type: 'disk'
                id: 0
                guid: 5892508334691495384
                path: '/dev/ada0s2'
                whole_disk: 1
                metaslab_array: 33
                metaslab_shift: 23
                ashift: 12
                asize: 983564288
                is_log: 0
                create_txg: 4
            children[1]:
                type: 'disk'
                id: 1
                guid: 296669430778697937
                path: '/dev/ada2p2'
                whole_disk: 1
                metaslab_array: 37
                metaslab_shift: 34
                ashift: 12
                asize: 2997366816768
                is_log: 0
                create_txg: 1489

Thnks in advance!
Evgeny.

*openzfs-developer* | Archives
<https://www.listbox.com/member/archive/274414/=now>
<https://www.listbox.com/member/archive/rss/274414/28015265-27880bfd> |
Modify
<https://www.listbox.com/member/?&;>
Your Subscription   [Powered by Listbox] <http://www.listbox.com>







-------------------------------------------
openzfs-developer
Archives: https://www.listbox.com/member/archive/274414/=now
RSS Feed: https://www.listbox.com/member/archive/rss/274414/28015062-cce53afa
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=28015062&id_secret=28015062-f966d51c
Powered by Listbox: http://www.listbox.com

Reply via email to