Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 21 Jul 2008 23:42:14 +0200
From:      John Morgan Salomon <john@zog.net>
To:        freebsd-questions@freebsd.org
Subject:   Re: Recover Lost Superblocks?
Message-ID:  <4B57CEBA-2F62-4979-B4CE-D3C9727A0E85@zog.net>
In-Reply-To: <1CD17FFA-7CE2-4C6A-A578-BA9542E4A9AE@zog.net>
References:  <2385.82.120.108.188.1216634229.squirrel@www.88.net> <20080721125225.956c3aa4.freebsd@edvax.de> <1CD17FFA-7CE2-4C6A-A578-BA9542E4A9AE@zog.net>

next in thread | previous in thread | raw e-mail | index | archive | help
OK, I have a followup question to this.

After some mucking around, I've managed to lose my partition again  
(although the data is still there, I installed testdisk and let  
photorec run; it looks like it's finding pretty much everything.)

Running newfs -N on /dev/aacd0 finds a ton of backup superblocks.

My filesystems were originally /dev/aacd0s1a, aacd0s1b and aacd0s1e.   
When I originally recreated the FreeBSD partition with the same  
geometry under my new rescue HDD, it added a device entry "aacd0s1c"  
but not any of the others.

Running fsck_ufs -b <any of the listed backup superblocks> doesn't  
seem to do much of anything.

I'd be grateful if someone could help me with the following questions:

1) when I run the above command, is it supposed to replace a  
filesystem's superblock with the backup superblock?
2) is there a way to look at the contents of the backup superblocks  
that newfs -N found?
3) is there a way to re-create aacd0s1a, aacd0s1b and aacd0s1e?  The  
rescue OS seems to only want to bother with aacd0s1c, which was not  
used by any of the partitions previously.

Thanks for any help,

-John

On Jul 21, 2008, at 1:04 PM, John Morgan Salomon wrote:

> Wow, a sympathetic ear, was expecting far more scorn than that :-)
>
> I am currently running TestDisk, which at least _appears_ to be  
> finding something filesystem-like (at least it's listed a few  
> "empty" "somethings" that look somehow reasonable, size-wise.)   
> Cross your fingers.  Gpart and TestDisk are entirely passive, i.e.  
> don't touch data on the disks.
>
> My plan, if this works out, is to buy a secondary backup consisting  
> of a RAID 1+0 NAS.  I don't have anything big enough to back up  
> everything to.
>
> I tried pretty much everything with fsck_ufs.  Like I said, though,  
> I am able to mount the entire partition from the bootable IDE  
> drive.  I see /, /etc/, /dev/ and all that, but since the "rescue"  
> OS can't see any additional superblocks, it has no devices for the  
> other filesystems.  I am not sufficiently well versed in UFS to  
> understand how an entire partition can be mounted as a filesystem if  
> that partition originally had multiple filesystems on it.  I'm a bit  
> wary of playing more with fsck until all else has failed.  :-)
>
> What also weirds me out is that FreeBSD constantly bitches about the  
> partition being larger than the physical disk (which it decidedly  
> isn't.)  I've tried setting geometry in fdisk any which way  
> (including using the RAID controller's provided values), and as I  
> said, the thing mounts the root partition of the array just fine.   
> I'm considering an exorcist.
>
> Best,
>
> -John
>
>
> On Jul 21, 2008, at 12:52 PM, Polytropon wrote:
>
>> Hi!
>>
>> On Mon, 21 Jul 2008 11:57:09 +0200 (CEST), "John Morgan Salomon" <john@zog.net 
>> > wrote:
>>> Before you ask, this was the backup server.  My primary box had  
>>> decided to
>>> die shortly before.  I had no backup backup server.  Murphy strikes.
>>
>> I completely do understand you, I'm suffering from a similar problem
>> at the moment, but much worse than yours...
>>
>> Buy tape drives! Buy tape drives! Buy tape drives! :-)
>>
>>
>>> Can someone recommend a way to manually scan the entire partition  
>>> (either
>>> aacd0, aacd0s1 or aacd0s1c) for formerly present filesystems?  I  
>>> am 99%
>>> sure that all the data is still present, and if I reinstall the
>>> superblocks I'll be able to boot the array, mount the filesystems  
>>> and get
>>> the data off before I continue.  I don't know whether I've missed  
>>> any
>>> gpart options (I have the impression it only scans for lost  
>>> partitions,
>>> not ufs filesystem signatures.)
>>
>> As far as I know - NB that I'm just starting to learn more about UFS,
>> shame on me that I'll do this just as every piece of data is gone -
>> there are more than one superblock present. According to "man  
>> fsck_ufs",
>> this could be a starting point:
>>
>>    -b      Use the block specified immediately after the flag as  
>> the super
>>            block for the file system.  An alternate super block is  
>> usually
>>            located at block 32 for UFS1, and block 160 for UFS2.
>>
>> This applies if just the first superblock is gone.
>>
>> Before you start experimenting, maybe it's a good idea to dd the
>> data out of the disks and run fsck on the images? I'm not sure...
>>
>>
>>> Any help, tips or pointers would be tremendously appreciated.
>>
>> Hope you're lucky.
>>
>>
>>
>>
>>
>>
>> -- 
>> Polytropon
>>> From Magdeburg, Germany
>> Happy FreeBSD user since 4.0
>> _______________________________________________
>> freebsd-questions@freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
>> To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org 
>> "
>
> _______________________________________________
> freebsd-questions@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org 
> "




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B57CEBA-2F62-4979-B4CE-D3C9727A0E85>