Re: "Unformat" a ReiserFS partition : a testimony :-)

[Yury Umanets <umka@namesys.com>]

Manuel Krause wrote:

> On 03/25/2003 02:26 AM, Yury Umanets wrote:
>
>> Manuel Krause wrote:
>>
>>> On 03/25/2003 01:16 AM, Yury Umanets wrote:
>>>
>>>> Nicolas Vanderavero wrote:
>>>
> [...]
>
>>>>
>>>>> while reinstalling a Debian on a new hard drive, I made the 
>>>>> mistake to format in reiserfs my good old /home which was on 
>>>>> /dev/sda7 instead of formatting the new /dev/hda7. Gasp ! Eight 
>>>>> gigabytes lost ...
>>>>>
>>>>> Or ... maybe it was not totally lost :-) I immediately made a dump 
>>>>> of the partition with 'dd' and started grepping some known text on 
>>>>> it. It seemed that no data was really lost. I didn't know what to 
>>>>> do. So I read the man pages and found the reiserfsck command quite 
>>>>> interesting :-)
>>>>
>>>>
>>>> You have lost the super block and probably old root node. I said 
>>>> "probably" because it might be not used in teh time you did format.
>>>
>>>
>>> Can you be more clear on this? In which cases does SB get used? Or, 
>>> how would this change Nicholas scenario in practical use? E.g. would 
>>> a blank mount (just mount+umount) change the story so that the 
>>> partition would not be recoverable? 
>>
>>
>> I'm sure that usual mount+umount cannot do bad things like this. But 
>> why did you ask me about that? I mean that Nicolas has explicitly 
>> formatted 
>
>
> No, sorry, for bothering again with my unclear questions.
>
> I meant mounts happening just after a mkreiserfs run in failure and
> wanted to make sure for myself that they don't destroy more.
> (Rest of the story: Acoustically a mount after mkreiserfs does MORE on
> disk than a mount of a previously mounted fs. So I just only thought of
> a first mount changing Nicholas' circumstances before successful 
> recovery.) 

Probably you mean journal replaying in new filesystem on mount? If so, 
then it is not big deal, due to journal is created empty.

>
>
>> his partition by means of using mkreiserfs. And I know that 
>> mkreiserfs creates new super block at 16 block and also leaf for root 
>> node, first bitmap and journal.
>>
>> As old bitmap is corrupted, --rebuild-tree -S will be apparently needed.
>
> [...]
>
>> So the plan is the following:
>>
>> (1) Backup your data.
>> (2) Rebuild super block if it had some special like relocated journal.
>> (3) Rebuild tree with partition scanning --rebuild-tree -S.
>>
>> This should reanimate your old partition but expect, that you will 
>> found some of your data in the /lost+found directory.
>
> [...]
>
> Thank you, 

You are wellcome.

>
>
> Manuel
>
>
>


-- 
Yury Umanets