Fwd: [help] how to deal with Incorrect metadata area header checksum in LVM
max
maximilianbianco at gmail.com
Fri Jun 13 22:00:28 UTC 2008
max wrote:
> Robin Laing wrote:
>> max wrote:
>>> Zhen Zhou wrote:
>>>> On Wed, Jun 11, 2008 at 8:45 PM, David Timms <dtimms at iinet.net.au>
>>>> wrote:
>>>>> Zhen Zhou wrote:
>>>>>> So what is the right path to recover these lvm?
>>>>>>
>>>>>> Any tips will be welcome! Thanks in advanced,
>>>>>>
>>>>>> BTW: the attached is lvm metadata file. More detail information
>>>>>> needed, pls tell me.
>>>>> You might find experts in LVM hanging out on:
>>>>> linux-lvm at redhat.com [subscription required].
>>>>>
>>>>> Good luck.
>>>> FTA, thanks for your tips.
>>>> and I really wonder whether these errors "Incorrect metadata area
>>>> header checksum,
>>>> Volume group VolGroup00 metadata is inconsistent" is impossible to
>>>> recover.
>>>>
>>>> Or what I need to deal, is to follow:
>>>> http://thread.gmane.org/gmane.linux.redhat.fedora.general/196481
>>>> to recover all lvm's data?
>>>>
>>>> Any tips will be welcome. TIA
>>>>
>>>> Zhou Zhen
>>>>
>>> I asked in a related thread on the developer's list and I'll ask
>>> here. Someone please let me know if my question doesn't make any
>>> sense. Is it harder to recover files from an LVM than with the older
>>> partioning scheme?
>>>
>>
>> FWIW, I did recover data from an LVM that was removed (on purpose) in
>> part of an upgrade. I couldn't mount the drive as the LVM recovery
>> procedures listed and I thought I was toast. If I left the drive in
>> the computer, the computer would try to add the drive back into the
>> LVM no matter what I tried.
>>
>> To complicate things even further, the drive was part of a RAID 1
>> array. I put the drive into a carrier. As the drive was SATA, I
>> ended up getting a Thermaltake carrier that supports SATA and IDE (end
>> product plug). :)
>>
>> It was many months ago but I used normal recovery tools to pull the
>> data off of the drive. I would have to look at my notes which are at
>> home. If I remember I will see if I can find my notes.
>>
>> Basically, it is possible but I don't know if it is harder as this was
>> my first time.
>>
> I suspect it is more difficult by comparison but I could easily be
> misinterpreting what I am reading. For instance this appears on
> wikipedia at the end of the LVM entry.
>
>> Caveats
>>
>> The current implementation does not support write barriers. This means
>> that any guarantee against filesystem corruption offered by journaled
>> file systems like ext3 and XFS is negated.
>
> It seems to imply that a lack of support for write barriers may be more
> likely to cause/lead to/result in/not protect against corruption and
> therefore by extension make data recovery more difficult in particular
> cases. Though whether support for write barriers is now present I don't
> know. I have learned a bit about file operations in searching but I am
> still not clear on what a write barrier is, though I think I am getting
> close, if you know of some place I can find a factual definition that
> would be great because I have found many instances of people discussing
> them but obviously both parties know what it is and so they never have
> to or bother to agree on a common definition.If you can explain what a
> write barrier is then that would be even cooler. I feel like I have most
> of the pieces but I am missing the lynch pin that will hold it all
> together. Though its just as likely that the answer will cause me more
> confusion but I won't know until I find it.
>
> Here is the url to the entry:(its not very long)
>> http://en.wikipedia.org/wiki/Logical_Volume_Manager_(Linux)
>
>
>
Funny but I think I have found my answer, right under my nose as usual : )
Thanks for responding though because I probably wouldn't have seen it
otherwise.
--
Fortune favors the BOLD
More information about the users
mailing list