raid 0 problems

[Michael Sammartano]

Shawn Badger wrote:

>I have found that with certain controller board problems that sticking
>the drive int the freezer for a couple hours will bring it back to life
>long enough to at least recover the data. It is a cheap and not nearly
>as extreme as swapping platters or the controller boards.
>  
>
That being said I would dd the drive at that point to another working 
drive and not try to access bad drive again unless neccessary. It is 
quite possible the heads are not lining up properly. If the motor is 
failing get the data off immediately and remove the drive for storage!

>
>On Sun, 2006-03-26 at 22:11 -0700, Technomage wrote:
>  
>
>>On Sunday 26 March 2006 21:45, Alan Dayley wrote:
>>    
>>
>>>-----BEGIN PGP SIGNED MESSAGE-----
>>>Hash: SHA1
>>>
>>>Technomage wrote:
>>>      
>>>
>>>>since I *know* this can be done, there has to be a way for someone in my
>>>>position to be able to do this without expending THOUSANDS of DOLLARS (in
>>>>money I don't even have) to do this.
>>>>        
>>>>
>>>One trick I have been successful doing is to find a working hard drive
>>>of the same model.  Then, assuming the platters and motors are good, you
>>>can replace the logic board on one drive with the logic board on the
>>>good drive.  Do this very carefully or you will end up with two dead
>>>drives!  But, if the logic board "contains" the problem, this can get
>>>the drive up and running again.  The description of the actual drive
>>>failure seems to point to a logic board problem so this may work for you.
>>>      
>>>
>>I concur. We are, however, without a proper sized torx bit to accomplish this 
>>(will wait and head off to the hardware store tomorrow to find one)
>>
>>    
>>
>>>RAID 0 with two drives means half the data is on one and half on the
>>>other.  The RAID logic will put, for example, 4 blocks on one and then 4
>>>blocks on the other and so on.  That means every 4th (or whatever number
>>>it was using) block is on the dead drive.  That is hard to recover from
>>>because the data doesn't make sense with only one of the drives.  There
>>>are other issues and possible complications but you probably are
>>>studying up on all of that.
>>>      
>>>
>>the above is true in HARDWARE RAD-0. However, given that this was a linux 
>>(suse 10+) software raid-0, it appears (and was proven by direct observation) 
>>that the data was written in serial fashion from 1 drive and spanning to the 
>>second. 
>>
>>    
>>
>>>This is why it costs so much to recover.  It's not easy and take
>>>experience to do efficiently.  I hope you can get the dead drive up
>>>since that would be the easiest way to recover.
>>>      
>>>
>>understandable.
>>
>>we are exploring all available options though (and I have contacted the 
>>developer though have not yet received a reply).
>>
>>
>>    
>>
>>>Good Luck!
>>>      
>>>
>>thank you!
>>---------------------------------------------------
>>PLUG-discuss mailing list - PLUG-discuss at lists.plug.phoenix.az.us
>>To subscribe, unsubscribe, or to change  you mail settings:
>>http://lists.PLUG.phoenix.az.us/mailman/listinfo/plug-discuss
>>
>>    
>>
>
>---------------------------------------------------
>PLUG-discuss mailing list - PLUG-discuss at lists.plug.phoenix.az.us
>To subscribe, unsubscribe, or to change  you mail settings:
>http://lists.PLUG.phoenix.az.us/mailman/listinfo/plug-discuss
>
>  
>

--