[mdlug] Recovering damaged EXT2

Robert Adkins radkins at impelind.com
Wed Sep 16 14:20:58 EDT 2009


"They" like to say that you need a cleanroom setup to do this work.

If you have a relatively dust free environment and do not care whether or
not you will ever be able to use the current busted or "recovery" drive, you
can do this at home. (Granted the proper skills, steady hands, etc.)

I have heard of it being done and I have even seen the results of someone
taking it upon themselves to open up a perfectly working HD by removing the
cover, machining a window into that cover, cleaning that cover nicely and
inserting a clear plexiglass window into the cover, adding led lights into
the drive, putting it back together and then using the modded drive without
problem... For a while at least.

-Rob

> -----Original Message-----
> From: mdlug-bounces at mdlug.org 
> [mailto:mdlug-bounces at mdlug.org] On Behalf Of Aaron Kulkis
> Sent: Tuesday, September 15, 2009 11:01 PM
> To: MDLUG's Main discussion list
> Subject: Re: [mdlug] Recovering damaged EXT2
> 
> Michael Rudas wrote:
> > --- David McMillan wrote:
> > 
> >> Now, when I try to mount it, I get a "bad superblock"
> >>  error.  If I try to run fsck or e2fsck, I get "the superblock is 
> >> corrupt" and a suggestion to run e2fsck with the "e2fsck -b 8192 
> >> <device".  But if I try that, I get a "device busy" error message, 
> >> even though the partition is not mounted and no other process is 
> >> accessing that drive.
> > 
> > It sounds like the drive is mechanically pranged.  If the platter 
> > spindle gets bent or pushed to one side, it can often read the 
> > partition table but hang up on the platter when it tries to track 
> > across the disc (usually noisy, though).  It's also 
> possible that the 
> > platter has been gouged by the head--or one of the (multiple) heads 
> > has been dislodged and no longer reads that side of its 
> platter.  The 
> > "Device busy" error may be the HD waiting for the data flow 
> from the missing head.
> >
> 
> 
> And if that is the case (which I suspect it is), then you'll 
> have to hire a data forensics service.  Worst case scenarios 
> are that they have to remove the platters from your current 
> drive chassis and put them into another drive chassis.
> 
> This is rather expensive (it requires a "clean room" although 
> the actual space could be a box on a table with both air 
> filters and an air evacuation system, and of course, a set of 
> air-proof gloves going into the box to do the actual work).
> 
> If this doesn't work, then they can go to scanning electron 
> microscopes, and that can detect each and every magnetic 
> field on the platters -- but, that's REALLY expensive ($10k typical).
> 
> So, before you proceed, you have to ask yourself some
> questions:
> 
> 1: How valuable are the files on this disk drive  (in terms 
> of what the business uses them for)?
> 
> 2: How much money (wages, etc.) and time will it
>    cost to re-generate the data?
> 
> 3: How much time is the business able to spend
>     to regenerate the data?
> 
> 4: In light of (1), (2) and (3) above, how much is the
>    business prepared to pay to recover these files?
> 
> 
> If the data on the disk is worth $1,000,000, then spending up 
> to $10k to recover the data is something you should agree to 
> without hesitation.
> 
> On the other hand, if the business value of the data is only 
> worth $100, then probably not.
> 
> These, of course, are extremes.  I'm sure that the value of 
> your data is somewhere in between.
> 
> 
> 
> 
> Lesson:  There's no substitute for a traditional backup 
> system -- be it tape drives, or network attached storage (so 
> that the disks are sitting in a larger computer chassis and 
> not likely to get dropped on the floor).
> 
> Businesses that skimp on data backup are prone to being put 
> out of business by accidents such as these.
> _______________________________________________
> mdlug mailing list
> mdlug at mdlug.org
> http://mdlug.org/mailman/listinfo/mdlug
> 




More information about the mdlug mailing list