Accidentally cloned disk

[matw2]

Hi, I think I made a massive screw up!

I had a laptop but the hard disk failed. The data is still there but in Windows the disk repeatedly connects and disconnects when connected via SATA. Someone suggested I try Ubuntu.

I connected it up and Ubuntu mounted it with no problems. My friend suggested I try and make a disk image, only he linked me to the wrong article. Instead of giving me a step by step guide to imaging the disk he sent me a step by step guide to cloning the disk. I, being somewhat new to all this, didn't realise the difference. So I hooked my failing HDD up to my new computer and in Ubuntu terminal typed "sudo dd if=/dev/sdd of=/dev/sda" The drive disconnected after about an hour and only 72 GB was transferred successfully.

Instead of making a copy of the partitions from the failing HDD on to the new HDD it seems to have overwritten everything on the new HDD. As it did not complete successfully I cannot boot to windows.

Is there any way back from this? I'm happy to get 72 GB of data from the old disk but not at the expense of everything that I have written to the new disk over the last 2 weeks!

I'll be sitting over in newbie corner, praying for a reply

[Vulcan]

Oh dear, several mistakes here e.g. for a flakey hard disk, you don't use that command line to clone it anyway.

Unfortunately I'm not completely understanding your story about the disks which are being used, and the PCs which are being used. Just to be clear: On this (new?) system, /dev/sdd was the "problem" disk (originally from a different system), and /dev/sda was the system disk in that new system, which usually booted Windows, and this is where you wanted an image file of the contents of /dev/sdd to be written. Is that correct? If not, then please explain clearly the corrections.

[Edited to correct my interpretation of what /dev/sda usually contained.]

[mr_spokk]

I would say everything above 72Gb is fairly intact...all below that is gone.

[matw2]

Thanks for the quick reply.

sdd was the old 320 GB problem hard disk. It had the name "Acer"
sda was the new 1 TB hard disk which has Windows Ultimate.

I had booted to Ubuntu using a USB flash drive. The instructions I followed are here: http://www.howtogeek.com/howto/19141/cl ... u-live-cd/

I was able to browse and access all the files that I wanted, and copied them to the functional disk. I thought that rather than just bin the disk I would image it in case there was anything that I had missed.

Unfortunately now I cannot boot to Windows and when booting into Ubuntu the 1 TB disk is now named Acer and none of the data that I had successfully copied is on the disk. Also any files that were on the disk before I started this are all gone too.

Hope that clarifies; I have my Windows recovery CD if all else fails.

[BlackST]

Forget what is below the 72 gb line.
That's why we - me first - usually use HW tools or protections to avoid these 'accidents' .

[Vulcan]

@matw2,

Thanks, yes that's clearer for me (I expected that things were as you explained, but I didn't want to assume and then make a mistake in subsequent suggestions).

Unfortunately I agree with mr_spokk & BlackST - you've replaced the first 72GB of the 1TB disk (which includes the volume name, hence why that has changed on the 1TB disk), with the first 72GB from the 320GB disk. Not only will that have overwritten some data on the 1TB disk, but also some of the original filesystem metadata (e.g. directory info) from the 1TB disk has likely been overwritten. I'm sorry to say this is a classic example of the risks of DIY recovery attempts, and the lack of protection against human error when using this type of Linux/Unix utility. I've seen it several times before, when Linux/Unix sys admins are under pressure...

You now have some choices about what you want to do:

(a) attempt to recover files (again) from the 320GB disk, and/or
(b) attempt to recover files from the 1TB disk (note the potential problem I explain below)

and whether you're going to try to do this yourself (using one of the available utilities and you'll need at least one more new disk, and not make any more mistakes!), or use the services of a DR company.

The ability to recover files successfully (i.e. uncorrupted) from the 1TB disk, will be reduced if (as seems likely) a raw recovery is needed, with reduced amount or none of the MFT available.

Also, as I mentioned briefly before, that dd command from your linked web page is not suitable for cloning source disks that aren't perfect (and as you said, it is for cloning disk to disk, not disk to image file). Also it misses some important steps.

Sorry for the bad news.

[matw2]

Thanks for the information.

I have only had the new computer a short period of time and since I haven't had a version of Office on the new computer I've been working entirely in google docs so everything that I've personally created is stored in cloud-based services so it won't be a huge loss.

I think I'll do a clean install of Windows in the circumstances.

Is there any chance you could suggest what might work for imaging the 320 GB disk? I suspect that it's a hardware failure that has caused it as it seems to connect pretty randomly. My only interest is salvaging any files that are still accessible (and I know that the majority of what I need is not corrupt as it was accessible earlier tonight, but the disk might be completely dead). There is nothing on there that warrants spending a massive sum on sending it to a lab so DIY is really all I can afford!

Is it even worth imaging the disk? I've not got successfully past 33% before the disk disconnects/unmounts. Perhaps I should just copy across whatever I can when I can.

Many thanks for your extremely informative replies.

[Vulcan]

everything that I've personally created is stored in cloud-based services so it won't be a huge loss.

That's good news. However you aren't mentioning the files which you had copied from the 320GB disk onto the 1TB disk - some of those may not have been overwritten on the 1TB disk.

I think I'll do a clean install of Windows in the circumstances.

The timing of doing that, depends on the value to you of the files from the 320GB disk (which may still exist on the 1TB disk at the moment) and what chance you want to take of the 320GB disk not failing before you can attempt further recovery. Personally I wouldn't touch that 1TB disk yet, if the files from the 320GB disk are important to you.

Is there any chance you could suggest what might work for imaging the 320 GB disk?

Suitable for beginners? No. Personally I use (a customised) GNU ddrescue when I need to do something like this in my day job (which isn't DR, but elsewhere in the data storage industry), but that has the same risks as dd, it requires some configuration and really needs some practice before using it. In your case, I suspect it would also require some interactive "babysitting" which needs skill gained through experience & practice, for the best possible result.

Other members have previously recommended different cloning software - search the forum for "cloning" or "clone" to read those previous threads. I would just say is that I wouldn't recommend cloning software under Windows, based on my experiences. Of course YMMV and you might be lucky using such software.

Or you could use the services of a reputable professional DR company to clone that 320GB disk using a hardware imager (e.g. member pcimage in Peterborough) - but you mention that you don't want to spend that money.

My only interest is salvaging any files that are still accessible (and I know that the majority of what I need is not corrupt as it was accessible earlier tonight, but the disk might be completely dead). There is nothing on there that warrants spending a massive sum on sending it to a lab so DIY is really all I can afford!

Is it even worth imaging the disk? I've not got successfully past 33% before the disk disconnects/unmounts. Perhaps I should just copy across whatever I can when I can.

Either choice (cloning the disk or doing a file copy) carries risks of the disk deteriorating during the process. You have less control of error recovery and retries, if you attempt a standard file copy using Windows or Linux/Unix. For small amounts of data, some people advocate doing a file copy. Other people advocate always making a clone. Your lack of experience with cloning software might guide you towards doing file copies, but without a crystal ball and not knowing how successful any cloning attempt by you would be, I can't predict what is best for you.

Many thanks for your extremely informative replies.

You're welcome