Sector by sector clone - how long should it take?

ITMAN

Member
Reaction score
0
Location
Nr Oxford, UK
Hi

I am doing a sector by sector clone for he first time to try and salvage a raw disk

The disk has errors so I expect it to be slow when covering these sectors but we are 4 days in and it's about 18%!

How long should it take with acronis 2014, and 500gb sata in laptop to ext usb 500gb sata?

Thanks
 
It depends on how you set it up. If you are forcing it to retry all bad blocks it might never finish. One recommended practice is to run a scan that skips bad blocks to grab what you can then do a more thorough one. Remember that it has problems and might die without warning.

If the customer is serious about getting back the maximum amount of data that has real monetary value you should stop doing anything and find a data recovery shop. Several data recovery shops on here have regularly mentioned that they get drives that someone had already tried to get data back and destroyed the drive in the process.

I just ran some scans on drives with no known hardware problems using R-Studio. 1TB each using USB to SATA bridges for both source and destination and they took around 20-30 hours each. But it was not a sector by sector image, just a data recovery.
 
Are you using Acronis in Windows or from a Boot device? If your in Windows it's usually hopeless. If your working from a boot CD it may complete. Personally, I haven't found Acronis to be very useful in the past when recovering from a bad drive, only good drives. For bad drives you should look at using ddrescue in Linux.

Your probably causing more harm than good by hammering that drive for 4 days for 18% using Acronis. 4-5 days for a recovery would be my absolute longest time limit before recommending professional recovery (Not including other factors such as read speed and amount of bad sectors vs. good ones that would cause me to stop sooner).
 
Fire up a linux disk use dd for it:

dd if=/dev/sd?? of=/home/user/drive.img conv=noerror,sync

That will skip bad sectors and save the image file in your home directory as drive.img. Fill in the ?? as per the drive to be imaged.

I have done sector by sector drives using hotswap bays and it took about 30 hours to complete also.

coffee
 
If the data is not worth $350, then use GNU ddrescue with a log file.

ddrescue --force {source} {destination} {logfile}

It certainly sounds like the drive is not in great shape and it is possible that most of the drive that you think you have read has just reported errors and skipped a lot of blocks.

If the data is worth $350 or more, I suggest you stop and outsource to a data recovery professional before you completely kill the drive or make the cost of recovery too high.

I've received projects like these, after the tech gives up and have found some of the following reasons for their problems:

- damaged PCB
- corrupt firmware
- weak and/or dead heads
- severe media damage
- failing spindle motor

It is very important to properly diagnose the point of failure, make the necessary corrections (reprogram donor PCB, fix firmware, change heads, fix/replace motor, ect) to stabilize the drive as best as possible and then to tweak the imaging process accordingly (modify read timeouts, block size, read mode, read direction, read look ahead, etc).

As for the OP topic question, "how long should it take?" A perfectly healthy 500GB drive should mirror in less than 2 hours. A failing hard drive could take months, if it is bad enough.

Good luck.
 
Thabks all. The data isn't worth that much but if I can recover something that would be good. I am using getdataback at the min which is working but I will look i to dd rescue

Usually I use macrium to create a backup image of disks before work and acronis to clone to a new disk if replacing so would like to try something better than acronis to do the job - sounds like dd rescue is the way to go but does it do normal clones too? Or is a whole day about right for a sector by sector clone skipping bad sectors?

Thanks again
 
Don't waste your time using data recovery software until you get a full sector-by-sector clone. Then run the data recovery software, if you need them, against the clone. Think of it this way, you should never read a single sector on a hard drive more than once. If you read it, copy it to another drive.

If you try to clone with software such as Acronis, GetDataBack and the likes, they usually read sectors in blocks, they don't log the missed sectors and there is no way to go back and fill in the holes left behind.

Here is an example:

Program reads a block size of 256 sectors at a time. The program hits 100,000 bad sectors throughout the drive. But, because the drive is reading in blocks of 256 sectors, the 100,000 is now 25,600,000 unread sectors. With each sector being 512 bytes, that is about 12.8GB of unread data vs 51.2MB of unread data.

Of course, it is possible that some of the bad sectors will be clustered together, making it possible that there may be more than one literal bad sector within each 256 sector block skipped.

With a program like ddrescue, it logs the skipped sectors and continually will decrease the block size until it has read all the good sectors around the bads. It is possible that multiple read retries (after the first pass is complete) will result in a sector that wasn't read the first time to read the second, third or 100th time.

Does this make any sense? Have I veered off topic? Anyway, if you don't have a professional data recovery imager like a DeepSpar Disk Imager, your best option is to use ddrescue when trying to recover data from failing hard drives.
 
Back
Top