Post Ddrescue Tips?

[dubesor]

(See my original post about my drive failure) After nearly a month waiting for ddrescue to finish rescuing data from a dying 1TB drive to a new 2TB drive, it's finally almost done ... and with just 2.5 GB worth of errors to report, not bad.

I've seen lots of recommendations to run fsck on the destination drive after the rescue process.

What is the equivalent command if I'm working with an NTFS filesystem? (the dying drive came from a Windows 7 PC)

And how long should I expect this error checking to take?

[drguy]

2.5GB of errors is an astronomical amount.
The first thing you need to do is make another clone of whatever you cloned, then run chkdsk.
If chkdsk clobbers your partition and you dont have another clone, get ready for another month of cloning and thats if your drive survives that long again.

[dubesor]

2.5GB of errors is an astronomical amount.
The first thing you need to do is make another clone of whatever you cloned, then run chkdsk.
If chkdsk clobbers your partition and you dont have another clone, get ready for another month of cloning and thats if your drive survives that long again.

I thought 0.25% of errors (of total drive size) wasn't so bad ... but you're probably right.

For rescue job #1, I had the -r 2 switch set in ddrescue. I'm guessing I don't need to for clone job #2 (from one physically brand new drive to another) ?

Any other things I can do to speed up that bit-by-bit copy process? Someone mentioned killing certain background monitoring/logging processes in Linux ... Or maybe there's a more optimized setting in BIOS?

[drguy]

1TB should take you about 4 hours to clone in linux.
Dont use ddrescue for this, use dd instead
dd if=/dev/src of=/dev/dst bs=512000 count=2000000
This will clone 1TB.

[dubesor]

1TB should take you about 4 hours to clone in linux.
Dont use ddrescue for this, use dd instead
dd if=/dev/src of=/dev/dst bs=512000 count=2000000
This will clone 1TB.

Are the bs= and count= switches necessary? What do they do?

[drguy]

run:
man dd

[dubesor]

Thx! Not being in front of my Linux, I checked here - http://www.netadmintools.com/html/dd.man.html

Appreciate the advice; will begin clone job #2 tonight

[drc]

dd will not work with bad sectors

[muu1]

I think its ok to use dd now because he is using dd to copy the ddrescue image to make a second copy to work on. I believe if ddrescue can't read bad sectors it fills those areas on its image with zeros.

I used ddrescue on another drive a few months ago and it is great program. After getting the image from ddrescue I used R-Studio program in Windows and files recovery was almost 100%.

[Agent]

On pretty light cases ddrescue can do a decent job, but on drives with heavier issues, it will barely touch the drive at all.

[muu1]

The drive I was successful with was an 80GB WD. ddrescue initially found about 300MB of errors and it kept retrying overnight until errors were down to about 500KB when I stopped it, I was pretty impressed. Might have been lucky with that drive though.

[drc]

dd will not work with bad sectors

I misread. Thought he was talking about running another pass on his bad drive.

[dubesor]

Now ddrescue is "trimming failed blocks". Errsize has dropped dramatically from about 2.5 GB to 670 MB ... but I'm also running out of patience.

What happens to the data if I terminate ddrescue with CTRL+C at this point before it's totally finished?

Is the filesize it's showing me in the 'rescued' row already copied to my healthy drive?

[Nick_CT]

Yes. If you used a logfile you can stop it with ctrl-c, check what it's copied to your 'healthy drive' and then carry on from where you left off if you want to continue running it.