Running ddrescue ( not dd_rescue ) for the past 10 hours
 

Is it safe to conclude ddrescue is not the right tool for this task? What other tool can I try? I already tried testdisk

http://imgur.com/gallery/JfBUpvs

I am trying to save a 900GB partition from a 1.5T hard disk

I am trying to save it into a 4TB external hard disk

Is that usb2 or 3?

When I write to a usb3 capable HD on a usb2 port, I get about 40MB sec. That's 2.4GB per minute. 375 minutes for 900GB. 6.25 hours.

ddrescue reads back and forth over and over again trying to get unreadable sectors. Is that what is says?

I've had ddrescue run for an hour on a DVD that had bad spots, until I finally stopped it.

https://wiki.archlinux.org/index.php...ynchronization

https://wiki.archlinux.org/index.php...ackup_programs

ddrescue continously reports how long it has been since the last successful read. If it has been running a long time without recovering any more data, it's time to stop it. (The definition of "long time" is subjective.)

dd's defaults for block size etc. are ~20 years old, so you probably need to boost them. I have seen people using up to 16MB (bs=16M).
Whenever I "burn" a distro to a USB, not setting the bs means it will take 3x as long.

"dd" and "ddrescue" are totally different programs.

I'm aware of that.
But the name must mean something.
Anyhow my argument still stands; a quick glance at the man page shows that ddrescue has settings for cluster and sector size. I am sure it would be beneficial to increase those on a terabyte operation.

It definitely makes a difference what sector size you specify for the dd's, and how long it takes to copy. I haven't tested that for years.

I thought that one sector at a time might be the fastest. So for a DVD 2048, for a HD 512 or 4096. I never thought of taking that big of a bite. I'll have to give that a try.

Or 20 sectors at a time makes sense. So that you are not splitting sectors.

Edit: Spellcheck gave me the wrong word.