DDRescue + Gui - RIG Config. Recommendations.

Mainstay

Well-Known Member
Reaction score
747
Hi All,

I used dd_rescue on a damaged drive about 2 years ago and it worked amazingly.

The downside is that I was so freaked out about the interface and the procedures that I basically stopped after that first attempt. There is no way I could safely train a staff member to perform these recoveries and still sleep at night.

I understand ddrescue is superior to dd_rescue and that there is even a GUI that helps things along.

First: is ddresuce + gui just as good? Maybe an expert could tweak it further with CLI but will this hit 95% of the scenarios you encounter?

Second: What OS do you run out of and / or recommend?

Definitely not a Linux guy, so hopefully someone says, "ya, Ubuntu with the repository installed will get you all you need."

Thanks all!

--Matthew


(btw: read the other posts on this and no-one mentioned exactly what flavor of Linux they were using)
 
I used to just have a junker desktop with a burned Parted Magic cd in and set as default boot device. Parted Magic comes with ddrescue preinstalled and is a great little diagnostic Linux with lots of utilities installed and "just works". Pretty sure it has the GUI for ddrescue too, but I just taught the other techs the command line and printed a cheatsheet that i taped on the wall next to the monitor.
https://partedmagic.com/
 
I used to just have a junker desktop with a burned Parted Magic cd in and set as default boot device. Parted Magic comes with ddrescue preinstalled and is a great little diagnostic Linux with lots of utilities installed and "just works". Pretty sure it has the GUI for ddrescue too, but I just taught the other techs the command line and printed a cheatsheet that i taped on the wall next to the monitor.
https://partedmagic.com/

It's like you're looking over my shoulder at my own workbench! Except my cheat sheet is taped to the side of my junker desktop, and I boot Parted Magic from usb drive rather than CD.

I've used ddrescue many more times than I can count. It's the only thing I ever do in Linux. I really don't have to understand it, I just follow my cheatsheet and do the same thing every time. It works.
 
  • Like
Reactions: GTP
Nothing special is required for the rig. Anything with SATA 3 would be fine and I even have one Linux machine still using ddrescue with SATA 2. Ddrescue with the GUI is the same program. It would be easy to write directions for the GUI. I prefer to run Linux Mint Mate but most Linux users have their preferred flavor (distro).
 
@mraikes and @Krynn72 - would you mind sharing your respective cheat sheets? If not, no worries. I just remember being pseudo-terrified each keystroke. It was a sensitive recovery that worked just wonderfully (I was working of a clone, so the original could have still been sent to the big boys).
 
@mraikes and @Krynn72 - would you mind sharing your respective cheat sheets? If not, no worries. I just remember being pseudo-terrified each keystroke. It was a sensitive recovery that worked just wonderfully (I was working of a clone, so the original could have still been sent to the big boys).

Certainly, I will tonight when I get back there.
 
I have a couple of data drives in my linux machines (I run Mint) for image targets. i found it useful to print the disk IDs out on labels and stick it to the front of the machine. Helps minimize the chance you'll mix them up if your patient drive happens to have the same capacity
 
I found it useful to print the disk IDs out on labels and stick it to the front of the machine. Helps minimize the chance you'll mix them up if your patient drive happens to have the same capacity

I find having a black cable on SATA 0 and a red cable on SATA 1 helps hugely. Data transfers always go from black to red, and always from Drive 0 to Drive 1.

It's foolproof (but not damn-foolproof) and in thirteen years I can count the number of times I've got it wrong on the fingers of one foot.
 
@mraikes and @Krynn72 - would you mind sharing your respective cheat sheets? If not, no worries. I just remember being pseudo-terrified each keystroke. It was a sensitive recovery that worked just wonderfully (I was working of a clone, so the original could have still been sent to the big boys).

Ok, this is it in all it's brief glory. Now that I think about it, it probably came from someone here years ago.

And a caveat - all I do is clone drive to drive, so that's what this does. It isn't for creating or working with images.

To see a list of devices currently attached to computer, run this command:
# lsblk -o name,label,size,fstype,model

Cloning directly to new disk:
Assuming that /dev/sda is the old drive and /dev/sdb is the new drive:

ddrescue -d -f /dev/sda /dev/sdb

On my junker/ddrescue computer, I have exposed sata connections that I've marked "A" and "B", so I know without running lsblk which drive is which.
 
@mraikes and @Krynn72 - would you mind sharing your respective cheat sheets? If not, no worries. I just remember being pseudo-terrified each keystroke. It was a sensitive recovery that worked just wonderfully (I was working of a clone, so the original could have still been sent to the big boys).
Don't have access to it anymore since I left that shop. However if you just get a big drive and just save image files rather than clone direct to another disk there's literally no risk involved. That's the way I did it unless I was just upgrading an HDD I didn't trust to be fully healthy. Any data recovery hdd was always turned into an image.

My cheat sheet was essentially just the relevant flags I use (-n or -a for instance) and when I'd suggest using them. Most of it was just based on this guide I found on some obscure website.
 
Awesome - thank you guys. Really appreciate it. I will be setting this up tomorrow and have a test drive in mind to practice on.

I hope my next foray into using this software goes as smoothly as you guys have had it and that it simply integrates into our practices.

It is yet another area that I feel we make due using other techniques, but I know there is a really good way that I am avoiding. Time to address these areas in 2018!
 
I used to use Parted Magic to run ddrescue, but I could not figure out how to make persistent changes in that environment (not exactly a Linux guru...). Used an old machine and set it up for dual boot with Win 7 and Linux Mint. It took some time, but I eventually got the various tools I used on Parted Magic installed on Mint. I connect the customer drive(s) via an eSATA docking station (machine is under the desk and I'm getting too old to crawl under there to hook up SATA cables.)

My cheat sheet is very similar to one posted above, but I use ddrescue to make a full image of the customer's drive before doing anything to it, so my ddrescue command is a little bit different:

To run ddrescue to create an image file

ddrescue -d /dev/sdX test.img test.logfile


If there are errors, then run this command to retry bad sectors

ddrescue -d -r3 /dev/sdX test.img test.logfile


Harry Z
 
@Haole Boy ; so you'd hot swap in a client's drive into the docking station and then clone it to an image on your linux / win box. How long would you hold onto the images? The capacity of your cloning station must have been significant, no?

Right now we clone to a rotating stack of shop drives that we write the latest ticket number and the date on a label. Very easy, but there is always the possibility of error.

I like the idea of it ALWAYS being A to B so that operator error is reduced. So making images would eliminate that (so long as you used a unique naming structure and don't overwrite another client's data).

I am also really interested in the recovery aspect of ddrescue as this would prove useful on those drives that are damaged but not quite at that recovery lab level (or have value that justifies the cost of the recovery effort).

I'm going to be using a hotswappable raid bay that we will mount client drives in a cartridge and then slot into the machine. Should make for a faster and safer way of transferring data than by going through a USB connection.
 
@Haole Boy ; so you'd hot swap in a client's drive into the docking station and then clone it to an image on your linux / win box. How long would you hold onto the images? The capacity of your cloning station must have been significant, no?

Aloha @Mainstay. My business is very small (sole proprietor) and most of my work is residential and done at the client's house. For machines that I find (or suspect to be) heavily damaged, then I bring them back to my place to do a full image before working on them. I have a Synology 4-bay enclosure with 4 x 2TB drives in it. And I'm running the Synology raid configuration (can't remember what they call it) which gives me 6 TB of storage space. I try to keep customer data for 30 days after the machine is returned to the customer, but don't always have the space to do so. If I have to clean up space for a current customer, I delete the oldest first and then get rid of the images where there was really not much wrong with the system.

As for using ddrescue for recovery purposes, I did a lot of reading of posts here on Technibble and then asking lots of questions that I did not find the answers to before trying this for my customers. While my data recovery success rate is not overly high, my customers do not want to spend the $400 minimum for professional data recovery. I have only had 1 customer in 8 years who agreed to go with the pro's. But, I did have a major success last month. I was only able to recover a couple of folder's in Documents, and it was what the customer needed!

Harry Z
 
I'm going to be using a hotswappable raid bay that we will mount client drives in a cartridge and then slot into the machine.

That sounds a little more robust than my setup - I use 2 of these guys in each machine. Only downside is you need to use a MB with a lot of SATA ports, and make sure your power supply has enough connectors. Two of those plus 2 internal drives plus a CD drive is 7 SATA ports. It makes a neater setup, I can close up the case and not have cables lying everywhere.
 
Hi All,

I used dd_rescue on a damaged drive about 2 years ago and it worked amazingly.

The downside is that I was so freaked out about the interface and the procedures that I basically stopped after that first attempt. There is no way I could safely train a staff member to perform these recoveries and still sleep at night.

I understand ddrescue is superior to dd_rescue and that there is even a GUI that helps things along.

First: is ddresuce + gui just as good? Maybe an expert could tweak it further with CLI but will this hit 95% of the scenarios you encounter?

Second: What OS do you run out of and / or recommend?

Definitely not a Linux guy, so hopefully someone says, "ya, Ubuntu with the repository installed will get you all you need."

Thanks all!

--Matthew


(btw: read the other posts on this and no-one mentioned exactly what flavor of Linux they were using)
I have a tower with one side removed on my bench and have a dual boot win 10 and ubuntu (with ddrescue installed) disk connected to it, have a couple sata connectors and power cables hanging out to attach the corrupt drive and backup drives to, it works good for me.
 
Since I run quite a few external SATA docks I put these external ports on all my bench machines instead of keeping them open and trying to hot swap off a motherboard.

ESATAPLATE2.Main.jpg
 
I'm sure you end up with a pile of old computer parts... here's what I did: 4bay SATA dock: http://amzn.to/2meNCgG
Shove that into a tower that has three 5.25" bays on the front. That way you don't have to mess with cables or open the side cover all the time. Make sure the mobo you use has hot-swap capability, usually need to enable that through the BIOS. Buy Parted Magic for the ridiculously low price that it is. I keep a Windows install on a drive with other tools, Paragon and Acronis. Parted Magic makes things difficult when going from something like 1TB to 256GB SSD. Paragon and Acronis through Windows make that a lot easier with higher success rate.
 
Back
Top