[SOLVED] Recovering data from a faulty HDD & choosing a replacement solution (+ what's the longest a drive has lasted you?)

patashnik

Distinguished
Mar 30, 2013
12
0
18,510
Hi folks, I have a few questions around a specific hardware situation. In short: (1) I have an old HDD suspected to be dead and whose data I'm looking to recover if possible, and (2) I'm trying to decide on a storage solution with which to replace it.

To preface, I've read the pinned threads in the Storage sub-forum and noted the advice given. I realise that, especially in more particular instances such as this, it's not as easy to recommend a single suitable product - which I'm not necessarily looking for as such but would appreciate any suggestions nonetheless - I'd be grateful just for any advice, specific or general, on either of these issues. Bear with me if it gets a bit War and Peace - I'd rather provide too much detail than too little.

The old drive:
I'd been using a WD Caviar Green as my main user data partition (with boot/OS/program data partitions on a separate SSD). It'll be 9 years old now, which is either a miracle, or I simply RMA'd it sometime and forgot. After rebooting the desktop containing this drive the other day, the system wouldn't pass POST, with the BOOT_DEVICE LED on the mobo a solid red. After a few hours of troubleshooting/deep cleaning/frustration, I narrowed the issue down to this drive, and the system can now POST and reach the bootloader.

I say all this because reconnecting the SATA cable to this drive just produces the same issue, and it's very important I rescue as much data as possible from it - for various reasons, the last backup ran was a couple of weeks ago, in which time I've done a lot of work and file ops. (Yup, I know, this my wake-up call to set up a cron job or something to better automate things.) I'm hoping I can hotplug it after POST/booting to OS and run a recovery tool if the partition table is janky or there are too many bad sectors to access the data otherwise.

My daily driver is Linux (LMDE, though I also have a W10 partition), so I'm comfortable using a *nix CLI method if needed. Any tips on certain tools to use/approaches to take?

The new drive:
It's clearly time to replace the ancient HDD in question, but rather than skimp and go for a typical halfway-decent HDD for the lower cost per GB, I'd like to decide on a more robust, long-term solution. The old disk is 2 TB, which is the minimum I need from the replacement though preferably 3 TB - I can be flexible depending on other factors such as performance/redundancy or if changing my current (haphazard) setup, which is:
  • a 240 GB SSD for OS installations and all programs except games
  • the faulty 2 TB HDD
  • another, similar 2 TB HDD used as the backup drive
  • an even older(!) generic 500 GB HDD only used for replaceable data because it still works
- all internal. As above, the new drive is to be my main storage device, in my case meaning, in size order: game data, music, images/video and docs/code/other files (however, what I do most often is word processing/light coding and rarely transcoding/rendering work).

While I'm leaning towards getting a single large SSD that meets the criteria as mentioned (high capacity and reliable, mainly), I'm also considering trying out a RAID - level 5 in particular, based on this post - in which case I'd definitely have to make changes to the above setup, since I've used up all the 3.5" drive bays in the case and would have to be convinced to go external (unless of course if for cold storage). I'm also wondering whether an SSHD would be suitable, or if a hybrid RAID of some description wouldn't be more hassle than it's worth, in this case. Finally, I'm not wild about the idea of cloud storage - ignoring the time spend for an initial vs. incremental/differential backup and cost effectiveness over time, I'd be hesitant re: security and ease of access (unless, for instance, Microsoft allows rsync - I'm frankly not up to date on what's available out there).

NB re: connectivity, especially if for an M.2 SSD, the mobo is a similarly dated Asus Sabertooth Z77 that has only one free PCIe 3.0 slot.

In summary, for the faulty drive's replacement as a 2-3 TB main user data storage, I'm trying to decide between:
  • one large SSD
  • an SSHD
  • a RAID config, level TBD and possibly hybrid, if suitable/practical.
Again, I'm looking to both recover the data on the old drive if possible and decide on the best type of storage to replace it, and would be grateful for any suggestions for either. Thanks in advance.

Finally - just out of interest/for fun - what's the oldest a drive (of any type of secondary storage) that you've owned/encountered has lived for and remained operational?
 
Solution
Hi folks, I have a few questions around a specific hardware situation. In short: (1) I have an old HDD suspected to be dead and whose data I'm looking to recover if possible, and (2) I'm trying to decide on a storage solution with which to replace it.

To preface, I've read the pinned threads in the Storage sub-forum and noted the advice given. I realise that, especially in more particular instances such as this, it's not as easy to recommend a single suitable product - which I'm not necessarily looking for as such but would appreciate any suggestions nonetheless - I'd be grateful just for any advice, specific or general, on either of these issues. Bear with me if it gets a bit War and Peace - I'd rather provide too much detail than...
Hi folks, I have a few questions around a specific hardware situation. In short: (1) I have an old HDD suspected to be dead and whose data I'm looking to recover if possible, and (2) I'm trying to decide on a storage solution with which to replace it.

To preface, I've read the pinned threads in the Storage sub-forum and noted the advice given. I realise that, especially in more particular instances such as this, it's not as easy to recommend a single suitable product - which I'm not necessarily looking for as such but would appreciate any suggestions nonetheless - I'd be grateful just for any advice, specific or general, on either of these issues. Bear with me if it gets a bit War and Peace - I'd rather provide too much detail than too little.

The old drive:
I'd been using a WD Caviar Green as my main user data partition (with boot/OS/program data partitions on a separate SSD). It'll be 9 years old now, which is either a miracle, or I simply RMA'd it sometime and forgot. After rebooting the desktop containing this drive the other day, the system wouldn't pass POST, with the BOOT_DEVICE LED on the mobo a solid red. After a few hours of troubleshooting/deep cleaning/frustration, I narrowed the issue down to this drive, and the system can now POST and reach the bootloader.

I say all this because reconnecting the SATA cable to this drive just produces the same issue, and it's very important I rescue as much data as possible from it - for various reasons, the last backup ran was a couple of weeks ago, in which time I've done a lot of work and file ops. (Yup, I know, this my wake-up call to set up a cron job or something to better automate things.) I'm hoping I can hotplug it after POST/booting to OS and run a recovery tool if the partition table is janky or there are too many bad sectors to access the data otherwise.

My daily driver is Linux (LMDE, though I also have a W10 partition), so I'm comfortable using a *nix CLI method if needed. Any tips on certain tools to use/approaches to take?

The new drive:
It's clearly time to replace the ancient HDD in question, but rather than skimp and go for a typical halfway-decent HDD for the lower cost per GB, I'd like to decide on a more robust, long-term solution. The old disk is 2 TB, which is the minimum I need from the replacement though preferably 3 TB - I can be flexible depending on other factors such as performance/redundancy or if changing my current (haphazard) setup, which is:
  • a 240 GB SSD for OS installations and all programs except games
  • the faulty 2 TB HDD
  • another, similar 2 TB HDD used as the backup drive
  • an even older(!) generic 500 GB HDD only used for replaceable data because it still works
- all internal. As above, the new drive is to be my main storage device, in my case meaning, in size order: game data, music, images/video and docs/code/other files (however, what I do most often is word processing/light coding and rarely transcoding/rendering work).

While I'm leaning towards getting a single large SSD that meets the criteria as mentioned (high capacity and reliable, mainly), I'm also considering trying out a RAID - level 5 in particular, based on this post - in which case I'd definitely have to make changes to the above setup, since I've used up all the 3.5" drive bays in the case and would have to be convinced to go external (unless of course if for cold storage). I'm also wondering whether an SSHD would be suitable, or if a hybrid RAID of some description wouldn't be more hassle than it's worth, in this case. Finally, I'm not wild about the idea of cloud storage - ignoring the time spend for an initial vs. incremental/differential backup and cost effectiveness over time, I'd be hesitant re: security and ease of access (unless, for instance, Microsoft allows rsync - I'm frankly not up to date on what's available out there).

NB re: connectivity, especially if for an M.2 SSD, the mobo is a similarly dated Asus Sabertooth Z77 that has only one free PCIe 3.0 slot.

In summary, for the faulty drive's replacement as a 2-3 TB main user data storage, I'm trying to decide between:
  • one large SSD
  • an SSHD
  • a RAID config, level TBD and possibly hybrid, if suitable/practical.
Again, I'm looking to both recover the data on the old drive if possible and decide on the best type of storage to replace it, and would be grateful for any suggestions for either. Thanks in advance.

Finally - just out of interest/for fun - what's the oldest a drive (of any type of secondary storage) that you've owned/encountered has lived for and remained operational?
No matter what type storage you get they all fail sooner or later.
If your data is important backups are not an option.
How often to backup depends on how much can you afford to lose.

I like to keep things simple so I don't use raid....your option.
You might want to limit the use of the problem disk until someone comes along with some stick time on a recovery method.
 
Solution
Yeah, RAID is a universally terrible idea for about 99% of consumers. It has very specific uses and none of them are "backups." RAID is for data availability, not data protection.

Since it's just data, the simplest solution and the one I would go with is a sufficiently large traditional hard drive and a duplicate as a daily backup. I'd also at a bare minimum also want the most crucial data backed up daily in the cloud. The 321 rule is a good rule of thumb: a minimum of at least three different copies of important files, on at least two different devices, with at least one off site.

As for recovering the old files, tools like Recuva and EaseUS will give you your best chance to get things back on your own. There are more effective means, involving very expensive data recovery services that use clean rooms, but only you know how important your lost data is.

There are lots of tools to set up automated recovery, depending on what you do. Some have Macrium Reflect simply backup an entire disk image. Personally, I use Bvckup2 and since I carefully have my files in specific directories, have it copy new and updated files nightly rather than the whole shebang. Only you will know what works best for your needs!
 
No matter what type storage you get they all fail sooner or later.
If your data is important backups are not an option.
How often to backup depends on how much can you afford to lose.

I like to keep things simple so I don't use raid....your option.
You might want to limit the use of the problem disk until someone comes along with some stick time on a recovery method.
Oh, I know whatever shiny new piece of kit I get this time will also give out eventually - no escaping that. Though, either the (most important) files were something I could actually afford to lose, or I was just that careless. If you can't serve as an example, serve as a warning and all that.

Anyway, I've not even been able to access the disk, so I've at least done something right by leaving the data alone till I'm absolutely sure of what I'll do.

Yeah, RAID is a universally terrible idea for about 99% of consumers. It has very specific uses and none of them are "backups." RAID is for data availability, not data protection.

Since it's just data, the simplest solution and the one I would go with is a sufficiently large traditional hard drive and a duplicate as a daily backup. I'd also at a bare minimum also want the most crucial data backed up daily in the cloud. The 321 rule is a good rule of thumb: a minimum of at least three different copies of important files, on at least two different devices, with at least one off site.

As for recovering the old files, tools like Recuva and EaseUS will give you your best chance to get things back on your own. There are more effective means, involving very expensive data recovery services that use clean rooms, but only you know how important your lost data is.

There are lots of tools to set up automated recovery, depending on what you do. Some have Macrium Reflect simply backup an entire disk image. Personally, I use Bvckup2 and since I carefully have my files in specific directories, have it copy new and updated files nightly rather than the whole shebang. Only you will know what works best for your needs!
To clarify, neither the old disk nor the replacement were/will be for backups, only live/source data.

I'm still not too keen about RAID - like Bob there, I'm for the KISS principle - but do want to keep my options open at a juncture like this. I understand your point re: data availability vs. protection, though, and I was looking to adhere to something like that 321 rule as a matter of course going forward.

I did see Recuva suggested elsewhere and looked it up, though unless something better suited for damaged/faulty disks comes up, I'll probably just stick with a CLI tool I used once years ago (the name escapes me right now, but it's an old and powerful for things like deleted partition tables).

As for automation, I'd been doing fine with a custom shell script for incremental backups using rsync, apart from the simple/stupid matter of frequency. Again, unless something more suitable crops up, I'll stick to it and just change my behaviour.

(Yup, could just have the system run the script for me routinely, but I preferred having the manual control over what/when in case I didn't want a backup copy written over. Time to up my strategy and maintain an aux backup location...)


Thanks for the initial responses, guys.
 
Update if anyone's curious (I hate troubleshooting threads that end on a cliffhanger...)

The BOOT_DEVICE LED issue:
happened to me again, even when the faulty HDD was disconnected. After another round of troubleshooting, I realised it was due to the printer's USB cable of all things. I still don't understand why, but I don't have any other explanation. Remember to unplug any peripheral cables, even unsuspecting ones, when troubleshooting boot issues!

The old drive:
thankfully was still accessible. Although the system still couldn't POST while the SATA cable was connected to it, I hotplugged before booting into a live USB of LMDE (Linux Mint Debian Edition) and was then able to access the drive as normal in gparted. Interestingly, the SMART checks reported the drive as OK - there's nothing to indicate there's a fault apart from the failure to POST when it's connected. I'm still waiting on its partition to be copied over to the new drive, but I've had no errors so far so am cautiously optimistic that it'll succeed unless it hits a bad sector or something.

The new drive:
if anyone's interested, I went for a 4 TB Samsung 870 QVO. (Secondary backup/cold storage TBD - either a half-decent HDD plus an enclosure for easy removal or simply an external portable drive.)
 
Longest lasting drive?
Dunno....20+ years old and still works.

Shortest?
5 weeks off the store shelf.

ALL storage devices are liable to die in the next 0.25 sec.
Plan accordingly.

A dead drive should never involve data loss. At most, "Aw crap, I have to buy a new one"