Don’t confuse 30-50 years ago with modern practices, do not confuse disks of the times with those since, and do not confuse ‘getting’ data with restoring the data for use.
In the days they were. So were memory cards from governmental systems.
As for the movies world some good mates lived the life being flown around blindfolded and eventually being unhooded in a computer facility they knew not where. I once spent a weekend locked into a secure facility to do my work, with a guard, and needed special dispensation to take out what I took in. I did not have ‘the clearance’ but I was the only one who could do ‘the work’. All the [washing machine sized] disks were physically detached and locked into a secure room except one for me to use and the facility was otherwise shut down as a clean site for the purpose.
Scoff if you wish. Those times were far from boring.
I have concerns with advice provided in this thread. These concerns being:
The age of the drive it is likely to be covered by a manufacturer’s warranty and highly likely to be covered by the Australian Consumer Law. Unless one is willing to throw away potentially several hundred dollars, taking a hammer to the drive should not be recommended. It will void any opportunity to make a warranty/consumer guarantee claim.
The original post indicates data wasn’t lost. It appears that the drive may be readable and writeable. I read the post differently the back up drive may still be readable and writeable. Hitting it with a hammer will insure any data is lost permanently.
It is assumed that the drive is a disk drive. It could be very well a SSD.
There appears confusion over how a disk under normal operating conditions reads and rewrites, and how cleansing or scrubbing software works. There is a significant difference. Under normal operating conditions old data sits on the media until fully overridden. This allows potential for future data recovery even if the files have been deleted by the operating system. Cleansing works by rewriting over every byte/sector over a whole drive/storage to remove any legacy data. Cleansing software removes opportunity to recover data. Even a friend who is a specialist data recovery analyst for an enforcement agency has indicated that it is important that those subject of raids aren’t aware of a pending raid, as they can scrub valuable data/evidence which can no longer be recovered by the agency. The data is lost for ever. This was said when we asked for advice …when we gave away an unneeded desktop to a education charity before moving to Tassie - we asked how to guarantee legacy data clusters were fully removed. The end result of a proper cleanse is a completely blank storage, just like it is new. OS etc needs to be reinstated.
If anyone is concerned with management of backup files, it is recommended that they are stored encrypted on the backup drive, so that data on a drive (if the drive is lost, stolen or disposed) can’t readily be seen by the prying eyes of others.
Edit: As there are a number of options and providers of disk scrubbing/cleansing/sanitization software, and being another useful tool for protecting data security, it may be worthy of Choice to do a review. I suspect that they are not all equal, and some may be more useful tools than others.
For those who are interested, the latest standard to irretrievable data sanitization is NIST 800.88 US. If you plan to use a data sanitizer, make sure it meets this standard. Sanitizing software with this standard is used by organisations such as US Defence, government departments and business to ensure stored data can’t be recovered. More information on the standard is available here.. Note: Standard DoD 5220.22-M has long been an industry standard when it comes to data sanitization, but drive technology has changed drastically since the standard was accepted. Avoid software that only complies with DoD 5220.22-M.
Many of us might do the same, knowing there is a cost in time and effort to do otherwise.
If one has the knowledge and toolset, is time blessed and money tight, a different choice, might perform a resurrection. But for how long?
I might persist for an hour or two to be sure before deciding on option 1. Experience of drives experiencing any failure it is not a random one off. Most progress to become more, and not less reliable.
Optionally?
Is it fact spinning disk and solid state drives mostly fail for physical reasons? Something software might be aware of but is unable to truely repair. Hence a downward spiral follows. While I might in a life time have experience of only ten or so failed drives, only one has ever presented with a failure not indicative of the active storage medium. It was still a physical failure occurred with the controller card loosing one of the factory parameter settings. Not one software can compensate for.
Is it an external disk or an internal disk? Is it a magnetic disk or a solid state disk? Giving the make and model number could clarify this question.
Is it completely dead or just giving errors or giving no errors but giving corrupted content?
The third question is: how sensitive is the content? That is something that only you can answer (and you may not want to answer publicly). You are concerned about it falling into whose hands? What would be the consequences if it did?
Without any of those details: I would say “incineration”.
Fortunately, there are published standards for deleting data. The US Department of Defence used to use multiple (up to seven) overwrites.
It was never issued for ‘civilian’ use, and as previously stated above has been replaced by NIST 800-88 revision 1.
The reality is that we think of computers as digital devices, but some of their operations reside in an analogue world. That means that early hard drive technology could have some ‘spread’ of the data written on it beyond the defined physical bounds of a block or sector, and overwriting just once did not necessarily get that spread. Modern hard drives are so efficient at compacting enormous amounts of data into tiny spaces that they need their own error-correction routines for when data is not written correctly.
With solid state drives, a charge can decay over time changing a 1 to a 0. Again, error-correction is required to ensure this does not occur or correct once it has occurred.
So methods of destroying/deleting data have to consider the strengths and weaknesses of the chosen storage device, the value of the data to an ‘adversary’, and the cost and benefit of ‘100% definitely wiped’ vs. ‘90%+ wiped’. Then of course there are environmental considerations. Modern storage devices contain all sorts of nasty chemicals, incineration of which may cause side effects such as headaches, fatigue and death. If symptoms persist, see your doctor.
Looks like the OP was a hit-and-run so we are just debating amongst ourselves but …
The point of my answer was: if a drive fails so that the content is intact but it is inaccessible (for example, there is a fault in the interfacing electronics within the drive - or in the case of a magnetic disk a mechanical failure that prevents movement) then for most people it will be technically impossible to repair the drive so that the content can be erased and/or financially unrealistic to do so … and hence if the content is highly sensitive
… your only reasonable choice is destruction of some kind. This is the worst case scenario and in the absence of answers to my questions, I will assume the worst case.
Of course if the content is corrupted but the drive is operable - so that the drive is considered too unhealthy to continue using on an ongoing basis - then I will recommend secure erase followed by delivery of the drive to your nearest eWaste centre.
I guess this highlights the benefits of having a drive with SMART capability (or anything functionally equivalent) and actually monitoring the stats over time - so that you can retire a drive before it fails completely.
Oh no! SMART is only as good as its implementation, which does not always tell you that a drive is heading for failure and can sometimes be quite opaque.
I have just been diverted by the status reserved for ‘Current Helium Level’ - Backblaze has published data on the helium drives it uses.
True, but 20 used M2 SSD’s can rattle around in a drawer and not take up any noticeable space. Are there innovative ways to fry each memory chip, Or is a deft hand with a sharp chisel and blow across each chip sufficient to shatter the silicon substrate rendering the device unusable?
Not so. The data on a disk is stored semi digitally, and read digitally. Each 0 or 1 is a magnetically encoded bit of the disk, but that area is not 100% defined. As long as it is ‘big enough’ to be distinguished by the read heads as a 0 or 1 then the drive works. If data is overwritten once, the size and magnetic intensity of areas is influenced by what was there before. A 0 written over a 0 will be slightly bigger with a larger magnetic field than a 0 written over a 1.
Forensic hardware reads the actual size and intensity of the data written as analog data. By knowing the characteristics of the write head, it is possible to subtract the effect of the last data to reveal the data written before. Reputedly this is possible for more than 2 layers of data.
This technique is not used normally because it is far more complex and costly than doubling the size of the platters.
Sounds about right. Except that the number of particles in each magnetic domain doesn’t really change. So the size of a bit remains constant. The degree of polarization does change.
If a bit position was already more polarized to be read as a zero than a one, then writing over with a zero would reinforce the degree of polarization.
If you could differentiate between polarized enough and more strongly polarized with some very sensitive read heads, then you could determine the previous bit setting.
So disk cleaning software would normally offer multiple passes of writes using different patterns each time.
That is true. That is another option for the original question: don’t.
Leave it to your executor, by which time the damage from the content’s falling into the wrong hands is reduced.
SSDs includes SSDs with a traditional SATA physical interface, in addition to SSDs with an M.2 physical interface. The former are basically the same form factor as the corresponding magnetic disk, so not as compact as M.2 disks.
Most people would recommend … multiple passes. Job done.
Which is why anyone serious about it physically shreds the drive (or any storage/memory component for that matter) into small fragments and then toss them into a high temperature furnace. Probably not viable for home user but the reality is that physical destruction is the only way to be sure.
For home users? yeah - hammer and drill - if you are interesting enough then the people who really want your data already have it anyway
… thats spooky - I still have a meme page open where I was going to add that to my post Of course, it wasn’t to be for them … “game over man, game over” !
Sorry but I cannot resist this invitation. I must say I fail dismally when it comes to restricting myself to a “modicum” of learning, as I also suspect most other posters in this thread would too.
But, you are right about the disks in the olden days of longitudinal magnetic recording.
However, what relevence that has to the last few decades is of no interest since it is long gone technology.
I destroyed a disk in 2017 that MIGHT have yielded my own archival Bitcoin recovery info. I bought over 100 when they were insanely cheap, mostly to (falsely -as it turns out) cover my tracks for certain medicinal herbal purchases. And no criminals DON’T use Bitcoin to cover their tracks, no matter what the media tell you. Fiat does that job with far less hassle!
And no criminals DON’T use Bitcoin to cover their tracks, no matter what the media tell you. Fiat does that job with far less hassle!