What Can Kill A Hard Drive?

[michaelahess]

That's what I'm talking about, now give us the chemical composition of the magnetic media and the physical properties of a spinning disk and the effect on the media

 

[nottheking]

While I see a lot of potentially valid means for making HDDs un-usable, I must say that in actuality, many of those occurances would be rare, as well as the fact that after most of them, it would be QUITE possible to recover all of the data... It'd just cost you a few cool grand.

However, in normal, SANE treatment, I've found that it appears only really ONE thing is out there that all PC users (and in particular, enthusiasts/gamers) should be wary of: heat.

Since a hard drive is simply using magnetism to read and record data to a chemical medium, heat can REALLY mess with things here, even though the metals chosen for modern storage mediums (Platinum, Palladium, and Cobalt, if memory serves me correctly) are known to be fairly temperature-stable, you can still mess them up.

As a general rule, one should make sure to keep their hard drive's temperature as cool as they can within their case; waterblocks (let alone phase-change solutions) aren't necessary, but good airflow is. Generally, at or under 40 degrees Celsius (104 Fahrenheit) is good enough. More than that, and you could consider things risky. Of course, that's largely my opinion; I KNOW many will disagree with me there.

As far as the topic on hand... Formating a drive to ANY FS will, obviously, cause all data to be lost. However, since it will typically (unless it's a "quick" format) go through EVERY cluster on the drive, it can increase the wear on it. Spinning may be the main source of wear, but reading or writing do pose the potential for a mishap, which could render a cluster useless.

As for this only being said when formatting AWAY from NTFS... I really couldn't guess as to that. Since it does seem to write parts of the file table to areas that AREN'T in the early sectors of the drive, it could be what it implies is that with the new FS, you may wind up writing data to sectors that previously held file table sections. Not quite sure what bad that might bring, though...

 

[tcdude]

EMP would be my favourite...

EMPs don't erase magnetic storage that easily. To erase a hard drive, you need a constant magnetic field of 20 Terra (Earth's field is only around 0.4 Terra, if I remember correctly), for around 2 hours.

ok, so this means I have to look around for other methods to kill my hard drives than the heart-shaped kitchen-magnets i have stored in my server-room

 

[ixion]

EMP can and does kill hdd's. EMP stands for Electromagnetic pulse in the same way you get electromagnetic waves, light, radio, x-rays etc. EMP knocks out electronics by producing very high electric fields many kV/m which induce currents in circuitry, these currents can be sufficiently high to blow integrated circuits up as a hdd has ICs on it, it is susceptile to EMP.

 

[sdrawkcaBgoD]

And nuclear weapons. Or is that overkill?

 

[tcdude]

And nuclear weapons. Or is that overkill?

not nearly is that overkill!! that's the way to go... i'd rather be hearing also about chemical agents that may attack and destroy a hdd.

but remind me again, is the topic confusing me a bit or should we be providing methods to kill a hdd or methods to keep them alive as long as possible? 8O

 

[ares_wargames]

put the HDD in a microwave

 

[tcdude]

put the HDD in a microwave

been there... i also had some bad expiriences with raw eggs in the microwave oven... cleaning it was a b.... though

edit: how would u call it? discs'n'chips...

 

[clue69less]

EMP can and does kill hdd's. EMP stands for Electromagnetic pulse in the same way you get electromagnetic waves, light, radio, x-rays etc. EMP knocks out electronics by producing very high electric fields many kV/m which induce currents in circuitry, these currents can be sufficiently high to blow integrated circuits up as a hdd has ICs on it, it is susceptile to EMP.

EMP also stands for Experience Music Project, a really cool place in Seattle that would never do a bit of harm to your HD.

 

[ASk]

The data is not stored on the IC
Given the platters, you can rebuild the data relatively easily

 

[sailer]

Has anyone ever formatted an NTFS drive with some other filesystem, like RiserFS or ext3? Within the formatter program, it gives a warning that this MAY hurt the drive. I too thought it was BS, until my primary (nothing important, just windows and SUSE) started randomly crashing and it got a very low score in speedfan's SMART analysis. Now I am beginning to think maybe there is something to this.

Woulda been nice not to have to buy a new hard drive out of my very meager budget though.

One time years ago I had an old drive and I wanted to format it for a newer computer. I got the warning that doing so might damage the hard drive, but I could try to force the format if I wanted. I tried to force it, and the hard drive died. Hoping that maybe it could be recovered I took it to a shop and was told that the old drive's sectors were different from those used on new drives and I had scrambled everything beyond recovery. That was years ago, the drive was only 6 meg or so, and probably very different from anything seen today. I can't say exactly what happened, only that it did happen, so yes, those warnings might be true.

 

[elpresidente2075]

One time years ago I had an old drive and I wanted to format it for a newer computer. I got the warning that doing so might damage the hard drive, but I could try to force the format if I wanted. I tried to force it, and the hard drive died. Hoping that maybe it could be recovered I took it to a shop and was told that the old drive's sectors were different from those used on new drives and I had scrambled everything beyond recovery. That was years ago, the drive was only 6 meg or so, and probably very different from anything seen today. I can't say exactly what happened, only that it did happen, so yes, those warnings might be true.

Hmm. That must be the problem that those warnings were talking about. Sounds kinda plausible taking into account the Linux community's affinity for support for outdated hardware. That must be why such a warning exists.

but remind me again, is the topic confusing me a bit or should we be providing methods to kill a hdd or methods to keep them alive as long as possible? 8O

Either way. It's fun seeing what kind of destructive manners people can devise for data/hardware destruction.

 

[ASk]

One time years ago I had an old drive and I wanted to format it for a newer computer. I got the warning that doing so might damage the hard drive, but I could try to force the format if I wanted. I tried to force it, and the hard drive died. Hoping that maybe it could be recovered I took it to a shop and was told that the old drive's sectors were different from those used on new drives and I had scrambled everything beyond recovery. That was years ago, the drive was only 6 meg or so, and probably very different from anything seen today. I can't say exactly what happened, only that it did happen, so yes, those warnings might be true.

This would be a case of CHS (Cylinder-Head-Sector) access versus LBA access. Your old drive was probably formatted under BIOS' CHS access mode, and you tried to format it under LBA mode. This would destroy all the data on the drive, but it shouldn't destroy the hardware.

 

[michaelahess]

Or they are really old MFM or RLL drives....

 

[mitch074]

It can indeed happen that a disc gets so scrambled that further formatting is useless - it appears dead.

However, there are several tools allowing one to, for example, do an inconditional erase of the whole drive (fill it with '0' on several passes). Some are made by the manufacturer (Hitachi provides Drive Fitness Test for their and IBM's drives), some others are generic (have a look at the Ultimate Boot CD, Disk repair section).

Most, if not all my drives are Hitachi now. And once every few months, I take one apart and check it. This actually allowed me to repair a 2.5" 5 Gb drive that wouldn't work at all, a 60 Gb that had bad sectors problems and a few others behaving erratically out of the box.

I also remap them several times; it is however true that sometimes, removing a NTFS partition to put an ext2/3 in its place can lead to erratic behaviour.

If that happens, just erase the MBR and remap the drive. It's long, but more often than not will do the trick. Say bye-bye to your data though.

 

[kyleawesome]

By a negligable margin at best, formatting is the same as accessing data across the entire drive at once. Not exactly different from what most users do anyway, just all at once instead of spread out. And that depends also on the type of format.

I would venture to say nobody has ever killed a drive by formating it to death.

And 4-5 years isn't a great run for a well kept hd, they should last upwards of 8-10 if taken care of properly, even under heavy use. Most HD manufacturers state a 5 year life span, same as car companies stating 10 year life spans, they only do this so you don't bitch when the drive craters right after 5 years of life.

BS Please don't mislead people if you don't what you're talking about.

3-5 years is far more accurate. I replace mine for my primary system every three years just to be safe. In an office setting they may last longer but generally don't do the trashing the my home computer does. The newer drives may also last longer but no drive with 5 years on it should be trusted for anything important and no drive that is 5 years old or older is worth having. 10 years ago drives were measure in Megabytes it couldn't even hold my windows folder...

Frequently defragging on these large drives will significantly reduce their life if not done incrementally over time.


Alright, show of hands, who still has a 98/95 box lying around? (me with a 3,200mb drive) Let me ask you this, does it still boot? 98 being almost a decade old, and 95 surpassing that length of time, I dont find it hard to believe a drive can last longer then 8 years...

Explain:

 

[kyleawesome]

However, there are several tools allowing one to, for example, do an inconditional erase of the whole drive (fill it with '0' on several passes). Some are made by the manufacturer (Hitachi provides Drive Fitness Test for their and IBM's drives), some others are generic (have a look at the Ultimate Boot CD, Disk repair section).

Hitachi = DFT
Seagate = Seatools
Maxtor = Maxblast
Western Digital uses a few either WDTools or some generics

 

[katmandude]

humans do... :mrgreen:

 

[the1pro]

ok, so let's go over what will and what won't kill a HDD

a hammer will
an EMP will
nuclears will
a microwave owen won't (nonono)
a regular owen will
operating constantly for more than 10^8 seconds will
in-case heat will
formats won't
dropping it will
often physical access to it will (due to static)
god will (if you are a bad boy)
you will (if it's too slow)
your girl will (if she sees that photos you got in it)
coke will
water will
dust will (if you take out the prottective cover)
lighniting/over-voltage-nasty psu will
drilling a hole in it will

 

[bmouring]

Or for any HDD:

Boot a Linux(or BSD) LiveCD of your choice.
[code:1:07c978e743]fdisk -l
//Find the listing that corresponds to the size of the drive you're looking for
//i.e. /dev/hda
dd if=/dev/null of=/dev/hda[/code:1:07c978e743]

Same effect, no artificial limitations that prevent it from working with other drives.

 

[sailer]

One time years ago I had an old drive and I wanted to format it for a newer computer. I got the warning that doing so might damage the hard drive, but I could try to force the format if I wanted. I tried to force it, and the hard drive died. Hoping that maybe it could be recovered I took it to a shop and was told that the old drive's sectors were different from those used on new drives and I had scrambled everything beyond recovery. That was years ago, the drive was only 6 meg or so, and probably very different from anything seen today. I can't say exactly what happened, only that it did happen, so yes, those warnings might be true.

This would be a case of CHS (Cylinder-Head-Sector) access versus LBA access. Your old drive was probably formatted under BIOS' CHS access mode, and you tried to format it under LBA mode. This would destroy all the data on the drive, but it shouldn't destroy the hardware.

Yeah, I know that with enough work, the drive might have been used again. Then again, it was a very old 6 meg drive, so it seemed much easier to just buy a new drive.

 

[elpresidente2075]

What about a stampede of bulls? Surely they might dent it enough to make it inoperable.

 

[ches111]

"One day at band camp" We had this HDD and we formatted it three times with three different operating systems. Man you should have been there

Can you say "Mechanical"? ANY mechanical device is prone to failure for ANY number of reasons. You simply mitigate these reasons through good design.

Heat can be an issue, but given the operating temperature specs of the newer drives this is becoming less of an issue (quite a large range). I have left my notebook computer in my truck during a full night of -13 degrees F. It booted no problem "right after I let it come back up to the operating temperature range". Kinda cool seeing the condensation on the notebook evaporate away!

There are obviously three main elements to any HDD. Platters, Spindles and Read/write arm. If any of those three items become misaligned there is a problem. Heat can cause a distortion of the platters or the read/right arm. Not likely in normal operations but could happen in a poorly designed machine with very poor ventilation.

What we need is a frictionless bearing like a Mag Lev bearing. With a fixed read/write arm with pickups the entire length of the platter so it could read each sector in a fixed fashion. Then all you would need is a synchronization of the platter speed with the fixed read/write arm. This is similar to beam steering.

But this would only work well in computers that DO NOT MOVE.

By the way Mag Lev would also be beneficial since after the initial spin up of the device (slow) you would never have to wait for spin up again because you could simply keep it spun up. And the platter speeds could be MUCH higher if the platters were gyroscopically stabilized!

Edited to correct horrible typos!!

 

[ches111]

K, now the short answer,

Heat and Vibration tend to be the main killers of HDDs.

 

[ares_wargames]

ok, so let's go over what will and what won't kill a HDD
a microwave owen won't (nonono)

Why wouldnt it?