Hardware How do HDD's die?

[gifi4]

Just curious, my 2 internals are fairly old (They were new when my parents got them, only has a capacity of 37GB) and my 1TB external is new but I'm afraid of my external failing on me cause I access it all the time, hell my torrents download directly onto there, so can it randomly die if accessed to much or what?

 

[Rayder]

It's a crap-shoot really. Some drives last forever, some die in a relatively short time. I chalk it up to bad components and shoddy workmanship.....some don't meet the specs they were supposed to be designed for......fluctuations in the manufacturing process. Unavoidable.

Overheating, power-spikes and knocking them around while in use can cause them to fail too.

Sometimes, you will have warning signs that they are dying.....flaking access, running slow, etc. But usually, if they are gonna die, POOF! They just crap out on you and you're screwed.

You would think that after all the years HDD's have been manufactured, they would have them perfected by now and they would last far longer than you would need them to, but nope. Of course, I think the same thing about a lot of things. It never works out that way.

 

[Originality]

Each sector has a read/write lifespan which, after so many cycles, will wear out and become bad sectors. Once flagged as a bad sector, the HDD is supposed to stop writing to it and work around it instead. That's one way for a HDD to slowly die - to get filled with bad sectors until it's pretty much unusable. Also vibration can cause the header to scratch the surface which will cause a lot of bad sectors.

The most common way for a HDD to die is for the spindle to wear out (which leads to friction, heat, vibration, and ultimately widespread damage). Also the motor can wear out, which stops the drive from being able to rev up. When you get hardware failure like this, the most common symptom is to hear a constant clicking sound.

It's also possible for the PCB to burn out (usually caused by heat, sometimes by power surges). Once this happens, accessing the HDD is mostly impossible without one of those expensive data recovery tools.

There are a couple other ways for a HDD to die, but I don't know enough about HDD hardware to explain them. My knowledge mostly comes from what I've experienced myself.

 

[Vulpes Abnocto]

The simplest answer is "friction"

Without that single factor many HDDs would last far longer than their current expected lifespan.

 

[Pong20302000]

all moving parts eventually break anyway

if your worried purchase SSD (Solid State Drive)
no moving parts so last alot longer
you get what you pay for

 

[Originality]

SSDs may not have any moving parts, but it is still subject to friction. Each of the chips used SSDs has a certain amount of read/write cycles expected before they die. SSDs tend to be overprovisioned precisely to compensate for any damage built up within the SSD. Also certain SSDs have a much lower life expectancy than the average HDDs. A bit like how Seagate HDDs have a lower life expectancy than other HDDs (my opinion, based on observation and experience).

With HDDs though, it's vibration more than friction that causes damage (even though friction itself causes vibration).

 

[AlanJohn]

Viruses.

 

[gifi4]

Thanks to everyone for chipping in information. I suppose it's pretty hard to prevent then.

 

[Originality]

Measures can be taken to protect HDDs to a certain extent. Rubber mounts absorb vibration (both from the HDD and the case). Many modern cases have a fan positioned either on or near the HDD mounts (keeping them cool prevents two of the causes of HDD failure). HDD mirroring (RAID) protects your data incase one of the HDDs fail. Also many HDDs will spin down when idle to make it last longer.

Picking Samsung or WD instead of Seagate or Maxtor (which is now Seagate anyway) means you'll have a more reliable drive. I've no opinion on Hitachi drives though...

 

[doyama]

Hard drives can fail for a number of reasons. There are the base electronics. Also there are mechanical pieces (spinning platters, floating heads) that can fail as well. There are SMART codes that a hard drive can spit out that 'hopefully' tell you things before a hard drive dies. Though to be honest this doesn't happen all that often. Usually the drive just dies spontaneously without warning.

There's not much a user can do in terms of making things better. You can try to maximize the life of your drive by

1) Having adequate cooling. Fans and whatever you can throw at it to make it cool will help. Heat is the enemy of all electronics.
2) Get rid of dust regularly. This is more to address the above heat problem. Dust traps heat thus reducing the life of your parts
3) Minimize movement. Not usually a problem for internal drives. For externals don't move it around when the drive is on and active.
4) Check warranty on all your drives. I have a small spreadsheet that lists the model, serial number, and the warranty date on it for all my drives. This allows you to assess risk of data loss vs time.

The things is the hard drive dying isn't the problem. It's the DATA LOSS that's the problem. Thus you need to think about data loss prevention more than hard drive failures. A few ideas

1) If your mobo supports it use a mirror RAID config.
2) You can use programs that will copy directories that you want onto another drive on a regular schedule

Personally I have a 1.5TB mirrored RAID config. I put stuff on there that are unrecoverable. For me that's stuff like my documents, personal pictures, and a few other things. I try to keep that drive with as minimal amount of what I think are 'critical'. Once you think about it, maybe you won't really be all that upset if you lose your torrent movies, anime collection, nds games, etc.

 

[marcus134]

Although RAID seems to be a good idea for preventing data loss, it's a single point failure.
raid 1,5,6,10 have their own weaknesses and must not be your only back-up option.

1. usually, raid arrays are built with a bunch of drive of the same model, bought and put into function at the same time, meaning that there is a possibility that if one drive dies, the other ones will follow in a relatively short span of time.

2.death of the raid controller. This is not always a problem if you use a pci or pcie raid controller as it is usually possible to buy another one identical to the first one.
However, with raid controller integrated on a mobo, if the controller or the mobo dies, you have to find, ideally, another identical mobo or a mobo with the same chipset (or controller)and cross your finger (sometimes, a more recent model of the controller may be compatible with arrays built with an older model). It may not seem like a big deal when your computer is brand new, but when it is 5 years old, the hopes of finding a compatible mobo are pretty slim.

Edit: point 2 doesn't hold for RAID 1 as most drives used in that configuration can be read like normal hdd

 

[Originality]

In the case of RAID 1, if one drive fails then you can expect the other to die soon after, but not straight away. You often have enough time to buy a new HDD to replace the old one (effectively backing up your data again) before the other drive fails.

In the event of the RAID controller failing, that still should not prevent you from using the drive as a standalone (again, assuming RAID 1 configuration). That should last long enough to purchase a new PCI/PCIe RAID card. The only problem comes from reconfiguring BIOS, but that shouldn't take long anyway.

Then again, it is true that your HDD should not be your only backup option. Having an additional computer/server in the house to backup files to, or having an online storage solution, or even just using external HDDs (even though external drives have a significantly higher failure rate than internal HDDs) are all valid ways of ensuring your data is safe.