what causes HDD failure?

[Red Squirrel]

One of the things I know is physical movements, like bumping, or moving the PC and it "drops" off the edge of the desk (the case feet sliding off and bottom of case hitting desk, basically a 1mm drop but still rough)

But other then that, any other common things? Like this one HDD of mine that is brand new randomly crapped out. It was connected to a cheapo $60 IDE controller. Could a controller cause that? I just want to ensure that if I buy another, the same thing won't happen.

 

[Gillbot]

Heat

A controller could cause a drive to fail if the card itself had an issue. I've seen many a card fail and take the drive with it.

 

[Red Squirrel]

Yeah heat was not the issue, in fact all the drives are cold. because theres such nice cooling in that case. the replacement drive will be sata so it wont be plugging into that controller so guess I don't have to worry if it is in fact the controller.

 

[n7]

I think people look at HDDs the wrong way.

The question is always why they fail.

I feel the better question is why they don't fail?

Considering the complexity of the mechanical components in a HDD, i'm amazed they last as long as they do.

Oh, & does heat actually endanger a HDD's lifespan?

Google study on just that: http://www.searchengineworld.c...pers/disk_failures.pdf

Short answer, not likely (at least within reason).
In fact, according to their study, HDD's with lower than 30C average temps are the most likely to fail to in first two years...

And the 30-35C range has lower failure rates than <30C for as long as they tested (4 yrs.)

Not saying this study is the be all end all, but it shows that keeping HDDs anywhere close to cool isn't ideal, or at least not necessary.

I do wonder about external HDDs though; i suspect temps are an issue with them based on the 3 & 4 yr results where high temps have higher failure rates.

 

[Mr Fox]

Originally posted by: n7
I think people look at HDDs the wrong way.

The question is always why they fail.

I feel the better question is why they don't fail?

Considering the complexity of the mechanical components in a HDD, i'm amazed they last as long as they do.

Oh, & does heat actually endanger a HDD's lifespan?

Google study on just that: http://www.searchengineworld.c...pers/disk_failures.pdf

Short answer, not likely (at least within reason).
In fact, according to their study, HDD's with lower than 30C average temps are the most likely to fail to in first two years...

And the 30-35C range has lower failure rates than <30C for as long as they tested (4 yrs.)

Not saying this study is the be all end all, but it shows that keeping HDDs anywhere close to cool isn't ideal, or at least not necessary.

I do wonder about external HDDs though; i suspect temps are an issue with them based on the 3 & 4 yr results where high temps have higher failure rates.

Shock, obviously, and bad Power/PSU will toast them.

I have one set of Raptor 1's 36Gig (8MB Cache) that is Five Years old... they still run great !!

Most of my older IDE drives died after 3 years unless they were in a USB Enclosure.

 

[Red Squirrel]

Hmm that server has a 700 watt PSU, but it came with the case. Probably a crappy one. I need to try and find a fuse for my multimeter (which I blew by forgetting to bypass fuse for testing amps... oops) and test out if the voltages are too jumpy.

Interesting note on heat. Is there such thing as "too cool" when its just air cooling? Mind you this is probably one of my first HDD failures on a true production box, so probably just a really bad case of bad luck. Teaches me to set backups to hourly not daily. Lost all of yesterday's work.

 

[imported_wired247]

Originally posted by: RedSquirrel
Hmm that server has a 700 watt PSU, but it came with the case. Probably a crappy one. I need to try and find a fuse for my multimeter (which I blew by forgetting to bypass fuse for testing amps... oops) and test out if the voltages are too jumpy.

Interesting note on heat. Is there such thing as "too cool" when its just air cooling? Mind you this is probably one of my first HDD failures on a true production box, so probably just a really bad case of bad luck. Teaches me to set backups to hourly not daily. Lost all of yesterday's work.

Unless your room temp is below 15C or so I'm sure your HDs will be fine on air cooling...

 

[jiffylube1024]

Hard drives involve rotating platters and moving read heads that hover above the surface of the disk to read the data. As n7 said, it's amazing that more HD's don't actually fail young.

One thing you could really hear go on older hard drives (older than about 4 years) is "platter whine" - before the invention of fluid bearings, the (ball?) bearings used in the rotation of the disk would start to get really loud and whiny after about 3 years of operation. I'm sure there's similar wear and tear on all of the parts involved in rotating the platters so fast for so long.
--------

A traditional hard drive "crash" involved the read head slipping off the pin or whatever holds it in place and actually bumping into the disc. When the drive is on, the disc is always rotating, and this crash would cause a scratch along the whole surface of the platter(s).
-------

More recently, the drives I've had that 'crashed' involved a loud clicking sound, almost as if something was blocking the rotation of the disc at certain times, or perhaps that the motion of the read heads wasn't perfectly smooth anymore and perhaps that was hitting something.

 

[Calculator83]

Originally posted by: Mr Fox

Originally posted by: n7
I think people look at HDDs the wrong way.

The question is always why they fail.

I feel the better question is why they don't fail?

Considering the complexity of the mechanical components in a HDD, i'm amazed they last as long as they do.

Oh, & does heat actually endanger a HDD's lifespan?

Google study on just that: http://www.searchengineworld.c...pers/disk_failures.pdf

Short answer, not likely (at least within reason).
In fact, according to their study, HDD's with lower than 30C average temps are the most likely to fail to in first two years...

And the 30-35C range has lower failure rates than <30C for as long as they tested (4 yrs.)

Not saying this study is the be all end all, but it shows that keeping HDDs anywhere close to cool isn't ideal, or at least not necessary.

I do wonder about external HDDs though; i suspect temps are an issue with them based on the 3 & 4 yr results where high temps have higher failure rates.

Shock, obviously, and bad Power/PSU will toast them.

I have one set of Raptor 1's 36Gig (8MB Cache) that is Five Years old... they still run great !!

Most of my older IDE drives died after 3 years unless they were in a USB Enclosure.

Yup u just had bad luck. Hard drives, small moving parts, gonna fail, they're doomed at production. The google studies are likely Not worth much because there are more drives in one temperature range than the other, and that would mess up the failure averages.

 

[NXIL]

Hey Red,

just read this article, which may be pertinent: pg 44 of Feb 2008 issue:

http://www.computerpoweruser.c...BC4BECA739868026113FF6

They talk about hard drive failures in one section of their lab, which stopped when they added a quality UPS (interruptible power supply). Flipping one light switch caused a voltage dip, which trashed hard drives.

Surge protectors: pretty lame.

http://www.apc.com/products/

HTH

NXIL

 

[PCTC2]

Yeah, I still have 3 harddrives that are about 1-4 years old each, and each run between 50-70C and none have failed.

I have killed about 2 dozen harddrives from bad PSUs and people throwing my computer around.

 

[Red Squirrel]

Yeah I got a UPS so I'm good in that department. It was a brand new drive so if any elements would of destroyed a drive it would probably hvae been an older one. So was probably bad luck. Time will tell, hopefully I don't get another failure.